Precision spectroscopy of narrow optical transitions underpins metrology, tests of fundamental physics, and spectral databases. But its progress is limited not by measurement but by analysis: assigning lines to quantum states is a model-dependent iterative process requiring effective Hamiltonian fitting and expert involvement. Even for methane, a five-atom molecule, the density of rovibrational lines is such that a significant fraction of measured transitions remains unassigned. This work sits alongside MARVEL and spectroscopic network methods, which convert measured transitions into empirical energy levels but require already assigned lines. The new approach removes this barrier by recovering levels from unassigned frequencies.
The key tool is the SCALS spectrometer (Sequential Cavity-mode Auto-Lock Spectroscopy), in which a tunable diode laser is locked to a high-finesse optical cavity. The cavity serves simultaneously as the sample cell, frequency reference, and scanning element: a piezoelectric actuator changes its length, and an auto-lock system sequentially switches the laser between adjacent cavity modes. This provides continuous scanning over tens of terahertz with a step of 160 kHz and absolute accuracy at the kilohertz level. Spectra of acetylene, methane, and deuterated water were recorded in the 1.537–1.691 μm range. Then, the inverse graph construction algorithm was applied to the set of 686 unassigned Lamb dips of water: pairwise frequency differences are clustered, repeated differences form hyperedges, and four-cycle closures (K2,2 motifs) reveal vertices, which are energy levels. Stitching fragments via union-find and propagating values from a chosen anchor yield the complete level network.
From 686 unassigned transitions of HD16O, the algorithm recovered 158 energy levels, grouped into two subnetworks. The accuracy of the recovered levels reaches the kilohertz level, two orders of magnitude better than the MARVEL database for HDO. Comparison with MARVEL showed agreement within stated uncertainties, except for eight levels whose deviations were traced to mixed Doppler spectra in the original data. For example, levels 44,1 and 44,0 differ by only 0.002 cm⁻¹, and their lines merge in the Doppler spectrum but are resolved by the kilohertz method. Additionally, two-photon transitions for acetylene and HD16O were observed for the first time in broadband spectra, as well as a crossover resonance of methane previously mistaken for a real transition.
This work differs from neighboring methods MARVEL and spectroscopic networks by completely removing the requirement of prior line assignment. MARVEL operates on already assigned transitions, while here assignment is not needed at all: levels are recovered from raw frequencies. This is the first demonstration of model-independent reconstruction of a molecular energy network with kilohertz accuracy. The approach opens the way to exploratory precision spectroscopy of polyatomic molecules without prior knowledge of transition frequencies.
The method can be extended to more complex molecules where line density and strong interactions make traditional analysis intractable. Extending the SCALS spectral range and increasing sensitivity will allow more complete level networks to be recovered. The algorithm can also be adapted for automatic construction of next-generation spectroscopic databases.
The method directly impacts precision molecular spectroscopy, spectroscopic databases, atmospheric research, and the search for new physics through precision measurements.
Apply the method to larger molecules such as methane or polyatomic organic compounds, and integrate it with automated spectral acquisition systems to create complete atlases.
This work directly addresses a long-standing problem in spectroscopy: line assignment in dense spectra of polyatomic molecules. It offers a bypass that does not require a Hamiltonian model, which is especially important for molecules with strong rovibrational interactions.
This section is addressed to the author of the paper and their colleagues. Neighbouring works were found by meaning-based search across our archive, not by word overlap. We do not review: we show the field around the work.
What the work does
The authors reconstructed molecular energy levels directly from unassigned transition frequencies using a graph-based construction built on repeated differences and four-membered cycles. On a water spectrum in the 1537–1605 nm range recorded with the SCALS spectrometer, the algorithm extracted 158 levels with kilohertz accuracy, which is two orders of magnitude more precise than literature values.
Where its strength is
The method removes the bottleneck of manual line assignment, allowing an energy network to be built from raw spectral data without a Hamiltonian. This is especially valuable for dense spectra of large molecules, where traditional analysis gets stuck.
Where it could go next
Worth trying
Why this may matter
If the method scales to more complex molecules, it could radically accelerate the creation of accurate spectral databases and open the way to automatic spectral analysis in astrophysics and atmospheric chemistry.
Nearby works in our archive
The archive is sparse around this work: the closest review is only 0.53, against a usual 0.64. It looks like few have drilled nearby.
The effective absorption path length in the SCALS cavity is about 28 km, which is like light passing through a sample the length of a marathon course, even though the cell itself is only 42 cm.
The idea of recovering hidden structure from indirect measurements recalls plots where a map of unknown territory is reconstructed from fragments of data. In Carl Sagan's novel Contact, the alien signal contains instructions for building a machine, and here too, a coherent network of levels is assembled from disparate frequencies.