新模型助力理解系外行星上云的形成和组成

新模型助力理解系外行星上云的形成和组成

Understanding the composition of clouds on exoplanets

行星形成和演化的秘密很有可能藏在系外行星的大气层中。透射光谱和辐射光谱中的分子线能够揭示行星的化学组成。然而，在系外行星大气中普遍存在的云为光谱的分析过程增添了不确定性。云会抹平分子的谱线，给证认谱线带来挑战；但也会有自己的吸收特征，并改变温度结构，为限制大气成分提供另一视角。因此，在对光谱进行回溯分析时，我们同时要求云的模型计算速度较快，也要求它在物理和化学上的自洽性。然而，当今使用的很多云的模型将云彻底参数化了，忽略了云由气体凝结而来、和气体有不可分割的联系。

最近，清华大学天文系博士生黄赫龙领衔开发了一个用于系外行星光谱回溯分析的云模型。该模型，ExoLyn（霖），求解了多种组分同时凝结时云的垂直输运方程。给定行星的温度结构和化学组成，ExoLyn能够计算出各个高度下云的大小、组成，并且研究云的形成对气体的反馈作用。

热木星的云层结构。左图的粗线展示了大气中云的质量分数，线的颜色代表了该高度下的主导成分。左图的虚线、点线和紫色实线代表了蒸汽、凝结核的质量分数和云的颗粒大小。对于该热木星，云在0.01bar以下为光学厚的（如灰色区域所示）。右边的图展示了不同高度下云的成分。

热木星是一种具有木星大小，但是离主星非常近的行星。ExoLyn被广泛地测试于它们上。这些行星的温度特别高，在深层大气中，即使硅和铁都会蒸发。这些蒸发的铁和硅在高层大气上凝结，形成硅酸盐云和铁云。黄赫龙发现在太阳金属丰度的热木星上，硅酸镁、镁橄榄石和二氧化硅都会同时存在于云中，但是富金属的热木星更倾向于形成二氧化硅云。ExoLyn形成的云的结构和更复杂的模型一致，但是只需不到2秒即可计算完成。云里面的硅酸盐或二氧化硅成分会在光谱的特定波长上留下痕迹，例如硅酸镁的特征吸收在10微米处，而二氧化硅的特征吸收在8和12微米处。因此，ExoLyn可以帮助光谱分析，告诉我们不同行星上云成分的不同。除了热木星之外，通过改变云的凝结反应，ExoLyn还能应用于自发光行星和亚海王星上。在亚海王星上，NaCl和KCl云会形成，将光谱抹平，就如同对于GJ1214 b的观测一样。

不同组分热木星的透射光谱。深色和浅色线分别代表有云和无云的光谱。对于太阳丰度的热木星（黑线），10微米处的吸收峰对应着硅酸盐云的吸收；当热木星金属丰度为10倍太阳时，8微米和12微米处会出现SiO2 云的吸收线；C/O=0.9的热木星上主要形成铁云，因此谱线会被抹平。

该工作发表于杂志《天文学和天体物理》。清华大学天文系2022级博士生黄赫龙是第一作者和通讯作者。这项工作的合作者有清华大学天文系副教授Chris Ormel和荷兰空间研究所（SRON）的Michiel Min。

本工作由国家自然科学基金资助。

文章链接：

•

Exoplanet's atmosphere contains the information where the planet is formed and how it accretes and evolves. Probing the atmosphere from the molecular lines in the transmission and emission spectrum unveils the chemical composition of the planet. However, an uncertain factor in retrieving exoplanet spectra is the presence of clouds. Clouds affect the transmission and emission spectrum by smearing out molecular features, displaying their own absorption features or inducing temperature inversion. To obtain the composition of the atmosphere from spectrum, forward cloud models are run millions of times, a process called retrieval. This requires the cloud model to run fast, preferentially computed within seconds. Currently many retrieval analyses simply impose clouds, neglecting the fact that clouds form by the condensation of gas species, and ignoring their connection to the composition of the exoplanet atmosphere gas.

Recently, a cloud model tailored to retrieval studies of exoplanet spectroscopic retrieval, have been developed by a team led by Helong Huang, a PhD student in the Department of Astronomy, Tsinghua University. The new cloud model, named ExoLyn, considers the simultaneous condensation of multiple species onto a cloud particle and follows the vertical diffusive and settling motions of particles in the atmosphere. Given the planet temperature and bulk composition, ExoLyn computes the cloud particles’ composition and sizes at every layer in the atmosphere and studies how the formation of clouds affect the concentration of vapor species.

Clouds in an atmosphere of a hot Jupiter planet. The thick line in the left panel shows the total cloud mass mixing ratio down the atmosphere. The color represents the dominant species. The colored dashed line, dotted line, and purple lines shows the vapor, nuclei concentration, and the particle size, respectively. The clouds are optically thick below 0.01 bar, as shown by the gray region. The right panel presents the composition of the cloud particles at each height.

ExoLyn has been most thoroughly tested on hot Jupiters, a type of exoplanet having Jupiter size and mass, but orbiting closer around their host stars. The temperature on these planets is so high that silicates and iron evaporated deep in the atmosphere and condense in higher layers, forming silicate clouds on the top of iron layer. For a hot Jupiter with solar abundance, the team found that silicate clouds are composed of a mixture of magnesium-silicates, e.g., MgSiO3, Mg2SiO4 and SiO2, while a metal-rich hot Jupiter tends to form SiO2 clouds. These characteristics of clouds generated by ExoLyn agree with more complex models, but only cost the computational time of 2 seconds. The silicate components in the clouds contribute to additional absorption at specific wavelengths, like MgSiO3 at 10 micron and SiO2 at 8 and 12 micron. As a result, ExoLyn can contribute to the interpretation of JWST spectra and enlighten us about what kind of clouds form on the exoplanets. Aside from hot Jupiters, ExoLyn can be applied to other types of planets, like self-luminous planets and sub-Neptunes, by selecting different condensation reactions. On sub-Neptune planets, NaCl and KCl clouds form, and can reproduce a flat near-infrared transmission spectrum, as observed on GJ1214 b.

Transmission spectra of hot Jupiters with different compositions. The cloudy and clear hot Jupiters are contrasted by the dark and light color lines. For the solar abundance case (black line), there is an absorption feature at 10 micron, due to the formation of silicate clouds. The 10 times solar metallicity planet (blue line) shows additional absorption features at 8 and 12 um, corresponding to SiO2 clouds forming in the upper atmosphere. For a hot Jupiter with C/O=0.9 (red line), the transmission spectrum becomes flat due to the absorption of light by Fe particles in the upper atmosphere.

The new work was published in Astronomy & Astrophysics. Helong Huang is the first and corresponding author. Other co-authors include Prof. Chris Ormel of Tsinghua University, and Michiel Min from the Netherlands Institute for Space Research (SRON).

The work of the Tsinghua team was supported by the Natural Science Foundation of China.

Paper Link:

供稿：黄赫龙

审稿：蔡峥 Chris Ormel

清华大学天文系