Thermal Annealing of Implanted 252 Cf Fission-Tracks in Monazite 1 2

Thermal Annealing of Implanted 252Cf Fission-Tracks in Monazite 1 2 Sean Jones, Andrew Gleadow, Barry Kohn 3 4 School of Earth Sciences, University of Melbourne, Victoria 3010, Australia 5 6 Correspondence: Sean Jones (seanj1@student.unimelb.edu.au) 7 8 Abstract 9 A series of isochronal heating experiments were performed to constrain monazite fission10 track thermal annealing properties. 252Cf fission-tracks were implanted into monazite crystals 11 from the Devonian Harcourt Granodiorite (Victoria, Australia) on polished surfaces oriented 12 parallel to (100) pinacoidal faces and perpendicular to crystallographic c-axis. Tracks were 13 annealed over 1, 10, 100 and 1000 hour schedules at temperatures between 30°C and 400°C. 14 Track lengths were measured on captured digital image stacks, and then converted to 15 calculated mean lengths of equivalent confined fission tracks which progressively decreased 16 with increasing temperature and time. Annealing is anisotropic, with tracks on surfaces 17 perpendicular to the crystallographic c-axis consistently annealing faster than those parallel 18 to the (100) face. To investigate how the mean track lengths decreased as a function of 19 annealing time and temperature, one parallel and two fanning models were fitted to the 20 empirical dataset. The temperature limits of the monazite partial annealing zone (MPAZ) 21 were defined as length reductions to 0.95 (lowest) and 0.5 (highest) for this study. 22 Extrapolation of the laboratory experiments to geological timescales indicates that for a 23 heating duration of 107 years, estimated temperature ranges of the MPAZ are -44 to 101°C 24 for the parallel model and -71 to 143°C (both ± 6 – 21°C, 2 standard errors) for the best fitting 25 linear fanning model (T0 = ¥). If a monazite fission-track closure temperature is approximated 26 as the mid-point of the MPAZ, these results, for tracks with similar mass and energy 27 distributions to those involved in spontaneous fission of 238U, are consistent with previously 28 estimated closure temperatures (calculated from substantially higher energy particles) of 29 <50°C and perhaps not much above ambient surface temperatures. Based on our findings we 30 estimate that this closure temperature (Tc) for fission tracks in monazite ranges between ~45 31 Deleted: (100) prismatic faces 32

timescales indicates that for a heating duration of 10 7 years, estimated temperature ranges 23 of the MPAZ are -44 to 101°C for the parallel model and -71 to 143°C (both ± 6 -21°C, 2 24 standard errors) for the best fitting linear fanning model (T0 = ¥). If a monazite fission-track 25 closure temperature is approximated as the mid-point of the MPAZ, these results, for tracks 26 with similar mass and energy distributions to those involved in spontaneous fission of 238 U, 27 are consistent with previously estimated closure temperatures (calculated from substantially 28 higher energy particles) of <50°C and perhaps not much above ambient surface 29 temperatures. Based on our findings we estimate that this closure temperature (Tc) for fission 30 https://doi.org/10.5194/gchron-2020-24 Preprint. Discussion started: 28 September 2020 c Author(s) 2020. CC BY 4.0 License.

Introduction 34
Fission track thermochronology is an analytical technique used to reconstruct the low- In monazite, studies have mostly focused on the U-Th-Pb and (U-Th)/He systems but only 52 limited research has been carried out into the potential of the fission track system, mainly 53 due to technological limitations. Conventional fission track dating relies on thermal neutron 54 irradiation of samples to obtain an estimate of 238 U content via the formation of 235 U fission 55 tracks, usually captured in an adjacent external solid-state track detector such as mica. This 56 approach, however, has hindered the development of monazite fission track dating for a 57 number of reasons. Monazite is highly unsuitable for irradiation due to massive self-shielding 58 by thermal neutron capture from gadolinium (Gd), which may reach abundances in excess of 59 2 wt%. Gd has an extremely high thermal neutron capture cross-section of 48,890 barns, 60 averaged over its constituent isotopes, compared to 580 barns for 235 U fission (Gleadow et 61 al., 2004;Weise et al., 2009). An even more serious issue is that neutron capture by Gd 62 https://doi.org/10.5194/gchron-2020-24 Preprint. Discussion started: 28 September 2020 c Author(s) 2020. CC BY 4.0 License.
induces substantial nuclear heating in monazite during irradiation, which may be sufficient to 63 melt the grains and would certainly anneal any fission tracks produced. The first published study of fission track dating in monazite was by Shukoljukov and Komarov 74 (1970), who reported very young ages for two monazite samples from Kazakhstan. The 75 unexpectedly young results obtained were the first to suggest that fission tracks in monazite 76 anneal at relatively low temperatures (Shukoljukov and Komarov, 1970). Since this study, the 77 majority of reported monazite fission track studies have been in conference abstracts (e.g. 78 Fayon, 2011, Gleadow et al., 2004, and Shipley and Fayon, 2006 anisotropic annealing properties to apatite in that tracks anneal faster perpendicular to the 99 c-axis compared to the c-axis parallel direction. All of these studies have suggested that fission 100 tracks in monazite have significant potential as a new ultra-low temperature 101 thermochronometer, but that further work is required to quantify the annealing kinetics. (1.67 GeV Nb, 3.54 GeV Pb and 2.38 GeV U) in mica, apatite and zircon, and concluded that 108 the activation energies for annealing the different energy ion tracks were identical in the 109 same mineral. Furthermore, they found that in the same mineral, the activation energies for 110 annealing of tracks formed by 252 Cf fission fragments were also identical to those from the 111 heavy ion tracks. These studies have shown that the minimum energy required to initiate 112 annealing is largely independent of the nature and energy of the ion source and rather is a 113 property of the detector mineral (Sandhu et al., 1990). Because the mass and energy 114 distributions of both light and heavy fission fragments from 252 Cf are similar to those 115 produced by spontaneous fission of 238 U, the annealing properties of fission tracks from either 116 source in monazite should be similar (Fleischer et al., 1975). 117

118
In this study, implanted 252 Cf fission tracks are used to constrain the thermal annealing 119 properties of monazite using a modified etching protocol (Jones et al., 2019). The new 120 annealing experiments cover a wider time-temperature range than previously reported. 121 Three alternative kinetic models are then developed that describe the reduction of fission 122 track lengths as functions of time and temperature. Extrapolation of these models then allows 123 estimates to be made of the temperature range over which fission-track annealing occurs on 124 geological timescales. 125 126 https://doi.org/10.5194/gchron-2020-24 Preprint. Discussion started: 28 September 2020 c Author(s) 2020. CC BY 4.0 License.

Experimental methods 127
Monazite crystals used in the thermal annealing experiments were separated from the Late 128 Devonian Harcourt Granodiorite (Victoria, Australia). This is a high-K, calc-alkaline granite 129 dated by zircon U-Pb and 40 Ar/ 39 Ar geochronology to ~370 Ma (Clemens, 2018  setting Struers Epofix epoxy. For each annealing experiment, two sample mounts were made, 142 one with grains orientated parallel and another perpendicular to the c-axis. Each sample 143 mount was then pre-ground using a Struers MD-Piano 1200 grinding disc and final polishing 144 with 6, 3, 1 and 0.25 µm diamond pastes. Polished grain mounts were then exposed to 145 collimated fission fragments approximately 2 cm from a thin 4mm diameter 252 Cf source 146 under vacuum for 7 hours to implant a density of ~5 x 10 6 tracks/cm 2 . Tracks were implanted 147 at an angle of approximately 30° to the polished surface which had been shown to be optimal 148 for measurement in previous experiments (Ure, 2010

Discussion 255
The average track length for the unannealed control samples across all analyses is 10.60 ± 256 0.19 µm which is slightly shorter but within error of the 11.30 ± 0.36 µm mean length reported 257   The results suggest an increase in standard deviation at short mean lengths, as is observed 340 for confined track length measurements in apatite during annealing (e.g. Green et al., 1986, 341

Parallel Linear Model 371
As a starting point, the annealing data of this study will be tested with the 'parallel model' 372 that has straight line contours (Laslett et al., 1987) Both single and double Box-Cox transforms were applied to Eqs. 10 and 11. A single Box-Cox 467 transformation was better suited to fit the data; however, it did not statistically improve the 468 models. A t-test found that Eq. 11 with a single Box-Cox transformation had a P-value of 0.096. 469 Generally, a P-value < 0.05 suggests strong evidence against the null hypothesis and that it 470 should be rejected. Whereas a r-value > 0.05 indicates weak evidence against the null 471 hypothesis, failing to reject it. In the case of Eq. 11 the null hypothesis is the equation without 472 https://doi.org/10.5194/gchron-2020-24 Preprint. Discussion started: 28 September 2020 c Author(s) 2020. CC BY 4.0 License. a transformation and the alternative is to include the single Box-Cox transformation. Using a 473 similar form of test for Eq. 10 found that the C3 constant produced a r-value of 0.123. This 474 high P-value suggests that the constant is not preferred and that the model from Eq. 11 is 475 more parsimonious. For these reasons, the final fanning models are presented with no 476 transformation (Eq. 10 and 11) and their assumptions can be checked in Figure 9.      In previous studies (e.g. Crowley et al., 1991;Laslett et al., 1987;Yamada et al., 1995), both 504 fanning models have a Box and Cox (1964) or similar type of transformation on the left-hand 505 side of the equation, but because they did not statistically improve them, they were 506 abandoned in this study. The fanning models, as they stand, explain the data very well, and 507 in general, when constructing empirical models to be used as the basis of prediction, simple 508 models with less fitted parameters are generally preferable (Laslett et al., 1987). Regardless 509 of using a transformation or not, all models presented in this study give a statistically 510 satisfactory description of the available data. 511 512 When comparing the models over laboratory timescales, little difference is observed between 513 them, particularly at length reductions < 0.80. The 0.90 track reduction contour shows the 514 largest difference over laboratory timescales, where both fanning models splay out to lower 515 temperatures. This suggests that fission tracks in monazite are even more sensitive to low 516 temperature annealing in the fanning models compared to the parallel model. As with all such 517 annealing studies, differences in annealing are magnified when the data are extrapolated to 518 geological timescales. The assumption underlying such extrapolations is that track annealing 519 results from the same physical mechanism under both laboratory and geological conditions. 520 All models show that significant reduction in the etchable lengths of fission tracks takes place 521 at ambient and lower temperatures (< 20°C) over geological timescales and that monazite is 522 particularly sensitive to low temperature thermal annealing. Considerably more track 523 Using the equations of Laslett et al. (1987), three empirical models describe the data 572 remarkedly well, with the parallel Arrhenius plot fitting the data slightly better than two 573 alternative fanning models. The differences between models are negligible, however, and, in 574 line with experience in other minerals, a fanning model is preferred. Extrapolation of the data 575 to geological timescales suggest that fission tracks in monazite are very sensitive to low 576 temperature annealing and that significant shortening of tracks occurs even at ambient 577 surface temperatures (~20°C) and below. Continued shortening of tracks occurs at 578 temperatures between ~50 -160°C when extrapolated to geological timescales, with few 579 tracks being recorded at lengths of l/l0 <~0.5. Closure temperatures for fission track retention 580 in monazite are estimated to be only 46 -25°C over geological timescales of 10 6 -10 7 years, 581 consistent with the <50°C estimate of Weise et al. (2009). 582 583 As highlighted in Laslett et al. (1987), there is no good reason why the contours in the fanning 584 Arrhenius plot need to be straight and an alternative fanning curvilinear model has been 585 proposed in the case of apatite by Ketcham et al. (2007Ketcham et al. ( , 1999. Further experiments to 586 increase the number of data points, especially for even longer heating schedules, would be 587 required to test this model in monazite. Factors that have not been considered in this study 588 and could possibly influence annealing kinetics are compositional effects (e.g. Green et al., 589 1985), radiation damage effects on etching (e.g. Gleadow, 1981)