Infrared luminosity functions based on 18 mid-infrared bands: revealing cosmic star formation history with AKARI and Hyper Suprime-Cam
Summary (3 min read)
1 Introduction
- Mid-infrared (mid-IR) is one of the less explored wavelengths due to the earth's atmosphere, and difficulties in developing sensitive detectors.
- To overcome these problems, the authors have newly obtained deeper optical data over the entire AKARI NEP wide field, using the Hyper-Suprime Cam on the Subaru telescope.
- Using the deeper optical data, in this paper, the authors measure mid-infrared galaxy LFs, and estimate total IR LFs (based on the mid-IR SED fitting) from the entire AKARI NEP field.
2 Data
- To rectify the situation and to fully exploit the AKARI's spacebased data, the authors carried out an optical survey of the AKARI NEP wide field (PI:Goto) using Subaru's new Hyper Suprime-Cam (HSC; Miyazaki et al. 2018) in five optical bands (g, r, i, z, and y, Oi et al. 2018 submitted) .
- The HSC has a field-of-view (FoV) of 1.5 deg in diameter, covered with 104 red-sensitive CCDs.
- It has the largest FoV among optical cameras on 8m-class telescopes, and can cover the AKARI NEP wide field (5.4 deg 2 ) with only 4 FoV (Fig. 1 ).
- See Oi et al. (2018, submitted) for more details of the observation and data reduction.
- Subaru telescope does not have u * -band capability, while it is critically important to accurately estimate photometric redshifts (photo-z) of low-z galaxies.
3 Analysis
- Uncertainties of the LF values include fluctuations in the number of sources in each luminosity bin, the photometric redshift uncertainties, the k-correction uncertainties, and the flux errors.
- To estimate errors, the authors used Monte Carlo simulations from 1000 simulated catalogs.
- A new flux is also assigned following a Gaussian distribution with the width of flux error.
- The smaller data points at the faint ends are adopted from the NEP deep field, where AKARI data are deeper (Goto et al. 2015) , and are included in the fit.
- Vertical arrows show the 8µm luminosity corresponding to the flux limit at the central redshift in each redshift bin.
4.1 The 8µm LF
- The authors first present monochromatic 8µm LFs, because the 8µm luminosity (L8µm) has been known as a good indicator of the TIR luminosity (Babbedge et al.
- Often in previous work, SED based extrapolation was needed to estimate the 8µm luminosity.
- This is not the case for the analysis present in this paper.
- The smaller data points at the faint ends are adopted from the NEP deep field, where AKARI data are deeper (Goto et al. 2015) , and are included in the fit.
- Vertical arrows show the 12µm luminosity corresponding to the flux limit at the central redshift in each redshift bin.
templates (2% from the sample).
- The authors corrected for the completeness using Kim et al. (2012) (25% correction at maximum, with their selection to the 80% completeness limits).
- Then, the 1/Vmax method was used to compensate for the flux limit.
- Various previous studies are shown with dashed lines for comparison.
- Interestingly, the 8µm LFs peaks in the 3rd bin (z∼1), then declines toward z∼2.
4.2 12µm LF
- The 12µm luminosity L12µm) is also known to correlate well with the TIR luminosity (Spinoglio et al.
- AKARI's advantage still holds in not needing extrapolation based on SED models.
- Various previous studies are shown in dash-dotted lines.
- Similar to the 8µm LF, the evolution becomes less evident between the two higher redshift bins.
4.3 Total IR LFs estimated from mid-IR SED fit
- The authors caution readers that estimation of the LTIR involves extrapolation to the far-IR wavelength range based on the SED models, and thus invites associated uncertainty, as they further discuss in Section 5.
- The L18W flux (Matsuhara et al. 2006 ) are used to apply the 1/Vmax method, because it is a wide, sensitive filter (but using the L15 flux limit does not change their main results).
- For clarity, the authors separated LFs in four different panels at each redshift bin.
- The TIR LFs show a strong evolution compared to local LFs, but again turns over at z > 1.2.
4.4.1 Total IR Luminosity Density from L8µm
- LFs First, the authors estimate Total IR Luminosity Density from L8µm LFs.
- Possible SED evolution, and the presence of AGN will induce further uncertainty.
- Murata et al. (2014) also reported that L8µm/LTIR is constant at below the main sequence, while it decreases with starburstiness at above the main sequence, concluding that starburst galaxies have deficient PAH emission compared with main-sequence galaxies.
- Overplotted previous studies are taken from Le Floc'h et al. ( 2005) in the dark-green, dash-dotted line, Magnelli et al. (2013) in the dark-red, dash-dotted line, Huynh et al. (2007) in the dark-yellow, dash-dotted line, Gruppioni et al. (2013) in the pink, dash-dotted line at several redshifts as marked in the figure.
4.4.2 Total IR Luminosity Density from L12µm LFs
- Due to the same reasons as L8µm (improved statistics, and availability of 140 and 160µm), the authors use the following conversion (Goto et al. 2011b) .
- The authors caution readers again here for the use of a single conversion for varieties of galaxies with different SFR at different redshifts.
- Results should be interpreted with this uncertainty in mind.
4.4.3 Integration to TIR density
- The derived total LFs are multiplied by LTIR and integrated to measure the TIR density ( ΩTIR).
- Following their previous work, the authors use a double-power law.
- With the lowest redshift LF, the authors first fit the normalization (Φ * ) and slopes (α, β).
- The authors also note that ΩIR from 12µm is sensitive to the faint-end slope of 12µm LFs.
- Much deeper observations are awaited to clarify the issue.
5 Discussion
- The conversions are based on local star-forming galaxies.
- The pink dashed line shows the total estimate of IR (TIR LF) and UV (Schiminovich et al. 2005) .
- Following the results in the literature discussed in Section 4.3, in this section, the authors compare LT IR estimated from L8µm and L12µm from equations 1 and 2 in three overlapping redshift ranges in Fig. 6 using their data.
- One can immediately notice that the relation deviates at logLTIR >12 (or equivalently at z > 1).
- 4 and 5 between midand far-IR measurements could be the result of the change in the SED, rather than incorrect measurements on either.
6 Summary
- Previously AKARI NEP wide field lacked deep optical photometry, and thereby, accurate photo-z, despite the presence of space-based 9-band mid-IR photometry from AKARI.
- To rectify the situation, the authors have obtained deep optical 5-band imaging covering the entire 5.4 deg 2 of the NEP wide field, using the new Hyper Suprime-Cam mounted on the Subaru 8m telescope.
- Thanks to the large area coverage, the brightends are better-determined.
Did you find this useful? Give us your feedback
Citations
134 citations
50 citations
34 citations
Cites background from "Infrared luminosity functions based..."
...…have been possible for some time for high-z galaxies (e.g., Hughes et al. 1998; Chapman et al. 2005; Magnelli et al. 2013; Gruppioni et al. 2013; Goto et al. 2019), they do not yet provide significant insights at very high redshifts (z>∼ 6), nor for low-mass galaxies which have low SFRs and low…...
[...]
30 citations
17 citations
Cites background from "Infrared luminosity functions based..."
...…galaxy population also known as the cosmic evolution of the star formation rate density (SFRD), is well established up to redshift of ∼9 (e.g. Lilly et al. 1996; Schiminovich et al. 2005; Bouwens et al. 2011, 2014; Madau & Dickinson 2014; Hagen et al. 2015; Alavi et al. 2016; Goto et al. 2019)....
[...]
References
553 citations
"Infrared luminosity functions based..." refers methods in this paper
...4.4 Total IR Luminosity density, ΩIR Using LFs in previous sections, we next compute the IR luminosity density, to estimate the cosmic star formation density (Kennicutt 1998)....
[...]
504 citations
"Infrared luminosity functions based..." refers background in this paper
...The 12µm luminosity L12µm) is also known to correlate well with the TIR luminosity (Spinoglio et al. 1995; Pérez-González et al. 2005)....
[...]
...Previous work also found a stronger evolution in luminosity than in density (Pérez-González et al. 2005; Le Floc’h et al. 2005)....
[...]
...L12µm is also reported to correlate with LTIR (Chary & Elbaz 2001; Pérez-González et al. 2005)....
[...]
483 citations
461 citations
"Infrared luminosity functions based..." refers result in this paper
...This may be qualitatively consistent with previous reports by Herschel that the dust attenuation peaks and declines at z>1.2 (Gruppioni et al. 2013; Burgarella et al. 2013)....
[...]
440 citations
"Infrared luminosity functions based..." refers background in this paper
...Pope et al. (2008) showed that z ∼2 sub-millimeter galaxies lie on the relation between LTIR and LPAH,7.7 that has been established for local starburst galaxies....
[...]
Related Papers (5)
Frequently Asked Questions (6)
Q2. Why do they use the L15 flux?
The L18W flux (Matsuhara et al. 2006) are used to apply the 1/Vmax method, because it is a wide, sensitive filter (but using the L15 flux limit does not change their main results).
Q3. What is the common wavelength of the AKARI NEP?
Mid-infrared (mid-IR) is one of the less explored wavelengths due to the earth’s atmosphere, and difficulties in developing sensitive detectors.
Q4. What are the uncertainties of the LF?
Uncertainties of the LF values includefluctuations in the number of sources in each luminosity bin, the photometric redshift uncertainties, the k-correction uncertainties, and the flux errors.
Q5. What is the largest FoV of a Subaru telescope?
It has the largest FoV among optical cameras on 8m-class telescopes, and can cover the AKARI NEP wide field (5.4 deg2) with only 4 FoV (Fig.1).
Q6. Why is IR from 12m LFs larger?
even with AKARI’s sensitivity, the observation might not be deep enough to reliably measure the faint-end slope of 12µm LFs, possibly because 12µm does not contain as luminous emission lines as in the case of 8µm.