Ph.D. Thesis - Physics

choices and heat sinking. The wire used for the dc connections and a portion of the rf
line is a 36-gauge phosphor bronze wire, LakeShore Part No. WSL-36-500. The constant
of thermal conductivity is 48 W/(K·m). For a temperature gradient of 294 K across a
wire length of 10 cm, the thermal conductivity is 2 mW. Since this is much less than
the cooling power at 4 K and 40 K (0.8 W and 35 W, respectively), heat gain due to
conduction should be negligible. Actually, the figure is better than this, since the dc wires
were wrapped repeatedly around the 40 K shield (increasing the length and heat-sinking)
before being taped down to the 40 K baffle and then heat-sunk to the 4 K shield using
StyCast thermally-conductive epoxy.
The rf wiring was purposefully made shorter than the dc wires to reduce stray ca-
pacitance. From the exterior feedthrough, a thin wire (to break thermal conductivity) was
taped to the 40 K baffle, then attached to a thicker wire (to increase electrical conductivity),
before being soldered to a thin wire which was heat-sunk with StyCast to the 4 K base-
plate and routed to the trap. This combination was designed to provide as much electrical
conductivity as possible while still breaking the thermal connections.
Upon finishing work inside the cryostat, the system was put under turbo pump vacuum,
leading to ultimate pressure in the high 10 −6 torr range. A 50 Ω heater is installed on the
cold head, and this was engaged to bring the system to a temperature of 380 K for around
24 hours, allowing some oils, water, and other residues to bake off and be removed from the
system. This is done partially to improve the ultimate pressure, and partially to remove
substances which might freeze onto the trap electrodes at 4 K. A pair of charcoal getters
were also installed on the 40 K shield, and were discharged during this time with a wire
whose resistance totaled 150 Ω. A current of 0.06 A through the getters was used.
Finally, a 20 l/s ion pump was installed directly on the outer octagon. This was turned
on before cooldown. There are two main advantages to this. The first is that the system is
completely closed from the “outside world” during cooldown. This prevents cryopumping
from actually pumping material from the turbo pump line into the cryostat. Furthermore,
in case of a malfunction of the turbo or roughing pumps, it prevents oils from being cryop-
umped into the system. The other advantage is that reading the ion pump current enables
one to upper-bound the pressure inside. Since the ion pump is connected directly to room
temperature, we expect the pressure inside the radiation shield to be considerably less than
that measured on the ion pump. It was impressive that the pressure reading on the ion
pump, when the system was fully cooled, read 0.0×10 −9 torr.
The system was equipped with two diode temperature sensors. Each requires a 10 µA
current, and the temperature is inferred from the voltage across the diode. One is anchored
to the cold head, and is called the control sensor. The other can be placed anywhere else,
and is called the free sensor. We chose to place the free sensor directly on the 4 K baseplate,
anchored with StyCast. The free sensor came to a temperature of 4-5 K normally, while
the control sensor was at 11-13 K. It is not known why the temperature was not lower;
176