Absolute Irradiance Measurement

[@jaydave5631]

Hi, Thanks for the video. After the calibration, how to start the measurement? Do we need to press absolute irradiance again and open it?

[@stephenlief846]

Hi Yvette! Once you take both a UV irradiance and then a VIS irradiance, can you combine them into one spectrum? We have the Ocean HDX.

[@stephenlief846]

We do have a combined calibration file covering 200-1100 nm.

[@ejipci]

I wouldn't call it a power meter (i.e., watts, flux, lumens) as much as you are measuring spectral irradiance [W/m2/lambda], power per area, spectrally. My small integrating sphere attachment is calibrated for power (W or lm).

[@RaphaelSilva-oq9kf]

Is it possible to measure solar absolute irradiance (field conditions) with the STS-VIS? If so, should all components shown in this video be used?

[@oceanoptics]

The new Ocean ST is a great option for solar irradiance measurements. Similar to the components shown in the video, these measurements would require an ST spectrometer, fiber, cosine corrector and OceanView software. Radiometric calibration of the system can be done by the Ocean Insight Calibration Lab or by the user with a radiometrically calibrated light source as shown in the video. Please reach out to [e-mail] to discuss your application and to receive a quote for a system to measure solar irradiance.

[@veronicasolis-y6e]

Hi! Is it possible to connect the spectrometer directly to the source? Thanks

[@oceanoptics]

You connect the light source to the spectrometer via a fiber. If you are asking about a direct attach light source we used to offer in the FLAME-CHEM systems., we still offer direct attach light source, but it only works with the FLAME-T and some other legacy spectrometers. Please reach out to our Technical Sales to discuss further.

[@arlinross1675]

Is an "intensity count" equivalent to uW/cm*2/nm? Also, how can I convert uW/cm*2/nm into more industry standard units such as lux or lumens? Does it just depend on the wavelength peaks?

[@oceanoptics]

In Ocean Insight OceanView software, QuickView mode is the graph view that allows users to see raw detector counts without any adjustments, calibration or compensation. While QuickView mode is useful for setting integration time and adjusting light levels, it is not a truly accurate representation of the incoming light because it does not take into account instrument response function, fixed pattern noise and baseline offset. Every Ocean Insight spectrometer has a unique instrument response function, or IRF. The IRF characterizes how the specific spectrometer responds to light across its wavelength range. This response is far from uniform: a spectrometer will produce a different response (here defined as the number of QuickView mode counts produced for a fixed number of photons) at every pixel. The IRF is non-uniform because of the cumulative effects of optical inefficiencies in the light path. These include, but are not limited to the following wavelength dependent effects: attenuation of light in the fiber optic cable; absorbance of light by the mirrors; grating efficiency; and detector response. The IRF for each spectrometer is unique, and cannot really be measured. However, it is possible to compensate for the IRF. The two common corrections are relative irradiance and absolute irradiance calculations. Irradiance is the measurement of radiant flux per unit area hitting or passing through a surface. An absolute irradiance measurement results in a spectrum that is accurate in both shape and magnitude. The y-axis scale becomes scaled in power or flux units like μW/nm or μW/cm2/nm, making it easy to calculate other power or energy values. Measurements in absolute irradiance mode require a calibration using a source with known power output. A spectrum is measured with the sampling optic (fiber tip, CC-3 cosine corrector, etc.) connected to the calibration light source, and is then compared to the known output power of the calibration light source. The calibration process generates a file with energy response data for each pixel in the CCD, given in μJ/count. Factoring in the surface area of the sampling optic and the integration time allows irradiance measurements in μW/cm2 to be reported (power=energy/time). Calibration is only possible if the absolute power output of the calibration light source is known, so if the light source calibration data is not given in the units μW/cm2/nm, it may not be a light source capable of calibrating for absolute irradiance measurements. Radiometric calibration can be done by the user with a radiometric calibration source like the one shown in the video or Ocean Insight can calibrate the spectrometer in our ISO/IEC 17025:2017-certified calibration lab using a NIST-traceable light source for which irradiance as a function of distance is known. This video walks the user through the steps required to radiometrically calibrate a spectrometer using one of Ocean Insight’s radiometric calibration sources. www.oceaninsight.com/products/light-sources/calibration-sources/radiometric-calibrated/

[@runrideski6003]

At about 3:00 min into the video, the setup parameters of integration time, scans to average and boxcar are discussed. Integration time is set first to maximize the tallest peak to "the blue line" (i.e., a little bit higher than 50,000 intensity counts). However, when boxcar is applied, this peak drops to ~20,000 intensity counts. Does this mean that the integration time could then be increased by ~2.5x to maximize S:N? Overall, this was a well-done video. I believe more information on how averaging & boxcar function, in the context of a calibration (i.e., CC3 in fixed position, stable light source), would be beneficial. Also some more info on the process of 'stitching-together' of the separate UV and Vis calibrations, which requires the selection an 'optimum' threshold wavelength for stitching. Finally, I think it should be more clearly stated that any absolute irradiance calibration is only valid for a given setup of the light path (e.g., slit, fibre, CC3 etc), and further is only valid while these components remain connected together - even if using the exact same components, if they are disassembled, the calibration must be redone. For my application, I use a small drop of (very carefully applied) nail polish or light-strength threadlocker on the male threads of the SMA connectors to ensure continuous integrity of the light path.

[@oceanoptics]

Thank you so much for taking the time to watch this video and provide your very valuable input! The tips you provided regarding absolute irradiance calibration clearly come from strong expertise in this area. You read our minds as we are currently working to create additional content around the topics you mentioned. For your question about increasing the integration time after applying the boxcar, I have asked the same question myself. We recommend setting the integration time before applying boxcar smoothing to the spectral data. When the boxcar is applied, even though it appears that we now have less intensity at the ~656 nm D-alpha deuterium line in this case, the light intensity hitting the detector has not changed. The signal levels only appear lower because we have applied a boxcar value that averages pixels with much lower intensity on either side of the intense D-alpha line. Even in the case where the detector behaves well with some saturation (very detector dependent), this saturation would still show up in the data as extreme non-linear behavior making the data in the region where saturation occurs inaccurate and not meaningful. If it is determined through testing that some level of saturation can be tolerated, the integration time could be increased as you described to improve SNR but the data at the saturated wavelengths would still inaccurate and not meaningful.

[@runrideski6003]

@@oceanoptics Thanks for your detailed response. I have 2 followup thoughts: 1) when I have used this calibration lamp spectrum for absolute irradiance calibration, I believe only the part of the spectrum that is below 400nm is 'kept' after stitching the UV calibration together with the Vis calibration. In this case, how much would it really matter if the detector was saturated at (some) wavelengths that are above 400 nm? 2) I always figured that it was important to NOT use Boxcar and (maybe less-so) scans to average in the calibration process. Given it uses a specific calibration radiation source, shouldn't it inherently be VERY stable in terms of spectral intensity? Or are the averaging functions only used to moderate 'noise' on the acquisition side of the process (i.e., the light path and internal electronics) in this context?