In Figure 2b,c and Figure 5c, in which we have presented the octahedral mismatches (Δμ), formation energies (E_f), and room-temperature vibrational anharmonic scores (σ⁽²⁾) as a function of the octahedral (μ̅) and tolerance (t) factors, respectively, for halide double perovskites, different markers were used aiming to differentiate the halide anions. This information was further presented in Figure S3 of the Supporting Information, where we had broken down the same plot as Figure 5c into four subplots, one for a group of halide perovskites with a specific halide anion. Originally, it was observed on these t – μ̅ plots that points corresponding to double perovskites with different halides were overlapping on each other and cannot be well differentiated. It was later discovered that this was caused by an unexpected behavior in the Python code that is used to generate these plots, in which the entire set of data is replicated, resulting in the data becoming identical across all four halides. The revised Figures 2 and 5 are now provided below, and Figure S3 is also updated in the revised Supporting Information, in which the points corresponding to different halides can be well distinguished. It can be clearly seen from the revised Figure S3 that as the mass of the halide anion increases, the points move toward the lower left part of the t – μ̅ plots, with decreasing number of structures that are geometrically fit into a cubic perovskite structure. However, please note that these corrections do not affect the main conclusions of this work. In particular, the general trends of how Δμ, E_f, and σ⁽²⁾ are determined by t and μ̅ did not vary with such a correction. Figure 2. Landscape of HDPs plotted against the octahedral (μ̅) and tolerance (t) factors. Each data point on the landscape represents an HDP that has been screened in this work, which is color-coded according to the (a) chemical identity of the A-site cation, (b) magnitude of the octahedral mismatch (Δμ), and (c) formation energy (E_f) of every HDP. For comparison, in (a), halide single perovskites that have been screened in the previous work (1) have also been included (in light blue crosses). Areas enclosed by the dashed lines are the maximal stability regions of the perovskites identified by Filip and Giustino. (2) (d) [Top row] t – μ̅ plots highlighting the positions of the top seven most stable A₂MM′X₆ (X = F, Cl, and Br) HDP systems with decreasing stabilities (from left to right), as predicted from the machine learning model reported by Li et al. (3) [Bottom row] locations of the fluoride double perovskites on the t – μ̅ plot, which are the systems with the lowest E_f, as shown in (c). Figure 5. Landscape of room-temperature vibrational anharmonicities σ⁽²⁾ as a function of formation energies E_f for (a) double perovskites and (b) single perovskites [adopted from Yang and Li]. (1) In (c), the σ⁽²⁾ values for double perovskites are plotted against the octahedral (μ̅) and tolerance (t) factors (as in Figure 2), where each point is color-coded with respect to the magnitude of σ⁽²⁾ for each compound. Different markers are used for differentiated HDPs with different halide anions. The corresponding data analysis Python code has been updated in our open-source repository (https://github.com/yangjackie/futuremat_public). A set of Jupyter notebooks have also been developed and are hosted in the open repository (https://github.com/yangjackie/halide_double_perovskite_anharmonicity_database) for readers interested in further exploring the data presented in this work interactively. We sincerely apologize to the Editors and readers of Chemistry of Materials for this mistake that we overlooked. We will pay stricter attention to the accuracy of our data analysis in the future. The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.chemmater.2c03276. Exemplary structures of cubic perovskites distorted along the imaginary phonon modes, phonon dispersion curves for Cs₂AgBiBr₆ calculated with two different K-point settings, individual halide, relationship between ⟨ω⟩_σ and E_f, heat plots of σ⁽²⁾ for individual HDPs categorized by the combinations of different A-/X-site ions, table containing the raw data, and plots of timedependent electronic dynamics (PDF) Most electronic Supporting Information files are available without a subscription to ACS Web Editions. Such files may be downloaded by article for research use (if there is a public use license linked to the relevant article, that license may permit other uses). Permission may be obtained from ACS for other uses through requests via the RightsLink permission system: http://pubs.acs.org/page/copyright/permissions.html. This article references 3 other publications. This article has not yet been cited by other publications. Figure 2. Landscape of HDPs plotted against the octahedral (μ̅) and tolerance (t) factors. Each data point on the landscape represents an HDP that has been screened in this work, which is color-coded according to the (a) chemical identity of the A-site cation, (b) magnitude of the octahedral mismatch (Δμ), and (c) formation energy (E_f) of every HDP. For comparison, in (a), halide single perovskites that have been screened in the previous work (1) have also been included (in light blue crosses). Areas enclosed by the dashed lines are the maximal stability regions of the perovskites identified by Filip and Giustino. (2) (d) [Top row] t – μ̅ plots highlighting the positions of the top seven most stable A₂MM′X₆ (X = F, Cl, and Br) HDP systems with decreasing stabilities (from left to right), as predicted from the machine learning model reported by Li et al. (3) [Bottom row] locations of the fluoride double perovskites on the t – μ̅ plot, which are the systems with the lowest E_f, as shown in (c). Figure 5. Landscape of room-temperature vibrational anharmonicities σ⁽²⁾ as a function of formation energies E_f for (a) double perovskites and (b) single perovskites [adopted from Yang and Li]. (1) In (c), the σ⁽²⁾ values for double perovskites are plotted against the octahedral (μ̅) and tolerance (t) factors (as in Figure 2), where each point is color-coded with respect to the magnitude of σ⁽²⁾ for each compound. Different markers are used for differentiated HDPs with different halide anions. This article references 3 other publications. The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.chemmater.2c03276. Exemplary structures of cubic perovskites distorted along the imaginary phonon modes, phonon dispersion curves for Cs₂AgBiBr₆ calculated with two different K-point settings, individual halide, relationship between ⟨ω⟩_σ and E_f, heat plots of σ⁽²⁾ for individual HDPs categorized by the combinations of different A-/X-site ions, table containing the raw data, and plots of timedependent electronic dynamics (PDF) Most electronic Supporting Information files are available without a subscription to ACS Web Editions. Such files may be downloaded by article for research use (if there is a public use license linked to the relevant article, that license may permit other uses). Permission may be obtained from ACS for other uses through requests via the RightsLink permission system: http://pubs.acs.org/page/copyright/permissions.html.

中文翻译：

---

无机卤化物双钙钛矿室温动态稳定性图的校正

在图 2b、c 和图 5c 中，我们展示了八面体失配 (Δμ)、形成能量 ( E _f ) 和室温振动非谐波分数 (σ ⁽²⁾ ) 作为八面体 (μ̅) 的函数和耐受性 ( t ) 因素，分别用于卤化物双钙钛矿，使用不同的标记来区分卤化物阴离子。此信息在支持信息的图 S3 中进一步介绍，其中我们将与图 5c 相同的图分解为四个子图，一个用于一组具有特定卤化物阴离子的卤化物钙钛矿。最初，它是在这些t上观察到的– μ̅ 图对应于具有不同卤化物的双钙钛矿的点相互重叠，无法很好地区分。后来发现，这是由用于生成这些图的 Python 代码中的意外行为引起的，其中复制了整个数据集，导致所有四种卤化物的数据变得相同。下面提供了修订后的图 2 和图 5，修订后的 Supporting Information 中也更新了图 S3，其中可以很好地区分不同卤化物对应的点。从修改后的图 S3 中可以清楚地看出，随着卤化物阴离子质量的增加，点向t的左下部分移动– μ̅ 图，几何上适合立方钙钛矿结构的结构数量减少。但是，请注意，这些更正不会影响这项工作的主要结论。特别是，Δμ、 E _f和 σ ^(2)如何由t和 μ̅确定的一般趋势并没有随着这种校正而变化。图 2. 根据八面体 (μ̅) 和公差 ( t ) 因素绘制的 HDP 景观。景观上的每个数据点代表已在这项工作中筛选的 HDP，根据 (a) A 位阳离子的化学特性、(b) 八面体失配 (Δμ) 的大小和(c) 形成能 ( E _f) 每个 HDP。为了进行比较，在 (a) 中，还包括了在之前的工作 (1) 中筛选出的卤化物单钙钛矿（浅蓝色十字）。虚线包围的区域是 Filip 和 Giustino 确定的钙钛矿的最大稳定性区域。(2) (d) [顶行] t – μ̅ 图突出显示稳定性降低的前七个最稳定的 A ₂ MM'X ₆ （X = F、Cl 和 Br）HDP 系统的位置（从左到右） ，正如 Li 等人报告的机器学习模型所预测的那样。(3) [底行] t – μ̅ 图上氟化物双钙钛矿的位置，这是具有最低E _f的系统，如（c）所示。图 5. 室温振动非谐性 σ ^{(2)与 (a) 双钙钛矿和 (b) 单钙钛矿的形成能}E _f的函数关系[摘自 Yang 和 Li]。(1) 在 (c) 中，双钙钛矿的 σ ^{(2)值根据八面体 (μ̅) 和容差 (} t ) 因子绘制（如图 2 所示），其中每个点都根据大小进行颜色编码σ ⁽²⁾对于每个化合物。不同的标记用于区分具有不同卤化物阴离子的 HDP。相应的数据分析 Python 代码已在我们的开源存储库 (https://github.com/yangjackie/futuremat_public) 中更新。还开发了一套 Jupyter 笔记本，并托管在开放存储库 (https://github.com/yangjackie/halide_double_perovskite_anharmonicity_database) 中，供有兴趣以交互方式进一步探索这项工作中呈现的数据的读者使用。向Chemistry of Materials的编辑和读者致以诚挚的歉意对于我们忽略的这个错误。以后我们会更加注意数据分析的准确性。支持信息可在 https://pubs.acs.org/doi/10.1021/acs.chemmater.2c03276 免费获得。沿虚声子模式扭曲的立方钙钛矿示例性结构，使用两个不同的K点设置计算的 Cs ₂ AgBiBr ₆的声子色散曲线，单个卤化物，⟨ ω ⟩ _σ和E _f之间的关系， σ ⁽²⁾的热图对于按不同 A 位/X 位离子的组合分类的单个 HDP，包含原始数据的表格和时间相关电子动力学图 (PDF) 大多数电子支持信息文件无需订阅 ACS 网络版即可获得。此类文件可以按文章下载以供研究使用（如果相关文章有链接的公共使用许可，则该许可可能允许其他用途）。可以通过 RightsLink 许可系统请求从 ACS 获得用于其他用途的许可：http://pubs.acs.org/page/copyright/permissions.html。本文引用了其他 3 篇出版物。这篇文章尚未被其他出版物引用。图 2. 根据八面体 (μ̅) 和公差 ( t)绘制的 HDP 景观）因素。景观上的每个数据点代表已在这项工作中筛选的 HDP，根据 (a) A 位阳离子的化学特性、(b) 八面体失配 (Δμ) 的大小和(c)每个 HDP 的形成能 ( E _{f )。}为了进行比较，在 (a) 中，还包括了在之前的工作 (1) 中筛选出的卤化物单钙钛矿（浅蓝色十字）。虚线包围的区域是 Filip 和 Giustino 确定的钙钛矿的最大稳定性区域。(2) (d) [顶行] t – μ̅ 图突出了前七名最稳定的 A ₂ MM′X _6的位置（X = F、Cl 和 Br）稳定性降低的 HDP 系统（从左到右），正如 Li 等人报告的机器学习模型所预测的那样。(3) [底行] t – μ̅ 图上氟化物双钙钛矿的位置，这是具有最低E _f的系统，如 (c) 所示。图 5. 室温振动非谐性 σ ^{(2)与 (a) 双钙钛矿和 (b) 单钙钛矿的形成能}E _f的函数关系[摘自 Yang 和 Li]。(1) 在 (c) 中，双钙钛矿的 σ ⁽²⁾值针对八面体 (μ̅) 和公差 ( t) 因子（如图 2 所示），其中每个点都根据每个化合物的 σ ⁽²⁾大小进行颜色编码。不同的标记用于区分具有不同卤化物阴离子的 HDP。本文引用了其他 3 篇出版物。支持信息可在 https://pubs.acs.org/doi/10.1021/acs.chemmater.2c03276 免费获得。沿虚声子模式扭曲的立方钙钛矿示例性结构，使用两个不同的K点设置计算的 Cs ₂ AgBiBr ₆的声子色散曲线，单个卤化物，⟨ ω ⟩ _σ和E _f之间的关系， σ ⁽²⁾的热图对于按不同 A 位/X 位离子的组合分类的单个 HDP，包含原始数据的表格和时间相关电子动力学图 (PDF) 大多数电子支持信息文件无需订阅 ACS 网络版即可获得。此类文件可以按文章下载以供研究使用（如果相关文章有链接的公共使用许可，则该许可可能允许其他用途）。可以通过 RightsLink 许可系统请求从 ACS 获得用于其他用途的许可：http://pubs.acs.org/page/copyright/permissions.html。