Charge management for gravitational-wave observatories using UV LEDs

Accumulation of electrical charge on the end mirrors of gravitational-wave observatories can become a source of noise limiting the sensitivity of such detectors through electronic couplings to nearby surfaces. Torsion balances provide an ideal means for testing gravitational-wave technologies due to their high sensitivity to small forces. Our torsion pendulum apparatus consists of a movable plate brought near a plate pendulum suspended from a nonconducting quartz fiber. A UV LED located near the pendulum photoejects electrons from the surface, and a UV LED driven electron gun directs photoelectrons towards the pendulum surface. We have demonstrated both charging and discharging of the pendulum with equivalent charging rates of $\sim {10}^{5}e/\mathrm{s}$, as well as spectral measurements of the pendulum charge resulting in a white noise level equivalent to $3\times {10}^{5}e/\sqrt{\mathrm{Hz}}$.

Figure 1

Schematic showing the geometry of the apparatus. A Si pendulum is suspended from a quartz fiber with split Cu plate and Al electrodes nearby. The torsional displacement of the pendulum is measured optically and is used in a feedback loop to control the pendulum angle. Also shown are the UV LED and the electron gun, which are used to remove and add electrons to the pendulum, respectively.

Figure 2

Torque sensitivity in feedback for a plate-pendulum separation of 1 mm with a tungsten (red) fiber, quartz fiber (dotted green), and quartz fiber under UV illumination (black). The tones in the quartz spectrum are produced from the charge measurement technique with a fundamental at 10 mHz. The instrumental limit when using the tungsten fiber (dashed) is a combination of thermal noise (at low frequencies) and measurement noise (at high frequencies). The thermal noise of the quartz fiber is 1.7 times higher than that for the tungsten fiber due to the differences in $\kappa$ and $Q$. The measured noise level of the quartz fiber is consistent with the measured charge fluctuations on the pendulum $\sim 2\mathrm{mV}/\sqrt{\mathrm{Hz}}$ [11].

Figure 3

Cross-sectional drawing of the electron gun. UV light from the LED creates photoemission from the magnesium coated cathode, biased at $V_B=-6\mathrm{V}$. Electrons accelerate towards the grounded nozzle of the electron gun, maintain their kinetic energy inside the field-free nozzle, and exit the gun towards the pendulum.

Figure 4

Charging and discharging the electrically isolated Si pendulum. The charge on the pendulum is driven negatively (to $-2.5\times {10}^{8}e$) using the electron gun and positively (to $+2.5\times {10}^{8}e$) using the UV LED.

Figure 5

The measured charge noise (solid) in our experiment with an electrically isolated pendulum. The LISA requirement (dashed) is the contribution of charge noise accounted for in the LISA acceleration budget, assuming an effective charging rate of $500e/\mathrm{s}$.