Thermal annealing characteristics of fission tracks in natural zircons of different ages

Fission track (FT) thermochronometry using zircon has widely been applied to the resolution of a variety of geologic problems, for which the understanding of FT annealing behaviour is essential. Thermal annealing experiments were conducted on FTs in natural zircons having different ages (ranging from ~0.6 to ~70 Ma) and radiation damage levels. We measured horizontal confined track lengths on nine zircon concentrates separated from rapidly cooled volcanic rocks, after 1 hr annealing at 400–700°C. As the annealing temperature increases, the observed tracks show a consistent and systematic length reduction in all samples, and the mean track lengths are hardly distinguishable among the nine samples for the same annealing step. Our results suggest that the thermal annealing characteristics at laboratory time‐scale are concordant among the zircons, regardless of their ages, and that identical annealing kinetics may work for Late Mesozoic to Cenozoic zircons.


| INTRODUC TI ON
Fission track (FT) thermochronometry using zircon has widely been applied to the resolution of geologic problems, such as to infer the exhumation history of orogenic belts and understand fault-zone processes (e.g., see reviews by Malusà & Fitzgerald, 2019;Tagami, 2005). Thermal annealing characteristics of FTs in zircon have primarily been studied using spontaneous FTs (Murakami, Yamada, & Tagami, 2006;Yamada, Tagami, Nishimura, & Ito, 1995; see also Tagami, Ito, & Nishimura, 1990). This is because (a) in contrast to apatite, the mean length of spontaneous FTs in zircon from rapidly cooled volcanics is indistinguishable from that of induced FTs, suggesting the absence of natural shortening under ambient conditions (Hasebe, Tagami, & Nishimura, 1994), and (b) thermochronologic analyses of geological samples are carried out using spontaneous FTs in natural zircons, for which thermal annealing behaviour should be described.
The thermal annealing kinetics were first determined on spontaneous FTs of the Nisatai Dacite (NST) zircon (Ketcham, 2019;Tagami, Galbraith, Yamada, & Laslett, 1998;Yamada, Murakami, & Tagami, 2007;. However, it is not well known how the annealing kinetics can vary among natural zircons, for example, as a possible consequence of radiation damage accumulation (Rahn, Brandon, Batt, & Garver, 2004). Kasuya and Naeser (1988) measured confined track length reductions during 1 hr isochronal laboratory annealing for both spontaneous and induced FTs in zircons, and found that (a) induced FTs in pre-annealed zircons are more resistant to thermal annealing than spontaneous FTs, (b) annealing behaviours of spontaneous FTs are indistinguishable between four samples of Palaeogene to Miocene ages, showing a range of spontaneous track densities from 0.9 to 10 × 10 6 cm −2 and (c) induced FTs in non preannealed zircons behave like spontaneous tracks.  also compared confined track length reductions during 1 hr isochronal laboratory annealing between spontaneous and induced FTs in the NST zircon. Experimental procedures were improved by applying analytical criteria and measuring crystallographic orientation of confined FTs to use L 60 (mean length for tracks with angles >60° to the crystallographic c-axis) instead of L all (mean length for all crystallographic directions) . It was found that induced FTs are more resistant to thermal annealing than spontaneous FTs at the advanced stage of annealing, even after correction of track length anisotropy that reflects anisotropic etching and annealing by adopting L 60 .
Here, on the basis of those previous results, we perform new laboratory annealing experiments using nine zircon samples of different ages and spontaneous track densities, in order to better reveal the possible variation of annealing behaviours among natural zircons.
Confined track lengths were measured on each sample aliquot, after 1 hr isochronal heating at elevated temperatures ranging from 400 to 700°C. Our attention was particularly focused on young natural zircons, in which accumulated radiation damage is minor. This may give a clue to better understand the annealing behaviour of spontaneous and induced FTs and to clarify any difference between them.

| E XPERIMENTAL PRO CEDURE AND SAMPLE S
Zircon grains were concentrated and separated from each host rock using conventional mineral separation procedures (Kohn, Chung, & Gleadow, 2019). Nine zircon samples separated from quickly cooled (and not later reheated) volcanics were adopted in this study, including the age standards Fish Canyon, Bishop and Buluk Member Tuffs (Table 1). The zircon samples range in age from 0.6 to 70 Ma and in spontaneous track density from ~0.05 to 7 × 10 6 cm −2 . For five zircon samples <3 Ma, grains were irradiated by thermal neutrons before thermal annealing at the Thermal Column Pneumatic Tube facility of the Kyoto University Research Reactor. This allowed to form induced FTs from 235 U next to spontaneous 238 U FTs, so that a sufficient number of confined tracks could be measured in euhedral grains even on these individual sample aliquots. In those five samples, the FTs under consideration for our annealing experiments are thus a mixture of spontaneous and induced FTs. The mixing ratio ρ i /(ρ s + ρ i ) ranges from ~0.29 to 0.99, where ρ s and ρ i denote the mean track density of spontaneous and induced FTs, respectively, observed on a grain-internal surface (4π geometry) (Table 1).
Thermal annealing experiments of individual sample aliquots were conducted for 1 hr using a conventional muffle furnace, controlled to within ±2°C. Annealing temperatures were monitored by a thermocouple (connected to a temperature monitor) that was inserted into the furnace and placed within 1 cm from the sample position, in order to monitor the exact temperature during annealing.
Zircon samples were annealed at ~400 to ~700°C in approximate 50 or 100°C temperature steps (Table 2). Note. All errors are quoted at 2σ. All values after references, except for data measured in this study (*) and tentative estimates in this study because the density is too low to be measured accurately (**). Abbreviations: ρ s = areal density of spontaneous tracks (10 6 cm −2 ); ρ s + ρ i = areal density of both spontaneous and induced tracks (10 6 cm −2 ); Hf = Hf content (mg/g); T = age (Ma).
The individual sample aliquots were then mounted into pieces of PFA Teflon sheets, and ground/polished using conventional procedures (Tagami, 2005). FTs were etched by conventional NaOH:KOH eutectic etchant (Gleadow, Hurford, & Quaife, 1976)  Abbreviations: L 60 = mean length for >60° to c-axis (µm); L all = mean length for all crystallographic directions (µm); n = number of grains scanned; N 60 = number of measured tracks for directions >60°to c-axis; N all = number of measured tracks for all crystallographic directions; r 60 = normalized value of L 60 by that of unannealed tracks; SD 60 = standard deviation for the length distribution of tracks for >60° to c-axis (µm); SE = standard error of L 60 (µm); t e = etching duration (hr); T = annealing temperature (°C). a Zircon grains of these samples were irradiated with thermal neutrons prior to laboratory annealing.  Table 2), because the impact of anisotropy is negligible for that azimuth angle interval (Hasebe et al., 1994;.
The Hf contents of zircon samples were measured with analytical precision of ±0.01% using the LA-ICP-MS system installed at Kanazawa University. Multiple analyses were performed on each sample aliquot using 3-5 grains, following the experimental procedure described by Morishita et al. (2009). Table 2 shows the analytical results of track length measurements of unannealed and annealed aliquots for the nine zircon samples. Figure 1 presents a plot of mean confined track lengths (L 60 ) against annealing temperature. The mean confined track lengths measured on unannealed aliquots range from ~10.5 to ~11.2 µm for the nine zircons, approximately concordant with each other, and also with those of previous studies (Hasebe et al., 1994;. With increasing annealing temperature, the observed track lengths show consistent and systematic reductions, as observed previously (Kasuya & Naeser, 1988;.

| Comparison among natural zircons
To better compare the results between the nine zircons of different ages, observed mean track lengths were plotted at three different annealing stages (unannealed, 600°C and 650°C) against (a) age and (b) spontaneous track density (Figure 2 & Naeser, 1988), this implies that identical annealing kinetics may apply for most of the Late Mesozoic and Cenozoic zircons.
Variability in FT annealing behaviour among natural zircons may also be introduced by Hf, which is commonly the most abundant minor element in this mineral. We performed LA-ICP-MS analyses of the nine zircons to determine their Hf content, which ranges from 5.5 ± 0.4 to 8.4 ± 1.1 mg/g (Table 1). We can thus conclude that identical kinetics are applicable to the Late Mesozoic and Cenozoic zircons having Hf composition within the observed range.

| Comparison between natural zircons versus pre-annealed zircons
Previous studies (Kasuya & Naeser, 1988; showed that induced FTs in pre-annealed zircons are more resistant to thermal annealing than spontaneous FTs in natural zircons, which was confirmed by the comparison of the results of this study with the data by . In addition, we observe that the thermal annealing behaviour of spontaneous FTs is concordant with that of the mixture of spontaneous and induced FTs, regardless of the mixing ratios ( Figure 2,  (Pupin, 1980), whereas the laboratory heating conditions to preanneal spontaneous FTs in zircons are 800°C for 24 hr (Kasuya & Naeser, 1988) or 1,009 ± 2°C for 2 hr . Although zircon formation is kinetically controlled and thus its formation temperature cannot be directly compared with that of laboratory annealing, those rather high temperatures of pre-annealing may have modified some crystallographic properties controlling the FT annealing rate.
Another possible factor is the difference of time periods that elapsed since the last cooling episode to ambient temperature until the time of laboratory annealing of FTs. While the pre-annealed zircons experience merely a few to several months after the last cooling at laboratory (subsequent to the pre-annealing), zircons without preannealing were kept unheated for geological time periods (Table 1).
During such extended periods of time, some temporal change of zircons, such as the accumulation of radiation damage, may have preceded to change the crystallographic ordering/structure and thus the FT annealing rate. Because of the approximately identical annealing behaviour among zircons that range in age from ~0.6 to ~70 Ma (or in spontaneous track density from ~0.05 to 7 × 10 6 cm −2 ), it is implied that the structural change may proceed quickly in zircons after their thermal resetting and is likely to reach saturation within 0.6 m.y. (or with accumulation of spontaneous tracks of <~0.05 × 10 6 cm −2 ).

| CON CLUS IONS
1. As a result of 1 hr isochronal heating, confined track lengths in nine natural zircons, separated from rapidly cooled volcanic rocks, show consistent and systematic reductions with increasing annealing temperature.

2.
No systematic difference was found for FT annealing characteristics among natural zircons of ~0.6 to ~70 Ma age (and spontaneous track density of ~0.05 to 7 × 10 6 cm −2 ).

3.
The present results, coupled with previous data, imply that identical annealing kinetics can work for many of the Late Mesozoic to Cenozoic zircons, with Hf contents ranging from 5.5 ± 0.4 to 8.4 ± 1.1 mg/g.

ACK N OWLED G EM ENTS
We thankfully acknowledge Meinert Rahn and Marco Malusà for their helpful and constructive reviews of the manuscript, and Chuck Naeser, Barry Kohn and Andy Carter for supplying zircon samples used in the present study. We also thank Akihiro Tamura and Shoji Arai for their support with LA-ICP-MS analyses at Kanazawa University. This study was supported by a visiting research program of the Kyoto University research reactor.