From CMB-S4 wiki
Back to Argonne 2018 main page
Parallel session P2-2: Cryostat and cold optics (chair: A. Vieregg)
Notes from session (Tuesday, March 6, 09:45-10:45)
- We won’t talk about telescope calibration today, but we will need to talk about it in future discussions.
- Cooling: Consensus is 100 mK DRs, because we can guarantee some amount of boresight rotation).
- We did not have consensus on how important boresight rotation is and how much we need. We have the technology to do it, but in Chile it’s a trade off between doing 360 and seeing Tau A. At Pole you might want to do 360 (unless you wanted to look lower in the sky for something else)
- Lenses: Baseline silicon because we are baselining an aperture diameter of less than 46 cm. If there is a compelling case for going larger than 46 cm, then baseline alumina. Alumina would stay on as an option in the appendix of the report (we want to keep developing this technology, and it may also need to be developed for an alumina filter). Bigger silicon would also be important to develop.
- There was a lens material discussion in the telescopes and mounts section. For the small aperture design we discussed adopting a BICEP3-like design. The aperture size is somewhat contingent on the choice of lens material. Silicon lenses work for less than 46 cm in diameter, and this is advanced technology that has been on the sky is dichroic receivers. We would need to scale up production to produce the number of lenses required for S4. It is easier to scale up the production of alumina lenses. However, we have an incomplete knowledge of cold alumina transmission at high frequencies. Available sapphire sizes limit HWPs to less than 51 cm, but we might not need them at lower frequencies or at all sites (see later discussion). Can baseline 46 cm diameters unless there is a strong reason to go larger. There may be a scientific driver.
- Prospects for larger silicon optics: it is not yet clear if the materials properties will be as good, so this would need more study.
- Production of silicon lenses could happen through a technology transfer to the national labs. Production for SO will be ~30 in about 1 year. Major investment would be needed to make this work for S4, but the problem seems manageable.
- Filters: Consensus is use a combination of metal mesh (hot-pressed low pass), alumina, and zotefoam stacks
- One plan that is easy to sell is to use scaled designs that have already been proven in the field. We can use ACT scaled for large apertures, and Bicep3 scaled for small apertures.
- Production of metal mesh filters is a concern, so we want to minimize the number of these if possible, but comes with the risk of large or unknown thermal load
- Alumina has the advantage of absorbing lots IR power at a given stage as possible, rather than reflecting it back into the cryostat through multiple bounces. S4 will have much more difficult requirements on cooling power than previous experiments, so we may need to take advantage of this.
- Spider’s experience is that with only shaders, the loads at 4 K and 40 K are not large, and so the in band instrument load is very low.
- The ACT design uses a combination of shaders and metal mesh (and the lenses are silicon so they don’t act as filters).
- BICEP-3 has eliminated all but one metal mesh filter and uses a zotefoam stack (which is fairly thick). The thickness may be challenging to fit into the optical design, but it is likely that the BICEP-3 stack thickness could be reduced by a factor of 2 because the spaces between the sheets are larger than they need to be. They saw much better performance when metal mesh filters were replaced with zotefoam (in-band transmission went from ~92% to 100%). At 280 GHz, the in band scattering is roughly 1% for a 5 inch thick window.
- These questions can be resolved with good measurements, and some experiments have data in hand that can be scaled (for instance, ACT).
- Windows: Consensus is polyethelene with hot pressed e-PTFE (multi-layer).
- Polarization modulators: Consensus is sapphire, because it has been demonstrated (and we are assuming a cold, continuously rotating HWP). Put them only in the small aperture telescopes in Chile, but the optical design at the Pole with allow them to be easily added.
- HWPs in large aperture telescopes cost you focal plane area (if you put it near the front of the cyrostat). You can still get atmospheric rejection if you put it at the Lyot stop, but it’s still costly because of the magnetic shielding.
- Is detector differencing good enough at the highest frequencies to not suffer a noise penalty without a HWP (particularly in Chile)? The loss in sensitivity from including the HWP might not make up for what you gain by adding the HWP. It might make sense to put modulators in only the high frequency receivers.
- For small aperture you can design the optics such that including the HWP is optional. This lets you decide to allow them later.
- The baseline number will affect budgeting because these are expensive
- In Chile do we need them at every frequency? Yes, for the baseline. If we decide to go bigger at lower frequencies, we can do 90 GHz and above instead.