Optical spectrometers and Raman microscopes including such spectrometers
A CMOS sensor with rectangular pixels in multiple rows addresses the challenge of high spectral resolution and noise in Raman spectrometers by enabling fast readout and reduced noise, enhancing performance in Raman microscopy and spectropolarimetry.
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- HORIBA FRANCE SAS
- Filing Date
- 2024-03-28
- Publication Date
- 2026-04-14
AI Technical Summary
Existing optical spectrometers, particularly Raman spectrometers, face challenges in achieving high spectral resolution while minimizing measurement noise and readout time, often exacerbated by astigmatism and the use of square-pixel sensors that require summing values over multiple pixels, leading to increased noise and prolonged readout times.
The use of a CMOS sensor with rectangular pixels arranged in multiple rows, tilted at an angle less than 10 degrees relative to the spectral diffraction direction, along with an optical system that forms images on the sensor, allowing for simultaneous imaging of multiple spectra and reducing readout time and noise through column-parallel analog-to-digital conversion.
This configuration enhances spectral resolution, reduces measurement noise, and significantly shortens readout time, making it suitable for applications like Raman microscopy and spectropolarimetry, while maintaining a high signal-to-noise ratio and reducing thermal or mechanical drift.
Smart Images

Figure 2026512090000001_ABST
Abstract
Description
Technical Field
[0001] The present invention generally relates to an optical spectrometer.
[0002] More specifically, the present invention includes an aperture configured to receive a light source light beam, a first optical system, a diffraction grating, a second optical system, and an image sensor. The first optical system is arranged and configured to receive the light source light beam and form a collimated light beam directed toward the diffraction grating. The diffraction grating is positioned to receive the collimated light beam and is configured to form a diffracted light beam. The second optical system is arranged and configured to form an image of the diffracted light beam on the image sensor, and relates to an optical spectrometer.
[0003] The present invention is particularly advantageously applied in a Raman microscope.
[0004] The present invention also relates to a Raman microscope.
Background Art
[0005] In a spectrometer, more precisely a Raman spectrometer, spectral resolution and imaging resolution are important parameters to be optimized. In many cases, spectral resolution is often more emphasized than imaging resolution. As a result, due to an optical aberration known as astigmatism, the spectrum diffuses in height (i.e., in a direction perpendicular to the direction of spectral dispersion).
[0006] Next, the spectrum is composed of spectral lines, which are generally recorded by a matrix sensor of the CCD or CMOS type consisting of square pixels. Therefore, the spectral lines spread over several pixels, and different pixel values are summed to obtain the value of the spectral signal.
[0007] This technology is costly in terms of time (in a CCD sensor, the time required to read out multiple pixels is longer than for a single pixel) and noise (for example, in a CMOS sensor, read noise B is generated for each pixel, and the total noise associated with reading out N pixels is B*√N).
[0008] Other types of sensors also exist, such as CCD sensors with a single row of rectangular pixels. These sensors consist of pixels with a large surface area (making it possible to capture the entire height of the spectrum for all wavelengths within a single pixel), but they generate considerable measurement noise because the dark current is proportional to the surface area of the pixel. Furthermore, these sensors are not suitable for high-performance Raman spectroscopy. In fact, performing Raman spectroscopy, or more specifically Raman microscopy, may require simultaneously imaging multiple spectra across the height of the detector. [Overview of the project] [Problems that the invention aims to solve]
[0009] There is a need for an imaging optical spectrometer that offers both high spectral resolution and the ability to image one or more spectra simultaneously, while reducing measurement noise and limiting readout time. [Means for solving the problem]
[0010] To improve upon the aforementioned shortcomings of the latest technology, the present disclosure relates to an optical spectrometer comprising an aperture configured to receive a light source beam, a first optical system, a planar diffraction grating, a second optical system, and an image sensor, wherein the first optical system is positioned and configured to receive a light source beam and form a collimated light beam directed toward the diffraction grating, the diffraction grating is positioned to receive the collimated light beam and configured to form a diffracted light beam, and the second optical system is positioned and configured to form an image of the diffracted light beam on an image sensor.
[0011] According to the present invention, the image sensor is a CMOS sensor including pixels arranged in N rows, oriented in a direction tilted by up to an angle alpha with respect to the spectral diffraction direction of the image of the diffracted light beam, where the angle alpha is less than 10 degrees, N is an integer of 3 or more, each pixel has a height h and a width w, defining a pixel aspect ratio R = h / w, all pixels have the same height h and width w, and the pixel aspect ratio R is greater than 2.
[0012] This disclosure proposes the use of a CMOS sensor containing multiple rows of rectangular pixels.
[0013] Therefore, the use of a CMOS sensor reduces readout time. The aspect ratio is adjusted to reduce measurement noise. Furthermore, using at least three rows of pixels makes it possible to use the spectrometer for imaging applications (e.g., Raman microscopy) or for spectral polarimetry applications.
[0014] Preferably, the aspect ratio R of the pixels is 20 or less.
[0015] In one exemplary embodiment, each row of the sensor contains M pixels, where M is between 512 and 4096.
[0016] An advantage is that the pixels are placed in column M.
[0017] According to a particular embodiment, the image sensor includes an electronic system configured to collect and sum the values of pixels in the same column.
[0018] According to one embodiment, the number of rows N is 256 or less, and even 128.
[0019] In other specific embodiments, the pixel height h is 6 micrometers to 300 micrometers, and the pixel width w is 2 micrometers to 50 micrometers.
[0020] In one exemplary embodiment, the aperture is a rectangular entrance slit having a slit height Hf. In other embodiments, the aperture includes at least one circular or square aperture.
[0021] Optionally, the optical spectrometer includes an optical polarization splitter arranged and configured to split the diffracted light beam into two polarization beams, and the optical spectrometer is configured to simultaneously form an image of each of the two polarization beams on the image sensor, and an image of one of the two polarization beams is separated from an image of the other of the two polarization beams by at least one row of pixels.
[0022] The present invention also relates to a Raman microscope including an optical spectrometer according to the present disclosure.
[0023] Advantageously, the aperture includes at least two confocal apertures, and the optical spectrometer is configured to simultaneously form an image of each of the at least two confocal apertures on the image sensor, and each image of the confocal aperture is separated from the other image of the other confocal aperture by at least one row of pixels.
[0024] Obviously, different features, alternatives, and embodiments of the present invention can be combined with each other according to various combinations as long as they do not conflict with each other and are not exclusive.
[0025] The following description of the accompanying drawings is provided as a non-limiting example to enable a good understanding of what the present invention consists of and how it can be implemented.
Brief Description of the Drawings
[0026] [Figure 1] It is a schematic diagram of an optical spectrometer according to a first embodiment. [Figure 2] It is a schematic diagram of an optical spectrometer according to a second embodiment. [Figure 3] It is a schematic diagram of an optical spectrometer according to a third embodiment. [Figure 4]Schematic diagram of a detector of a spectrometer according to any one of the embodiments. [Figure 5] It is a figure of an example of a spectroscopic measurement image obtained by a spectrometer provided with a CMOS sensor according to the present disclosure. [Figure 6] The figure shows an aperture including three confocal apertures on the left side and a corresponding spectral image obtained by a spectrometer provided with a CMOS sensor according to the present disclosure on the right side. [Figure 7] Schematic diagram of a Raman microscope described by the present invention.
Mode for Carrying Out the Invention
[0027] In FIG. 1, an optical spectrometer 10 according to a first embodiment is shown. With the optical spectrometer 10, it is possible to generally analyze a light beam from a light source 12 which is generally external light.
[0028] The light source 12 can be obtained, for example, from the end of a cable including several overlapping optical fibers that provide light to the optical spectrometer 10.
[0029] The optical spectrometer 10 includes a casing 11 having an aperture 13 (or port hole) configured to receive a light beam. The aperture 13 can take the form of a rectangular or circular hole, or a slit. The aperture 13 has a height Hf in a direction perpendicular to the plane of FIG. 1. [[ID=2E]]
[0030] The light source 12 can be generated from any light source including part or all of the spectrum. Depending on the application, the light source 12 here emits light having a discrete or continuous optical spectrum ranging from ultraviolet to infrared (260 nm to 2 μm), for example.
[0031] Light from the light source 12 enters the casing in the form of an incident beam 16 that diverges from the incident point 14 to the first optical system 18. Here, the first optical system 18 is a concave collimating mirror with spherical curvature. As shown in Figure 1, the first optical system 18 redirects the light in the form of a collimated beam 20 to the diffraction grating 22.
[0032] The diffraction grating 22 is planar and is formed by straight, parallel, equally spaced lines 24. The lines of the diffraction grating 22 are perpendicular to the plane in Figure 1. The diffraction grating 22 is reflective in this case, receiving and reflecting the collimated light beam to form light beams 20 that are diffracted in different directions depending on the different wavelengths in the spectrum of the light source 12.
[0033] In Figure 1, schematically, each light ray incident on the diffraction grating 22 is dispersed into three rays, forming a spectrally diffracted beam. Naturally, diffraction depends on the light source 12 and is not limited to three wavelengths. After reflection from the diffraction grating 22, the diffracted beam 26 is collimated. The diffracted beam 26 is directed towards the second optical system 28.
[0034] The second optical system 28 is a concave focusing mirror. The second optical system 28 focuses the light beam 26 into an output beam 30, which is directed towards the image sensor 32. In other words, the second optical system 28 forms an image of the diffracted beam on the image sensor 32.
[0035] In the second embodiment of the optical spectrometer 10 shown in Figure 2, the diffraction grating 22 is a transmission grating. Furthermore, the first optical system 18 is a refractive optical system including, for example, a collimating lens, and the second optical system 28 is a refractive optical system including, for example, a lens for focusing the output beam 30 to the image sensor 32.
[0036] In the third embodiment shown in Figure 3, the optical spectrometer 10 is of the Czerny-Turner type. The diffraction grating 22 is a reflective grating. The first optical system 18 is a reflective optical system, and the second optical system 28 is also a reflective optical system. This configuration allows the optical path to be folded, resulting in a more compact spectrometer.
[0037] Of course, there are other known configurations of optical spectrometers, and these are also included in the scope of this disclosure. In particular, the optical spectrometer can be a non-planar configuration. The optical spectrometer 10 may also include a plurality of diffraction gratings 22 arranged in series along the path of the light beam in order to increase the spectral dispersion of the light beam.
[0038] As shown in Figures 1-3, each wavelength is focused to different image spots 31 along the image sensor 32 in the spectral diffraction direction. In the case of a planar optical spectrometer 10, the spectral diffraction direction lies in a plane perpendicular to the lines of the diffraction grating 22. Each image spot 31 is called a spectral line. The size of the image spot 31 on the image sensor 32 with respect to a particular wavelength of light, i.e., perpendicular to the spectral diffraction direction, depends on the size of the incident spot 14, as well as the ratios of the various shape dimensions of the optical components and their positions in the optical system. The width of the image spot in the spectral diffraction direction (which is the same direction as the length of the image sensor 32) determines the spatial resolution of the optical spectrometer 10. The optical design of the optical spectrometer 10 is generally defined to maximize the spectral resolution.
[0039] Furthermore, when the light beams 16, 20 and 26, 30 are reflected or transmitted by the curved off-axis optical systems 18, 28, the effective focal lengths of the optical systems 18, 28 in the "tangential" or "meridian" planes of Figures 1, 2, or 3 are shorter than the effective focal lengths of the optical systems in the "sagittal" plane perpendicular to the plane of the drawing. Therefore, when the image point is focused in the tangential plane, it cannot be perfectly focused onto the sagittal plane by the second optical system 28. As a result, the light beam 30 forms lines, rather than points, of light of each wavelength perpendicular to the plane of Figure 1 on the detector. This is known as astigmatism. The spectrum, including astigmatism, imaged on the image sensor 32 is shown in Figure 6.
[0040] In Figure 6, three light source points 15, 17, and 19 are shown, positioned on the entrance slit 13 of the optical spectrometer. The light source points 15, 17, and 19 also correspond to images of, for example, three optical fibers or spatially extended light sources. Each light source point 15, or 17, 19 generates spectra 55, or 57, 59 on the image sensor 32 of the optical spectrometer 10. For example, spectra 55, 57, and 59 are continuous spectra within the spectral domain under consideration. For each spectrum 55, 57, and 59, the spread of the image spot due to astigmatism in the direction across the spectral diffraction direction, i.e., in the direction of pixel height, is indicated by two dotted lines.
[0041] Astigmatism depends on several parameters, including wavelength. In other words, the height of the aberration, and therefore the height of the image on the sensor, depends on the wavelength. For example, in Figure 6, it can be seen that astigmatism is higher at both ends of each spectrum. The maximum height of the aberration, i.e., the spectral line, under the operating conditions of the optical spectrometer 10 (in particular, the wavelength range or the orientation of the grating) is called Hmax.
[0042] If the light rays exceed the height of the image sensor 32, excessive light is lost, and the sensitivity of the optical spectrometer 10 decreases. The image sensor 32 is selected in this case such that its total height H, defined by the maximum height Hmax, is higher than a threshold.
[0043] Figure 4 schematically shows the image sensor 32. Here, the image sensor 32 includes pixels 34 arranged in N rows, where N is an integer. N is 3 or greater. Preferably, N is 128 or less. The image sensor 32 has a height H and a width L. For example, the height H and width L are H=6mm and L=25mm. Typically, the height H is 1mm to 10mm and the length L is 6mm to 30mm.
[0044] The row of pixels in the image sensor 32 is oriented in a direction tilted by a maximum angle alpha with respect to the spectral diffraction direction of the image of the diffracted light beam, where angle alpha is less than 10 degrees, preferably less than 5 degrees. Advantageously, the row of pixels 34 is oriented parallel to the spectral diffraction direction of the image of the diffracted light beam, in other words, angle alpha is zero.
[0045] Preferably, each row contains M pixels 34, where M is between 512 and 4,096. Furthermore, the pixels 34 are generally arranged in columns, forming a matrix of the image sensor 32. Preferably, all pixels 34 of the image sensor 32 have the same width w and height h. For example, the height h of a pixel 34 is between 6 micrometers (μm) and 300 μm, and the width w of a pixel 34 is between 2 μm and 50 μm.
[0046] Each pixel 34 has a height h and a width w, defining the aspect ratio R = h / w of the pixel 34. The aspect ratio R of the pixel 34 is greater than 2.
[0047] The advantage of using such an image sensor 32 is that, unlike conventional sensors with square pixels, the aspect ratio of the pixels 34 allows each pixel 34 to receive the spectral lines 31. Consequently, the read noise associated with spectral measurement corresponds to the read noise of a single pixel 34.
[0048] In other words, because the height h of the pixel 34 is large, even though there is a large aberration in the vertical direction, the light from the spectral lines can be reliably measured in the height direction on the pixel 34, thereby reducing read noise.
[0049] Alternatively, spectral lines can be measured on two or three pixels in the same column by summing the intensities detected on those two or three pixels in the same row. The read noise associated with this sum for two or three pixels is still smaller than that of conventional detectors, which require summing at least twice as many pixels.
[0050] Since the spectral width of a spectral line is somewhat fine depending on the line being measured and the characteristics of the optical spectrometer 10, it is also possible to measure spectral lines using one or more pixels of the same line.
[0051] The image sensor 32 is a CMOS sensor in this case. The height of the pixels 34 should not be too large in order to allow efficient operation of the CMOS circuit. The aspect ratio R is less than 20 in this case. The height h of the pixels 34 and the aspect ratio R allow the image sensor 32 to maintain its performance in terms of signal-to-noise ratio.
[0052] Furthermore, CMOS sensor technology allows for the acquisition of 2D spectra several times faster than conventional CCD cameras with the same number of pixels (34). In practice, while it takes approximately 4 seconds to acquire an image on a CCD detector with 2048 x 2048 pixels (34), acquiring an image on a CMOS detector with the same number of pixels (34) takes only 20ms, virtually instantaneous.
[0053] A particularly advantageous feature is that the image sensor 32 includes a measurement system that includes a column-parallel analog-to-digital converter, meaning that the image sensor 32 includes one analog-to-digital converter (ADC) per column. Such a device makes it possible to read out all M pixels 34 in the same row simultaneously, reducing the readout time to approximately 1 / M of the original time.
[0054] Furthermore, to reduce data processing time, the measurement system can also sum the values of pixels 34 across a specific portion of several adjacent pixels 34 in the same column.
[0055] Such an image sensor 32 having rectangular pixels 34 with an aspect ratio of 2 to 20 therefore has improved speed and / or signal-to-noise ratio compared to an optical spectrometer 10 using a matrix sensor having square or low aspect ratio pixels 34.
[0056] CMOS detectors have comparable sensitivity to CCD detectors, and while CCD detectors themselves have higher electron noise than photomultiplier (PM) detectors, their quantum efficiency is far superior. Furthermore, the dynamic range of CMOS detectors is much smaller than that of PM detectors. Finally, CMOS detectors are less expensive than CCD detectors.
[0057] Furthermore, arranging the pixels 34 in multiple rows makes it possible to compensate for thermal or mechanical drift, which may be one of the performance limitations of the optical spectrometer 10. For example, in the case of an optical spectrometer 10 using several diffraction gratings, the orientation of the lines (which determines the direction of spectral diffraction) may differ between the gratings or change over time. Therefore, it is advantageous to have several rows of pixels so that the spectrum can be obtained by summing the pixels 34 according to a profile that is not precisely longitudinal. For example, the spectral lines can be tilted by an angle beta with respect to the longitudinal direction defined by the columns of the image sensor 32, where beta is less than 10 degrees (see Figure 6). This aspect of the present invention makes it possible to obtain superior spectral resolution compared to the use of a sensor having a single row of rectangular pixels 34.
[0058] Furthermore, it is advantageous to set the number of rows p of rectangular pixels 34 that are read out and summed to obtain each spectrum, where p is an integer greater than or equal to 1. In fact, depending on the wavelength and the configuration of the optical spectrometer 10, the height of the spectrum on the image sensor 32 can vary (see Figure 6). Then, of course, it is beneficial to take into account the total height of the p rows of pixels L5, or L7, L9, on the image sensor 32, on which the spectrum 55, or 57, 59, resides. It is also advantageous that from this sum, we can exclude the group of rows L6, L8, L10 of unilluminated pixels 34, on which only read noise may be added. This is similar to that of a conventional sensor with square pixels 34, but this operation has the advantage of using fewer rows p than conventional sensors, and therefore a better signal-to-noise ratio is obtained. Since the readout is very fast, it is possible to select the row L5, or the group of pixels L7, L9, that contain the spectral signal by the first measurement.
[0059] Arranging the pixels 34 as rows also has the advantage of allowing several spectra to be measured simultaneously in different rows of the image sensor 32, as shown in Figure 6. This application is particularly advantageous in spectropolarimetry, where different polarization components of different beams can be imaged into different rows of pixels. This application can also be used, for example, in a Raman microscope to form spectral images of different points of a sample under analysis into different rows of the optical spectrometer 10.
[0060] Figure 7 schematically shows a Raman microscope 100 including an optical spectrometer 10 according to, for example, a second embodiment.
[0061] The Raman microscope 100 includes a laser light source 44 that emits a laser beam. The laser beam is reflected by a first mirror 47 and then focused onto the sample 42 under study by a collimation lens 43. The sample 42 is excited by the laser and emits light, which then passes through the collimation lens 43 again to form a collimated beam.
[0062] The collimated beam is reflected by the second mirror 46 and focused onto the aperture 13 of the optical spectrometer 10. The aperture 13 of the optical spectrometer 10 here includes, for example, a confocal aperture.
[0063] The collimated beam passes through the first optical system 18. The first optical system 18 is, in this case, a collimation lens, which allows the light beam to be collimated after it has passed through the aperture 13 of the optical spectrometer 10.
[0064] The output beam of the first optical system 18 passes through a polarization splitter 41 configured to split the beam into two orthogonal polarization beams. The B-splitter 41 can be, for example, a Wollaston prism. Alternatively, the B-splitter 41 can be a semi-reflector or a Rochon prism.
[0065] Subsequently, the two polarized space-splitting beams pass through the diffraction grating 22. Here, the diffraction grating 22 is a transmission grating. The diffraction grating 22 can diffract the polarized beams.
[0066] The diffracted beam is focused onto the image sensor 32 by a second optical system 28 (not shown in Figure 7 for simplification). The second optical system 28 is here a spherical or slightly toric mirror. The image sensor 32 receives two polarized images from the aperture 13 of the optical spectrometer 10. The two polarized images are formed on different rows of pixels in the image sensor 32.
[0067] To easily separate two spectra, the optical spectrometer 10 is configured such that the two polarization images are separated by at least one row of pixels 34 on the image sensor 32. Preferably, the two polarization images can be separated by three rows of pixels 34 (each spectrum is surrounded by one row of black pixels 34, i.e., a row of pixels that do not receive light, to ensure that the entire spectrum is displayed, and by other rows of black pixels to ensure that the spectra are separated). The positions of the rows of illuminated pixels and rows of black pixels can be determined by prior measurements, and these basically depend on the optical design rather than the sample being considered.
[0068] Optionally, the Raman microscope 100 has the advantage of including a waveplate 48 positioned in front of the aperture 13 of the optical spectrometer 10 (in the direction of light) to rotate the polarization at the entrance of the optical spectrometer 10.
[0069] Figure 5 shows an example of spectral measurement of sample 42 using the Raman microscope 100 shown in Figure 3. Two polarization spectra 51 and 53 are observed, separated by a space 52 corresponding to one or more rows of pixels, for example, three rows of pixels. The polarization spectra 51 and 53 have the same spectral components, but each spectral component does not have the same intensity. Each spectrum 51 and 53 is surrounded by two black pixel spaces. The first spectrum 51 (top) is surrounded by two black pixel regions 50 and 52. The second spectrum 53 (bottom) is surrounded by two black pixel regions 52 and 54.
[0070] Alternatively, aperture 13 may include at least two confocal apertures. In this case, at least two beams from sample 42 can be observed simultaneously.
[0071] The optical spectrometer 10 is configured to image two confocal apertures onto the image sensor 32. The two images are separated by at least one row of pixels 34, preferably three rows of pixels 34.
[0072] Similarly, aperture 13 can contain more than two confocal apertures, for example, three, as illustrated in the example in Figure 6. The same theory applies to the number of confocal apertures that can exist.
[0073] When analyzing P separate incident light spots using the optical spectrometer 10, the image sensor 32 preferably includes at least 2P+1 rows of rectangular pixels 34, so that the spectral images of the P spots are sufficiently separated to avoid confusion, and the spectral images are surrounded by two black columns (a lower row of black pixels and an upper row of black pixels) so that the full height of the spectrum is clearly obtained.
[0074] The present invention is not limited in any way to the embodiments described and illustrated, and those skilled in the art will be able to see how to apply any of the modifications provided herein.
Claims
1. An optical spectrometer (10) includes at least one aperture (13) configured to receive a light source beam, a first optical system (18), a planar diffraction grating (22), a second optical system (28), and an image sensor (32), wherein the first optical system (18) is positioned and configured to receive the light source beam and form a collimated light beam directed toward the diffraction grating (22), the diffraction grating (22) is positioned to receive the collimated light beam and configured to form a diffracted light beam, and the second optical system (28) is positioned and configured to form an image of the diffracted light beam on the image sensor (32), The optical spectrometer (10) is characterized in that the image sensor (32) is a CMOS sensor including pixels (34) arranged in N rows oriented in a direction tilted by up to an angle alpha with respect to the spectral diffraction direction of the image of the diffracted light beam, the angle alpha is less than 10 degrees, N is an integer of 3 or more, each pixel (34) has a height h and a width w, a pixel aspect ratio R = h / w is defined, all pixels (34) have the same height h and width w, and the pixel aspect ratio R is greater than 2.
2. The optical spectrometer (10) according to claim 1, wherein the aspect ratio R of the aforementioned pixels is 20 or less.
3. The optical spectrometer (10) according to any one of claims 1 to 2, wherein each row of the image sensor (32) contains M pixels, and M is between 512 and 4096.
4. The optical spectrometer (10) according to claim 3, wherein the pixels (34) are arranged in column M.
5. The optical spectrometer (10) according to claim 4, wherein the image sensor (32) includes an electronic system configured to collect and sum the values of pixels (34) in the same row.
6. The optical spectrometer (10) according to any one of claims 1 to 5, wherein the number of rows N is 256 or less.
7. The optical spectrometer (10) according to any one of claims 1 to 6, wherein the height h of the pixel (34) is 6 micrometers to 300 micrometers, and the width w of the pixel (34) is 2 micrometers to 50 micrometers.
8. The optical spectrometer (10) according to any one of claims 1 to 7, wherein the at least one aperture (13) is a rectangular incident slit having a slit height (Hf), or the at least one aperture includes at least one circular or square aperture.
9. An optical spectrometer (10) according to any one of claims 1 to 8, comprising an optical polarization splitter (41) arranged and configured to split the diffracted light beam into two polarized beams, configured to simultaneously form an image of each of the two polarized beams on the image sensor (32), wherein the image of one of the two polarized beams is separated from the image of the other of the two polarized beams by at least one row of pixels.
10. A Raman microscope (100) comprising an optical spectrometer (10) according to any one of claims 1 to 9.
11. The Raman microscope (100) according to claim 10, wherein the at least one aperture (13) includes at least two confocal apertures (15, 17, 19), and the optical spectrometer (10) is configured to simultaneously form images (55, 57, 59) of each of the at least two confocal apertures (15, 17, 19) on the image sensor (32), and each image of a confocal aperture is separated from other images of other confocal apertures by at least one row of pixels.