Hello,

Thank you everyone for joining this webinar!

In this thread below, we will answer the questions we didn’t get to during the live event.

If you have any further questions, please ask!

@Andrew.Mckie @Ying.Wang 

Q: Will the recording be available later?

A: Yes, the webinar is available online from now on, please find the direct link below:
Simulating Infrared thermal detectors in OpticStudio – Zemax

@Lan.SUN 

Q: Is the macro (zpl) available?

A: In order to take advantage of the additional programming capabilities, instead of the ZPL macro language, we used the ZOS-API, more specifically an Interactive Extension from Python .NET, to automatically simulate the thermal calibration process. The python code is available as an article attachment in the following Knowledgebase article:
Infrared thermometer and thermal camera simulation using OpticStudio – Knowledgebase (zemax.com)

@Michael.Weber 

Q: Is it best to use the parallel rays and scattering surface than the Lambertian source under the DLL source?

A: The main purpose for launching parallel rays and using an additional scattering surface was to be able to apply Importance Sampling, which unfortunately cannot be directly applied on source objects. Even though Lambertian distribution can indeed be achieved with multiple source types, even only within a given cone angle by using for example the built-in Cone or Lambertian_overfill DLL sources, the disadvantage of these solutions is that the axis of the cone in which the rays are launched is always perpendicular to the source, so it is less efficient in cases when the source is not oriented directly towards the detector, e.g. a tilted on-axis source, or an off-axis source. In contrast, Importance Sampling remains efficient in all these cases by forcing the rays to propagate towards a desired object. For further information about Importance Sampling, please check this knowledgebase article:
How to use importance sampling to model scattering efficiently – Knowledgebase (zemax.com)

@Andrew.Carlson

Q: I heard your assumption regarding emissivity.  This is rarely the case.  Do you have techniques for when the emissivity < 1.0?  When you have multiple blackbodies (lenses?) that are < 1.0, it can be quite complicated to model. Even if we came up with a technique for dealing with emissivity < 1.0, I'm not sure how to validate such a model.  While it is pretty clear how to validate emissivity = 1.0 with examples in the literature, I can't think of any examples in the literature for emissivity < 1.0 where reflectivity > 0 and absorption > 0.  Can you?

A: Indeed, the literature is limited about how to take into account emissivity < 1.0. While one source could be modelled by an intensity filter to scale the emitted energy, even on a wavelength dependent basis, the validation of such a model would be challenging.
One example of how to measure and take emissivity into account is discussed in this article:
Calibration and Measurement Procedures for a High Magnification Thermal Camera (nist.gov)

@Andrew.Carlson

Q: Is there a Knowledge Base Article associated with this talk?

A: Yes, this knowledgebase article discusses the same topic:
Infrared thermometer and thermal camera simulation using OpticStudio – Knowledgebase (zemax.com)

@Andrew.Carlson

Q: I need to know more about Importance Sampling.

A: You may find more details about Importance Sampling, in this knowledgebase article:
How to use importance sampling to model scattering efficiently – Knowledgebase (zemax.com)

@Andrew.Carlson

Q: How many rays were traced in your model when you used Importance Sampling? How long did it take to run your thermal calibration using ZOS-API / Python?

A: For the thermometer example we traced 1E7 rays with Importance Sampling and Simple Ray Splitting for each of the 7 calibration temperatures and 1 additional verification temperature, and with this setup on my laptop with 6 cores (12 logical processors), the thermal calibration process completed in 10 mins. The temperature error in this case was less than 1 K.

For the thermal camera example the calibration process took much longer, a couple of hours, because we traced 1E8 rays in order to reduce the SNR of our detector with 50x50 pixels and because the curve fitting process was performed on a pixel-by-pixel basis.

Q: When going to higher wavelengths like 40 - 100 um. Is the calculation of spectral responsivity extrapolated or simply not possible?

A: When extending the blackbody source’s emission range further e.g. up to 100 microns, first we need to adjust the Maximum Wavelength in our material catalog to be able to trace those rays through the system. This means that the refractive index calculations are extrapolated based on the Sellmeier formulas up to the given Maximum Wavelength. Generally, care should be taken with this and the validity of this extrapolation should be verified, however in our model, as we filtered the results to be between 8 and 14 microns according to the spectral sensitivity of our detector, this did not cause any adverse effects besides enabling accurate thermal radiation modelling. The applied settings may be adjusted if necessary.

Q: How do we shield the system from IR stray light?

A: The design and efficiency of the cold shield is the key. You may find some advanced techniques about how to shield IR detectors from stray light in these articles:
Stray light in infrared detector (spiedigitallibrary.org)
pdf (iop.org)
Research of suppressing stray light of cold shield with different black coating (spiedigitallibrary.org)

Q: If I understood correctly: did you recommend to simulate the responsivity of the detector in the model by applying the filter on the captured rays on the detector?

A: Yes, we used the X_WAVERANGE(n, a, b) filter string to take into account the spectral responsivity of the detector by identifying rays that hit object #n within the target wavelength range (a, b). This is a simplified model, where the responsivity of the detector is 1 within the given range and 0 outside. You may find more details about how to apply filter stings in this knowledgebase article:
Identifying specific rays using filter strings – Knowledgebase (zemax.com)
An alternative solution can be to apply a TABLE coating on the detector and define the intensity transmission as a function of the wavelength, as well as the incident angle and polarization state if desired. With this solution we may take finer details of the spectral responsivity into account. Further details about the syntax and definition of TABLE coatings can be found in the Help files at:
The Libraries Tab > Coatings Group > Defining Coatings > The TABLE Data Section

Thank you so much for sharing this amazing work on simulating thermal detector. I have been trying to integrate my lumerical metasurface with zemax optistudio for infrared bolometers. Will definitely get to you with querries. 

@designerguy13-photonics 

Thanks for your follow-up message!

Integrating metasurface models from Lumerical to Zemax OpticStudio is an interesting topic. Lumerical’s FDTD simulation enables you to model the emissivity of your source objects based on Kirchhoff’s law which states that the absorptivity and emissivity are equal at equilibrium. Therefore, the emissivity of target objects can be modelled by measuring their absorptivity in Lumerical. For further discucssion, please take a look at this Lumerical knowledgebase article:
Thermal emission from a micro-hole array – Lumerical Support

Once the emissivity data is available from Lumerical, you can use the blackbody source model in OpticStudio and add a TABLE coating on an extra object right after the source to accurately model its thermal emission. This way thermal emission is taken into account by multiplying the emissivity of the object with the blackbody emission spectrum. 

Finally, you may also find these articles and webinars useful in general for Lumerical and Zemax OpticStudio integration and interoperability:
Export EM data to ray tracers – Lumerical Support
Zemax and Lumerical: Part 1 - from nano-scale to macro-scale optics and back - webinar – Knowledgebase
Zemax and Lumerical: Part 2 - from nano-scale to macro-scale optics and back - webinar – Knowledgebase

Q: What would be the best way to include the signal originating from camera's own thermal emission and its effect on temperature measurement accuracy ? 

A: The emissivity of the device can be estimated based on Kirchoff’s law, and then extra sources can be added to model the thermal radiation. Finally, Filter Strings can be used to identify specific ray paths to post-process the ray trace results and analyze the effect of self-emission, as well as stray light and ghosts in the system.

Further discussions and advanced modelling techniques can be found in these articles:
Stray Light and Thermal Self-Emission Minimization at the ELT (eso.org)
(17) (PDF) Antarctic Survey Telescope(AST3-3) NIR camera for the Kunlun Infrared Sky Survey (KISS): Thermal Optimization and System performance (researchgate.net)
Clever tricks in optical engineering (spiedigitallibrary.org)

Thanks for providing these references and for the interesting presentation.

In some of my optical simulation projects I also look on the effects of thermal radiation from for example a black body source on the sensors package/housing parts besides the absorbed irradiance of the sensors active area itself since the absorbed heat on the package level also might influence the overall temperature measurements. Here the “Object is A Detector”-feature for CAD-parts to read out the absorbed irradiance in combination with different face coatings and their respective absorption properties is my way of approach usually. Actually, the results from these optical simulations is then used by a colleague of mine with Ansys in order to simulate the thermal impact on the system level. Since Zemax is now part of Ansys, do you think that in the future optical and thermal simulations might be performed within a single simulation environment? That would be great for this kind of work-flow! 

Thank you mam for a quick reply, The resources are amazing. Also I checked in that source modelling for unpolarized blackbody radiiaion lumerical link is not working now even after official sign out.

@Sven.Stöttinger 

Thanks for your comments!

When simulating thermal effects, besides using the Object Is A Detector option to measure the absorbed flux on all optical as well as mechanical surfaces in the system, you can also add Detector Volume objects to record the absorbed flux inside the volumes of for example lenses. Further information about how to use Detector Volume objects, and visualize their measurements in OpticStudio can be found in this knowledgebase article:
How to show Detector Volume data in 3-D – Knowledgebase (zemax.com)

Then, as you said, these results can be imported into FEA software to simulate the thermal impact on the system. While unfortunately currently there is no single software solution for optical and thermal analyses, OpticStudio STAR Module provides the first step towards multi-physics simulations by enabling thermal as well as structural deformation data import from FEA packages directly into OpticStudio. This solution already simplifies and optimizes workflows between FEA packages and OpticStudio. If you would like to read more about the OpticStudio STAR module, please take a look at these articles:
OpticStudio STAR Module: Ansys Data Export Extension – Knowledgebase (zemax.com)
OpticStudio STAR Module - Data import and analysis tutorial – Knowledgebase (zemax.com)
The entire workflow is demonstrated on a high-power laser example in this whitepaper:
WP_Thermal_Lensing_Sept_2021.pdf (shopify.com)

@designerguy13-photonics

Thanks for your comment about the incorrect links in the Thermal emission article:
Thermal emission from a micro-hole array – Lumerical Support  
I have contacted my colleagues in the Lumerical team to update the links.
Apologies for the inconveniences.

Your welcome mam. I feel overwhelmed to be a part of this community.