FIELD OF THE INVENTION

This invention relates to a particle sensor and sensing method.

BACKGROUND OF THE INVENTION

Particulate matter (PM) suspended in the air, often referred to as atmospheric aerosol, has been attracting tremendous public interest due to its testified adverse impacts on human health. Either outdoors or indoors, PM can be natural (pollen, sea salt, mineral dust, volcanic ash, etc.) or artificial (soot and organic droplets from automobiles or from industry, cigarette smoke, cooking fumes, etc.).

In addition, PM can also be categorized as primary or secondary depending on the transformation it undergoes during its transport. An example of primary PM is pollen or mineral dust. A typical example of secondary PM is polyacrylonitrile (PAN) and propranolol (PPN) which are formed by NOx and VOC (volatile organic compound) precursors via a photochemical reaction in the presence of sunlight and moisture.

Evidence and statistics have shown that prolonged exposure to an environment with elevated concentration of fine PM (PM2.5 with diameter less than 2.5 micrometers) poses particular risks to human health. Depending on the specific characteristics and compositions (primarily the chemical compositions), different categories of PM can affect human health from different perspectives. For example, water droplets and dissolved sea salt have a neutral health effect; pollen causes allergies due to allergen release; NH4NO3 (a secondary aerosol formed by NH3 and NO/NO2) and acidic aerosols typically result in airway irritations, coughing and breathing difficulties; organic aerosols (formed via photochemical reactions and directly emitted primary organic aerosols) in general increase the risk of heart attack as well as lung and cardiovascular diseases.

Many efforts have been made in developing PM sensors for either scientific or commercial applications. Generally, such sensors are optical sensors, using either infrared LEDs or laser diodes as the light source. In all cases, these optical PM sensors all apply the same underlying principle based on Mie Theory. Mie Theory, named after Gustav Mie, is an analytic solution of Maxwell's Equations describing the scattering of an electromagnetic plane wave by a homogeneous dielectric spherical particle. This theory relates to the light scattering pattern (angular light intensity distribution) off particles with diameter in the range 0.1λ to 20λ, where λ refers to the wavelength of the incident light. When using a visible red laser diode (λ~600 nm) as the light source, the applicable range of Mie Theory (60nm to 12μιη) covers PMio as well as PM2.5, which provide the greatest threats across the whole PM size spectrum.

The scattering pattern of light intensity, calculated by Mie Theory, is represented by two scattering phase functions Si and S2 which are functions of the particle size parameter (2πά/λ), particulate refractive index (m) and scattering angle (Θ).

The first two arguments 2πά/λ and m reflect the properties of a particle itself - size and composition, while the third argument Θ translates the ID incident light into a 2D pattern in the scattering plane. A schematic representation of the parameters relevant to Mie Theory is shown in Figure 1.

Figure 1 shows a particle 1 , an incident light beam 2 and a scattered light beam 4 with a scattering angle Θ.

The scattering forward calculation is given by:

The two vector components E|| _s and E± _s are the scattered electric field intensities parallel and perpendicular to the scattering plane, and the two vector components E||i and E_u are the incident electric field intensities parallel and perpendicular to the scattering plane. The light intensity is proportional to E ² .

The scattering pattern calculated by Mie Theory contains information on both size and refractive index of a particle. However, almost all available PM sensors can only interpret particle number concentration, mass concentration and/or size distribution, while ignoring the information embedded in the particle refractive index.

The ignorance of refractive index retrieval in the available PM sensors can largely be attributed to insufficient measurement of Mie scattering pattern. Due to cost control and size restriction reasons, the sensors are not able to measure scattered light intensity over a continuous angular range, but instead at either one angle with a single photon detector (e.g. 60°) or discrete angles with a photon detector array. To compensate for the influence of particle refractive index during data processing, low cost sensors apply regressed data via dimension reduction with empirical refractive index values; while scientific sensors may require users to input a tested/arbitrary refractive index value for a group of particles with known composition.

The common practice of overlooking particle refractive index in particle sensing prevents retrieving a critical piece of information - the particle composition.

Refractive index, determined by a material's electric permittivity (ε) and magnetic permeability (μ), is a strong indicator of particle composition. Health effects brought about by a certain category of PM are largely determined by the chemical composition, thus understanding the composition of the particle passing through the sensor will provide more robust information about human health implications beyond particle mass concentration. The table below shows organic water soluble acids, with different solution concentrations, with the refractive index measured at 600nm.

The table below shows minerals, again with the refractive index measured at 600nm.

The table below shows different forms of carbonaceous aerosol, again with the refractive index measured at 600nm.

The table below shows the corresponding measurement for pear poll with the refractive index measured at 600nm.

Pear Pollen 1.25 - 1.28

The process of determining particle size and refractive index from the Mie scattering pattern involves an inverse Mie calculation. Compared to a Mie forward calculation, the biggest challenge of the inverse algorithm lies in the forward analytical solution itself, in which the presence of the Bessel and Hankel functions makes it impossible to derive the corresponding inverse function with an explicit expression.

To perform the calculation, an input of sufficient information from the scattering pattern is required to find the best fit within certain boundary conditions (particle size parameter, refractive index), and the insufficient measurement of the Mie scattering pattern in available sensors makes it infeasible to retrieve more information beyond particle number/mass concentration or particle size distribution.

There is therefore a need for a system which enables particle size parameters and refractive index information to be obtained, in particular via an inverse Mie calculation.

SUMMARY OF THE INVENTION

The invention is defined by the claims.

According to examples in accordance with an aspect of the invention, there is provided a particle detection system, comprising:

a flow channel;

an illumination source for providing a collimated light output towards the flow channel;

a light detector;

a focusing reflector for focusing scattered light from a particle in the flow channel towards the towards the light detector, to provide a sweep of scattering angles of greater than 90 degrees to the light detector for independent detection of the light intensity for each scattering angle within the sweep, the scattering angles being spaced by 2 degrees or less;

a transmission control element positioned in the path of the scattered light to the light detector to vary periodically the intensity of the scattered light as a function of the particle's travelled distance along the flow path and thus of its scattering angle; and

a controller for processing the light detector signals,

wherein the controller is adapted to obtain a function of detected light intensity with respect to scattering angle and to perform an inverse Mie calculation to determine a particle size and particle refractive index.

In this system, when a particle is in a flow channel, a light detector receives scattered light. By using a focusing reflector, the system is arranged such that for each scattering angle within a range, scattered light of each scattering angle (wherein the sweep is a set of discrete scattering angles) reaches the light detector for independent detection. In this way, a range of scattering angles is obtained, so that a function of scattering intensity versus scattering angle is obtained for that range of scattering angles. This provides much more information than detection at a single or small set of scattering angles, and it enables inverse Mie calculations to be performed so that both particle size and particle composition information may be obtained. By varying periodically the intensity of the scattered light as a function of its distance travelled along the flow path and thus of its scattering angle, the system can be largely independent of any variations in speed of the particle as it progresses along the flow path, and knowledge of the particle speed is unnecessary.

The sweep may be a set of angles with a constant spacing such as 0.2 to 2 degrees to provide a set of values. The angular spacing is thus 2 degrees or less. The angular spacing may be constant across the set, but this is not essential. It is continuous in the sense that there are no large gaps (e.g. no gaps larger than 2 degrees). The sweep is larger than 90 degrees, and has a resolution sufficiently high (i.e. a small angular spacing) to enable the data to be used for inverse Mie calculations with sufficient data to extract a particle size and refractive index value.

The calculated results of particle size and also composition can be used as control parameters for further control systems, for example for selective indoor pollution filtration, air quality health risk evaluation, healthy air management, etc.

The flow channel may extend through free space, i.e. it does not need to be defined by an outer enclosure. Thus, the term "channel" should be understood accordingly, as essentially relating to a controlled flow path. The channel may be defined between a pair of ducts which control the path of the flow channel.

The illumination source for example comprises a laser diode and a collimator or collimating lens. This ensures light reaches the particle with a known angle of incidence.

In a preferred set of examples, for each particle location along the flow channel, scattered light of a corresponding scattering angle is focused by the focusing reflector to the light detector.

In this example, as a particle moves along a flow channel, a light detector receives scattered light, with the scattering angle changing as the particle moves. By using a focusing reflector, the system is arranged such that for each location along the flow path, scattered light of a particular scattering angle reaches the light detector. In this way, the path of a particle along the flow path yields the set of detector signals for a continuous range of scattering angles.

The focusing reflector for example may then comprise an ellipsoidal mirror. The ellipsoidal mirror preferably has a first focal point and second focal point, further from the mirror than the first focal point, with the flow channel between the first and second focal points and the light detector at the second focal point. This arrangement provides the desired condition that from each location of the particle along the flow channel, light reaches the light detector from the particle with a different scattering angle.

The system for example comprises a delivery duct and a removal duct with the flow channel defined between them. Particles are delivered to the flow channel from the delivery duct, and they then pass along the flow channel. The delivery duct for example has an output nozzle and the removal duct has an input funnel to catch the flow. A flow generating device such as a pump or fan is preferably coupled to the removal duct. This is used to control the flow along the flow channel.

The system may further comprise a trigger for detecting the arrival of a particle from the delivery duct, wherein the controller is adapted to perform a measurement cycle in response to detection of the arrival of a particle by the trigger. In this way, the location of a particle is known by the trigger, and this functions as a calibration for the detection of the scattered light from the particle as it moves along the flow channel.

The trigger for example comprises a light source, a focusing lens for focusing the light source output to a detection region of the flow channel, a collecting lens and a light detector, wherein light collected from the detection region of the flow channel is focused to the light detector by the collecting lens. A change in intensity at the light detector is indicative of a particle present in the detection region.

The system may comprise an enclosure in which the flow channel, illumination source, light detector and focusing reflector are contained, wherein the enclosure comprises a filter portion between the inner volume of the enclosure and the ambient surroundings. By drawing air through the filter (rather than though the delivery duct) the air contained within the enclosure can be cleaned in preparation for a particle detection cycle.

Examples in accordance with another aspect of the invention provide a particle detection method, comprising:

providing a collimated light output towards a flow channel using an illumination source;

focusing scattered light from a particle in the flow channel towards a light detector to provide a sweep of scattering angles of greater than 90 degrees to the light detector, the scattering angles being spaced by 2 degrees or less;

providing independent detection of the light intensity for each scattering angle within the sweep;

varying periodically the transmission of the scattered light to the light detector as a function of the particle's travelled distance along the flow path and thus of its scattering angle;

processing the light detector signals to obtain a function of detected light with respect to scattering angle; and

performing an inverse Mie calculation to determine a particle size and particle refractive index.

This method enables particle size and particle refractive index to be obtained, by obtaining light intensity information for a range of scattering angles.

The method may comprise creating a flow between a delivery duct and a removal duct, wherein the flow channel is defined between them. In this way, the particles fo llo w a known traj ectory.

The method may comprise detecting the arrival of a particle into the flow channel and performing a measurement cycle in response to detection of the arrival of a particle. This provides an initiation function for the measurement cycle.

The method may comprise providing filtered air into an enclosure in which the illumination source, flow channel and light detector are provided before performing particle detection.

The method may comprise controlling the flow to the flow channel from the delivery duct using a flow valve at the delivery duct and/or controlling the flow rate thereby to provide a single particle in the flow channel at a time. By using a valve, the flow from the delivery duct can be reduced. If a constant flow rate is present at the removal duct, the reduction in flow to the delivery duct may be compensated by filtered flow from the ambient surroundings. Alternatively, the flow rate may be controlled.

The signal analysis may be performed at least in part in software.

BRIEF DESCRIPTION OF THE DRAWINGS

Examples of the invention will now be described in detail with reference to the accompanying drawings, in which:

Figure 1 shows the parameters involved in Mie scattering calculations;

Figure 2 shows the main functional components of a first example of a particle detection system;

Figure 3 shows the main functional components of a second example of a particle detection system; Figure 4 shows a trigger for determining when a particle enters the flow channel of the system of Figure 2 or Figure 3;

Figure 5 shows an overall particle detection system;

Figure 6 shows an example of a Mie scattering pattern showing the angular dependence of the scattered light intensity off a particular particle;

Figure 7 shows the scattered intensity versus scattering angle without the transmission control element (the screen) of Figures 2 or 3 and shows the results of the inverse Mie calculation algorithm for a known particle;

Figure 8 shows how the scattered intensity versus scattering angle changes with the screen provided of Figures 2 or 3;

Figure 9 shows a pattern mapping approach for performing the inverse Mie calculations;

Figure 10 shows the results of the calculations of Figure 7 performed on scattered intensity measurements obtained without the screen of Figures 2 or 3;

Figure 11 shows the results of the calculations of Figure 8 performed on scattered intensity measurements obtained with the screen of Figures 2 or 3; and

Figure 12 shows a particle detection method.

DETAILED DESCRIPTION OF THE EMBODIMENTS

The invention provides a particle detection system in which particles travel along a flow path and are illuminated by collimated light. Scattered light is collected by a focusing reflector and light detector to provide a sweep of scattering angles of greater than 90 degrees and with scattering angles separated by 2 degrees or less. There is independent detection of the light intensity for each scattering angle within the sweep. A function of detected light intensity with respect to scattering angle is thus obtained. An inverse Mie calculation can then be performed to determine a particle size and particle refractive index.

Figure 2 shows the main functional components of a first example of a particle detection system.

A particle flow is defined in a flow channel 10 which extends between a delivery duct 12 and a removal duct 14.

The delivery duct 12 has a nozzle feature at its end and the removal duct 14 has a hood-type feature at its opening where the flow is collected. Valves such as needle valves 12a, 14a are formed in the delivery duct (valve 12a) and the removal duct (valve 14a) so that the flow channels may be opened and closed. The delivery duct and the removal duct are arranged as coaxial tubes.

An illumination source 16 comprises a laser diode 18 and a collimating lens 20 (or other collimating arrangement) and it provides a collimated light output towards the flow channel 10. The laser diode is located at the focal point of the collimator. A red laser diode with wavelength -600 nm may be used to cover particle size ranges of most interest, <10 μιη.

A focusing reflector 22 focuses scattered light from a particle in the flow channel towards a light detector 24. For each particle location along the flow channel 10, scattered light of a corresponding scattering angle is focused by the focusing reflector to the light detector.

By way of example, a first particle location is shown as reference 26. The light path from the particle at this location to the light detector 24 comprises scattered beam 26a and reflected beam 26b. The scattering angle for the beam 26a is Θ1.

A second particle location is shown as reference 28. The light path from the particle at this location to the light detector 24 comprises scattered beam 28a and reflected beam 28b. The scattering angle for the beam 28a is Θ2.

The locations 26 and 28 are at the ends of the flow channel 10. Thus, when a particle moves along the flow path 10, the range of scattering angles which reaches the light detector 24 is Θ1 to Θ2. This range of angles may for example cover an overall angular width of between 90 and 120 degrees, which as a result of the symmetry of the scattering function covers most of the possible scattering angles. This continuous angular dependence of the scattered light intensity is used to function as the input of inverse Mie calculation for particle size and refractive index retrieval.

To avoid interference in scattering light measurements, the un-scattered incident light should not reach the reflector mirror and hence does not enter the light detector 24.

The reflector 22 is a belt-shaped ellipsoidal reflection mirror. It is belt-shaped in the sense that it comprises a slice of a 3D shape, in particular part of a 3D ellipsoid. The system is for measuring scattering in essentially only one plane, i.e. in a narrow angular range of planes corresponding to the thickness of the slice. This means the overall system is lighter and hence more portable.

The reflector 22 has two focal points.

A first focal point 22a is closer to the reflector 22 than the flow channel 10 and a second focal point 22b is further from the reflector than the first focal point and on the opposite side of the flow channel 10. Thus, the flow channel is between the first and second focal points 22a, 22b. The light detector 24 is at the second focal point 22b.

In an alternative embodiment not shown, the reflector is parabolic, but it could be of any shape with an axis of symmetry from its apex on which an object point is focussed to an image point. The reflector can have a narrow belt shape in order to reduce sensor size, or it can occupy a larger section of the abovementioned 3D shape in order to capture more of the scattered light to increase sensor sensitivity.

The orientation of the delivery duct 12 and the removal duct 14 is perpendicular to the line between the two focal points 22a, 22b.

The light detector 24 is a photon detector with a light entry aperture for light intensity measurement. The aperture minimizes potential interference during light intensity measurements.

The light detector signals encode a function of detected light intensity with respect to scattering angle. This information may then be used to perform an inverse Mie calculation to determine a particle size and particle refractive index.

The data analysis is based on a knowledge of the particle trajectory along the flow path.

There is therefore a need to know when a particle enters the flow path. There is also a need to know when the particle leaves the flow path so that the position along the flow path between those time instants can be interpolated. Alternatively, a fixed flow speed is required.

In particular, the system is designed so that the location of the particle is known with respect to time. The scattering pattern is recorded at the photodiode in a time series, but in fact it is function of scattering angle (i.e. the location in the flow channel). It is therefore necessary to know the speed of the flow (i.e. the location at a function of time) to perform a temporal to spatial deconvolution.

To allow for the particle speed in the flow channel to vary, in accordance with the invention, the light from the particle is spatially modulated with respect to the flow channel, so that the image intensity is modulated over time as a given particle flows along the flow channel. This modulation pattern provides information on the position of the particle and thus its scattering angle.

In the embodiment shown in Figure 2, a screen 25 with slits (i.e. a perforated shield) is added in front of the focusing reflector. In this way, light reaches the detector as a series of pulses as described below with reference to Figure 8. The spatial position from which each successive pulse originated is then known. In this way, the position of the particle is known, to a precision determined by the period of the slits in the screen 25, even when the flow speed in the flow channel is not constant. This approach limits the amount of information for the inverse Mie calculation by providing a set of spaced angular

measurements instead of a continuous set. However, the slits are designed so that the desired number of measurements is obtained.

In this example, the slits are generally rectangular and of equal width, but this is not essential. At large scattering angles especially, which is the latter part of the trajectory, where the scattering pattern detected at the sensor contains more information, the width of the slits and the gaps between them can be made smaller so that more information will be provided for use with the algorithm for determining particle size and refractive index. Also, in this example the screen is parallel to the flowpath, and the dwell time for each slit, for a particle travelling at constant speed, would vary as a quadratic function reaching a minimum in the centre, so the slit widths could be varied to compensate for this, starting relatively narrow and widening to the centre.

The flow speed in the channel is for example around 0.5m/s. With a channel length of 5 cm the duration of one cycle of measurements takes approximately 100ms.

There is also a need for a single particle to be provided along the flow path at a time. The needle valves in the delivery duct and removal duct are adjusted to ensure that no more than one particle exists in the measurement region during any measurement cycle. To ensure that only one particle is monitored at a time, two approaches may be adopted

(separately or together).

A first approach is to decrease the flow rate in the flow channel (by decreasing the negative pressure connected to the removal duct) to a point where the flow can still be controlled as laminar (particles flow in approximately straight line along the flow channel). The decrease in flow rate will reduce the number of particles entering the system, and thus satisfy the one-particle requirement in a less polluted environment with low PM number concentration.

A second approach is to maintain the original flow rate, but partly close the delivery duct (via the needle valve 12a) to impede the incoming particle flow. In this case, the difference in air flow between the removal duct 14 and delivery duct 12 will be compensated by the clean air entering the chamber through the filter, which will dilute the incoming particle concentration to satisfy the one-particle requirement. A second embodiment is shown in Figure 3, which differs from that of Figure 2 in that the spatial light modulation is performed by a curved slitted screen 25 parallel to the surface of the reflector 22. In a section 100 of the flow path, the scattering angle of light that is focussed at the detector 24 of the sensor varies from Θ1 to 9a, and is not blocked by the curved screen 25. Over the next section of the flow path, the scattered light that would have been focussed to reach the sensor is blocked by a band of the screen 25. The next section corresponding to a slit is not blocked, and so on. The sensed intensity as it varies over time is shown in Figure 8.

In an alternative embodiment not shown, the screen 25 may be replaced with a form of engraving on the reflector 22 so that the reflector has alternating reflective and non- reflective sections corresponding to specific scattering angular ranges. The engraving can be fine enough to reduce the angular range within each section and thus to increase the total number of sections along the flow path. The reverse Mie algorithm described below may be adjusted with this input information dimension. Any form of periodic variation of

transmissivity or of reflectivity could be formed on the surface or formed within the thickness of the reflector 22.

The screen 25 could be formed of any suitable material such as plastics or metal, and the slits could be relatively transmissive regions of a continous laminar material instead of openings.

Any equivalent arrangement of a transmission control element for periodically spatially modulating the focussed light, i.e. modifying the transmissivity of the focussed light, from the particle to the sensor as it progresses along the flow path could be substituted for the screen 25 or engraving, to provide the spatial modulation that gives the positional information. It is not necessary for the screen to be flat or to be curved with the same shape as the reflector. For simplicity and low cost, the screen is passive and fixed with respect to the overall detection system, but it is envisaged that in some applications it may be appropriate to use an active modulator such as an electrically controlled spatial light modulator.

Figure 4 shows a trigger for determining when a particle enters the flow channel 10.

The trigger comprises a light source 30, a focusing lens 32 for focusing the light source output to a detection region 34 of the flow channel 10, a collecting lens 36 and a light detector 38. Light collected from the detection region 34 of the flow channel is focused to the light detector 38 by the collecting lens 36. The light source is a laser diode or infrared LED, and the light detector is a photon detector.

When a particle flows across the beam between the light source 30 and detector 38, the photon detector detects the change in light intensity and then sends a signal to trigger one particle sensing cycle for light intensity measurement and data processing. The trigger thus initiates the measurements and data recording when a particle enters the flow channel.

Another purpose of the trigger is to decrease the load/effort in data recording because the light intensity data (the scattering pattern) is recorded only when a particle is present; other data, probably noise data, will be discarded. However, it is still possible to obtain satisfactory measurements without the trigger. For example, the same laser diode 18 and light detector 24 could be used for triggering as well as sensing. When idling, the laser 18 could be in a low power mode and could be used to detect the presence of a particle in its first scattering section, e.g. with a scattering angle of 10-20 degrees. This detection would act as a trigger to power up the laser diode 18 to measure scattering over all the subsequent sections defined by the screen 25 or engraving. The information from the first section would then be used only for triggering and not for sensing.

Figure 5 shows the overall system for use with the arrangement of either of Figures 2 and 3. The components described above are housed in an enclosure 40 for example formed of acryl glass. The removal duct 14 is connected to a negative pressure system 42, which may be a fan or a vacuum pump for particle removal and for control of the flow path.

The enclosure 40 has a filter arrangement, such as two HEPA filters 44 mounted on a pair opposite sidewalls of the enclosure 40.

The flow induced by the negative pressure system 42 is used to guide the moving particles discharged from the delivery duct along approximately a straight line parallel with its long axis. In addition it is used to remove the particles to avoid

contamination of the environment inside the enclosure and to generate and maintain a clean environment using the filter arrangement before and during the particle detection cycle. Furthermore, the filter can also generate clean air flow which can dilute the incoming particle flow to satisfy the one-particle requirement as explained above.

Figure 5 also shows a controller 46 which receives timing information from the trigger 48, controls the illumination source 16 and receives the data from the light detector 24. The light detector collects and measures the intensity of scattered light when particles pass from the delivery duct to the removal duct across the measurement region.

The system enables monitoring of a continuous range of scattering angles but with no moving parts. The particle movement itself is tracked in order to generate scattering measurements at successive points in time, for the same particle. The range of angles which can be analyzed is continuous. In practice, measurements will be taken at successive discrete time points, such that the resolution of the measurements for example corresponds to a 1 degree interval. For example there may be 115 measurements for scattering angles from 5 to 120 degrees. More generally, there may be between 50 and 500 individual measurements, each corresponding to a different scattering angle. The angle resolution is for example between 0.2 and 2 degrees. The greater the number of measurements, the more accurate the particle size and composition determination, but with a requirement for greater processing resource.

The light source can be a single wavelength source since no spectral analysis is required. Combined with the lack of moving parts, the system is simple to implement at low cost.

In the system above, only relative intensity values or normalized intensity values are needed, or simply a distribution pattern. Thus, absolute values of scattered light intensity are not needed for the inverse Mie calculation. Normalized values can eliminate interference factors including laser power, photon detector sensitivity, instrument dimensions, etc. However, the response of photon detector to light intensity should be linear or the response curve should be known.

Before the test, the removal duct may be opened and the delivery duct may be closed, so that the negative pressure of the negative pressure system 42 replaces any contaminated air in the enclosure with filtered clean air. After a number of cycles of air exchange, the air inside the enclosure is clean enough for the measurement. This avoids interference of the scattering signals by other unwanted particles.

The Mie Theory calculations present certain limitations on the accuracy of the calculations. In particular the boundary condition restrictions for the analytical solution of Mie Theory give the limitations as explained below.

The inverse Mie calculation does not apply to metal particles as the interaction of electromagnetic wave with free electrons in the metallic particle is beyond the scope of Maxwell's Equations. The inverse Mie calculation does not apply to particles with shape

significantly deviating from perfect spheres. Analytical solutions have been obtained for Mie Scattering of non-spherical particles with axis of rotational symmetry (e.g. a cylinder), named the T-matrix solution. This solution adds an additional three degrees of freedom on top of the current spherical particles solution, in particular azimuth and altitude angles denoting the particle orientation, and also the aspect ratio representing another dimension in describing the size/shape of a particle with rotational symmetry. However, considering that the orientation of a non-spherical particle along flight path is randomized, the introduction of a more complicated T-matrix solution is not necessary for atmospheric PM sensing, and the algorithm used may be designed under the assumption that particles for detection are spherical.

Figure 6 shows an example of a Mie scattering pattern showing the angular dependence of the scattered light intensity off a particle with size parameter 20.34 and refractive index 1.234. To reveal detailed information for angles with small scattering light intensity, a logarithmic scale is applied to the polar plot.

The inverse Mie calculation will now be described in more detail as well some calculation results. The calculations involve the following steps:

The angular range for detection and inverse calculation is first obtained. This is the range Θ1 to Θ2. By way of example, 91 = 5° and Θ2 = 120°.

The Mie scattering pattern is then calculated for the angular range based on the measurement cycle for a particle, for example at a resolution of 1°, and with a random angle error less than ± 0.5°. The random angle error represents the error in actual angle

measurements of the system.

A random intensity error is assumed as ±10% for the light intensity data calculated at each angle of the calculated scattering pattern. The random intensity error represents the error in actual light intensity measurements at the photon detector.

The measured Mie scattering pattern data, with the applied intensity error and angular error, is then provided as input for the inverse Mie calculation algorithm. The algorithm provides the first three best determinations of the data pairs (of size parameter a, and refractive index m) that are the best fits to the measured scattering pattern.

Figure 7 shows the results of the inverse Mie calculation algorithm for a known particle with parameters a = 20.35, m = 1.234. These are unknown parameters to the algorithm. Based on a measured scattering pattern, the best fit data pairs provided by the algorithm are (a: 20.46, m: 1.230), (a: 20.46, m: 1.235) and (a: 19.80, m: 1.240). This shows accurate conformity to the known value.

Figure 7 shows the scattered intensity versus scattering angle, and plots the measured results as well as the three best fit theoretical results.

As explained above, a screen 25 such as a flat or curved plate with slits may be added in front of the focusing reflector, or the reflector may have engraved regions, to providing timing calibration information. The effect on the detected signal of this type of spatial modulation is shown in Figure 8, which plots intensity as a function of time (and hence scattering angle). The width in time of each receiving window and each non-receiving window enable the path of the particle to be known as a function of time, even if the flow has a non-constant speed.

After obtaining the scattering pattern, an algorithm is needed to calculate the particle size and refractive index. The forward calculation of the Mie solution contains the Legendre function, the spherical Bessel function, the spherical Hankel function and others, which make it impossible to develop a corresponding inverse function in explicit form.

To decipher the scattering pattern, algorithms may be obtained using two possible approaches.

A first approach is to implement a fitting algorithm, which requires large computational resource but yields accurate and stable outcomes.

A second approach is to use a characteristic parameter algorithm, which does not require extensive computation but the accuracy largely depends on the choice of characteristic parameters.

The use of a fitting algorithm will be discussed in more detail.

Figure 9 shows a 2D space of refractive index and size parameter, forming a grid of nodes. Each node is associated with a particular scattering pattern, of the type shown in Figure 7.

Plot 50 is the received pattern. To retrieve the particle size a and refractive index m, a fitting algorithm searches for that scattering pattern, in a pre-calculated database, which best matches the scattering pattern 50 obtained from the sensor. The database is stored in the sensor and for example contains the scattering pattern of particles with (a, m) pairs in size parameter range 1.0 to 50.0 in steps of 0.5 and refractive index range 1.01 to 1.99 in steps of 0.01.

In this example, 9801 scattering patterns are stored in this database. When a scattering pattern 50 is measured, the fitting algorithm searches through all patterns the database and determines the scattering pattern at location (a = 10, m = 1.54 in this example) as the best match.

With this information, the particle just captured can be determined as most probably having a size parameter of 10 and a refractive index of 1.54. If enough computation resources are available, a finer database with higher size parameter and refractive index resolutions can be generated for more accurate fitting algorithm calculation.

To test the algorithm stability, scattering patterns have been provided as a test procedure, with a random error within ±10% before being sent to the fitting algorithm.

Figure 10 shows the result of 2000 trials without the screen 25, and Figure 11 shows equivalent trials with the screen 25. Each figure shows the frequency density of calculated values as a function of the deviation from the true value. The top images show the deviation of the size parameter in histogram and scatter diagram form and the bottom images show the deviation of the refractive index in histogram and scatter diagram form.

It has been found that without the screen 98% of the size parameter retrievals and 95% of the refractive index retrievals fall within ±5% deviation of their true values. With the screen, 91% of the size parameter retrievals and 87% of the refractive index retrievals fall within ±5% deviation of their true values. Deviations of up to 0.5 are seen, this being the price paid for reducing the information input into the algorithm. Nonetheless, the fitting algorithm is both accurate and stable even with the screen. With the screen or an alternative form of spatial modulation, the system is less sensitive to vaiations in the speed of the particles.

Figure 12 shows a method of conducting the measurements.

In step 70, the removal duct valve is opened and the negative pressure system is turned on, with the delivery duct value closed. This drives air through the enclosure filter, providing clean air inside the enclosure. A period of time is allowed until the environment inside the enclosure becomes clean enough for testing.

In step 72 the delivery duct valve is opened for testing, and a measurement cycle is carried out.

In step 74, the delivery duct valve is closed so that the enclosure is refilled with clean filtered air.

In step 76 the removal duct valve is closed and the negative pressure system is turned off, which helps to maintain a clean environment inside the enclosure until the next operation.

The measurement cycle of step 72 comprises: step 72A of providing a collimated light output towards a flow channel using an illumination source; and

step 72B of focusing scattered light from a particle in the flow channel towards a light detector to provide a sweep of scattering angles of greater than 90 degrees to the light detector, wherein the light detector provides independent detection of the light intensity for each scattering angle within the sweep. The angles within the sweep are separated by 2 degrees or less.

In step 78, the light detector signals are processed to obtain a function of detected light with respect to scattering angle and this data is used to perform an inverse Mie calculation in step 80 to determine a particle size and particle refractive index.

As discussed above, embodiments make use of a controller for performing the data processing functions and the control functions. The controller can be implemented in numerous ways, with software and/or hardware, to perform the various functions required. A processor is one example of a controller which employs one or more microprocessors that may be programmed using software (e.g., microcode) to perform the required functions. A controller may however be implemented with or without employing a processor, and also may be implemented as a combination of dedicated hardware to perform some functions and a processor (e.g., one or more programmed microprocessors and associated circuitry) to perform other functions.

Examples of controller components that may be employed in various embodiments of the present disclosure include, but are not limited to, conventional microprocessors, application specific integrated circuits (ASICs), and field-programmable gate arrays (FPGAs).

In various implementations, a processor or controller may be associated with one or more storage media such as volatile and non-volatile computer memory such as RAM, PROM, EPROM, and EEPROM. The storage media may be encoded with one or more programs that, when executed on one or more processors and/or controllers, perform the required functions. Various storage media may be fixed within a processor or controller or may be transportable, such that the one or more programs stored thereon can be loaded into a processor or controller.

The invention is of interest for particle size and particle composition sensing, for example for pollen sensing (the pollen type and concentration) for virus or bacteria sensing (the type and concentration), for indoor/outdoor air pollution healthy risk assessment, or for indoor air quality management, for example by providing selective indoor air filtration/cleaning.

Other variations to the disclosed embodiments can be understood and effected by those skilled in the art in practicing the claimed invention, from a study of the drawings, the disclosure, and the appended claims. In the claims, the word "comprising" does not exclude other elements or steps, and the indefinite article "a" or "an" does not exclude a plurality. The mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measures cannot be used to advantage. Any reference signs in the claims should not be construed as limiting the scope.