Updated Design of the CMB Polarization Experiment Satellite LiteBIRD

Recent developments of transition-edge sensors (TESs), based on extensive experience in ground-based experiments, have been making the sensor techniques mature enough for their application on future satellite cosmic microwave background (CMB) polarization experiments. LiteBIRD is in the most advanced phase among such future satellites, targeting its launch in Japanese Fiscal Year 2027 (2027FY) with JAXA’s H3 rocket. It will accommodate more than 4000 TESs in focal planes of reflective low-frequency and refractive medium-and-high-frequency telescopes in order to detect a signature imprinted on the CMB by the primordial gravitational waves predicted in cosmic inflation. The total wide frequency coverage between 34 and 448 GHz enables us to extract such weak spiral polarization patterns through the precise subtraction of our Galaxy’s foreground emission by using spectral differences among CMB and foreground signals. Telescopes are cooled down to 5 K for suppressing thermal noise and contain polarization modulators with transmissive half-wave plates at individual apertures for separating sky polarization signals from artificial polarization and for mitigating from instrumental 1/f noise. Passive cooling by using V-grooves supports active cooling with mechanical coolers as well as adiabatic demagnetization refrigerators. Sky observations from the second Sun–Earth Lagrangian point, L2, are planned for 3 years. An international collaboration between Japan, the USA, Canada, and Europe is sharing various roles. In May 2019, the Institute of Space and Astronautical Science, JAXA, selected LiteBIRD as the strategic large mission No. 2.

as its origin. The LiteBIRD (Lite (Light) satellite for the studies of B-mode polarization and Inflation from cosmic background Radiation Detection) aims to detect signatures of these primordial gravitational waves in the form of specific curl patterns [4,5] of polarization angle distribution of the Cosmic Microwave Background (CMB). The sizes of expected spiral patterns in the sky are characterized by Hubble lengths at the electron scattering eras of the CMB, since primordial gravitational waves that enter the horizon in these eras most effectively produce tensor anisotropies. It is therefore essential to cover the whole sky to investigate these large Hubble lengths at the recombination era as well as the reionization era.
LiteBIRD is focused on this point: targeting [6] both the recombination era with the multipole moment l between 11 and 200 and the reionization era with l between 2 and 10, optimizing the angular resolution. The other important concepts of this satellite are a warm launch without the requirements of heavy vessels/tanks and use of multichroic detectors for the effective exploitation of finite focal-plane areas. Advantages of measurements from space are being free from atmospheric effects, providing high sensitivity, stability with less systematic uncertainties [e.g., 7], and no restrictions on observing band selection. Space measurements also give no pickup from the ground. The Sun-Earth L2 point has been selected, since the Sun, the Earth, and the Moon are all located in almost the same direction, which makes it easier to avoid facing them in terms of optical and thermal aspects. Care should be taken, however, on cosmic ray effects [8] because the satellite is more directly exposed to them. Sky observations are planned for 3 years: The presently guaranteed cooling-chain lifetime is 3.5 years, in which 0.5 year is assigned to the transitional period to the normal observation phase on course to L2. A scanning strategy with a combination of boresight spin angle of 50° around the satellite axis and its precession-like rotation around the anti-Sun direction of 45° is used (Fig. 1), since this combination provides not only a fairly uniform sky coverage but also the minimization of instrumental systematic uncertainties in polarization measurements.

Status of LiteBIRD
In May 2019, the Institute of Space and Astronautical Science (ISAS), JAXA, confirmed that LiteBIRD completed activities planned during Pre-phase A2 (previously called Phase-A1 and dedicated to the concept study and development of key technologies) and selected LiteBIRD as the strategic large mission No. 2, with modifications to the cooling chain and subsequently the focal-plane designs: We have chosen to use adiabatic demagnetization refrigerators (ADRs) in series, removing ideas [6,9] of using a JAXA-provided 1-K J-T cooler and of having 2-K telescope apertures. To recover the degraded sensitivity with 5-K apertures instead, we have increased the number of detectors. With the international collaboration between Japan, the USA, Canada, and Europe, the satellite is planned to be launched with JAXA's H3 rocket in 2027FY (Fig. 1). Tables 1 and 2 summarize the updated basic parameters and current baseline design.

Cooling Chain
With the design modification required, we have selected the following baseline combination for the cooling chain: (i) down to 5 K a sunshield and passive cooling with V-grooves [10], 15-K pulse tube coolers, and a 4-K J-T with 2ST precoolers; (ii) from 5 to 1.75 K a parallel three-stage ADR for providing continuous cooling at 1.75 K; and (iii) from 1.75 K to 100 mK a multistaged ADR with continuous cooling at 300 mK and 100 mK. Detailed schemes for this cooling chain are described in [11].

Telescopes and Polarization Modulator Units
The accurate, precise, and robust foreground cleaning requires 15 frequency bands covering a wide range of 34 to 448 GHz. This will be achieved with two kinds of telescopes (Table 2): a reflective one for the lower-frequency range (LFT) and refractive ones for medium-and high-frequency ranges (MHFT [12]). While the LFT uses crossed-Dragone reflective optics [13,14] to minimize effects from multireflection among refractive surfaces, the MHFT uses lens optics with the advantage of compactness. For the LFT, the F#3.0 and the crossing angle of 90° have been selected, since this combination suppresses the straylight [9,15] (Fig. 2), and aluminum is used as the mirror material [16]. All the telescopes are cooled down to 5 K for suppressing thermal noise. Each telescope has a polarization modulator unit (PMU) [12,[17][18][19], which consists of a transmissive half-wave plate (HWP) system rotated with a superconductive magnetic bearing (SMB), as the first optical element, close to the aperture stop to distinguish the sky and instrumental polarization components and to reduce the instrumental 1/f noise. The SMB enables contactless operation of the HWP system, which minimizes heat generation, through magnetic levitation by using YBCO bulk and permanent SmCo magnets. The LFT HWP system consists of multilayer stacked sapphire HWPs with optical axes shifted relative to each other to achieve high polarization efficiency over the wide frequency range [20], while the MHFT employs the metal-mesh HWPs [19]. At the top and bottom surfaces of the LFT HWP system, we adopt anti-reflection with sub-wavelength structure produced by laser machining. A cryogenic testbed is designed to study interactions, such as between spaceoptimized detectors and the PMU, consisting of a magnetically levitated and rotating HWP system [21].

Detectors and Readout
Transition-edge sensor (TES) bolometers are used to form multichroic pixels, based on maturity from successful ground-based and balloon experiments [22]. TES bolometers are coupled with a silicon lenslet and a sinuous antenna for individual broadband pixels for LFT and MFT, while silicon platelet-based corrugated horn and orthomode transducers are used for HFT [23,24]. The noise equivalent power of TES bolometer is proportional to √(P sat T b ), with the T c /T b ratio being optimized.
Here, P sat , T c , and T b are the saturated power, transition temperature, and thermal bath temperature, respectively. The TES parameter optimization for LiteBIRD, toward low-saturation power detectors for the satellite environment, has been carried out, as well as a sensor impedance that readily couples to the frequency-domain multiplexer. The electro-thermal time constants are controlled by slowing down with the additional heat capacity. Details of these experimental results are described in [25]. Cosmic ray mitigation methods have been developed by reducing propagation with palladium structures to absorb phonons and with the removal of bulk silicon to block phonons [26]. Readout will be carried out through digital frequency multiplexing, with 68 bolometers connected to each SQUID array amplifier [27,28]. The updated focal-plane designs and the expected sensitivities are shown in Fig. 2.

Foreground-Cleaning and Systematic Uncertainty Studies
The focal-plane designs have been determined through iterations with sensitivity calculations based on thermal studies [10], as well as with foreground-cleaning (by using methods described in [29][30][31][32]) and systematic uncertainty studies. The baseline foreground-cleaning results have been obtained based on an eight-dimensional parameterization of the foregrounds, including synchrotron power-law spectral index plus curvature and effective temperature of dust, plus power-law index of dust emissivity for Q and U Stokes parameters, each of which is allowed to vary across the sky. Systematic uncertainties have also been studied [33], including beam systematics [14], instrumental polarization and HWP harmonics, polarization efficiency, relative and absolute gain, pointing, polarization angle, time-correlated noise, cosmic ray glitches, bandpass mismatch [34], transfer function, nonlinearity, and non-uniformity in HWP, with realistic ground/in-flight calibration methods taken into account [e.g., 15,35]. Through all these studies, it has been shown that the Journal of Low Temperature Physics (2020) 199:1107-1117 updated focal-plane designs satisfy LiteBIRD's full success of the total uncertainty in the tensor-to-scalar ratio of less than 0.001, with its uncertainty budgets distributed comparably into a statistical part (including foreground residual and lensing B-mode), a systematic part, and a margin.