We would like to thank Greg Ludvigson for the comments.

1. We will add the requested literature Gulbranson et al., 2022 to the revised manuscript.
2. Yes, this is the same manuscript we reference and we will update the status on submission of the revised manuscript.
3. We agree that the CL image in Figure 1b is too dark and will enhance the image quality for a revised version of the manuscript so that more information becomes visible. As this is a stitched image of about 30 individual images, I suspect that the stitching software decreased the brightness to make it balanced, and I as user did not correct it. An image enhancement is no problem.
4. –

We Thank Niki Sota for the detailed review and the valuable comments that will help to improve the manuscript.

L13: We agree that the U–Pb terminology should be consistent throughout the manuscript, and we will make sure the revised manuscript is, and as suggested we will use the dash (–) instead of the minus (-).

L43: We agree that the abbreviation of Laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) should be consistent throughout the manuscript and we will adjust it.

L59: There is no Google earth image and we do not intend to add a google map / earth image as everyone can enter the coordinates and look at the location directly in google earth. Additionally, the google satellite image is not perfectly accurate to the reel coordinates and might change in future. That is also why there are two set of coordinates, the “real” ones in the (EPSG:4258 (ETRS89) and the ones for google earth/maps image that shows the location on the satellite image. To prevent any confusion, we will change the revised manuscript and just report the real coordinates.

L77: The correction factors for the 23 sessions can be added but we think that this information is of limited help. Specifically:

• The mass bias for the 207/206 correction is found to be very small and stable for different sessions (usually less than 1% variation).
• The correction factor for 238U/206Pb is highly variable due to variations in ionisation efficiency and strongly depends on the daily tuning of the ICP-MS instrument. In Wu S, et al. 2022, as an example the correction factors are given and vary between 1.063 and 1.189 for 25 sessions. We observed similar values, but other instruments might have different values and different ranges.
• The correct assessment of downhole fractionation using calcite is difficult due to variable initial Pb content, possible surface Pb contamination, and the generally small amount of change in the ratio due to the low drill rate and large craters. Again, an example is given in Wu S, et al. for WC-1, we observe similar results.

The present manuscript focuses on the characterisation of RA-138 with ID-TIMS and LA-ICP-MS including the repeatability and does not specifically deal with the data reduction method and matrix effects. This was already described in Guillong et. al 2020 and Wu, S et al 2022 and we do not think repeating this content would add value to this manuscript.

L112: We agree and will change in the revised version of the manuscript.

L137: We acknowledge that the choice of ²³⁸U/²³⁵U measured in accessory minerals by Hiess et al. (2012) is not ideal. This value corresponds to δ²³⁸U = -0.19 +/- 0.3 ‰ (2 sigma, relative to CRM 112a) and comes with a large uncertainty which is propagated into every date. We did not measure ²³⁸U/²³⁵U in our aliquots and have no independent knowledge of the appropriate value; however, compilations of δ²³⁸U in modern and ancient marine carbonates show significant variability in natural systems. For example, Chen et al. (2021) compiled data where Phanerozoic carbonates are on average δ²³⁸U = -0.37 ‰, with a significant spread of ca. 0.6 ‰ (2 SD). We do not know whether this could be random (i.e. compositions vary for each of our aliquots) but suspect that U isotopic composition should be coherent at least within the same cement generation. As such, a deviation of δ²³⁸U from the Hiess value should result in a systematic error, as suggested by the reviewer.

The magnitude of this error can be estimated. Assuming that RA138 has δ²³⁸U = -0.37 of average Phanerozoic carbonate instead of -0.19 results in a negligible shift of the isochron age of 10 ka (30 ppm relative) towards younger values and has no effect on the value of the common Pb intercept. As such, in this particular case and at the currently achievable level of precision in ID-TIMS, the exact ²³⁸U/²³⁵U of the carbonate can be neglected.

We note, however, that future studies employing high-precision U-Pb geochronology of (particularly old) carbonates should consider directly analysing ²³⁸U/²³⁵U in the unknowns to establish this value. In younger samples, estimates of initial ²³⁴U/²³⁸U disequilibria are likely to be the dominant limitation to accuracy; however, in developing RA138 as a reference material we are interested in the raw date before any disequilibrium corrections rather than its true age.

Chen, F. L. H. Tissot, M. F. Jansen, A. Bekker, C. X. Liu, N. X. Nie, G. P. Halverson, J. Veizer, N. Dauphas, The uranium isotopic record of shales and carbonates through geologic time. Geochimica et Cosmochimica Acta 300, 164-191 (2021).

L204: We agree that for the ID-TIMS results showing some clear evidence for containing non-C1 components within the D16 domain would be preferable, however this is difficult to achieve as all the solutions have either been used or were disposed so a direct analysis for elevated Mg, Mn and Fe is not possible. A high-resolution image of the sample prior to the micro drill sampling does not exist and would not help as we selected the sampling location based on the CL image (high resolution) presented in Figure 3 and we (obviously) target the reliable domains. However, the CL image is from the surface only, and the micro drill sampling removes several 100 to 1000s of micrometres into the sample. On the CL image there is a partial cut from the sawing between sampling locations 4 and 6 revealing a finite depth of the C1 phase. Based on this observation we think it is valid to suggest that the micro drilling might have touched a different phase. Another CL image would reveal this, unfortunately the sample was used for other experiments and extensive micro drilling.

Fig 2. B: We will extend the x scale to include the intercept of the regression with the Concordia curve for the revised version of the manuscript