Black hole accretion with a tilt

Size: px

Start display at page:

Download "Black hole accretion with a tilt"

Blaise Cook
5 years ago
Views:

1 Black hole accretion with a tilt Alexander (Sasha) Tchekhovskoy Northwestern University Koushik Chatterjee Matthew Liska

2 M87 Jet: Acceleration and Collimation over 5 Orders in Distance Chandra XRC Relativistic motions on pc-scales Pushchino (Biretta+ 1999) John Biretta, Space Telescope Science Institute!2

3 <latexit sha1_base64="rwwgwqsc9kiuushatelv4fhbmfm=">aaacbxicbvc7tsmwfhxkq5rxciolryxegkieizwxgowxipqq2la5jtnadwlldopk1jmvyiwjdbhyhqz8c27aavqozononht1fu4ggfxadb+swsrq2vpgcbo0tb2zu2ex95ukpxktbuamy3aafge0iq1nnsntiqmka0zawfbq6rdgrcrkkzs9fssput+hecvig6lnlx9hv0gunie395nnvcc9u+i6bg64tlw5qya56j37uxtynmyk0zghptqek7sfiakpzmrs6qakcishqe86hiyojuo0hfghcupneyojpdzmccmuzus0znxfyxmklrrhgzmmkr6orw8q/ud1uh1d+blnrkpjgmehoprbk3lacqypjfizsseis2q+dfeasys1ka5k+vaw0y+t5pnjuy53c16pxc6bkyjdcarogaeqoaauqr00aayp4bm8gffryxqz3q2p2wjbmu8cgd+wpn8aq2kx9q==</latexit> <latexit sha1_base64="rwwgwqsc9kiuushatelv4fhbmfm=">aaacbxicbvc7tsmwfhxkq5rxciolryxegkieizwxgowxipqq2la5jtnadwlldopk1jmvyiwjdbhyhqz8c27aavqozononht1fu4ggfxadb+swsrq2vpgcbo0tb2zu2ex95ukpxktbuamy3aafge0iq1nnsntiqmka0zawfbq6rdgrcrkkzs9fssput+hecvig6lnlx9hv0gunie395nnvcc9u+i6bg64tlw5qya56j37uxtynmyk0zghptqek7sfiakpzmrs6qakcishqe86hiyojuo0hfghcupneyojpdzmccmuzus0znxfyxmklrrhgzmmkr6orw8q/ud1uh1d+blnrkpjgmehoprbk3lacqypjfizsseis2q+dfeasys1ka5k+vaw0y+t5pnjuy53c16pxc6bkyjdcarogaeqoaauqr00aayp4bm8gffryxqz3q2p2wjbmu8cgd+wpn8aq2kx9q==</latexit> <latexit sha1_base64="rwwgwqsc9kiuushatelv4fhbmfm=">aaacbxicbvc7tsmwfhxkq5rxciolryxegkieizwxgowxipqq2la5jtnadwlldopk1jmvyiwjdbhyhqz8c27aavqozononht1fu4ggfxadb+swsrq2vpgcbo0tb2zu2ex95ukpxktbuamy3aafge0iq1nnsntiqmka0zawfbq6rdgrcrkkzs9fssput+hecvig6lnlx9hv0gunie395nnvcc9u+i6bg64tlw5qya56j37uxtynmyk0zghptqek7sfiakpzmrs6qakcishqe86hiyojuo0hfghcupneyojpdzmccmuzus0znxfyxmklrrhgzmmkr6orw8q/ud1uh1d+blnrkpjgmehoprbk3lacqypjfizsseis2q+dfeasys1ka5k+vaw0y+t5pnjuy53c16pxc6bkyjdcarogaeqoaauqr00aayp4bm8gffryxqz3q2p2wjbmu8cgd+wpn8aq2kx9q==</latexit> <latexit sha1_base64="rwwgwqsc9kiuushatelv4fhbmfm=">aaacbxicbvc7tsmwfhxkq5rxciolryxegkieizwxgowxipqq2la5jtnadwlldopk1jmvyiwjdbhyhqz8c27aavqozononht1fu4ggfxadb+swsrq2vpgcbo0tb2zu2ex95ukpxktbuamy3aafge0iq1nnsntiqmka0zawfbq6rdgrcrkkzs9fssput+hecvig6lnlx9hv0gunie395nnvcc9u+i6bg64tlw5qya56j37uxtynmyk0zghptqek7sfiakpzmrs6qakcishqe86hiyojuo0hfghcupneyojpdzmccmuzus0znxfyxmklrrhgzmmkr6orw8q/ud1uh1d+blnrkpjgmehoprbk3lacqypjfizsseis2q+dfeasys1ka5k+vaw0y+t5pnjuy53c16pxc6bkyjdcarogaeqoaauqr00aayp4bm8gffryxqz3q2p2wjbmu8cgd+wpn8aq2kx9q==</latexit> M87 Jet: Acceleration and Collimation over 5 Orders in Distance Chandra XRC Relativistic motions on pc-scales Parabolic Jet over 5 Orders of Magnitude Pushchino (Biretta+ 1999) John Biretta, Space Telescope Science Institute!2 z / R 1.7 (Nakamura and Asada 2013)

4 <latexit sha1_base64="rwwgwqsc9kiuushatelv4fhbmfm=">aaacbxicbvc7tsmwfhxkq5rxciolryxegkieizwxgowxipqq2la5jtnadwlldopk1jmvyiwjdbhyhqz8c27aavqozononht1fu4ggfxadb+swsrq2vpgcbo0tb2zu2ex95ukpxktbuamy3aafge0iq1nnsntiqmka0zawfbq6rdgrcrkkzs9fssput+hecvig6lnlx9hv0gunie395nnvcc9u+i6bg64tlw5qya56j37uxtynmyk0zghptqek7sfiakpzmrs6qakcishqe86hiyojuo0hfghcupneyojpdzmccmuzus0znxfyxmklrrhgzmmkr6orw8q/ud1uh1d+blnrkpjgmehoprbk3lacqypjfizsseis2q+dfeasys1ka5k+vaw0y+t5pnjuy53c16pxc6bkyjdcarogaeqoaauqr00aayp4bm8gffryxqz3q2p2wjbmu8cgd+wpn8aq2kx9q==</latexit> <latexit sha1_base64="rwwgwqsc9kiuushatelv4fhbmfm=">aaacbxicbvc7tsmwfhxkq5rxciolryxegkieizwxgowxipqq2la5jtnadwlldopk1jmvyiwjdbhyhqz8c27aavqozononht1fu4ggfxadb+swsrq2vpgcbo0tb2zu2ex95ukpxktbuamy3aafge0iq1nnsntiqmka0zawfbq6rdgrcrkkzs9fssput+hecvig6lnlx9hv0gunie395nnvcc9u+i6bg64tlw5qya56j37uxtynmyk0zghptqek7sfiakpzmrs6qakcishqe86hiyojuo0hfghcupneyojpdzmccmuzus0znxfyxmklrrhgzmmkr6orw8q/ud1uh1d+blnrkpjgmehoprbk3lacqypjfizsseis2q+dfeasys1ka5k+vaw0y+t5pnjuy53c16pxc6bkyjdcarogaeqoaauqr00aayp4bm8gffryxqz3q2p2wjbmu8cgd+wpn8aq2kx9q==</latexit> <latexit sha1_base64="rwwgwqsc9kiuushatelv4fhbmfm=">aaacbxicbvc7tsmwfhxkq5rxciolryxegkieizwxgowxipqq2la5jtnadwlldopk1jmvyiwjdbhyhqz8c27aavqozononht1fu4ggfxadb+swsrq2vpgcbo0tb2zu2ex95ukpxktbuamy3aafge0iq1nnsntiqmka0zawfbq6rdgrcrkkzs9fssput+hecvig6lnlx9hv0gunie395nnvcc9u+i6bg64tlw5qya56j37uxtynmyk0zghptqek7sfiakpzmrs6qakcishqe86hiyojuo0hfghcupneyojpdzmccmuzus0znxfyxmklrrhgzmmkr6orw8q/ud1uh1d+blnrkpjgmehoprbk3lacqypjfizsseis2q+dfeasys1ka5k+vaw0y+t5pnjuy53c16pxc6bkyjdcarogaeqoaauqr00aayp4bm8gffryxqz3q2p2wjbmu8cgd+wpn8aq2kx9q==</latexit> <latexit sha1_base64="rwwgwqsc9kiuushatelv4fhbmfm=">aaacbxicbvc7tsmwfhxkq5rxciolryxegkieizwxgowxipqq2la5jtnadwlldopk1jmvyiwjdbhyhqz8c27aavqozononht1fu4ggfxadb+swsrq2vpgcbo0tb2zu2ex95ukpxktbuamy3aafge0iq1nnsntiqmka0zawfbq6rdgrcrkkzs9fssput+hecvig6lnlx9hv0gunie395nnvcc9u+i6bg64tlw5qya56j37uxtynmyk0zghptqek7sfiakpzmrs6qakcishqe86hiyojuo0hfghcupneyojpdzmccmuzus0znxfyxmklrrhgzmmkr6orw8q/ud1uh1d+blnrkpjgmehoprbk3lacqypjfizsseis2q+dfeasys1ka5k+vaw0y+t5pnjuy53c16pxc6bkyjdcarogaeqoaauqr00aayp4bm8gffryxqz3q2p2wjbmu8cgd+wpn8aq2kx9q==</latexit> M87 Jet: Acceleration and Collimation over 5 Orders in Distance Chandra XRC Relativistic motions on pc-scales Parabolic Jet over 5 Orders of Magnitude Pushchino (Biretta+ 1999) John Biretta, Space Telescope Science Institute!2 z / R 1.7 (Nakamura and Asada 2013) What physics leads to the continuous power-law shape of the jet? How do such jets become relativistic?

5 How Do Jets Form and Koushik Chatterjee (U. of Amsterdam) Accelerate? (Beskin & Nokhrina 2006, Komissarov , Alexander AT (Sasha) 09-10, Tchekhovskoy Lyubarsky 10)

6 How Do Jets Form and Koushik Chatterjee (U. of Amsterdam) Accelerate? Black hole + disk = jets + outflows Span >5 orders of magnitude in distance: directly compare to observations largest extent disk-jet simulations (Beskin & Nokhrina 2006, Komissarov , Alexander AT (Sasha) 09-10, Tchekhovskoy Lyubarsky 10)

7 How Do Jets Form and Koushik Chatterjee (U. of Amsterdam) Accelerate? Black hole + disk = jets + outflows Span >5 orders of magnitude in distance: directly compare to observations largest extent disk-jet simulations (Beskin & Nokhrina 2006, Komissarov , Alexander AT (Sasha) 09-10, Tchekhovskoy Lyubarsky 10)

8 How Do Jets Form and Koushik Chatterjee (U. of Amsterdam) Accelerate? Black hole + disk = jets + outflows Span >5 orders of magnitude in distance: directly compare to observations largest extent disk-jet simulations Jets accelerate as they collimate: to relativistic Lorentz factors match shape/acceleration of M87 jet BUT: slow down if collimate too much (Beskin & Nokhrina 2006, Komissarov , Alexander AT (Sasha) 09-10, Tchekhovskoy Lyubarsky 10)

9 How Do Jets Form and Koushik Chatterjee (U. of Amsterdam) Accelerate? Black hole + disk = jets + outflows Span >5 orders of magnitude in distance: directly compare to observations largest extent disk-jet simulations Jets accelerate as they collimate: to relativistic Lorentz factors match shape/acceleration of M87 jet BUT: slow down if collimate too much See Koushik s poster! (Beskin & Nokhrina 2006, Komissarov , Alexander AT (Sasha) 09-10, Tchekhovskoy Lyubarsky 10)

10 How Do Jets Form and Koushik Chatterjee (U. of Amsterdam) Accelerate? Black hole + disk = jets + outflows Span >5 orders of magnitude in distance: directly compare to observations largest extent disk-jet simulations Jets accelerate as they collimate: to relativistic Lorentz factors match shape/acceleration of M87 jet BUT: slow down if collimate too much Jets efficiently convert magnetic to kinetic energy: reach matter domination See Koushik s poster! (Beskin & Nokhrina 2006, Komissarov , Alexander AT (Sasha) 09-10, Tchekhovskoy Lyubarsky 10)

11 Are We Missing Anything Important? We already understand the basics of disks and jets MS NASA (McNamara et al. 2009)

12 Are We Missing Anything Important? We already understand the basics of disks and jets Most of the work: aligned systems MS NASA ~a BH outflow inflow jet (McNamara et al. 2009)

13 Are We Missing Anything Important? We already understand the basics of disks and jets Most of the work: aligned systems Expectation (AGN, TDEs, BH-NS mergers): tilted disk MS NASA ~a BH inflow jet outflow (McNamara et al. 2009)

14 Are We Missing Anything Important? We already understand the basics of disks and jets Most of the work: aligned systems Expectation (AGN, TDEs, BH-NS mergers): tilted disk Challenge: understand the physics of the most common, tilted, accretion flows from first principles MS NASA jet?? ~a BH inflow (McNamara et al. 2009)?? outflow YES: disks are tilted No: we do not understand them (yet)

15 Tilted Disk Physics Thick disks precess due to general relativistic frame dragging by BH spin precessing tilted disk sims could not handle jets (Fragile et al. 2005, 2007) Do tilted disks produce jets at all? Do jets precess or point along BH spin? (McKinney, AT+2013) ~a BH inflow

16 Tilted Disk Physics Thick disks precess due to general relativistic frame dragging by BH spin precessing tilted disk sims could not handle jets (Fragile et al. 2005, 2007) Do tilted disks produce jets at all? Do jets precess or point along BH spin? (McKinney, AT+2013) Thin disks can align due to Bardeen-Petterson (1975) effect Seen only in pseudo- Newtonian simulations and at small inclinations (Hawley and Krolik 2015) At larger inclinations disks predicted to break (Nixon et al. 2012) Do thin disks align in GR? Or do they break? Challenge: enormous dynamical range. Need to resolve thin streams over long run times. How?! ~a BH ~a BH inflow

17 H-AMR: What s Your Nail? Multi-GPU 3D H-AMR ( hammer, Liska, AT, et al. 2018): Based on HARMPI 85% parallel scaling to 4096 GPUs (MPI, OpenMP, OpenCL, CUDA) x speedup on 1 GPU vs 1 CPU core Matthew Liska (U of Amsterdam)

18 H-AMR: What s Your Nail? Multi-GPU 3D H-AMR ( hammer, Liska, AT, et al. 2018): Based on HARMPI 85% parallel scaling to 4096 GPUs (MPI, OpenMP, OpenCL, CUDA) x speedup on 1 GPU vs 1 CPU core 100M Matthew Liska (U of Amsterdam) 3M

19 H-AMR: What s Your Nail? Multi-GPU 3D H-AMR ( hammer, Liska, AT, et al. 2018): Based on HARMPI 85% parallel scaling to 4096 GPUs (MPI, OpenMP, OpenCL, CUDA) x speedup on 1 GPU vs 1 CPU core Matthew Liska (U of Amsterdam)

20 H-AMR: What s Your Nail? Multi-GPU 3D H-AMR ( hammer, Liska, AT, et al. 2018): Based on HARMPI 85% parallel scaling to 4096 GPUs (MPI, OpenMP, OpenCL, CUDA) x speedup on 1 GPU vs 1 CPU core Advanced features (extra few - 10x speedup): Adaptive Mesh Refinement (AMR) Local adaptive time-stepping Matthew Liska (U of Amsterdam)

21 H-AMR: What s Your Nail? Multi-GPU 3D H-AMR ( hammer, Liska, AT, et al. 2018): Based on HARMPI 85% parallel scaling to 4096 GPUs (MPI, OpenMP, OpenCL, CUDA) x speedup on 1 GPU vs 1 CPU core Advanced features (extra few - 10x speedup): Adaptive Mesh Refinement (AMR) Local adaptive time-stepping Ideal for getting computational time: 5M GPU-hours/yr = 5B CPU core-hours/yr on NSF Blue Waters supercomputer Science is no longer limited by computational resources! Matthew Liska (U of Amsterdam)

22 0 Thick Disks Precess a = h/r = i = Liska, Hesp, AT+2018a -8 The first demonstration that tilted thick disks produce tilted jets tilted jets precess Longest GRMHD tilted disk simulation, 120,000 rg/c Highest resolution GRMHD simulations: convergence verified at 2x resolution: first ever billion cell run (Fragile+07,09; McKinney, AT+13)

0 Thick Disks Precess 200 0 100-4 0-4 -100-20 -10-200 0 10 20-8 -200-100 0 100 200 Liska, Hesp, AT+2018a -8 The first demonstration that tilted thick disks produce tilted jets tilted jets precess

23 0 Thick Disks Precess Liska, Hesp, AT+2018a -8 The first demonstration that tilted thick disks produce tilted jets tilted jets precess Longest GRMHD tilted disk simulation, 120,000 rg/c Highest resolution GRMHD simulations: convergence verified at 2x resolution: first ever billion cell run (Fragile+07,09; McKinney, AT+13)

24 BUT: precession slows down 2x lower resolution fiducial resolution (Liska, Hesp, AT+2018a) At 2x higher resolution: results are similar -> convergence

25 BUT: precession slows down 2x lower resolution fiducial resolution (Liska, Hesp, AT+2018a) At 2x higher resolution: results are similar -> convergence

26 Thick-ish Disks Precess and Align a = 0.93 i = 45 h/r = 0.1 Casper Hesp (University of Amsterdam)

27 Thick-ish Disks Precess and Align Casper Hesp (University of Amsterdam)

28 Thick-ish Disks Precess and Align Precession and alignment increase the probability of GRB detection Casper Hesp (University of Amsterdam) in BH-NS mergers

29 Thin Weakly Misaligned Disks Align a = i = 10-3 h/r = Liska, Hesp, AT+2018b Thinnest disk simulations to date: h/r = 0.03 general relativistic MHD simulation of a thin disk Effective resolution , 3 AMR levels First demonstration of (Bardeen-Petterson?) alignment in a

Thin Weakly Misaligned Disks Align 2 80 2 40-3 0-3 -40 Liska, Hesp, AT+2018b -8-8 -80-8 -4 0 4 8-80 -40 0 40 80 Thinnest disk simulations to date: h/r = 0.

30 Thin Weakly Misaligned Disks Align Liska, Hesp, AT+2018b Thinnest disk simulations to date: h/r = 0.03 general relativistic MHD simulation of a thin disk Effective resolution , 3 AMR levels First demonstration of (Bardeen-Petterson?) alignment in a

31 Thin Strongly Misaligned Disks Align and Break a = i = 45-3 h/r = Liska, Hesp, AT+2018b First demonstration of (Bardeen-Petterson?) alignment and disk breaking in GRMHD! Effective resolution , 3 AMR levels

32 Thin Strongly Misaligned Disks Align and Break Liska, Hesp, AT+2018b First demonstration of (Bardeen-Petterson?) alignment and disk breaking in GRMHD! Effective resolution , 3 AMR levels

33 Thin VERY Strongly Misaligned Disks Tear Disks can tear up into individual segments Extra dissipation and luminosity Completely different luminosity profile Larger observed disk size than expected? (Blackburn+2011)

34 EM Counterparts to Binary Mergers BH BH BH NS NS NS detected! detected as well!

35 EM Counterparts to Binary Mergers BH BH BH NS NS NS detected! detected as well! but:

36 EM Counterparts to Binary Mergers BH BH BH NS NS NS detected! but: detected as well! and have EM counterparts: GRB (beamed) GRB jet outflow inflow kilonova (unbeamed) (Lattimer & Schramm 1974)

37 EM Counterparts to Binary Mergers Effect of composition on kilonova light curves BH BH BH NS NS NS Electron fraction detected! L but: Y e = n e n B detected as well! and have EM counterparts: Y e > 0.25 Y e < 0.25 day week high Ye = short blue luminous transient t GRB (beamed) GRB jet outflow inflow kilonova (unbeamed) (Lattimer & Schramm 1974)

38 EM Counterparts to Binary Mergers Effect of composition on kilonova light curves BH BH BH fully NS forms in ~5 secondsns Electron fraction detected! L but: Y e > 0.25 day Y e = n e n B Y e < 0.25 week t Merger disk mass outflow: over this time, studied in 2D, neglecting detected GR as and well! magnetic fields (Fernandez+15; but see Siegel & Metzger 17) Crucial and have to EM include counterparts: both in 3D GRB (beamed) GRB jet outflow inflow kilonova (unbeamed) high Ye = short blue luminous transient (Lattimer & Schramm 1974)

39 Magnetic Fields Double Mass Outflow (AT, Fernandez+ 2018, in prep) MBH = 3 Msun Mdisk = 0.03 Msun a = 0.8 Bp = G Implemented into HARMPI: neutrino emission nuclear recombination

40 Magnetic Fields Double Mass Outflow (AT, Fernandez+ 2018, in prep) Ṁ Ṁ out MHD t [s]

41 Magnetic Fields Double Mass Outflow (AT, Fernandez+ 2018, in prep) Ṁ Ṁ out MHD hydro t [s]

42 Magnetic Fields Double Mass Outflow The longest 3D GRMHD simulation ever: cost ~5M CPU-hours

43 Magnetic Fields Double Mass Outflow The longest 3D GRMHD simulation ever: cost ~5M CPU-hours

44 Magnetic Fields Double Mass Outflow Ejecta composition hydro: 22% ejected MHD: 44% ejected Magnetic effects lead to: 2x increase in ejecta mass brighter kilonova broader ejecta composition The more longest heavy 3D element GRMHD enrichment simulation ever: cost ~5M CPU-hours (see also Siegel+17)

Magnetic effects lead to: 2x increase in ejecta mass brighter kilonova broader ejecta

45 Magnetic Fields Double Mass Outflow Ejecta composition hydro: 22% ejected MHD: 44% ejected Long-term goal: Compute kilonova light curves from first principles. Magnetic effects lead to: 2x increase in ejecta mass brighter kilonova broader ejecta composition The more longest heavy 3D element GRMHD enrichment simulation ever: cost ~5M CPU-hours (see also Siegel+17)

46 Jets without Poloidal Fields?

47 Jets without Poloidal Fields?

48 Jets without Poloidal Fields? Ṁ Ṁ out MHD jet hydro

Black Hole Disk Wrap-up Koushik Chatterjee (University of

likelihood of GRB detection from a nearby BH-NS merger

disks first seen in GRMHD essentially unexplored

merger remnant disks: universe enrichment w/heavy elements

49 Black Hole Disk Wrap-up Koushik Chatterjee (University of Amsterdam) Jets precess together with tilted disks higher likelihood of GRB detection from a nearby BH-NS merger Bardeen-Petterson-like alignment and breaking of thin disks first seen in GRMHD essentially unexplored observational manifestations Longest simulations of binary merger remnant disks: universe enrichment w/heavy elements Matthew Liska (University of Amsterdam) Casper Hesp (University of Amsterdam)

Acceleration and Collimation of Relativistic Magnetized Jets

Acceleration and Collimation of Relativistic Magnetized Jets Alexander (Sasha) Tchekhovskoy with Ramesh Narayan and Jonathan McKinney Harvard University Chandra X-ray Observatory Magnificent Galaxy Jets