1. Summary of other reasonable theoretical ideas
Paul Romatschke
Thank the organizer for giving me this remarkable title for the talk. The fact that my talk is the last before the conference dinner confirmed by suscipions about the nature of this talk, so I decided to rise up to the challenge. My understanding of my charge in this talk is summarized in the next slide.
&

2. …but my client’s idea is perfectly reasonable, your honor!

3. “other”: no comments will be made on other ongoing trials…
CGC!

4. I’ll defend myself
It’s much more fun to be the devil than the devil’s advocate!

5. How to “fake” flow

6. Definition:
To “fake” flow: produce results that look like they had come from hydrodynamics, when they in fact had not

7. List of known suspects accused of faking flow:
Hadron cascades (UrQMD,…)
AMPT
The escape mechanism
Free streaming+hadronization
Different kinds of flow: radial, elliptic, v3, …
Some easier to fake than others

8. Many more very reasonable theoretical ideas…
…but my time in this court today does not allow me to discuss them!

9. Suspect #1: Hadron Cascades
[ ]
The first trial suspect we discuss in this court today is that of …
Mass ordering (“radial flow”)
Elliptic flow

10. Suspect #1: Hadron Cascades
Hadronic interactions alone can generate some “radial flow” (mass-ordering) in p+A, but not enough (?)
Hadronic interactions alone can generate some elliptic “flow” in p+A, but not enough (?)
Hadronic interactions generate some v3, but not enough
Large non-flow effects in p+A (c2{4}>0)!

11. Suspect #2: AMPT Mass ordering (“radial flow”): 1501.06880
σ=1.5mb
[ ]
Mass ordering (“radial flow”):
Elliptic flow & triangular flow: ,

12. Suspect #2: AMPT
AMPT can generate elliptic “flow” in p+A in agreement with data even with almost no interaction (σ=1.5mb)
Same for radial & triangular flow
Works even for Cu+Au [ ] ! ? !?
See Jamie Nagle’s talk on Tue!

13. Suspect #3: Escape mechanism
[ ]
Surface bias effect, w/o interactions!
Elliptic flow

14. Suspect #3: Escape mechanism
Anisotropic “flow” signal from surface emission bias
Contributes to v2, v3; significant if parton interactions small
See Zi-Wei Lin’s talk on Wed!

15. Suspect #3: Free-streaming
radial flow equal or larger than hydro
Elliptic flow, Triangular flow, HBT, …

16. The defense would like to present a little more detail….
I’ll defend myself
It’s much more fun to be the devil than the devil’s advocate!
The defense would like to present a little more detail….

17. “Flow” in free-streaming
3D case, initial condition:
Exact FS solution
Energy-Momentum tensor

18. Decomposition If a local rest frame exists (not restricted to hydro!)
Solve for ur in Free-streaming case:
“Flow” without interactions!
See U. Heinz’ talk on Mon!

19. FS flow similar to hydro flow

20. The importance of freeze-out
Without interactions, particle pT does not change
If there is a change in d.o.fs, such as hadronization, we generically use switching hypersurfaces to get final particle spectra (Cooper-Frye)
Because ur is non-vanishing in FS, Cooper-Frye leads to a change in pT

21. Free-streaming: flow in p+Pb

22. Free-streaming: flow in Pb+Pb

23. Free-streaming: HBT in Pb+Pb
Unequal particles? I don’t know. Need to calculate.

24. Free-streaming: mass-ordering in Pb+Pb

25. If final particles originate from constant T hypersurface, you get radial flow no matter what (hydro, FS, AMPT,…)

26. Does radial flow ever turn off?
Radial flow proportional to the lifetime in the “hot” phase. Does turn off if no cell exceeds QCD phase transition temperature.

27. Also: small v2, v3,… without hydro when you have a hadronic afterburner. Let me tell you about it…

28. In this case, your honor, I will summarize…

29. Summary
Different “reasonable” ideas for explaining exp’ data in p+A collisions
One common finding: can “fake” radial flow, mass ordering and HBT simply by having a T=const freeze-out surface
You can even get this from free-streaming!
Radial flow, mass ordering and HBT are not indicative of the presence of hydrodynamics in p+A or even A+A!

30. I like hydro, but it’s really fun to be the devil and pick it apart
I like hydro, but it’s really fun to be the devil and pick it apart. We might learn a lot about QCD
this way.

31. Bonus Material

32. Equation of State (EoS)
Massive particle EoS with Z dofs
Used in both hydro and free-streaming model
m,Z chosen to latch onto hadronic EoS at TS=170 MeV (“freeze-out”)

33. Equation of State (EoS)

34. Initial Conditions (1/2)
Event-by-Event MonteCarlo-Glauber
MC position of nucleons using pdfs (WS for A, vanHulthen/GFMC for d/3He)
Positions (x,y) of nucleons undergoing at least one collision recorded (“participants”)
Assumed to contribute to energy density at (x,y) in form of a Gaussian with width 0.4 fm
Total energy density is sum over all Gaussians

35. Initial Conditions (2/2)
Assume Tab(τ=τ0,x,y) to be isotropic
Then energy density fixes Tab
For simplicity, velocities and viscous stresses at τ=τ0 set to zero

36. Hydro evolution Solved numerically in 2+1d using vh2+1
Cooper-Frye freeze-out for cells reaching T=170 MeV

37. Free-streaming evolution
Analytic solution:
Cooper-Frye freeze-out for cells reaching T=170 MeV

38. Hadron Cascade
Once all cells have cooled below T=170 MeV, evolution simulated using hadron cascade (B3D)
All stable particle resonances up to M=2.2 GeV
Conversion fully matching Tab incl. viscous stresses (not only ideal fluid part)
Typically 100,000 B3D events per hydro/FS event
Particle spectra, vn calculated from B3D events

39. Importance of correct switching

40. Radial flow in p+A

41. Radial flow in p+A

42. Radial flow in p+A

43. Radial flow in p+A
Finding: radial flow in almost ideal hydro (η/s=0.08) and free-streaming (η/s=∞) essentially identical for p+A

44. Radial flow in A+A

45. Radial flow Conclusions
Finding 1: radial flow in viscous hydro (η/s=0.08) almost the same as in free-streaming (η/s=∞) for p+A
Finding 2: radial flow in free-streaming larger than in almost ideal hydro
My conclusion: observation of radial flow is no indication of hydro phase

46. Elliptic flow in p+A

47. Elliptic flow in p+A

48. Elliptic flow in p+A

49. Elliptic flow in p+A
Finding 1: there is some elliptic flow generated in free-streaming (η/s=∞) + hadron cascade dynamics in p+A
Finding 2: the amount of elliptic flow in free-streaming (η/s=∞) + hadron cascade dynamics is only about 1/3 of that in almost ideal hydro (η/s=0.08) +hadron cascade in p+A

50. Elliptic flow in A+A

51. Elliptic flow Conclusions
Finding 1: big differences of elliptic flow between viscous hydro (η/s=0.08) and free-streaming (η/s=∞) for p+A and AA
My conclusion: observation of large elliptic flow is evidence for hydro phase

52. Triangular flow in p+A

53. Triangular flow in p+A

54. Triangular flow in p+A

55. Triangular flow in p+A
Finding: triangular flow in free-streaming (η/s=∞) +hadron cascade similar to (p+Pb) or larger than (d+Au, 3He+Au) in almost ideal hydro (η/s=0.08)+cascade

56. Triangular flow in A+A

57. Triangular flow Conclusions
Finding 1: triangular flow in viscous hydro (η/s=0.08) almost the same as in free-streaming (η/s=∞) for p+Pb
Finding 2: triangular flow in free-streaming larger than in almost ideal hydro for d+Au and 3He+Au at RHIC
My conclusion: v3,v4 measurements in p+A can easily be generated by hydro-imposters; do not use as viscometer!

58. HBT radii in p+A

59. HBT radii in p+A

60. HBT radii in p+A

61. HBT Radii Conclusions
Finding 1:HBT Radii are essentially identical in viscous hydro (η/s=0.08) and free-streaming (η/s=∞) dynamics in p+A and also A+A
My conclusion: measurements of HBT Radii are not sensitive to the presence or absence of a hydro phase

62. Remarks about
Apples-to-apples comparison of models for p+A collisions (hydro vs. free-streaming)
Radial flow, HBT radii are essentially the same for hydro and FS; do not use as evidence for hydro phase
v3 is similar in hydro/FS in p+A; do not use as ‘viscometer’
v2 is the gold standard: hard to fake without hydro; “large” v2 is evidence for hydro phase