Inside the Laser: Scientists Unravel Hidden Biases in Zircon Dating

Every time a geologist fires an ultraviolet laser at a tiny grain of zircon, a remarkable piece of planetary timekeeping unfolds in a matter of minutes. The technique, known as laser ablation inductively coupled plasma mass spectrometry, or LA-ICP-MS, has transformed Earth science by allowing researchers to date minerals in situ at scales of tens of micrometres, generating vast quantities of geochronological data that would have taken months or years to produce by older methods. Yet behind this speed lies a stubborn problem: the measurements carry biases that are large, variable, and only partially understood. A new study by Donald W. Davis and Heriberto Rochin-Banaga of the University of Toronto, published in the journal Geochronology, takes one of the most detailed looks yet at where these biases come from and what can realistically be done about them.

The stakes are considerable. Zircon is the workhorse mineral of geochronology because it incorporates uranium but rejects lead when it crystallizes, meaning that nearly all the lead found in an ancient zircon grain is the product of radioactive decay. By measuring the ratio of lead-206 to uranium-238, scientists can calculate when the crystal formed. But the measured ratio is never the true ratio. The authors identify a cascade of distortions: mass fractionation in the mass spectrometer, differences in how efficiently lead and uranium atoms are ionized in the plasma, and, most troublesome of all, a bias introduced by the laser ablation process itself. Together, the plasma-related effects alone suppress the measured lead-206 to uranium-238 ratio by roughly 25 to 35 percent, the largest single source of error, and the ablation bias piles further distortion on top.

The Toronto team worked with an Agilent 7900 mass spectrometer coupled to an NWR193UC nanosecond excimer laser, firing pulses at a frequency of 10 hertz with an energy fluence of 3.5 joules per square centimetre. To probe the earliest moments of ablation, they also ran experiments at a glacial 0.2 hertz, firing a single pulse every five seconds so that the signal from each individual pulse could be isolated and integrated. Their standards included NIST glass reference materials and several well-characterized zircons, among them the Keuhl Lake megacryst, which shows very little radiation damage, and the higher-uranium, more heavily damaged zircon DD91-1, dated at about 2.68 billion years.

The heart of the study is a physical picture of what actually happens when a laser pulse strikes zircon. A rough energy calculation shows that each pulse delivers roughly thirty microjoules to a thirty-micrometre target, and if that energy is absorbed in a layer about a tenth of a micrometre thick, the temperature rise is on the order of ten thousand degrees Celsius. At those energies, the mineral does not simply vaporize neatly; it boils, producing a transient ablation cloud of melt droplets and vapour. Because lead is far more volatile than uranium, lead preferentially evaporates from the molten material, leaving behind a melt pool at the base of the pit that is depleted in lead and a vapour phase that is enriched in it. The researchers confirmed this experimentally: when they scanned a laser beam across the floor of a large ablation pit, they found the solidified melt there carried lead-206 to uranium-238 ratios up to forty percent lower than the surrounding pristine zircon.

What happens next determines the shape of every measured age. Some of the ablated cloud escapes the pit and reaches the plasma, carrying a lead-enriched signal. But part of the cloud is deposited back onto the walls of the deepening pit or falls back around its rim as so-called fallback material. Because this deposited material is lead-depleted, its removal drives the measured ratio upward, and the deeper the pit, the more material is sequestered. The authors traced this behaviour using the signal from zirconium oxide, a proxy for the total mass being ablated. As the zirconium oxide signal declines pulse by pulse, the measured lead-206 to uranium-238 ratio rises in a mirrored pattern, exactly what would be expected if the loss of signal represents lead-depleted material being locked away where the laser cannot efficiently re-ablate it.

The first ten or so pulses behave in a surprisingly chaotic way. In theory, the earliest signal should start high and decrease as the melt pool becomes progressively more lead-depleted and approaches equilibrium. Instead, the measured ratios from the first ten pulses remain roughly constant, apparently because two opposing effects, the decreasing lead enrichment of the vapour and the deposition of lead-depleted fallback, nearly cancel each other out. Crucially, the average of these early ratios differed between standards in ways that could not be reconciled, with implied instrument biases of about minus 31 percent for both Keuhl Lake and DD91-1 in one session but variable intercepts across samples and sessions. This means the tempting strategy of extrapolating early data back to zero time to recover an unbiased ratio simply does not work reliably, particularly because radiation damage in the crystal structure appears to influence the outcome.

Radiation damage turns out to be a central complication. As uranium and thorium decay, alpha recoil events gradually shatter the zircon lattice, eventually producing a disordered, metamict state that melts more easily and ablates differently. The team built a mathematical model of the fractionation process, driven by the zirconium oxide signal decay, with two key free parameters: the fraction of lead lost from deposited material and the proportion of that material recycled by subsequent pulses. The best fits required substantial lead loss, around 78 percent for Keuhl Lake and 90 percent for DD91-1, and essentially no recycling, consistent with the idea that the steep walls of the pit prevent the laser from efficiently re-ablating deposited material. But because these parameters cannot be independently constrained, the model, however illuminating, cannot by itself deliver a bias-free age.

So what does work? The authors compared several calibration strategies and found that the most robust approach is also the most direct: measure the full lead-206 to uranium-238 profile of a standard and of the unknown sample under identical conditions, then multiply the standard’s profile by a calibration factor chosen to minimize the mismatch between the two curves. When they treated DD91-1 as an unknown and Keuhl Lake as the standard, omitting a final anomalous segment likely caused by the beam hitting an older zircon core, the inferred ratio was within 0.15 percent of the accepted value. Comparing simple averages of the profiles performed nearly as well, differing by only 0.2 percent. They packaged this method into a software tool called UTILAZ, written in Visual Basic for Applications, which processes and calibrates data automatically.

Applying the method to Precambrian zircons of known age delivered encouraging results. Samples with varying degrees of radiation damage, but short of the fully metamict state, yielded calibrated lead-206 to uranium-238 ages within about one percent of their established lead-207 to lead-206 ages, suggesting that moderate radiation damage introduces little bias into discordance. Fully metamict zircon was another matter: its profiles were flatter than the standard’s and its calibrated data showed reverse discordance ranging from minus 6 to minus 19 percent, with the magnitude roughly proportional to uranium concentration. That correlation might eventually serve as a first-order correction, even though its underlying cause remains unclear, since average zirconium oxide signals showed no relationship with uranium content or damage.

The study’s conclusions are refreshingly candid. For nanosecond excimer laser systems, no amount of clever data processing is likely to push the accuracy of lead-206 to uranium-238 ages much beyond about one percent at the one-sigma level without a far better understanding of the ablation process. The most promising route to improvement lies in instrument design, specifically shrinking the ablation chamber and improving aerosol transport efficiency to reduce fallback and the fractionation it causes. The authors also issue a plea to the community: because methods and understanding will keep evolving, researchers should preserve raw time-resolved data files in accessible repositories so that published ages can be reprocessed as calibration techniques improve. In a field where millions of years can hinge on a few percent of a ratio, knowing exactly where the bias hides is the first step toward making every laser shot count.

Subject of Research: Causes and mitigation of uranium-lead fractionation during laser ablation ICP-MS dating of zircon

Article Title: Causes and mitigation of U–Pb fractionation during LA-ICP-MS analyses of zircon using nanosecond excimer laser systems

Article References: Davis, D. W., & Rochin-Banaga, H. (2026). Causes and mitigation of U–Pb fractionation during LA-ICP-MS analyses of zircon using nanosecond excimer laser systems. Geochronology, 8(3), 529-545. https://doi.org/10.5194/gchron-8-529-2026

Image Credits: AI Generated

DOI: 10.5194/gchron-8-529-2026

Keywords: zircon, U-Pb geochronology, LA-ICP-MS, laser ablation, elemental fractionation, downhole fractionation, radiation damage, metamict zircon, excimer laser, mass spectrometry, geochronology, calibration