FIELD OF THE DISCLOSURE

[0001] The present disclosure relates to particulate matter sensors.

BACKGROUND

[0002] Airborne particulate matter can be generated, for example, by different forms of combustion, chemical processes, or mechanical wear. The size of the particles varies over a wide range, with some particles settling onto surfaces quickly in still air, whereas smaller particles may remain suspended for longer periods of time. Exposure to particulate matter can be harmful to human health. Further, some particulates act as abrasives or contaminates, and can interfere with the performance of equipment.

[0003] Some techniques for measuring the presence, amount and/or size of particulate matter in the air rely on optical techniques in which particles are illuminated with an optical signal and light scattered by the particles is detected.

SUMMARY

[0004] The present disclosure describes particulate matter sensor modules that operate based on sensing light scattered by the particulate matter. In applications such as smartphones and other portable computing devices, space is at a premium. In some instances, to help achieve compact particulate matter sensor modules, one or more metalenses are integrated into the particulate matter sensor.

[0005] According to a first aspect of the present invention, there is provided a particulate measurement device, comprising: a light source; a particulate-light interaction chamber; a light detector; an optical element; wherein, the optical element is arranged in a light path between the light source and the particulate-light interaction chamber and the optical element is operable to provide an asymmetric light intensity profile within the particulate-light interaction chamber, and the asymmetric intensity profile comprises a first part and a second part, the first part having a higher intensity than the second part, wherein the first part is further away from the light detector than the second part. The asymmetric profile may account for the intrinsic bias in conventional sensor modules, which are skewed by particulates close to the detector.

[0006] The optical element of the particulate measurement device may comprise a metamaterial, or a graduated density filter. These optical elements are configured to generate the compensated intensity profile.

[0007] The particulate measurement device may further comprise one or more apertures disposed in the light path.

[0008] The optical element (e.g., metamaterial or graduated density filter) may be disposed downstream or upstream of the one or more apertures.

[0009] The optical axis of the metamaterial may be substantially collinear with the light path.

[0010] The particulate measurement device may further comprise a reflective surface operable to redirect light emitted by the light source toward the particulate-light interaction chamber.

[0011] The metamaterial may be operable to function as a lens.

[0012] More specifically, the optical element (e.g., the metamaterial or graduated density filter) may be disposed between the reflective surface and the light source.

[0013] The particulate measurement device may further comprise a metalens disposed in the light path between the reflective surface and the light source.

[0014] The particulate measurement device may further comprise a further aperture, disposed in the light path. Optionally, the metalens being located downstream of this further aperture.

[0015] The light source may be a single VCSEL.

[0016] The metalens may be integrated with the light source (e.g., VCSEL).

[0017] The particulate measurement device may further comprise a fluid conduit operable to transport particulates into the particulate light-interaction chamber.

[0018] The particulate measurement device may further comprise a light trap configured to collect light exiting the particle-light interaction chamber along a further light path (an exit path from the particle-light interaction chamber). At least a first and a second aperture may be disposed between the particulate-interaction chamber and the light trap along this exit light path.

[0019] A further metalens may be disposed between the first and second aperture in the exit light path.

[0020] According to a second aspect of the invention, there is provided a mobile computing device, comprising the particulate measurement device as set out above, an application executable on the mobile computing device and operable to conduct air quality testing and a display screen operable to display a test result of the application.

BRIEF DESCRIPTION OF THE DRAWINGS

[0021] FIG. 1 shows a schematic diagram of a particulate matter sensor module.

[0022] FIG. 2 shows a schematic ray diagram of a particulate matter sensor module.

[0023] FIGs. 3A, 3B show a schematic diagram of a metalens.

[0024] FIG. 4 shows a schematic ray diagram of a particulate matter sensor module.

[0025] FIGs. 5A, 5B show a schematic diagram of the scattering of a particulate in a section of the sensor module.

[0026] FIG. 6 shows a schematic diagram of an intensity profile.

[0027] FIG. 7 shows a schematic diagram of a particulate matter sensor module.

[0028] FIG. 8 shows a schematic diagram of a particulate matter sensor module.

[0029] FIG. 9 shows a schematic diagram of a mobile computing device.

DETAILED DESCRIPTION OF THE DRAWINGS

[0030] As shown in FIG. 1, a particulate matter sensor module 20A includes a light source 22 (e.g., one or more vertical cavity surface emitting lasers (VCSELs); light emitting diodes (LEDs); or laser diodes) operable to emit light toward a reflective surface 28 (e.g., a mirror), which redirects the emitted light along a first path 30 through one or more light apertures 34A, 34B such that the light path 30 passes through a particle-light interaction chamber 40. Light emitted from the light source 22 can pass through one or more light apertures 34C before reflecting off the reflective surface 28. A lens, such as a metalens 100 can be disposed along the light path 30 between the light source 22 and the reflective surface 28. For example, immediately downstream of the aperture 34C. In some implementation, light apertures 34A, 34B, 34C have a width in the range 10 to 100pm. Fluid (e.g., an aerosol) is pumped through a fluid flow conduit 32, which can be substantially perpendicular to the light path 30. Thus, in the illustrated example, the light path 30 is in the x-direction, and the fluid flow conduit 32 is in the z-direction. As a fluid flows through the conduit 32, the light beam interacts, in the particle-light interaction chamber 40, with particulate matter in the fluid. The interaction scatters some of the light along a second path toward a light detector 24 (e.g., a photodiode) operable to detect the scattered light. In some implementations, a light pipe or other waveguide 42 (not illustrated) can be provided to guide the scattered light toward the light detector 24 and to reduce the effective distance from the particlelight interaction chamber 40 to the detector 24. Light that does not interact with the particular matter continues to travel along a third path 31 into a light trap chamber 36 to prevent such light from being reflected back toward the detector 24. In addition, some of the light scattered from the particulates in the interaction chamber 40 can be collected by the light trap chamber 36 provided the scattering angle, relative to the light path 31, is small (e.g., less than 20 degrees). The light source 22 and detector 24 can be mounted on, and electrically connected to a substrate 26 (e.g., a printed circuit board).

[0031] In some implementations, a second light detector 44 can be mounted on the substrate 26 which can be used to monitor the light power emitted from the light source 22. The second detector 44 can be placed, for example, next to the light source or below an aperture in the light trap chamber 36.

[0032] The detector 24 can be implemented, for example, as an optical photosensor that is operable to measure the signal of a single particulates. In such instances, the pulse height is proportional to the particle size and the pulse count rate corresponds to the number of detected particulates. The concentration can be derived, for example, from the number of detected particles, if the analysed volume is known. For example, from a fluid flow rate and measurement time. The mass can be calculated based on an assumed refractive index and density. The detector 24 can be integrated into a semiconductor chip that also may include electronics for reading, amplifying and processing the signals. In some cases, the processing circuity may reside in a separate chip. [0033] FIG. 2 shows a ray diagram, calculated using commercially-known software, of a particulate matter sensor module in use.

[0034] As illustrated in the FIG. 2, scattering occurs as light passes along the light path 30 from the light source 22 to the interaction chamber 40. As described above, in the interaction chamber 40, light is scattered toward a detector 24, which is operable to detect the scattered light. The signal the detector 24 measures is therefore light scattered from the particulates. FIG. 2 shows that a significant portion of light that reaches the detector 24 is scattered in regions of the sensor module, other than the interaction chamber 40. These portions of light do not originate from the particulates and therefore act as noise to the signal.

[0035] As the skilled reader will appreciate, there are a number of different elastic scattering mechanisms, such as Mie scattering, Rayleigh scattering, Raman scattering, geometric scattering and Thompson scattering, which can operate in the interaction chamber 40 and possibly elsewhere. Generally speaking, elastic scattering is strongest when the scattering centre is similar in magnitude to the wavelength of incident light. The dominant regime (i.e., the mechanism which dominants scattering) is therefore largely determined by these size factors, in particular whether the size of the scattering centres are smaller or larger than the wavelength of the incident light. That being said, as the skilled reader will appreciate, the dominant regime can be determined based on other parameters, such as the number concentration of the scattering centres and the like. As an example, scattering can arise from dust particles or other particulates.

[0036] Scattering can also arise through unwanted reflections. Referring back to FIG. 2, this type of scattering dominates the region of the sensor module comprising the reflective surface 28. One source of these reflections is that the width of light apertures 34A, 34B, 34C is smaller than the incident beam width. Accordingly, some light emitted from the light source 22 is incident on screens containing the light apertures 34A, 34B, 34C and reflected. To minimise this reflection, the screens can be designed to be light-absorbing in the wavelength range of the illumination light, but the absorbed light may subsequently be emitted by the screens in a different part of the spectrum. The light reflected from the light apertures 34A, 34B can then reflect off other surfaces, for example, the reflective surface 28. This process can repeat, causing significant scattering. As reflective surface 28 is designed to be highly reflective (i.e., with a reflection coefficient close to unity), any reflected light from light apertures 34A, 34B can be efficiently re-reflected. As shown in FIG. 2, a considerable number of rays are able to reach the detector via highly convoluted light paths. Despite the convoluted light paths, the scattered light can have appreciable intensity because the sensor module is small and the reflective surface 28 ensures efficient reflection. As detailed above, this unwanted scattered light that reaches the detector 24 causes noise, disrupting the signal of the detector. Further unwanted back-reflections can occur at the light trap chamber 36 and within the region containing the interaction chamber 40, as shown in FIG. 2. This arrangement therefore has relatively poor signal-to-noise ratio (SNR). The relatively poor SNR limits the resolution and accuracy of the sensor module.

[0037] In order to minimise this noise source, the light apertures 34A, 34B, 34C can be approximately 10 to 100pm in width to reduce the intake of scattered light into the interaction chamber 40. These light apertures 34A, 34B, 34C therefore prevent a significant proportion of power emitted from the light source 22 reaching the detector (the screens containing the apertures 34A, 34B, 34C absorb the light). In some cases, only around 15% of the power emitted by the light source 22 reaches the interaction chamber 40. An even smaller fraction of light is then scattered in the interaction chamber 40 along the second light path toward the detector 24. For this reason, conventional particulate sensor modules use multiple light sources 22 to compensate for this power loss. In some applications, such as smartphones and other portable computing devices (e.g., laptop computers, tablet computers, wearables, personal digital assistants (PDAs)), power and size reduction is desired and including multiple light sources 22 is sub-optimal. In addition, the screens, containing light apertures 34A, 34B, may require better thermal management systems in order to prevent overheating. Such system adds size and complexity to the sensor module, which is not desirable in some applications.

[0038] Therefore, in conventional sensor modules there is trade-off between using a large number of relatively large light apertures 34A, 34B, 34C with strong signal (but higher noise level) and using fewer relatively small light apertures 34A, 34B, 34C with weak signal (but lower noise level). Both of these options have their disadvantages. As the light paths are convoluted, it is difficult to determine an optimal aperture size. [0039] Turning back to FIG. 1, a metalens 100 is disposed downstream of the first light aperture 34C. In other implementations, the metalens can be disposed upstream of the first light aperture 34C, for example, placed directly onto the light source, or integrated with the aperture. The metalens 100 can be an integral part of the light source 22. The metalens 100 focuses or shapes the light beam emitted from the light source 22, thereby reducing the degree of scattering in the region comprising the reflective surface if the beam width of the light beam is smaller than the apertures. As the degree of scattering is reduced, fewer apertures 34A, 34B, 34C are required. In some implementations, the sensor module comprises one aperture 34A. The metalens 100 can shape light emitted from the light source 22 (e.g., a VCSEL) into a beam with smaller light beam width. The light beam width refers to the FWHM of a Gaussian beam.

[0040] In some implementations, the width of the light apertures 34A, 34B can be substantially equal to the light beam width, w. Reflection and/or absorption at apertures 34A, 34B can therefore be minimised, further improving the SNR. As the metalens 100 controls the light path, the aperture size can be optimised effectively.

[0041] An advantage of using a metalens 100 as opposed to a conventional lens is that the metalenses 100 can be fabricated much smaller. The thickness of a metalens can be in the order of 500nm, whereas conventional lenses are much thicker. The metalens 100 therefore allows miniaturisation of the sensor module, which is advantageous in some applications.

[0042] Whereas conventional lens shapes light by modulating the phase across the lens by, for example, continuous variation of the lens thickness, metalens 100 can achieve the same effect by spatially varying optical media properties, such as the refractive index, permittivity and/or permeability of the lens instead. All-dielectric metalenses are easier to fabricate and the remainder of this description refers to refractive index and permittivity only. As the skilled reader would appreciate, metalens comprising spatially varying permeability are also possible.

[0043] The field, which describes this behaviour, is known as “transformation optics” and the basis of the field is derived from the invariance of the Maxwell equations under coordinate transformations. A corollary of this invariance is that a medium of relative permittivity E and permeability p with a particular ray path can be transformed to any arbitrary ray path in another medium of relative permittivity E’ and permeability p’. The relationship between these optical propi according to transformation optics is given by Equation 1A.

Equation

Where J is the Jacobian matrix,

[0044] As the skilled reader would appreciate, the coordinate transformation (x, y, z) to (x’, y’, z’) can be chosen according to the operational requirements. In the technical field, a structure generated based on a spatially varying refractive index is known as a metamaterial. A metalens 100 is an example of a metamaterial.

[0045] Referring now to FIG. 3A and 3B, a metalens 100 can comprise pillars 302 or holes (not illustrated) etched into an amorphous silicon layer 304. The metalens 100 can be fabricated using standard semiconductor processing methods, as the skilled reader would appreciate. The pillars 302 are therefore “high” permittivity regions and the recesses surrounded by the pillars are “low” permittivity regions. In FIG. 3A, the pillars 302 are cylindrical, but they can also be hexagonal, rectangular or any other shape, as the skilled reader would envisage. The pillars 302 or holes form a patterned surface. The width of the pillars 302 and the spacing between them is smaller than the wavelength of incident light (the direction of which is indicated in FIG. 3B). The patterned surface can therefore be referred to as a sub-wavelength structure. As light passes through the metalens 100, the lens induces a phase delay (relative to, for example, the incident wave), which varies spatially across the sub-wavelength structure. In FIG. 3A, the induced phase delay is greatest in the centre of the metalens because the centre comprises mostly “high” permittivity pillar material. Conversely, the induced phase delay at the edges of the metalens is the lowest because the edges comprise mostly “low” permittivity air (absence of a pillar). As the patterned structure varies on a sub-wavelength scale, the structure generate/defines a grading in the induced phase delay between the centre and the edges. This grading can be designed to emulate a lens. The grading can be achieved by varying, for example, the size, shape, spacing and material of the sub-wavelength structure. There are therefore very many possible sub-wavelength structures which are capable of emulating a lens. The metalens 100 shown in FIG. 3A should not be viewed as limiting in any way. Commercially-available software, such as COMSOL and Zemax, are capable of generating such structures. For example, COMSOL may be used as a tool to simulate the phase delay induced by the subwavelength-structure (e.g., single pillars), and Zemax is used to select pillars into the lens.

[0046] In some implementations, the phase delay induced by the metalens 100 varies between TT and - TT.

[0047] In some implementations, the metalens 100 can be mounted on a glass substrate. The glass substrate can have a thickness in the range 100 to 400pm.

[0048] In some implementations, the micropillars can substantially comprise titanium dioxide or silicon dioxide etched in a silicon-containing substrate. The thickness of such a metalens 100 can be on the order of 500nm. The metalens 100 preferably, but not necessarily, exhibits transmission levels greater than 90%.

[0049] In some implementations, the metalens 100 can be integrated directly onto a VCSEL. As described above, metalens 100 can comprise silicon or silicon-compatible based materials. On the other hand, VCSELs typical comprise group lll-V materials. To simplify processing of the metalens, in some implementations, the metalens 100 can be fabricated from lll-V materials. Group lll-V materials generally exhibit poorer transmissive properties compared to silicon-based materials, which can reduce the power efficiency of the sensor module. This can set a practical limit on the thickness of a metalens fabricated from group lll-V materials. For this reason, the metalens 100 can comprise an additional layer (not shown) between the sub-wavelength structure and the substrate, which acts as a classical phase delay layer. The interface between the substrate and this additional layer can comprise steps. Preferably, this layer is highly transmissive and designed to induce a phase delay equal to, for instance, TT radians, depending on the height of the steps. Such conventional phase delay layers are well- known to the skilled reader. In this way, a metalens comprising a group lll-V material can be fabricated with high transmissivity (greater than 90%).

[0050] The effect of a metalens 100 disposed upstream of the first light aperture 34C is shown in FIG. 4. As shown, the majority of rays reaching the detector 24 originate from the shaped beam. Comparing the ray diagram to FIG. 2, the amount of scattered light reaching the detector is reduced significantly, thereby improving the SNR. A similar effect is seen when the metalens 100 is located downstream of the first light aperture 34C. In addition, as less light is scattered in the region comprising the reflective surface 28, the intensity of “unscattered” light reaching the interaction volume 40 increases, thereby improving the power efficiency of the sensor module. In some implementations, the metalens 100 improves power efficiency by 15 to 65%. In addition, less “scattered” light, from the region comprising the reflective surface 28, reaches the interaction volume 40 and the detector 24 because of this beam shaping. The SNR is therefore greatly improved, which, in turn, improves the resolution of the sensor module. The inventors have realised that this power efficiency increase means that a single VCSEL is sufficient to produce an appreciable SNR, instead of the multiple light sources required previously. The chip size of the sensor module can therefore be reduced and thermal management systems can also be simplified (cf one VCSEL to a plurality of VCSELs). This is especially advantageous in applications where size and power design limitations are important.

[0051] In other implementations, the metalens 100 can be replaced or used in combination with a micro-lens collimator. The micro-lens collimator can be, for example, placed upstream of the metalens 100 in the light path 30 to collimate the light before it passes through the metalens 100. In implementations of the sensor module which comprise both a metalens 100 and a microlens, the optical axis of the metalens 100 is preferably aligned with the collimated light. In some advantageous implementations, the metalens 100 is integrated directly onto a VCSEL and therefore controlled alignment of the metalens and/or the micro-lens collimator is not required.

[0052] Referring now to FIG. 5A, a columnar section of the sensor module, extending from the interaction chamber 40 to the detector 24, is shown. As detailed above, light emitted from the source 22 and reflected off surface 28 follow a light path along the x- direction into the interaction chamber 40. The light is then scattered by particulates in the interaction chamber onto the detector 24. As an approximation, each particulate 502 is assumed to scatter light in all directions equally. Therefore, when a particulate 502 scatters light, the particulate 502, in effect, acts like a point source radiating light radially outwards in all directions. This approximation is relatively accurate for Rayleigh scattering, but is less accurate for Mie scattering, which mainly comprises forward scattering. These forward scattering effects can however be taken into account, as the skilled reader would appreciate. In some implementations, the light path 30 is configured to be perpendicular to the surface normal of the detector in order to reduce the effects of forwardly scattered light reaching the detector.

[0053] In FIG. 5A, a selection of scattered rays originating from particulate 502, which reach the detector 24 are shown. In practice, the particulate 502 scatters light radially in all directions, but this is not shown for clarity. Therefore, only a fraction of the power of light scattered by the particulate 502 reaches the detector. The fraction is approximately equal to the ratio between the surface area of the detector 24 and the surface area of a sphere of radius, R, where R is the mean distance between the particulate 502 and the detector 24. Equivalently, the fraction is approximately equal to the ratio between the solid angle, Q, and 4TT. The solid angle, Q, for the detector 24 subtended at a point, defined by the position of the particulate 502 (xo,yo,zo), is equal to the solid angle of the projection of the surface S to the unit sphere centred on the particulate 502. The solid angle can be calculated according to the surface integral shown in Equation 2A.

Equation

Where r is the distance of the particulate to different spatial locations on the detector 24 and n is the surface normal of the detector.

[0054] Using the coordinate system defined in FIG. 1 and FIG. 5A, equation 2A can be written as Equation 2B. The expression can be calculated numerically.

Equation where a is the width and length of a planar square detector. As the skilled reader would appreciate, the detector 24 can be any shape and is not necessarily planar.

[0055] The results of this numerical calculation are shown in FIG. 5B for the plane labelled BB’ in FIG. 5A with a surface normal collinear to the light path 30 incident to the interaction chamber 40.

[0056] The results show that the solid angle Q (denoted as 52 on FIG. 5B) is small when the particulate 502 is far away from the detector 24 (larger y value). This corresponds to a relatively small fraction of power being delivered to the detector 24 because light is also scattered radially in all other directions. The intensity of scattered light, reaching the detector 24, originating from a particulate 502 further away from the detector 24, is therefore weaker. Likewise, the results show that the solid angle Q is large when the particulate 502 is close to the detector 24. This corresponds to a larger fraction of power being delivered to the detector 24 because a larger proportion of the light scattered radially from particulate 502 close to the detector 24, reaches the detector 24. The intensity of the scattered light, reaching the detector 24, originating from a particulate 502 close to the detector 24, is therefore stronger. The intensity of scattered light (i.e. , signal for the sensor module) reaching the detector 24 is therefore stronger for particulates closer to the detector.

[0057] Accordingly, in conventional systems, the detected signal is biased or skewed by particulates 502 which are proximal to the detector. This is a problem as the particulates 502 are not necessarily homogeneously dispersed in the interaction chamber 40. In practice, flow effects, buoyancy effects and/or gravitational effects cause variations in the size of particulates across the chamber 40. The accuracy of the measurement by the sensor module is then limited by the dispersion of the particulates 502 within the interaction chamber 40 which is difficult to control. Depending on the type of scattering and the type of measurement, the intensity of the light received by the detector may be used to estimate the size of the particulates, and a particulate close to the detector may incorrectly be estimated to be larger in size than a particulate further away from the detector due to the difference in solid angle.

[0058] The inventors have appreciated this problem and propose modulating the intensity of the incident light to compensate for this effect. Such a compensated intensity profile is shown in FIG. 6. The intensity of light is denoted as “62” in the Figure. The compensated intensity profile is asymmetric. In FIG. 6, the intensity of light incident to particulates 502 further away from the detector 24 is greater than the intensity of light incident to particulates 502 closer to the detector 24. This intensity profile can thereby account for the intrinsic bias that conventional sensor modules exhibit.

[0059] For design purposes, absolute values for the compensated intensity profile are not required. The intensity profile can for example be normalised and adjusted according to operational requirements. For example, if the signal from a particulate 502 decreases over distance according to an inverse square law, the intensity value in the compensated intensity profile quadruples as the distance from the detector doubles. As the degree of biasing from the particulates 502 is proportional to the calculated solid angle profile and absolute values of the intensity profile are not required, the compensated intensity profile can be determined by inversing this solid angle profile and optionally, normalising the resulting profile. For example, if the solid angle at yi is equal to TT and the solid angle at y2 is equal to TT/4, the compensated intensity values are 1 and 4. That is, the intensity of incident light along the light path (y= y2) is four times larger than the intensity of incident light along the light path (y=yi).

[0060] As mentioned above, scattering at particulates 502 is not necessarily isotropic and can, for example, exhibit forward biasing. When the light path 30 is perpendicular to the surface normal of the detector 24, the effects of forward bias are relatively weak because most of the forwardly biased light does not reach the detector. However, the effects of forward biasing on the fractional power delivered to the detector 24 from a particulate 502 can be modelled, as the skilled reader would appreciate.

[0061] In some implementations, the metalens 100 in FIG. 1 shapes or focuses light emitted from the light source and is operable to generate the intensity profile shown in FIG. 6. Such a metalens 100 can be referred to as a metamaterial and be located either downstream or upstream of apertures 34A, 34B, or on aperture 34C. In some examples, the metalens and/or metamaterial is a metasurface. References to metalens and metasurface should be construed accordingly. As detailed above, according to transformation optics, light can be shaped using sub-wavelength structures via grading a refractive index or relative permittivity across the metalens 100. In the same way, the metalens 100 can be designed to produce the compensated intensity profile in FIG. 6. As described above, as the skilled reader knows, there are many possible subwavelength structures that can achieve such an effect. Solutions can be calculated directly by applying a coordinate transformation according to Equation 1 A or by using commercial available software, such as Zemax. A desired output profile can be compared against a profile generated by an initial guess of a lens design in a cost function and the design parameters can be varied iteratively until the cost function has been minimised to an acceptable threshold. An exemplary workflow for fabricating a metalens is as follows:

Determine the metalens functional requirements; calculate the phase profile required to achieve the desired functionality; determine unit cell for the metalens, which discretize the calculated phase function; assemble the units into a layout to generate the discretized phase profile.

The steps may be iterated until the required functional requirements have been met to an acceptable limit.

[0062] An alternative arrangement for achieving the compensated intensity profile in FIG. 6 comprises a graduated neutral density filter which is placed in the light path 31. The filter is typically placed in the direction perpendicular to the main propagation direction, but can also be placed at an angle, for example to reduce back reflections. The light transmission gradually increases in one of the directions of the filter to modulate the intensity of the transmitted light accordingly.

[0063] A corresponding variation of the illumination light in the interaction chamber with the particulates provides an increase of illumination intensity in the direction away from the detector surface. The increase may be linear, which already provides an improvement over a homogeneous profile, or the increase as a function of distance to the detector surface may be inversely proportional to the solid angle as mentioned earlier.

[0064] The beam profile before metalens or neutral density filter is typically Gaussian, or could have a different standard beam profile, and the transmission profile of the metalens or filter is preferably adjusted accordingly such that the desired output profile is achieved. Given that each light source, or at least each type of light source such as a VCSEL with certain design parameters, has a unique output profile, the corresponding metalens or filter may also have a unique transmission profile.

[0065] In some implementations, the metalens 100 in FIG.1 can operate substantially as a lens and an additional metamaterial, operable to generate the compensated intensity profile in FIG. 6, can be located along the light path 30. The metamaterial can be located downstream or upstream of the apertures 34A, 34B.

[0066] In alternative implementations, a metamaterial operable to modulate light to the compensated intensity profile can be replaced with an array of microlenses and an array of light sources.

[0067] Referring to FIG. 7, in some implementations, the particulate sensor module 20B further comprises light apertures 34E, 34F, located along the light path from the interaction chamber 40 to the light trap chamber 36. These apertures 34E, 34F collect unscattered, or forwardly scattered light from the interaction chamber 40. The optical axis of apertures 34A, 34B, 34E, 34F can therefore be substantially collinear. Preferably, the screens containing the apertures 34E, 34F absorb light efficiently in order to minimise the intensity of light being reflected back into the interaction chamber 40, which would act as noise if it reached the detector 24. The first aperture 34E, located closer to the interaction chamber 40, is larger than the second aperture 34F in order to efficiently collect unscattered or weakly scattered light from the interaction chamber 40.

[0068] Preferably, the sensor module comprises a metalens 100, or equivalent lens structure which reduces scattering at apertures 34A, 34B, 34C. This reduction means that the apertures 34E, 34F can be smaller. This is advantageous because the light trap 36 can be reduced in size, thereby allowing miniaturisation of the sensor module.

[0069] In some implementations, as shown in FIG. 8, the particulate matter sensor module 20C includes a metalens 106 disposed in the light path 31 between the exit of the particle-light interaction chamber 40 and the entry of the light trap 36. The metalens 106 helps focus the light passing through the chamber 40 through a small aperture 34F provided in the light path 31 just after the metalens 106. The combination of the metalens 106 and the small aperture 34F can help reduce the amount of light that is reflected back onto the light detector 24. In some instances, the aperture 34F takes the form of a slit having a width in the range of 10 - 50 pm.

[0070] As shown in FIG. 9, a particulate matter sensor system 450 including a particulate matter sensor module (e.g., module 20A, 20B or 20C) can be incorporated into a mobile or handheld computing device 452, such as a smartphone (as shown), a tablet, or a wearable computing device. The particulate matter sensor system 450 can be operable by a user, e.g., under control of an application executing on the mobile computing device 452, to conduct air quality testing. A test result can be displayed on a display screen 454 of the mobile computing device 452, e.g., to provide substantially immediate feedback to the user about the quality of the air in the user’s environment.

[0071] The particulate matter sensor systems described here can also be incorporated into other devices, such as air purifiers or air conditioning units; or used for other applications such as automotive applications or industrial applications.

[0072] Various modifications will be readily apparent and can be made to the foregoing examples. Features described in connection with different embodiments may be incorporated into the same implementation in some cases, and various features described in connection with the foregoing examples may be omitted from some implementations. Thus, other implementations are within the scope of the claims.

[0073] Reference numerals should not be construed as limiting the scope. Features recited in separate dependent claims may be combined advantageously. References to “comprising” should not be viewed as excluding other elements or steps. References to “a” or “an” should not be viewed as excluding a plurality. A single processor or other unit can fulfil the functions of several means recited in the appended claims. LIST OF REFERENCE SIGNS

Particulate matter sensor module 20A, 20B, 20C

Light source 22

Light detector 24, 44

Substrate 26

Reflective surface 28

Light path 30, 31

Fluid flow conduit 32

Light apertures 34A, 34B, 34C, 34E,

34F

Light trap chamber 36

Particle-light interaction chamber 40

Waveguide 42

Solid angle 52

Intensity 62

Metalens 100, 106

Pillars 302

Amorphous silicon layer 304

Particulate matter sensor system 450

Handheld computing device 452

Display screen 454

Particulate 502