The metallicity of the intergalactic medium and its evolution

1
The metallicity of the intergalactic medium and
its evolution

• Anthony Aguirre
• UCSC

2
The intergalactic medium
The Lya forest
3
The intergalactic medium
The Lya forest
4
The intergalactic medium
Metals in the IGM!
5

• IGM metallicity provides information on
• History of star/galaxy formation.
• Formation of unobservably early stars/galaxies.
• UV ionizing background.
• Feedback in galaxy formation processes.

6
Ways to get enriched

7
Ways to get enriched two straw-man models

• 1. Late enrichment by 2 lt z lt 6 galaxies.
  Strong feedback during galaxy-formation epoch.
• Observed z 3 galaxies drive winds that seem
  likely to escape.
• Semi-analytics and simulations gas removal seems
  necessary during galaxy formation.
• Most of cosmic star formation at z lt 5.

8
Ways to get enriched two straw-man models

• 2. Early enrichment at z gtgt 5. Metals just
  sprinked in with no effect on galaxies or IGM
  at z lt 5.
• Easier escape from small potential wells.
• Larger filling factor?
• Would not disrupt IGM (as not observed).

9
Signatures of early vs. late in observed IGM.

• Look for evolution in Z at z lt 5.
• Check temperature of gas (late enrichment should
  come with/in hot gas).
• Compare amount of metals with expectations.
• Look at spatial distribution of metals.
• Look at abundance ratios for info. on
  nucleosynthetic sources.

Pixel statistics
All this and more can be done with
10
Pixel method (short version)
UVB model
Results
19x
HI, CIV, SiIV pixel optical depths
See Aguirre et. al. 2002 2004 Schaye et al. 2003
Hydro. simulations
11
Results Carbon metallicities from CIV

• 1. The carbon metallicity is inhomogeneous.
• At fixed d and z, p.d.f. for C/H is gaussian,
  i.e. carbon metallicity distribution is
  lognormal.
• Characterize by C/H and s(C/H)

12
Results Carbon metallicities from CIV

• 1. The carbon metallicity is inhomogeneous.
• Primordial enrichment is ruled out.
• But early vs. late will require detailed
  modeling.

13
Results Carbon metallicities from CIV

• 2. The median carbon metallicity C/H changes
  with density.

So does scatter s(C/H)
14
Results Carbon metallicities from CIV

• The median carbon metallicity C/H changes with
  density.
• Expected and reasonable, but never observed.
• But again, early vs. late will require detailed
  modeling.

15
Results Carbon metallicities from CIV

• 3. There is Carbon in underdense gas.
• 2.4s detection in medians
• 3.4s detection in higher
  percentiles.
• Most information from z gt 3.5.

16
Results Carbon metallicities from CIV

• 3. There is Carbon in underdense gas.
• The filling factor of metals is high tens of
  percent (depending on metallicity threshhold).
• May be difficult for late enrichment.

17
Results Carbon metallicities from CIV

• 4. The median carbon metallicity C/H does not
  evolve (for our fiducial UVB) from z4 to z2.

Neither does s(C/H)
18
Results Carbon metallicities from CIV

• 4. The median carbon metallicity C/H does not
  evolve (for our fiducial UVB) from z4 to z2.
• Clearly favors enrichment at z gt 4.
• But there is some room for more.

19
Results Carbon metallicities from CIV

• 5. C/H depends on UVB model.

But very different UVBs can be ruled out.
20
Results Carbon metallicities from CIV

• 5. C/H depends on UVB model.
• Inferences are sensitive to assumed UVB (and its
  history).
• But density-dependence, scatter are robust, and
  evolution fairly robust.

21
Gas temperature from CIII, SiIII

• 6. CIII/CIV, SiIII/SiIV provide thermometer.
• Bulk of SiIV gas at Tlt104.9K
• Little scatter in gas temp.
• But some evidence for hotter gas? (lt 30)
• Similar results using
  CIII/CIV.

22
Gas temperature from CIII, SiIII

• 6. CIII/CIV, SiIII/SiIV provide thermometer.
• Observed metals are in photoionized, warm gas,
  not the collisionally ionized warm/hot gas
  expected from winds.

23
Gas temperature from CIII, SiIII

• 6. CIII/CIV, SiIII/SiIV provide thermometer.
• Observed metals are in photoionized, warm gas,
  not the collisionally ionized warm/hot gas
  expected from winds.
• But slight evidence for some missing SiIII, and
  suggestions of collisionally ionized gas from OVI
  (in progress).

24
Silicon metallicities from SiIV, CIV

• 7. SiIV/CIV vs CIV ratios depend on d,
  reproduced by simulation.
• Si/C0.77/-0.05
• Si/C varies w/UVB hardness.
• No scatter in inferred Si/C

25
Silicon metallicities from SiIV, CIV

• 7. SiIV/CIV vs CIV ratios depend on d,
  reproduced by simulation.
• Suggests Pop. II enrichment, which can have
  Si/C 0.5.
• If Si/C0.77 taken seriously, could point to
  Pop. III contribution as per Heger Woosley.
• Lack of scatter -gt Si and C from same sources
  later C production not important.

26
Silicon metallicities from SiIV, CIV

• 8. SiIV/CIV vs CIV ratios depend little on z,
  reproduced by simulation.
• No jump in UVB hardness at z 3.
• No evolution in Si/C for usual UVB

27
Silicon metallicities from SiIV, CIV

• 8. SiIV/CIV vs CIV ratios depend little on z,
  reproduced by simulation.
• Again, more lack of evidence for anything
  evolving.

28
Adding up global C, Si abundances.

• 9. Medianscatter -gt mean metallicity, and
  contribution to cosmic C, Si abundance.
• C/H -2.8, Si/H -2.0
• -gt stars hold only lt 60-70 of cosmic
  Si rest is in Lya forest.
• Lots of metals in the forest!

29
Adding up global C, Si abundances.

• 9. Lots of metals in the forest.
• Metal dispersal into IGM is quite efficient
  before z 3-4. (also note most metals escape
  cluster galaxies)
• Could z gtgt 6 enrichment really provide enough
  metals?

30
The scorecard
31
The real picture early and late?

• Some questions/considerations
• Metals sprinked in non-feedback simulation
  reproduce all current observations. But
• Do the observed winds escape? If so, where do
  the metals go?
• If not winds, how to we fix baryon fraction in
  galaxies?
• Clusters, z 0 observations indicate Z 0.1
  Zsol. How do we close the gap?
• Metal from late galaxies may be hidden in
  unobservably hot gas, with low filling factor.
• Metal and H absorption does not have to come
  from same gas.
• Data allows some evolution, esp. using freedom
  in UVB.

32
To Do

• Complete OVI analysis, look for NV UVB has
  opposite effect on O inferences than on SiIV.
  Also, hotter gas can be seen in OVI.
• Looks at metallicity vs. distance from
  absorber.
• Look at correlations in PODs. See if simulations
  reproduce observations.
• Compare observed PODs in detail to hydro
  simulations with feedback.
• Try to connect these with simulations of
  individual galaxies.

33
Conclusions

• We can learn a lot from the Lya forest and the
  pollution in it.
• Evidence from galaxies suggests that they enrich
  the IGM.
• Evidence from the IGM suggests it was already
  enriched.
• Next step of detailed model/observation
  comparison holds great promise.