The study of Xiao et al.(1) deals with the thermal stability of perfluoroalkyl substances (PFAS), in particular with the decomposition pathways of PFAS during thermal reactivation of spent granular activated carbon (GAC). The study provides valuable insights into the behavior of perfluorinated carboxylic acids (PFCAs) and sulfonic acids (PFSAs). It shows clearly that, contrary to common belief, PFCAs start to decompose at relatively low temperatures such as 200 °C, whereas PFSAs are much more stable (decomposition at ≥400 °C). Beside these general statements, we see, however, a number of deficits and misinterpretations in the original Letter that need to be clarified, in particular: (i) unclear speciation of the substrates, (ii) inappropriate data evaluation, and (iii) misleading chemical interpretation of reaction pathways. (i) First, the authors specify their substrates as “acids” throughout the entire original Letter. According to Table S1, PFCAs were applied as free acids whereas PFSAs were applied as potassium salts. One can expect that protonated species and salts behave differently, with respect to volatility and thermal stability. Furthermore, it is likely that PFCAs and PFSAs as strong acids (pK_A = 0–1 and approximately −3, respectively(2−5)) do not occur as protonated species, in aqueous solution or in the adsorbed state. Nevertheless, the authors present TGA data in their Figure 2 for free acids. In our opinion, it would be more appropriate to investigate PFAS salts rather than to compare “artificial” acids with salts. Furthermore, it is very likely that the thermodesorption behavior of adsorbates may depend on their initial loading. Unfortunately, such data are missing for all experiments. (ii) Reliable analytical determination of the target compounds is a basic requirement for quantitative studies. Figure 1 presents recoveries of 11 PFAS obtained by extraction of prespiked GAC samples with methanol under various extraction conditions. The extraction with alkaline methanol yields recoveries between 132% and 261% for all PFCAs. The narrow scattering ranges of approximately ±5% point to quite reproducible results. These “over-recoveries” are neither discussed nor explained in the original Letter. Without a reasonable explanation, the reader cannot be expected to accept all of the other PFCA data. It is a challenge of this study that thermal treatments of PFAS samples are, in most cases, a superposition of physical volatilization and chemical decomposition. The authors describe the outcome of this overlap, named “destabilization”, by means of first-order kinetics (no data shown) and the observed temperature dependence in terms of Arrhenius-like and Eyring-like plots, i.e., ln(k) and ln(k/T) versus 1/T (panels e and f of Figure 2, respectively). For PFOA “destabilization”, the presented data (100–150 °C) reflect, however, pure volatilization rates without any chemical decomposition. We cannot recognize any benefit of applying the Eyring equation to simple volatilization kinetics, because volatilization has nothing to do with chemical transition states. Apart from the scientific issue, the ordinates in panels e and f of Figure 2 are mathematically incorrect. The correct ordinate designation would be ln(k × s) and ln(k/T × s × K). The abscissa units labeled as “°K” are also incorrect. (iii) The chemical evaluation of the experimental findings is summarized in Figure 4, which shows thermal decomposition pathways of perfluorooctanoic acid (PFOA) in an inert atmosphere, based upon thermodesorption pyrolysis GC-MS data. It appears to the reader that the authors interpret GC-detected products in terms of radicals (e.g., “the single sharp peak ... corresponds to the radical of 2H polyfluorocarboxylic acid (2HPFOA)”, page 347 of ref (1)), while it is common knowledge that GC analysis does not detect radicals. After careful inspection of the presented chromatograms and related mass spectra in Figure S3, it appears to be probable that the first (and only) product peak refers to perfluoroheptene (C₇F₁₄) and the second peak refers to the substrate PFOA. The presented mass spectra fit very well to library spectra of these substances. Perfluoroheptene is a plausible primary and highly volatile decomposition product of PFOA. However, the only clearly detected organic product does not appear in the proposed reaction scheme in Figure 4. A closer inspection of Figure 4 reveals that radicals, diradicals, and carbenes play major roles in the proposed reaction scheme. Apparently, the authors interpret fragmentation patterns in the observed mass spectra rather than thermal decomposition products. This impression is supported by given m/z values in the reaction scheme (Figure 4c) that match those from the mass spectra in Figure S3. However, the MS fragmentation patterns are caused by electron-impact ionization rather than by thermal activation. Charged species with meaningful m/z numbers are not involved in low-temperature pyrolysis. Briefly, the reaction scheme in Figure 4 appears to us highly misleading. Altogether, the Letter by Xiao et al.(1) contains a number of interesting findings but also some deficits and misconceptions that should be addressed and clarified. Furthermore, it would be desirable to compare the thermal stabilities of PFAS (salts) in the pure state and as adsorbates on activated carbon and to investigate the mechanism behind thermal degradation. This could be the topic of a full research paper. The authors declare no competing financial interest. The authors declare no competing financial interest.
This article references 5 other publications.

中文翻译：

---

关于“用过的颗粒状活性炭的全氟烷基物质的热稳定性和分解”的评论

Xiao等人（1）的研究涉及全氟烷基物质（PFAS）的热稳定性，特别是用过的粒状活性炭（GAC）热活化过程中PFAS的分解途径。该研究为全氟化羧酸（PFCA）和磺酸（PFSA）的行为提供了有价值的见解。它清楚地表明，与通常的看法相反，PFCA在相对较低的温度（例如200°C）下开始分解，而PFSA则更稳定（在≥400°C时分解）。但是，除了这些一般性陈述外，我们还发现原始信件中存在许多缺陷和误解，需要加以澄清，特别是：（i）基材的种类不明确，（ii）数据评估不当，以及（iii）误导反应途径的化学解释。（我先来，作者在整个原始信函中都将其底物指定为“酸”。根据表S1，PFCA以游离酸的形式施用，而PFSA以钾盐的形式施用。可以预期，就挥发性和热稳定性而言，质子化的物质和盐的行为不同。此外，PFCA和PFSA可能是强酸（pķ_一质子化的物质在水溶液或吸附状态下分别不为0-1和近似-3（2-5））。尽管如此，作者还是在图2中显示了游离酸的TGA数据。我们认为，研究PFAS盐而不是将“人造”酸与盐进行比较会更合适。此外，被吸附物的热脱附行为很可能取决于它们的初始负载。不幸的是，所有实验都缺少这些数据。（ii）对目标化合物进行可靠的分析测定是定量研究的基本要求。图1给出了在各种萃取条件下用甲醇萃取预先加标的GAC样品所得的11种全氟辛烷磺酸的回收率。对于所有PFCA，用碱性甲醇萃取的回收率在132％至261％之间。大约±5％的窄散射范围表明了可重复的结果。这些“过度恢复”未在原始信函中讨论或解释。没有合理的解释，就无法期望读者接受所有其他PFCA数据。这项研究的挑战在于，在大多数情况下，PFAS样品的热处理是物理挥发和化学分解的叠加。作者通过一阶动力学（未显示数据）和观察到的温度依赖性（以类似于阿伦尼乌斯和艾尔环的曲线图，即ln（），描述了这种重叠的结果，称为“稳定化”）。不能期望阅读器接受所有其他PFCA数据。这项研究的挑战在于，在大多数情况下，PFAS样品的热处理是物理挥发和化学分解的叠加。作者通过一阶动力学（未显示数据）和观察到的温度依赖性（以类似于阿伦尼乌斯和艾尔环的曲线图，即ln（），描述了这种重叠的结果，称为“稳定化”）。不能期望阅读器接受所有其他PFCA数据。这项研究的挑战在于，在大多数情况下，PFAS样品的热处理是物理挥发和化学分解的叠加。作者通过一阶动力学（未显示数据）和观察到的温度依赖性（以类似于阿伦尼乌斯和艾尔环的曲线图，即ln（），描述了这种重叠的结果，称为“稳定化”）。k）和ln（k / T）与1 / T（分别为图2的面板e和f）。对于PFOA的“稳定化”，所提供的数据（100–150°C）反映了纯挥发速率，没有任何化学分解。我们无法认识到将Eyring方程应用于简单的挥发动力学的任何好处，因为挥发与化学转变态无关。除科​​学问题外，图2的面板e和f中的纵坐标在数学上是不正确的。正确的纵坐标指定为ln（k ×s）和ln（k / T×s×K）。标为“°K”的横坐标单位也不正确。（iii）实验结果的化学评估总结在图4中，该图基于热解吸热解GC-MS数据显示了全氟辛酸（PFOA）在惰性气氛中的热分解途径。在读者看来，作者用自由基来解释GC检测到的产物（例如，“单个尖峰...对应于2H多氟羧酸（2HPFOA）的自由基”，参考文献（1）的第347页），众所周知，气相色谱分析不能检测到自由基。在仔细检查了图S3中给出的色谱图和相关质谱后，似乎第一个（也是唯一一个）产物峰是指全氟庚烯（C ₇ F ₁₄），第二个峰值是指底物PFOA。所提供的质谱图非常适合这些物质的谱库。全氟庚烯是PFOA的可能的主要且高挥发性的分解产物。但是，在图4的拟议反应方案中没有唯一清楚地检测到的有机产物。仔细观察图4可以发现，自由基，双自由基和卡宾在拟议的反应方案中起主要作用。显然，作者解释了观察到的质谱中的碎裂模式，而不是热分解产物。给定的m / z支持此印象反应方案（图4c）中的值与图S3中质谱的值匹配。但是，MS碎片图谱是由电子碰撞电离而不是热活化引起的。带电物种具有有意义的m / z低温热解过程中不涉及数量。简而言之，我们认为图4中的反应方案具有高度误导性。总体而言，Xiao等人的信（1）包含许多有趣的发现，但也存在一些缺陷和误解，应加以解决和澄清。此外，希望比较纯态的PFAS（盐）的热稳定性以及在活性炭上的吸附物的PFAS（盐）的热稳定性，并研究热降解背后的机理。这可能是完整研究论文的主题。作者宣称没有竞争性的经济利益。作者宣称没有竞争性的经济利益。
本文引用了其他5个出版物。