Doctor von der Wense

Yesterday, Lars von der Wense of LMU Munich successfully passed his PhD exam, and he did it with summa cum laude distinction! Lars is the first PhD student funded by the nuClock project to finish his thesis. Congratulations!

A snapshot of the PhD exam, the title reads “On the direct detection of the Th-229 nuclear isomer”.

Lars has been supervised by LMU group leader Peter G. Thirolf. The board of examiners also included fellow nuClock group leaders Adriana Palffy from MPIK Heidelberg and Thomas Udem from MPQ Munich. After the 90-minute torture, Lars was rewarded with the traditional doctoral hat prepared by hisPhD colleagues: look at his smile!

Lars wearing his new hat, fully equipped with an alarm clock linked to a Th-229 nucleus!

The celebration continued with a large crowd of family & friends in a nearby restaurant.

His thesis is now available online here.

Re-evaluation of the Beck et al. data

The currently most accepted value of the isomer energy is 7.8(5) eV, obtained by the so-called “Beck et al. measurement”. In this experiment, researchers from Lawrence Livermore employed a NASA microcalorimeter for a high-resolution gamma measurement of the U-233 decay. Using a clever differencing scheme, they indirectly inferred the energy of the isomer. The final result includes correction terms related to unknown branching ratios.

This measurement is now 10 years old, and in the meantime, no experiment was successful in refining the isomer energy further (there is evidence from the LMU experiment, however, that the isomer energy is between 6.31 and about 18 eV). In particular, two independent experiments at PTB in Germany and UCLA/ALS searched for an optical signal of the isomer using synchrotron radiation to excite nuclei on/in wide bandgap materials, but found no signal. E. V. Tkalya et al. were the first to revisit the Beck et al. measurement and showed that branching ratios markedly different from the ones assumed in the original publication would lead to an isomer energy much larger than 7.8 eV (see their paper here).

Many current experiments have a limited search window, contrained e.g. by the ionization thresholds of Th ions or the transmission window of crystals. Inspired by the mysteriously short lifetime of the Th+ isomer in the LMU experiment, the TU Wien team set out to re-evaluate the original Beck et al. data, aiming to check if the 2007 measurement would be compatible with a much higher isomer energy. Their analysis is now available on the arXiv.

The authors find no major flaw in the original 2007 data analysis. They expand the statistical error of 0.5 eV, which appears to be underestimated, into a two-dimensional contour plot, constructing confidence regions for any given branching ratio. They find that the isomer energy depends only mildly on the branching ratio, in contrast to an earlier analysis by S. L. Zakharov. To give a rough estimate, an isomer energy above 10 eV (below 125 nm) can be excluded at the 95% confidence level.

If you have comments and would like to join the discussion, please share your thoughts with us!

Contour plots of branching ratio b (the probability of the 29-keV state to decay out-of-band onto the ground state) vs isomer energy. The horizontal lines represent various values of b that can be found in the literature. The grey shaded region has been scanned in the UCLA@ALS experiment. The three plots assumes different values of the energy splitting of the 42-keV doublet: (d) 198.22 eV, (e) 198.44 eV, and (f) 198.66 eV.

LMU arXiv paper: feasibility study for electron spectroscopy of the IC electron

A series of recent breakthrough experiments at LMU Munich was able to confirm theoretical and experimental predictions on various properties of the Th-229 isomer. Specifically, the LMU team was able to show (1) that the isomer exists at all, (2) that its energy is somewhere between 6.3 and 18 eV, (3) that the isomer lifetime in Th2+ and Th3+ is longer than a few minutes, (4) that the lifetime in the neutral atom is very short, about 10µs, and (5) that internal conversion (IC) is the dominant decay channel in the neutral atom. The next major topic to address is a refinement of the value of the isomer energy. The original plan of the LMU team was to measure the wavelength of the VUV gamma emitted upon de-excitation (read the corresponding paper here), but as IC seems to be dominant even on large-bandgap surfaces, the LMU team switched gears and now prepares to measure the energy of the IC electron.

In a theoretical feasibility study, LMU PhD student Benedict Seiferle looked into surface effects of the catcher plate and the spectrometer, the attainable resolution, and the expected signal/noise. He concludes that a resolution of 0.1 eV can be reached. Also, he suggests a calibration scheme in which a light source near 160 nm with well-known wavelength is used to, via the photoeffect, knock electrons out of the catcher. This technique could also be used to optimize the spectrometer. Generally, the measured energy spectrum does not depend on the work function of the catcher plate, but only on the retarding voltage applied between the catcher and the spectrometer, and the work function of the spectrometer itself.

The manuscript had been submitted to EPJ D and is now available on the arXiv.

Th-229 isomer lifetime measured in the neutral charge state

The preferred de-excitation pathway of low-energy excited nuclear states is internal conversion (IC). Instead of de-exciting to the ground state via emission of a gamma, they couple to electronic states of similar energy. Their excitation energy is transferred to the electron, which typically leaves the atom and carries away the excess energy. Such a process is common for nuclear states up to 1 MeV, and its probability generally increases with decreasing excitation energy. For the U-235 isomer at 70 eV, IC is virtually the only de-excitation pathway. The nucleus does not only couple to electronic states of the atom itself, but also the the environment: the IC rate depends on the electronic states in the neighborhood (charge state, work function of the material the atom is placed on, molecule/chemisty, …). Indeed, for U-235, a mild dependence of the isomer half-life on the nature of the host material has been observed.

The nuclear optical clock, envisioned by nuClock researchers, relies on optical excitation and optical detection of the isomeric transition. As such, IC is thought of as a competing, undesired process that needs to be suppressed. The IC rate (or equivalently, the half-life of the isomer) is expected to depend dramatically on the charge state of the Th-229 atom or ion. For the neutral thorium atom, the first ionization energy is smaller than the isomer energy: the isomer may pass its energy to a valence electron quickly, which then leaves the atom. This process was used for the first direct detection of the isomer in the earlier work of the LMU group. For atoms in higher charge states (more specifically, with an ionization energy way larger than the isomer energy), the IC channel should be heavily suppressed, and the lifetime of the isomer is expected to reach its radiative limitation of about 1000 s.

The LMU group set out to measure the isomer half-life in neutral Th-299. Using much of the technology available from their earlier work, they accumulate about 400 ions of Th2+ or Th3+ in an ion trap (about 2% of them in the isomeric state), and gently slam this bunch of ions onto an MCP detector. The ionic impact (neutralization of the ions on the detector surface) produces a rather large signal, which decays in a fraction of a microsecond. Afterwards, the isomeric nuclei are free to de-excite via coupling to the electrons of the Th atom (or of the detector surface). The electrons are ejected with a small kinetic energy (e.g. the isomer energy minus the work function of the material) and generate a signal on the MCP. By summing over many 10’000 bunches, the signal of the IC electrons forms a very beautiful exponential decay with a half-life of 7(1) µs, corresponding to a lifetime of about 10 µs. Such an exponential decay was observed only for Th-229, but not for any other Th or U isotope.

The half-life of the isomer matches theoretical expectations very well (predicted reduction of the 1000-seconds radiative half-life by a factor of 10^9). The half-life of other charge states (e.g. Th+ or Th3+) is of interest as well, but cannot be measured with the present technique. The layout of the experiment, however, allows to deduce certain bounds on these values: < 10 ms for Th+, and > many minutes for Th2+ and Th3+. It is still unclear how the surprisingly short half-life in Th+ is related to the isomer energy.

Learning that the isomer half-life is at the expected order of magnitude is reassuring for the international research community working towards the nuclear clock. Although a more precise measurement of the half-life would not aid the development of the nuclear clock much, it would add to similar work on the isomer in U-235. So far, the LMU group worked with only one surface material. It would be insightful to replace the present nickel alloy by a range of other materials, e.g. metals with different work functions, or insulators. Such a measurement would explain whether the IC is dominated by the Th atom’s electronic shell, or by the surface around it. Ultimately, researchers at LMU plan to perform the experiment in free space (without any surface), e.g. by neutralizing the ion bunch with a beam of cesium atoms.

The lifetime of the Th-229 isomer was measured in the neutral state. Bunches of (a) Th2+ and (b) Th3+ ions were placed on an MCP surface, where they neutralize and give rise to a large signal peak, clearly visible at about 82 µs (figure a) and 68 µs (figure b). For the Th-230 isotope (blue curves), the signal drops to zero within less than a µs. For Th-229 (red curves), however, an exponential tail is clearly visible. This additional signal is interpreted unambiguously as the de-excitation of the isomer. An exponential fit to the data gives a half-life of 7(1) µs.

The findings of this experiment have been published today with Phys. Rev. Lett. (even as an Editors’ Suggestion!) and can be found here.

The corresponding LMU press releases are available in German and in English.

Transportable optical Sr clock presented by PTB

Building an optical clock based on the Th-229 nuclear transition is the ultimate goal of the nuClock project. We like to claim that such a nuclear clock will be less sensitive to perturbations (because the nucleus is so much smaller than the orbits of valence electrons), offer a supreme quality factor (because the transition energy is so large, and the lifetime of the isomer is so long), and outperform existing clocks in flat-out all respects (well, we become a little bit emotional sometimes). We also like to claim that the nuclear clock will be very robust in operation, ideally suited for geodesy and space applications. In terms of robustness, there is now an experiment that sets a new standard.

The PTB group of Christian Lisdat built, operated, and characterized a transportable optical clock (TOC, as compared to SOC (space optical clock) and NOC (nuclear optical clock)). Such TOCs are required to compare distant optical clocks where no suitable fiber link exists. There are two other strategies for the comparison of clocks: satellite links for inter-continental comparisons, and transportable Cs clocks for same-continent comparisons, but both of these are too weak to fully exploit the accuracy of today’s best optical clocks. Transportable clocks would also be able to measure the geoid via the gravitational redshift, and they would do so with higher spatial resolution compared to satellite missions, and potentially with higher precision.

The TOC presented by PTB is based on Sr-87 and operates very much like a standard laboratory-based Sr lattice clock. A few adaptations have been applied: the Zeeman slower is made of permanent magnets to reduce heat dissipation and power consumption by coils, and a total of eight temperature sensors have been placed near the atoms (both inside and outside the vacuum) to calculate BBR corrections. Following a general trend, the diode/TA laser system for the optical lattice has been replaced by a Ti:Sapph model to avoid ASE-related issues.

The clock reaches a systematic uncertainty of 7 x 10^-17, limited by lattice Stark shifts (which have not yet been fully characterized) and cold collisions. The clock averages down as 1.3 x 10^-13 per root-tau, reaching an agreement with a stationary Sr clock in the 10^-17 range within about two minutes. These two parameters are about two orders of magnitude better than Cs microwave standards and make this TOC the best optical clock that actually left the laboratory.

The transportable optical clock at PTB. Left: the trailer that houses a miniature laboratory (2.2 by 3.0 by 2.2 meters), and right: the inside of the trailer. Both pictures taken from the publication arXiv:1609.06183.

A short sidenote: it seems that the PTB team forgot to show that the clock actually is transportable… we miss that YouTube video showing the PTB trailer stuck in traffic.

The publication is already available on the arXiv preprint server and has recently been accepted by Phys. Rev. Lett. for publication.