We have measured and characterized the thermal crosstalk in two different arrays of transition-edge sensor (TES) X-ray microcalorimeters with frequency-domain multiplexing (FDM) readout. The TES arrays are fabricated at SRON and are a 8 × 8 and a 32 × 32 'kilo-pixel' uniform array. The amount of crosstalk is evaluated as the ratio between the averaged crosstalk signal and the X-ray pulse amplitudes. The crosstalk ratios (CRs) for our detectors are compliant with the requirements for future X-ray space missions, such as Athena X-IFU (CR < 10^{-3 for first-neighbor pixels): We measured a nearest neighbor thermal CR at a level of 10^{-4, with a highest CR of 4× 10^{-4 for the kilo-pixel array (worst case, center of array) and 1× 10^{-4 for the 8 × 8 array, with a margin of improvement achievable by optimizing the Cu metallization and the width of the Si supporting structures (muntins) in the backside of the TES array chip. Based on the measured CRs, we have estimated the impact on the spectral resolution by means of noise equivalent power considerations and a Monte Carlo simulation, finding an average degradation in quadrature of less than 40 meV, compliant with the <0.2 eV requirement for Athena X-IFU.

We are developing a kilo-pixels Ti/Au TES array as a backup option for Athena X-IFU. Here we report on single-pixel performance of a 32 × 32 array operated in a Frequency Division Multiplexing (FDM) readout system, with bias frequencies in the range 1-5 MHz. We have tested the pixels response at several photon energies, by means of a 55Fe radioactive source (emitting Mn-Kα at 5.9 keV) and a Modulated X-ray Source (MXS, providing Cr-Kα at 5.4 keV and Cu-Kα at 8.0 keV). First, we report the procedure used to perform the detector energy scale calibration, usually achieving a calibration accuracy better than ∼0.5 eV in the 5.4-8.9 keV energy range. Then, we present the measured energy resolution at the different energies (best single pixel performance: ΔEFWHM = 2.40 ± 0.09 eV @ 5.4 keV; 2.53 ± 0.10 eV @ 5.9 keV; 2.78 ± 0.16 eV @ 8.0 keV), investigating also the performance dependency from the pixel bias frequency and the count rate. Thanks to long background measurements (∼1 d), we finally detected also the Al-Kα line at 1.5 keV, generated by fluorescence inside the experimental setup. We analyzed this line to obtain a first assessment of the single-pixel performance also at low energy (ΔEFWHM = 1.91 eV ± 0.21 eV @ 1.5 keV), and to evaluate the linearity of the detector response in a large energy band (1.5-8.9 keV).