The precise determination of the 3He/3H mass ratio, and hence the tritium β-decay endpoint energy E 0, is of relevance for the measurement of the electron anti-neutrino mass performed by the Karlsruhe Tritium Neutrino experiment (KATRIN). By determining this ratio to an uncertainty of 1 part in 1011, systematic errors of E 0 can be checked in the data analysis of KATRIN. To reach this precision, a Penning Trap Mass Spectrometer was constructed at the University of Washington and has been transferred to the Max Planck Institute for Nuclear Physics in Heidelberg at the end of 2008. Since then it is called MPIK/UW-PTMS. Special design features are the utilization of an external ion source and a double trap configuration. The external Penning ion source efficiently ionizes the helium and tritium gas and can give superior elimination of unwanted ion species compared to the previously utilized in-trap-ionization by electrons from a field-emission point. The design as a double Penning trap allows a faster measurement procedure. This should help to avoid problems resulting from long-term drifts in the experimental conditions. Additionally, the laboratory in Heidelberg was carefully prepared to have very stable environmental conditions. Experimental challenges and the first Heidelberg results with the new spectrometer are presented.

A voltage reference has been developed to bias ring electrodes of two Penning traps between -90 and 0 V. For output voltages near -90 V, the Allan deviation of the system's voltage instability is less than 1 part in 10(8) over all time scales shorter than 10(4) s. For averaging times longer than several seconds, the system's stability is determined almost completely by the noise, drift, and aging of the zener diodes in the array of voltage reference integrated circuits. For shorter averaging times, active filters built into the new system significantly reduce the intrinsic noise of the zener diodes. The system makes it possible to continuously adjust the ring voltages for frequency locking the axial motion in the two Penning traps. By keeping electrical noise highly correlated between the two traps, measurement uncertainty should be reduced for precision experiments such as Penning trap mass spectrometry.

The University of Washington Penning Trap Mass Spectrometer (UWPTMS) is now producing measurements with uncertainties approaching 10 parts per trillion (ppt). We have recently published (VanDyck, Jr. et al., Int. J. Mass Spectrom. 251: 231-242, 2006) detailed analysis of several systematic shifts which can be important at this level of accuracy. Experimental studies of these effects in our older PTMS, combined with preliminary analysis of H-2 data, and re-analysis of the previously reported He-4 (Van Dyck, Jr. et al., Phys. Rev. Lett. 92: 220802/1, 2004) and O-16 (Van Dyck, Jr. et al., Hyperfine Interact. 132: 163-175, 2001) data, gives more accurate atomic mass values for O-16, He-4, and H-2. Currently we are taking data for a new measurement of the He-3 atomic mass, and working on some improvements to the PTMS, including a new amplifier system for phase-sensitive detection of the ion's axial motion, and a new computer-controlled ultra-stable voltage source for the Penning trap's ring electrode, used to adjust the ion's axial frequency. These new systems will allow us to simultaneously manipulate individual ions in two nearby Penning traps, and some sources of noise will be the same for both traps. We plan to investigate several techniques which should reduce measurement time and improve accuracy by working with the two ions simultaneously.

The University of Washington Penning Trap Mass Spectrometer (UW-PTMS) is now producing measurements with uncertainties approaching 10 parts per trillion (ppt). We have recently published (Van Dyck, Jr. et al., Int. J. Mass Spectrom. 251:231-242, 2006) detailed analysis of several systematic shifts which can be important at this level of accuracy. Experimental studies of these effects in our older PTMS, combined with preliminary analysis of 2H data, and re-analysis of the previously reported 4He (Van Dyck, Jr. et al., Phys. Rev. Lett. 92:220802/1, 2004) and 16O (Van Dyck, Jr. et al., Hyperfine Interact. 132:163-175, 2001) data, gives more accurate atomic mass values for 16O, 4He, and 2H. Currently we are taking data for a new measurement of the 3He atomic mass, and working on some improvements to the PTMS, including a new amplifier system for phase-sensitive detection of the ion’s axial motion, and a new computer-controlled ultra-stable voltage source for the Penning trap’s ring electrode, used to adjust the ion’s axial frequency. These new systems will allow us to simultaneously manipulate individual ions in two nearby Penning traps, and some sources of noise will be the same for both traps. We plan to investigate several techniques which should reduce measurement time and improve accuracy by working with the two ions simultaneously.

The UW-Penning trap mass spectrometer (UW-PTMS) is now able to generate measurements with uncertainties near 10 ppt, making it necessary to address several major systematics, both experimentally and theoretically, in order to improve our present accuracy. These effects range from the image charge shift, originally investigated two decades ago, to the recently investigated limits due to residual magnetron energy. Using the knowledge gained from these studies, the atomic mass of O16 has been completely reviewed and the results of this re-analysis are presented. There is also an adjustment given for the atomic mass of He4 and we present a preliminary result for the atomic mass of H2. In addition, we present our plans for replacing the present spectrometer with a new one containing a cylindrical storage trap that will be loaded using an external ion source and two hyperbolic traps whose dimensions are identical to the Penning trap in our present spectrometer. Thus, it is expected that the new spectrometer will have essentially the same systematics as those described in this paper.

The atomic masses of the alpha particle and 4He have been measured by means of a Penning trap mass spectrometer which utilizes a frequency-shift detector to observe single-ion cyclotron resonances in an extremely stable 6.0 T magnetic field. The present resolution of this instrument approaches 0.01 ppb [10 ppt (parts per trillion)] and is limited primarily by the effective stability (<5 ppt/h) of the magnet over hundreds of hours of observation. The leading systematic shift [at -202(9) ppt] is due to the image charge located in the trap electrodes. The new value for the atomic mass of the alpha particle is 4 001 506 179.147(64) nu and the corresponding value for the mass of 4He is 4 002 603 254.153(64) nu (nu=10(-9) u). The 16 ppt uncertainty is at least 20 times smaller than any previous determination.

The atomic masses of the $\ensuremath{\alpha}$ particle and $^{4}\mathrm{He}$ have been measured by means of a Penning trap mass spectrometer which utilizes a frequency-shift detector to observe single-ion cyclotron resonances in an extremely stable 6.0 T magnetic field. The present resolution of this instrument approaches 0.01 ppb [10 ppt (parts per trillion)] and is limited primarily by the effective stability ($<5\text{ }\mathrm{ppt}/\mathrm{h}$) of the magnet over hundreds of hours of observation. The leading systematic shift [at $\ensuremath{-}202(9)\text{ }\mathrm{ppt}$] is due to the image charge located in the trap electrodes. The new value for the atomic mass of the $\ensuremath{\alpha}$ particle is $4\text{ }001\text{ }506\text{ }179.147(64)\text{ }\mathrm{nu}$ and the corresponding value for the mass of $^{4}\mathrm{He}$ is $4\text{ }002\text{ }603\text{ }254.153(64)\text{ }\mathrm{nu}$ ($\mathrm{nu}={10}^{\ensuremath{-}9}\text{ }\mathrm{u}$). The 16 ppt uncertainty is at least 20 times smaller than any previous determination.