Reply to Comment on “In Situ ion microprobe U‐Pb dating and REE abundances of a Carboniferous conodont” by R. Romer
Reply to Comment on “In Situ ion microprobe U‐Pb dating and REE abundances of a Carboniferous conodont” by R. Romer
复制标题
回复 R. Romer 对“石炭纪牙形刺的原位离子微探针 U-Pb 定年和 REE 丰度”的评论
DOI：
10.1029/2002gl014671
复制
发表时间：
2002
影响因子：
5.2
通讯作者：
K. Terada
中科院分区：
文献类型：
--
作者：
Y. Sano;K. Terada
[1] 我们报道了石炭纪牙形刺的原位离子微探针 U-Pb 定年和稀土元素 (REE) 丰度 [Sano 和 Terada, 2001]。 Romer [2002] 在对该论文的评论中声称，样品中计算出的初始 Pb 同位素组成（2 秒时的 Pb/Pb = 13.0 ± 2.5 和 Pb/Pb = 12.2 ± 1.5）出人意料地具有较低的放射性，不应在石炭纪海洋化石中发现。此外，一些 Pb/Pb 比率的放射源性低于峡谷 Diablo 硫铁矿中的原始 Pb，这与地球的任何 Pb 演化模型相冲突 [例如，Stacey 和 Kramers，1975；卡明和理查兹，1975 年；扎特曼和多伊，1981 年]。 Romer [2002] 认为显着低的 Pb/Pb 比率可能源于分析问题，可能是由于对 Pb 信号的高估，并且有必要检查 Hg 干扰，因为在 Sano 和 Terada [2001] 实验中采用的 5800 质量分辨率下，Hg (203.97348) 无法从 Pb (203.97303) 中分离出来。 [2] 化学溶解和分离后，离子微探针技术相对于传统同位素稀释热电离质谱 (ID-TIMS) 的优势在于保留了样品的重要文本背景，并为其他同步分析工作（例如 REE 丰度）提供了机会 [Sano 等人，1999]。我们使用完全相同的样品进行了离子微探针实验。为了进一步减少氢化物干扰，在广岛大学 SHRIMP II 的样品锁中对环氧树脂圆盘中的石炭纪牙形刺进行了抽真空。 2-nA 的初级 O2 光束聚焦到样品坑直径 30 毫米的区域，Sano 和 Terada [2001] 已经在该区域进行了 U-Pb 分析。使用 10 kV 提取二次离子。磁体从质量 200 (Hg) 循环峰值步进到质量 204 (Pb + Hg)，分别包括 Hg 的背景和 202。 [3] 表 1 列出了石炭系牙形刺的 A/A 和 A/A 比值，但 Cono1.1.3 除外，由于 Sano 和 Terada [2001] 图 3 中明显的异常特征，我们没有对其进行分析。观察到的 A/A 比值与 Hg/Hg 比值 1.29 一致，在 2 sigma 的实验误差范围内，表明样品中可能存在 Hg 同位素，正如 Romer [2002] 所声称的那样。质量 204 处的 Hg 贡献通过 0.296 的 Hg/Hg 比率和观察到的 A/A 比率计算，并列于表 1 中。表观 Hg 干扰在 4% 到 28% 之间变化很大。 Sano 和 Terada [2001] 表 2 中列出的干扰与报告的 Pb 浓度之间没有相关性，表明样品中的 Hg 污染可能不均匀。 [4] 表 2 列出了基于 A/A 比值校正的汞干扰 Pb/Pb 和 Pb/Pb 比值。校正后的 Pb/Pb 比值与 Stacey 和 Kramers [1975] 进化模型估计的 350-3.63 亿年前的常见 Pb 值 15.6 一致，实验误差在 2 sigma 范围内，尽管少数点（Cono1.1.7、Cono1.1.8 和 Cono1.1.10）的 Pb/Pb 比值相对较低。这些铅同位素在地球化学上是合理的值，表明汞干扰是可能的。 [5] 我们根据校正后的Pb同位素，采用York方法计算了Pb* U等时线年龄。由此得出的 323 ± 79 Ma 年龄（2s；MSWD = 0.73，误差相关性 = 0）与化石的沉积年龄（350-363 Ma）一致，实验误差在 2 sigma 之内，尽管与未修正的 323 ± 72 Ma 年龄相同 [Sano 和 Terada，2001]。估计的初始 Pb/Pb 比率为 15 ± 7 (2s)，显示出较大的误差，并且在考虑 Hg 干扰时没有用处。值得注意的是，MSWD 值从原始数据的 1.03 下降到校正数据的 0.73，这意味着拟合概率增加。我们还根据三维线性回归计算总 Pb/U 等时线年龄，该回归被限制为与 Tera-Wasserburg 协和体相交 [Ludwig, 1998]，使用校正的 Pb 同位素数据（图 1）。最佳拟合线得出的年龄为 335 ± 44 Ma (95%CL; MSWD = 0.94)，这与沉积年龄 (350–363 Ma) 一致。该年龄再次与未校正年龄 332 ± 44 Ma 相同，表明高达 28% 的 Hg 污染不会影响 U-Pb 年龄。另一方面，初始 Pb/Pb 比为 15.0 ± 3.1 (2s) 和 Pb/Pb 为 14.0 ± 2.0 (2s)，分别显着大于基于原始数据的 13.0 ± 2.5 和 12.2 ± 1.5 [Sano 和 Terada，2001]。它们与 350–363 Ma 中常见的 Pb/Pb 比值 18.1 和 Pb/Pb 比值 15.6 略有吻合 [Stacey 和 Kramers, 1975]。 [6] Romer [2002] 认为样品的 Pb/Pb 比率应受到 Hg 干扰的影响，并转变为低于 Pb 演化比率的值，即 350–363 Ma 中的 37.9 [Stacey 和 Kramers, 1975]。在 Sano 和 Terada [2001] 的实验中，我们使用 60 秒的长积分时间测量了第一和第二点（Cono1.1.4 和 1.1.5）的 Pb。然后我们显着缩短了时间，其他点只有一秒，因为看起来 Th 丰度极小，无法获得 Th-Pb 年龄。 Cono1.1.4和1.1.5的原始Pb/Pb比分别为35.3±15.8（2s）和30.5±18.2（2s）。当校正 9% 和 27% 的汞干扰后，它们分别变为 38.8 和 41.8。校正后的Pb/Pb比值与常见的Pb 37.9一致。 [7] 最后，Romer [2002] 提出了两个主要问题。首先，使用分析上不正确的数据来反演 U-Pb 年龄是不合适的。我们同意这一点，并且当 Pb 丰度非常小时，使用离子微探针检查质量 200 和 202 是否存在 Hg 干扰非常重要。第二个年龄的实验误差（335±44Ma），关于地球物理研究快报，卷。 29、没有。 12, 1598, 10.1029/2002GL014671, 2002
[1] We have reported in situ ion microprobe U-Pb dating and rare earth element (REE) abundance of a Carboniferous conodont [Sano and Terada, 2001]. In his comment on the paper, Romer [2002] claims that the initial Pb isotopic compositions (Pb/Pb = 13.0 ± 2.5 and Pb/Pb = 12.2 ± 1.5 at 2s) calculated in the sample are unexpectedly less radiogenic and should not be found in Carboniferous marine fossil. Furthermore a few Pb/Pb ratios are less radiogenic than that of primordial Pb in Canyon Diablo troilite, which is in conflict with any Pb evolution models of the Earth [e.g., Stacey and Kramers, 1975; Cumming and Richards, 1975 Zartman and Doe, 1981]. Romer [2002] suggests that the significantly low Pb/Pb ratios may originate from analytical problems, probably due to the overestimation of Pb signal and it is necessary to check Hg interference since Hg (203.97348) is not resolved from Pb (203.97303) at the mass resolution of 5800 adopted in Sano and Terada [2001] experiment. [2] The advantage of the ion microprobe technique over the conventional isotope dilution thermal ionization mass spectrometry (ID-TIMS) after chemical dissolution and separation is preservation of the important textual context of sample and to provide an opportunity for other simultaneous analytical work such as REE abundances [Sano et al., 1999]. We have carried out ion microprobe experiment using exactly the same sample. The Carboniferous conodont cast in epoxy resin discs was evacuated in the sample lock of the SHRIMP II at Hiroshima University overnight in order to further reduce hydride interference. A primary O2 beam of 2-nA was focused to a 30-mm-diameter area of the sample pit where the U-Pb analysis were already made by Sano and Terada [2001]. Secondary ions were extracted using 10 kV. The magnet was cyclically peak-stepped from mass 200 (Hg) to mass 204 (Pb + Hg), including the background and 202 for Hg, respectively. [3] Table 1 lists the A/A and A/A ratios of Carboniferous conodont except for Cono1.1.3 which we did not analyze because of apparent outlier characteristics in Figure 3 of Sano and Terada [2001]. Observed A/A ratios are consistent with the Hg/Hg ratio of 1.29 within experimental error at two sigma, indicating that there may exist Hg isotopes in the sample as claimed by Romer [2002]. Contribution of Hg at mass 204 is calculated by the Hg/Hg ratio of 0.296 and observed A/A ratio and listed in Table 1. The apparent Hg interferences vary significantly from 4% to 28%. There is no correlation between the interferences and reported Pb concentrations listed in Table 2 of Sano and Terada [2001] suggesting that Hg contaminations are probably not uniform in the sample. [4] Table 2 lists corrected Pb/Pb and Pb/Pb ratios for Hg interference based on the A/A ratios. The corrected Pb/ Pb ratios are consistent with that of common Pb value, 15.6 in 350–363 million years ago estimated by Stacey and Kramers [1975] evolution model within experimental error of two sigma, even though the ratios of a few spots (Cono1.1.7, Cono1.1.8, and Cono1.1.10) show relatively lower Pb/Pb ratios. These Pb isotopes are geochemically reasonable values, suggesting that the Hg interferences are possible. [5] We calculate the Pb* U isochron age based on the corrected Pb isotopes by using the York method. Resultant age of 323 ± 79 Ma (2s; MSWD = 0.73, error correlation = 0) is consistent with the depositional age of the fossil (350–363 Ma) within experimental error of two sigma and even though identical to the uncorrected age of 323 ± 72 Ma [Sano and Terada, 2001]. The estimated initial Pb/Pb ratio of 15 ± 7 (2s) shows large error and is not useful when considering the Hg interference. It is noted that the MSWD value decreases from 1.03 of original data to 0.73 of corrected ones, implying the increase of probability of fit. We also calculate a total Pb/U isochron age based on threedimensional linear regressions conducted as constrained to intersect Tera-Wasserburg concordia [Ludwig, 1998] using corrected Pb isotope data (Figure 1). A best-fit line yields the age of 335 ± 44 Ma (95%CL; MSWD = 0.94), which agrees with the depositional age (350–363 Ma). Again the age is identical to the uncorrected age of 332 ± 44 Ma, suggesting that the Hg contamination of up to 28% can not affect the U-Pb ages. On the other hand, initial Pb/Pb ratio of 15.0 ± 3.1 (2s) and Pb/Pb of 14.0 ± 2.0 (2s) are significantly larger than those of 13.0 ± 2.5 and 12.2 ± 1.5 based on the original data [Sano and Terada, 2001], respectively. They show marginal agreement with the common Pb/Pb ratio of 18.1 and Pb/Pb of 15.6 in 350–363 Ma [Stacey and Kramers, 1975]. [6] Romer [2002] suggests that the Pb/Pb ratios of the sample should be affected by the Hg interference and are shifted to lower values than the ratio of Pb evolution, 37.9 in 350–363 Ma [Stacey and Kramers, 1975]. In the experiment of Sano and Terada [2001] we have measured Pb for the first and second spots (Cono1.1.4 and 1.1.5) using a long integration time of 60 sec. Then we have reduced the time significantly, only one second for the other spots, since it has appeared that Th abundances were extremely small and the Th-Pb age would not be obtained. Original Pb/Pb ratios of Cono1.1.4 and 1.1.5 were 35.3 ± 15.8 (2s) and 30.5 ± 18.2 (2s), respectively. They become 38.8 and 41.8 when corrected for the Hg interference of 9% and 27%. The corrected Pb/Pb ratios are consistent with the common Pb of 37.9. [7] Finally two major concerns are given by Romer [2002]. First it is not appropriate to use analytically incorrect data to retrieve the U-Pb age. We agree with the point and it is important to check mass 200 and 202 for Hg interference by using ion microprobe when the Pb abundance is significantly small. Second the experimental error of the age (335 ± 44 Ma), about GEOPHYSICAL RESEARCH LETTERS, VOL. 29, NO. 12, 1598, 10.1029/2002GL014671, 2002