文章详情
Comparison of the Arcoptix’s ARCSpectro HT with a standard grating spectrometer
日期：2025-04-28 12:58
浏览次数：1666
摘要：
北京锦坤科技有限公司
www.jon-kon.com
In this document we present a comparison of our high throughput ARCSpectro HT Fourier Transform Spectrometer and a standard grating spectrometer with similar resolution and spectral range when measurin g diffusing light sources. It is impossible to find a grating spectrometer with exactly the same performances than the ARCSpectroHT (because of the difference of its working principle). However we decided to choose for the comparison a very popular grating spectrometer produced by a well established manufacturer. It ha s comparable detector (and read outelectronic) quality (dark noise and saturation level), a little weaker resolution (slit of 25 μm), a little smaller spectral range, similar size, both USB read out electronic. Despite of the inherent trade-offs between resolution and spectral range of the grating spectrometers, we estimate the chosen devic(600 grooves) is the closest to our ARCSpectro HT-HR.
Of course the big difference between both spectrometers (which explains the measurements shown in the figures below) is the entrance aperture of 4mm of the ARCSpectro HT which permits to collect 50 times more light than its grating homologue! Notice that similar results can be obtained with the ARCSpectroHT-2D version which has a better SNR than the HT-HR version but lower resolution.
The following table summarizes the main performances of both spectrometers:
Optical setup
As shown in figure 1, we build a very simple setup constituted of a collimated halogen light followed by a filter (interference filter at 630nm or a broadband green filter). The light traverse a 1 cm thick cuvette containing water with a few drops of cream diluted in it. This mixture can be considered optically as a strong volume scatterer. The light is then focalized at his best with two lenses (but this does not help much because we are in the case of a extended diffusing source) on a measuring spot. In the measuring spot we put:
1) A 1 meter long 600 μm core fiber (with minimal light losses) connected to the USB.
2) Directly the entrance of the ARCSpectro HT.
In both case we tried to optimize fiber position (or entrance position) and lens position for optimal coupling intensity. Notice that is possible (if necessary because lack of space) to use a 4mm light guide to bring the light to the entrance of the ARCSpectro HT, this would deteriorate the
measurement shown in this document by about 40%.
We made three series of measurements:
1) with a interference filter at 630nm placed after the halogen lamp
2) with a broadband filter placed after the halogen lamp.
3) We replaced the halogen lamp with HgAr lamp (supplied with a 9V battery)
HQ-1 from Ocean Optics.
Figure 1: Filtered collimate d Halo gen la mp shines o n a volume scatterer (milky water) the scattered light is colle cted on a meas ur ement spot. 1) the light is collected via a 600mm op tical fiber to the grating sp ectrometer. 2) Th e ARCSpectro HT isdire ctly placed on the measur ement s pot.
Results
The figures below compares the spectra measured with the ARCSpectro HT and with t h e USB 2000 in strict ly similar measuremen t condit ion s described in f igure 1. We treated her e the case o f t h e spectr al analysis of a highly sca ttering sources. Th is is of course a pa rticular (but important) case where the ARCSepctro is clearly superior to a grating spectrometer (as one can see in th e figu res belo w ) . Th is is simply explained by its higher so called “étendue” (or throughput) tha t i s abou t 100 x larger than for a comparable grating spe ctrometer. In the ca se where only ver y few lights enters the grat ing spectr ometer, it is limited by its dark noise (signal is weaker than dark noise) and no useful spectrum can be obt a ined even if increasing the integr ation time (as we can see in figure 4). Indeed (in the low ligh t case) increasin g th e in t egration increases the dark noise as much as the signal so the SN R stays below 1 and no thing can be measured. On th e other hand t h e ARCSpectr o HT c o llec ts 50X more light and is no t necess aril y (for t he same setup) dark noise limit ed and has a much better SNR.
The measurements in the figur es bel o w demonstrate ver y clearl y (with completely di fferent types of li ght s o urces) how useful it ca n be to have high throughput spectrometer if de aling with extended diffusing light sources.
Figure 2: Measurement of the spectrum of the light emitted by a diffusing milky diffusing mixture illuminated with a halogen lamp filtered with 630nm interference filter. A bove the spectrum is measured with Arcoptix’s Arcspectro HT and below the spectrum measured with the garting spectrometer. Both spectra have been measured in strictly identical conditions as described in figure 1 and with an identical integration time of 10ms (no averaging or boxcar averaging).
Figure 3: Measurement of the spectrum of the light emitted by a diffusing milky diffusing mixture illuminated with a halogen lamp filtered with a broad band green filter which transmits the green light and the NIR spectral region. Above the spectrum is measured with the Arcspectro HT and below the spectrum is measured with the grating spectrometer. Both spectra have been measured in strictly identical cond itions as described in figure 1 and with an identical integration time of 10ms. A Boxcar of 5 pixels averaging (5 neighbor pixel averaging) is performed on the grating spectrometer spectrum and a Gaussian apodization of 400 pixels (in the interferogram) has been performed in the first measured spectrum (Fourier transform equivalent of boxcar averaging) .
Figure 4: Measurement of the spectrum of the light emitted by a diffusing milky diffusing mixture illuminated with a HgAr calibration HG-1 from oc ean optics. Above the spectrum is measured with the ARCSpectro HT and below the spectrum is measured with the grating spectrometer. Both spectrum have been measured in strictly identical conditions as described in figure 1 but with different integration times (200ms for the ARCSpectro HT and 800ms for the grating spectrometer). The plotted spectra is the result s of 4 averaged spectra (measured consecutively in time). Notice that the grating spectromet er gives a useless spectrum even for a higher integration time.
北京锦坤科技有限公司
www.jon-kon.com
电话：010-62530053,52494037
邮箱：jon_kon@163.com
www.jon-kon.com
In this document we present a comparison of our high throughput ARCSpectro HT Fourier Transform Spectrometer and a standard grating spectrometer with similar resolution and spectral range when measurin g diffusing light sources. It is impossible to find a grating spectrometer with exactly the same performances than the ARCSpectroHT (because of the difference of its working principle). However we decided to choose for the comparison a very popular grating spectrometer produced by a well established manufacturer. It ha s comparable detector (and read outelectronic) quality (dark noise and saturation level), a little weaker resolution (slit of 25 μm), a little smaller spectral range, similar size, both USB read out electronic. Despite of the inherent trade-offs between resolution and spectral range of the grating spectrometers, we estimate the chosen devic(600 grooves) is the closest to our ARCSpectro HT-HR.
Of course the big difference between both spectrometers (which explains the measurements shown in the figures below) is the entrance aperture of 4mm of the ARCSpectro HT which permits to collect 50 times more light than its grating homologue! Notice that similar results can be obtained with the ARCSpectroHT-2D version which has a better SNR than the HT-HR version but lower resolution.
The following table summarizes the main performances of both spectrometers:
Optical setup
As shown in figure 1, we build a very simple setup constituted of a collimated halogen light followed by a filter (interference filter at 630nm or a broadband green filter). The light traverse a 1 cm thick cuvette containing water with a few drops of cream diluted in it. This mixture can be considered optically as a strong volume scatterer. The light is then focalized at his best with two lenses (but this does not help much because we are in the case of a extended diffusing source) on a measuring spot. In the measuring spot we put:
1) A 1 meter long 600 μm core fiber (with minimal light losses) connected to the USB.
2) Directly the entrance of the ARCSpectro HT.
In both case we tried to optimize fiber position (or entrance position) and lens position for optimal coupling intensity. Notice that is possible (if necessary because lack of space) to use a 4mm light guide to bring the light to the entrance of the ARCSpectro HT, this would deteriorate the
measurement shown in this document by about 40%.
We made three series of measurements:
1) with a interference filter at 630nm placed after the halogen lamp
2) with a broadband filter placed after the halogen lamp.
3) We replaced the halogen lamp with HgAr lamp (supplied with a 9V battery)
HQ-1 from Ocean Optics.
Figure 1: Filtered collimate d Halo gen la mp shines o n a volume scatterer (milky water) the scattered light is colle cted on a meas ur ement spot. 1) the light is collected via a 600mm op tical fiber to the grating sp ectrometer. 2) Th e ARCSpectro HT isdire ctly placed on the measur ement s pot.
Results
The figures below compares the spectra measured with the ARCSpectro HT and with t h e USB 2000 in strict ly similar measuremen t condit ion s described in f igure 1. We treated her e the case o f t h e spectr al analysis of a highly sca ttering sources. Th is is of course a pa rticular (but important) case where the ARCSepctro is clearly superior to a grating spectrometer (as one can see in th e figu res belo w ) . Th is is simply explained by its higher so called “étendue” (or throughput) tha t i s abou t 100 x larger than for a comparable grating spe ctrometer. In the ca se where only ver y few lights enters the grat ing spectr ometer, it is limited by its dark noise (signal is weaker than dark noise) and no useful spectrum can be obt a ined even if increasing the integr ation time (as we can see in figure 4). Indeed (in the low ligh t case) increasin g th e in t egration increases the dark noise as much as the signal so the SN R stays below 1 and no thing can be measured. On th e other hand t h e ARCSpectr o HT c o llec ts 50X more light and is no t necess aril y (for t he same setup) dark noise limit ed and has a much better SNR.
The measurements in the figur es bel o w demonstrate ver y clearl y (with completely di fferent types of li ght s o urces) how useful it ca n be to have high throughput spectrometer if de aling with extended diffusing light sources.
Figure 2: Measurement of the spectrum of the light emitted by a diffusing milky diffusing mixture illuminated with a halogen lamp filtered with 630nm interference filter. A bove the spectrum is measured with Arcoptix’s Arcspectro HT and below the spectrum measured with the garting spectrometer. Both spectra have been measured in strictly identical conditions as described in figure 1 and with an identical integration time of 10ms (no averaging or boxcar averaging).
Figure 3: Measurement of the spectrum of the light emitted by a diffusing milky diffusing mixture illuminated with a halogen lamp filtered with a broad band green filter which transmits the green light and the NIR spectral region. Above the spectrum is measured with the Arcspectro HT and below the spectrum is measured with the grating spectrometer. Both spectra have been measured in strictly identical cond itions as described in figure 1 and with an identical integration time of 10ms. A Boxcar of 5 pixels averaging (5 neighbor pixel averaging) is performed on the grating spectrometer spectrum and a Gaussian apodization of 400 pixels (in the interferogram) has been performed in the first measured spectrum (Fourier transform equivalent of boxcar averaging) .
Figure 4: Measurement of the spectrum of the light emitted by a diffusing milky diffusing mixture illuminated with a HgAr calibration HG-1 from oc ean optics. Above the spectrum is measured with the ARCSpectro HT and below the spectrum is measured with the grating spectrometer. Both spectrum have been measured in strictly identical conditions as described in figure 1 but with different integration times (200ms for the ARCSpectro HT and 800ms for the grating spectrometer). The plotted spectra is the result s of 4 averaged spectra (measured consecutively in time). Notice that the grating spectromet er gives a useless spectrum even for a higher integration time.
北京锦坤科技有限公司
www.jon-kon.com
电话：010-62530053,52494037
邮箱：jon_kon@163.com
