R154 - NbTiN I-129 spectra, uniformity, pulse falltimes

Device info:

This note gives initial results for data taken in R154 with the following devices:

• Ch1: B101027.1, 50nm NbTiN, DMLE2. Illuminated with I-129 source. Glued to box with Ge varnish (8 mm^2 contact area). Tc ~ 0.65K.
• Ch2: Remounted same device as R153 from B110606.2, 75nm TiN, DMLE2. Illuminated with Cd-109 source. Glued to box with Ge varnish (8 mm^2 contact area). Tc ~2.4 K.

1. Pulse falltimes

The primary goal of this run was to see if the phonon lifetime increased for the device on Ch2 after remounting using Ge varnish instead of clamping with nylon washers. However, we don't see a significant change in the falltimes. Fig. 1 compares an averaged pulse for each of the mounting schemes.

Fig. 1 - Comparison of falltimes for different mounting

This indicates that mounting the device with washers last time was not the cause of the short falltimes. Instead, it seems that the phonon lifetimes are ~30 us or less, even when mounted with Ge varnish. For lower Tc devices, we have seen falltimes which are an order of magnitude larger (~200 us) for devices mounted with Ge varnish. Given the above results, it's likely that the quasiparticle lifetime dominates in the lower Tc devices, not the phonon lifetime.

Figs. 2a/b show the temperature dependence of the falltime for pulses from the I-129 source in the NbTiN device on Ch1. At 60mK, falltimes ~150 us are seen, which fall off rapidly with temperature.

Fig. 2a	Fig. 2b

Figs. 3a/b show the same thing for pulses from the Cd-109 source on the 2.4K TiN device on Ch2. The time constants are much shorter for the 2.4K film at base temp. The fall time is seen to fall off sharply above Tc/8.

Fig. 3a	Fig. 3b

Figs. 4-5 plot the livetimes (or inverse livetimes) vs. frequency shift (&delta f₀/f₀=(f₀(T)-f₀(0))/f₀(0)) or change in Q (&delta Q/Q = (Q(T)-Q(0))/Q(0)). As Sunil has pointed out, since &tau should be inversely proportional to the qp density, and f₀ and 1/Q are proportional to n_qp, we should see a linear relationship if we're seeing the quasiparticle lifetime. At high temperatures (leftmost points), we do see the expected linear relationship for both the NbTiN device and the TiN device. However, there is a change in slope near 130 mK (tau~80 us) for the NbTiN device.

Fig. 4a	Fig. 4b

Fig. 5a	Fig. 5b

2. NbTiN uniformity

Figs. 6-8 repeat the same plots as from the previous note on film uniformity, but adding the results for the 0.7K NbTiN film. The frequency shift vs. temperature for this device is more uniform than the comparable Tc TiN film, but less uniform than the higher Tc films. The frequency spacing is the best seen for any device, with all 20 resonators within 220 MHz of bandwidth (200 MHz design).

Fig. 6 - Comparison of frequency shift vs. temperature

Fig. 7 - Comparison of internal Qs

Fig. 8 - Comparison of array spacing

3. NbTiN spectrum

We were able to read out 17/20 resonators with the ROACH for this device. We lost one resonator due to a very low Qi, and two additional resonators to collisions in frequency. An initial analysis of the spectrum from the I-129 source is shown in Fig. 9. The &sigma~1.7 keV is very similar to the resolution measured for the other NbTiN device from this wafer in December. The better uniformity and larger number of resonators read out may make it easier to perform a position correction on this data.

Fig. 9 - I-129 spectrum