Skip to main content
Sandbox Physics
Stellar spectra from dark lines to abundance models

AC03 · L4 · Paper guide

Read a 3D non-LTE abundance paper

Sandbox Physics ·

Open Amarsi, Liljegren and Nissen, arXiv:2209.13449v3, beside this guide. Follow section 2 and figures 2, 3, 8 and 10 to understand why the same lines can imply different iron abundances. This is a source-reading exercise; the three-dimensional atmosphere and non-LTE solver have not been rerun here.

In section 2, separate two model changes

Make two columns in your notes. One describes the atmosphere: one dimension compresses a stellar surface into a structure varying with depth, whereas three-dimensional fluid models allow bright rising regions, cooler descending regions and spatially varying speeds. The other describes atomic populations: LTE sets them using local temperature, whereas non-LTE balances radiative and collisional transitions among levels. These choices can be varied separately.

Read the atmospheric setup in section 2.1, the model atom in 2.2 and radiative transfer in 2.3. Record parameter coverage, atomic-data sources and convergence criteria in separate places. Fe I means neutral iron; Fe II means iron missing one electron. Their differing responses help explain why the later figures distinguish them.

Keep microturbulence on its own line. A one-dimensional calculation uses a prescribed small-scale speed to represent unresolved motions. A three-dimensional calculation has its own velocity field. If the one-dimensional speed changes as well, line width and saturation change too. Record that contribution separately when interpreting the abundance difference.

Reading: Amarsi, Liljegren & Nissen (2022), 3D non-LTE iron abundances in FG-type dwarfs, arXiv:2209.13449v3

Start figure 2 with one line shape

Check the horizontal axis first. The paper plots air wavelength λAir\lambda_{\rm Air} in nanometres. Our teaching model and CALSPEC excerpt use vacuum wavelengths. Overlaying those coordinates without conversion would introduce a wavelength mismatch. The vertical axis is normalized flux, with continuum at one.

Within one panel, hold color fixed and compare dashed with solid curves to isolate LTE versus non-LTE. Next hold line style fixed and compare atmospheric dimensions. Only then move between temperature or metallicity panels. This sequence gives each comparison a clear changing condition.

Notice the integration boundary. A wider interval may collect missing light from neighboring transitions. Section 2.4 discusses blending and the line selection. When recording an equivalent width yourself, record both integration limits too. This makes the light assigned to the feature explicit.

Reading: Amarsi, Liljegren & Nissen (2022), 3D non-LTE iron abundances in FG-type dwarfs, arXiv:2209.13449v3

Check the subtraction before interpreting figure 3

The defined difference is ΔA1L−3N=A1D,LTE−A3D,nonLTE\Delta A_{\rm 1L-3N}=A_{\rm 1D,LTE}-A_{\rm 3D,nonLTE}. To transform a one-dimensional LTE abundance into the three-dimensional result, subtract that difference. A negative plotted value means the three-dimensional abundance is higher.

Work an example before reading the colors. If A1L=5.20A_{\rm 1L}=5.20 and ΔA=−0.30\Delta A=-0.30, then A3N=5.50A_{\rm 3N}=5.50. An increase of 0.3 dex corresponds to roughly twice the atom-number ratio. Writing the subtraction explicitly prevents reversing every correction in an analysis.

Figure 3 plots log⁡(W/λ)\log(W/\lambda) horizontally: equivalent width divided by wavelength, followed by a base-ten logarithm. The vertical axis is the abundance difference in dex, a difference of decimal logarithms. Compare the microturbulence panels and locate where lines become partly saturated. Different models can bend their growth curves at different places, requiring different atom columns to reproduce one equivalent width.

A(Fe)=log⁡10(NFeNH)+12,A3N=A1L−ΔA1L−3NA(\mathrm{Fe})=\log_{10}\left(\frac{N_{\mathrm{Fe}}}{N_{\mathrm H}}\right)+12,\qquad A_{\rm 3N}=A_{\rm 1L}-\Delta A_{\rm 1L-3N}
N denotes atom number. The convention assigns hydrogen abundance twelve. The teaching model’s column multiplier uses a separate artificial opacity scale.

Reading: Amarsi, Liljegren & Nissen (2022), 3D non-LTE iron abundances in FG-type dwarfs, arXiv:2209.13449v3

Use figure 8 for interpolation and figure 10 for stars

The upper panel of figure 8 compares interpolated abundance differences with the grid calculations; the lower panel distributes interpolation residuals. Ask which samples were held out, then identify where errors grow. Agreement with a diagonal checks how well a substitute predicts its underlying grid. Atomic collision data and atmospheric structure need other evidence.

Figure 10 changes to line-by-line stellar analysis. The horizontal lower-level excitation energy describes the atom’s energy before absorption. The vertical abundance is inferred separately from each line. Within one star, compare Fe I and Fe II scatter and the slope with excitation energy. A useful model brings these different lines into better agreement.

A flatter regression is insufficient on its own: temperature changes excitation populations and can change the slope too. Return to section 5.1 and identify the external stellar parameters. The stronger comparison asks whether one atmosphere can explain different lines while respecting information measured by other methods.

Reading: Amarsi, Liljegren & Nissen (2022), 3D non-LTE iron abundances in FG-type dwarfs, arXiv:2209.13449v3

Carry a short reading procedure to the next paper

First record atmosphere dimension, population treatment, atomic data and air or vacuum wavelengths. Then note the measurement windows, continuum and instrument or velocity broadening. Expand every abundance correction into a subtraction with the model names written out, and perform one numerical sign check.

Separate observational uncertainty, interpolation error and physical-model differences. The author-linked grid is a starting point for further study; inspect its release and parameter coverage before extrapolating. Full three-dimensional transfer belongs offline. The browser slab helps explain one mechanism within that larger calculation.

Reading: Amarsi, Liljegren & Nissen (2022), 3D non-LTE iron abundances in FG-type dwarfs, arXiv:2209.13449v3 · Amarsi et al., public non-LTE grids

Take the growth curve back to the paper

  1. Increase column in the cool-layer model and find how saturation bends its growth curve.
  2. Change temperature and instrumental resolution. Separate changes of intrinsic area from changes of recorded shape.
  3. Return to figure 3 and explain why changing the one-dimensional microturbulence changes the abundance difference at fixed equivalent width.

Exercises with explanations

  1. For the paper’s ΔA1L−3N=−0.2\Delta A_{\rm 1L-3N}=-0.2, what is added to the one-dimensional LTE abundance?

    Read the explanation

    Add 0.2 dex. The target abundance equals the original minus the defined negative difference.

  2. Does a tiny interpolation residual establish an accurate abundance for every star?

    Read the explanation

    It checks the substitute against held-out grid calculations. The underlying atomic data, atmosphere and applicability to a target still need independent assessment.

Sources and calculation scope

  1. Amarsi, Liljegren & Nissen (2022), 3D non-LTE iron abundances in FG-type dwarfs, arXiv:2209.13449v3

    Version three, 9 December 2022: sections 2.1–2.4 and 3.1, figures 2, 3, 8 and 10. The paper-reading guide checks conventions and sensitivity; it does not rerun the 3D solver.

  2. Amarsi et al., public non-LTE grids

    Author-linked model-data entry for later study; files are not loaded or reproduced by the browser Lab.

FeedbackSomething broken, confusing, or missing on this page?Tell us
What kind of feedback?

Sent with this page’s address, its Lab ID, the site language and your device type. Please leave out personal details. Feedback is kept for up to 12 months.Privacy policy