Time-reversal experiments at the University of Washington have uncovered unexpected changes in the light absorption by rubidium vapor. These changes depend on the relative orientation of the atomic polarization and external electric fields. The same effect is also present in mercury and cadmium. To gain a better understanding of possible systematic effects in the mercury time-reversal experiments and to provide an important check of atomic theory, I propose to measure this effect in both mercury and cadmium. The experimental techniques developed may pave the way for the next generation of time-reversal tests in atoms.
The anomalous decay of the neutral Kaon (K[sub 0]), which violates the combined symmetries of charge conjugation and parity (CP) is the first and only evidence for time-reversal (T) violation[1]. Almost 30 years after its discovery, the origin of the CP violation in the system remains a mystery. Virtually all theories put forth to explain the CP violation in the Kaon system predict a permanent electric dipole moment (EDM) on atoms. Researchers have used the upper limits set on the EDM of the neutron and of atomic systems to test T symmetry on a broad front. In recent work at the University of Washington, we have set the smallest limit to date for the EDM of any system[2,3]. As this experiment becomes even more precise, greater care must be taken to guard against spurious frequency shifts which could mimic or cancel a true EDM in the system. In some cases, these effects which are unwanted in the EDM experiment are important and interesting in their own right, and merit further investigation. The topic of this proposal falls into this category. A full understanding and measurement of the effect is important for the continued progress of the mercury EDM experiment, but in addition, the measurement will be a useful and important test of the atomic theory used to calculate its size.
The effect was originally observed unexpectedly during an experiment designed to measure the EDM of rubidium[4]. They observed a very small change in the absorption of resonant D1 light by the rubidium vapor which was correlated with an externally applied electric field. The shift in atomic energy levels due to applied electric fields is well known (Stark effect), but for electric dipole (E1) transitions between states of well defined parity only the second order effect is non-zero (quadratic Stark effect). The signature of this new effect is that, unlike the quadratic Stark effect, it is linear in the applied electric field. A theoretical model was proposed by Fortson [5], in which the applied electric field mixes states of opposite parity with the normal E1 states, adding magnetic dipole (M1) and electric quadrupole (E2) amplitudes to the electric dipole transition. It is the interference among these amplitudes which causes the absorption to depend linearly on the external electric field and upon the spin polarization of the atoms. Fortson et al., [5], find that the fractional change in absorptivity of the E1 transition has the form
For [sup 199]Hg, which has a [sup 1]S[sub 0] ground state electron configuration, the atomic polarization resides in the nuclear spin so P above represents the nuclear polarization of the ground state. The E1 transition of interest is the 254-nm intercombination line which is the same transition used to optically pump the mercury in the EDM experiment. The amplitudes have also been calculated for mercury[6], yielding a =(a[sub M1] + a[sub E2] = -6.6E{-8}(kV/cm), where the accuracy of the result is estimated to be about 30%. Although this effect is very small, preliminary noise measurements indicate that it can be measured with reasonable averaging time. The EDM search in mercury has the potential to improve the accuracy by another factor 10--100 if systematic effects can be understood and controlled. Improvements to the current mercury experiment require a measurement of the linear-Stark interference effect in mercury. I also plan to explore the possibility of using the effect as a direct calibration of the electric field within the vapor cells. (Currently the magnitude of the fields are inferred from other measurements.)
[sup 111]Cd can be optically pumped via the 326-nm intercombination line. Although cadmium is relatively insensitive to time-reversal violation effect because of its small Z, the development of techniques for optical pumping of cadmium in the presence of externally applied electric fields could have important ramifications for the future time-reversal experiments in mercury. The current mercury EDM experiment eliminates many spurious frequency shifts by looking for a relative shift in frequency between two adjacent mercury vapor cells. A major limit to this design is that spurious frequency shifts do not subtract off at some level because the atoms in the two cells do not sample the same volume in space. Using two spin-1/2 species in the same optical pumping cell would constitute a substantially improved system because one species is available to measure (and stabilize) the magnetic field in the cell, while the other species is sensitive to the EDM. One likely candidate for this new EDM search is cadmium which is compatible with mercury[7]. As a long term research goal (3--4 years), I intend to continue developing an optical-pumping system, for which I have already done some preliminary work, which uses cadmium as the `co-magnetometer' with Hg. Two major components must be developed before this goal can be achieved: a highly stable light source for the cadmium line at 326 nm, and vapor cells for Cd which allow application of external electric fields. Measuring the Stark interference effect for the 326 nm line in cadmium would serve as a the perfect testing ground for the Cd optical pumping system before combining the two species in an EDM measurement. In second part of this proposal I plan to use the expertise that I gained in developing the Hg EDM cells and apparatus to develop and construct these components, then use the mercury interference measurement system to measure the interference effect in cadmium. If the cadmium/mercury system is eventually built, it would be the first apparatus to employ two spin 1/2 species in the same cell which are both directly optically-pumped and transmission monitored. My choice of projects has been influenced greatly by future experiments that I might do with the same basic apparatus. This system would have many interesting and important applications in addition to its importance in the area of time-reversal measurements. Some general areas of interest include: measurement of DC Stark shift and buffer gas induced pressure shifts of the Hg 254-nm resonance line, investigation of photo-sensitized chemical reactions and atomic polarization gradients within the cells, study of optical pumping processes in dense vapors, and measuring the variation of spin relaxation lifetimes with application of electric field. Optical pumping experiments could also be done with rubidium and cesium (using diode lasers as light sources) with few adjustments to the apparatus. Another interesting application would be a search for coupling between particle spin and the gravitational force which is currently receiving much interest [8]. The proposed project nicely complements the mercury EDM work currently being done at the University of Washington. I will continue to collaborate with the fundamental measurements group there, and if significant progress is made in the cadmium/mercury comparison system at the University of Montana the collaboration will be mutually beneficial to both groups.
The general-purpose optical-pumping/spectroscopy apparatus required for this experiment consists of a three layer molypermalloy magnetic shield with magnetic field coils capable of generating fields in three perpendicular directions and a vessel holding the quartz cells containing enriched isotopes of Hg or Cd. The resonance radiation is generated either by a microwave discharge lamp (Hg) or a diode laser system (Cd). A relatively high intensity light source is used to polarize the atoms, and a second less intense source is used as a probe to measure the change in absorption of the vapor. Thus two complete sets of focusing and polarizing optics are required for each measurement.
Fabrication of the vapor cells to be used in the mercury and cadmium Stark interference effect measurements, and eventually in the cadmium/mercury comparison system requires a high vacuum system (10$^{-7}$ torr). The system will consist of a fore pump, a 3 stage oil diffusion pump and thermocouple and ionization gauges with controllers. The fused silica tubing is high purity fused quartz which serves as the raw material for the vapor cells, which will be blown by the glass blower (cell blanks) at the University of Washington, but otherwise assembled and filled here at the University of Montana.
The development of the cadmium laser will require an optical table, mount for external grating feedback, enhancement cavity mirrors, mode matching lenses between the laser and the enhancement cavity, polarizing beam splitter, 2 wave plates with rotating mounts, dielectric beamsplitter and piezo-control circuitry for locking the enhancement cavity. The most critical elements of the enhancement cavity are the doubling crystal, its mount and its temperature stabilization circuitry. Stabilization and tuning of the laser will be accomplished with an external grating optical feedback system [9]. The stabilized 652 nm laser will be combined with an external doubling enhancement cavity (described in reference [10]) to achieve the desired radiation at 326 nm. The power in the fundamental is ``recycled'' in the ring cavity producing higher output power of the doubled light. The external enhancement cavity is locked to the input laser using a scheme described in [11]. Normally this system is used in situations where high power is needed at the second harmonic; we only need a few $\mu$W of second harmonic light so the scheme is feasible with only a 3-7 mW input laser. It can be shown \cite{enhan} that the second harmonic power, P[sub 2], generated in the doubling crystal goes as P[sub 2] = KP[sub c][sup 2] where K is the non-linear conversion coefficient and P[sub c] is the power in the cavity. Typical values that they observed for P[sub c] were 20--25 times the input power to the enhancement cavity, so with 5 mW input power, we should be able to achieve the desired of approximately 2 microWatts with a doubling crystal having K on the order of 10E{-4}/Watts.