generalizedslowroll* and cmb*constraintson*the*inflaton*poten8al · 2011. 12. 15. ·...

16
Generalized Slow Roll and CMB Constraints on the Inflaton Poten8al 1 Cora Dvorkin University of Chicago/KICP June 2010, Great Lakes Cosmology Workshop References: C. Dvorkin and W. Hu, astroph/0910.2237, PRD (2009). M. Mortonson, C. Dvorkin, H.V.Peiris, W.Hu, astroph/0903.4920 , PRD (2009). C. Dvorkin and W. Hu (2010, in preparaHon).

Upload: others

Post on 25-Apr-2021

2 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: GeneralizedSlowRoll* and CMB*Constraintson*the*Inflaton*Poten8al · 2011. 12. 15. · Ordinary!slow!roll!(brief!background)!!Technique!for!compuMng!the!iniMal!curvature! power!spectrum!for!inflaMonary!models!where!

Generalized  Slow  Roll  and    

CMB  Constraints  on  the  Inflaton  Poten8al  

1  

Cora  Dvorkin  University  of  Chicago/KICP  

June  2010,  Great  Lakes  Cosmology  Workshop  

     References:  -­‐   C.  Dvorkin  and  W.  Hu,  astro-­‐ph/0910.2237,  PRD  (2009).  -­‐   M.  Mortonson,  C.  Dvorkin,  H.V.Peiris,  W.Hu,  astro-­‐ph/0903.4920  ,  PRD  (2009).  -­‐   C.  Dvorkin  and  W.  Hu  (2010,  in  preparaHon).  

Page 2: GeneralizedSlowRoll* and CMB*Constraintson*the*Inflaton*Poten8al · 2011. 12. 15. · Ordinary!slow!roll!(brief!background)!!Technique!for!compuMng!the!iniMal!curvature! power!spectrum!for!inflaMonary!models!where!

Ordinary  slow  roll  (brief  background)  

 Technique  for  compuMng  the  iniMal  curvature  power  spectrum  for  inflaMonary  models  where  the  scalar  field  potenMal  is  sufficiently  flat  and  slowly  varying.  

�H ≡ 12

�φ̇

H

�2

ηH ≡ −�

φ̈

Hφ̇

δ2 ≡...φ

H2φ̇

Slow-­‐roll  parameters:  

Linked  to  the  shape    of  the  potenHal  

<<  1  and  slowly  varying  

Slow  roll  approximaMon:      ∆2R ≈

�(1− (2C + 1)�H − CηH)

H2

2π|φ̇|

�2

kη≈1

Page 3: GeneralizedSlowRoll* and CMB*Constraintson*the*Inflaton*Poten8al · 2011. 12. 15. · Ordinary!slow!roll!(brief!background)!!Technique!for!compuMng!the!iniMal!curvature! power!spectrum!for!inflaMonary!models!where!

Why  go  further?  

•  These  models  require  order  unity  variaMons  in  the  curvature  power  spectrum:  slow-­‐roll  parameters  are  not  necessarily  small  or  slowly  varying.  

3  

•   Glitches  in  the  temperature        power  spectrum  of  the  CMB                  have  led  to  interest  in  exploring          models  with  features  in  the          inflaton  potenMal.  

   L.  Covi,  J.Hamann,  A.  Melchiorri,              A.  Slozar  and  I.  Sorbera,  (2006).     M.  Mortonson,  C.  Dvorkin,              H.V.  Peiris  and  W.  Hu,  PRD    (2009).  

Page 4: GeneralizedSlowRoll* and CMB*Constraintson*the*Inflaton*Poten8al · 2011. 12. 15. · Ordinary!slow!roll!(brief!background)!!Technique!for!compuMng!the!iniMal!curvature! power!spectrum!for!inflaMonary!models!where!

InflaMonary  features  A  step  in  the  inflaMonary  potenMal  generates  an  oscillatory  iniMal  

curvature  power  spectrum.  

The  slow-­‐roll  parameters  are  neither  small  nor  slowly  varying.  4  

C.  Dvorkin,  W.Hu,  PRD  (2009)  

Page 5: GeneralizedSlowRoll* and CMB*Constraintson*the*Inflaton*Poten8al · 2011. 12. 15. · Ordinary!slow!roll!(brief!background)!!Technique!for!compuMng!the!iniMal!curvature! power!spectrum!for!inflaMonary!models!where!

INFLATIONARY  POTENTIAL   POWER  SPECTRUM  

TEMPERATURE  

POLARIZATION  

Page 6: GeneralizedSlowRoll* and CMB*Constraintson*the*Inflaton*Poten8al · 2011. 12. 15. · Ordinary!slow!roll!(brief!background)!!Technique!for!compuMng!the!iniMal!curvature! power!spectrum!for!inflaMonary!models!where!

INFLATIONARY  POTENTIAL   POWER  SPECTRUM  

TEMPERATURE  

POLARIZATION  

OBSERVABLES  

Page 7: GeneralizedSlowRoll* and CMB*Constraintson*the*Inflaton*Poten8al · 2011. 12. 15. · Ordinary!slow!roll!(brief!background)!!Technique!for!compuMng!the!iniMal!curvature! power!spectrum!for!inflaMonary!models!where!

How  does  the  Generalized  Slow  Roll  approximaMon  work?    

7  

•   Field  equaMon:    

•   Perfect  slow  roll:  

•   GSR  approximaMon:  

d2y

dx2+

�1− 2

x2

�y =

g(lnx)x2

y

d2y0

dx2+

�1− 2

x2

�y0 = 0

d2y

dx2+

�1− 2

x2

�y =

g(lnx)x2

y0

SoluMon  can  be  constructed  with  Green  funcMon  technique.  

Source  funcMon  (linear  in  slow-­‐roll    parameters)  

(y =√

2kuk; x = kη)

E.  Stewart,  PRD  (2002).  

Page 8: GeneralizedSlowRoll* and CMB*Constraintson*the*Inflaton*Poten8al · 2011. 12. 15. · Ordinary!slow!roll!(brief!background)!!Technique!for!compuMng!the!iniMal!curvature! power!spectrum!for!inflaMonary!models!where!

How  does  the  Generalized  Slow  Roll  approximaMon  work?    

8  

•   Field  equaMon:    

•   Perfect  slow  roll:  

•   GSR  approximaMon:  

d2y

dx2+

�1− 2

x2

�y =

g(lnx)x2

y

d2y0

dx2+

�1− 2

x2

�y0 = 0

d2y

dx2+

�1− 2

x2

�y =

g(lnx)x2

y0

SoluMon  can  be  constructed  with  Green  funcMon  technique.  

•   Nodes  in  the  power  spectrum.  •   Curvature  is  not  constant  for  modes  outside  the  horizon.  

BUT…  

Source  funcMon  (linear  in  slow-­‐roll    parameters)  

(y =√

2kuk; x = kη)

E.  Stewart,  PRD  (2002)  

Page 9: GeneralizedSlowRoll* and CMB*Constraintson*the*Inflaton*Poten8al · 2011. 12. 15. · Ordinary!slow!roll!(brief!background)!!Technique!for!compuMng!the!iniMal!curvature! power!spectrum!for!inflaMonary!models!where!

Generalized  Slow  Roll  for  Large  DeviaMons  

9  

The  curvature  power  spectrum  only  depends    on  a  single  source  funcMon:  

C.  Dvorkin,  W.  Hu,  PRD  (2009)  

ln ∆2R(k) = G(ln ηmin) +

� ηmax

ηmin

ηW (kη)G�(ln η)

+ ln

�1 +

12

�� ηmax

ηmin

ηX(kη)G�(ln η)

�2�

Page 10: GeneralizedSlowRoll* and CMB*Constraintson*the*Inflaton*Poten8al · 2011. 12. 15. · Ordinary!slow!roll!(brief!background)!!Technique!for!compuMng!the!iniMal!curvature! power!spectrum!for!inflaMonary!models!where!

Generalized  Slow  Roll  for  Large  DeviaMons  

10  

The  curvature  power  spectrum  only  depends    on  a  single  source  funcMon:  

C.  Dvorkin,  W.  Hu,  PRD  (2009)  

ln ∆2R(k) = G(ln ηmin) +

� ηmax

ηmin

ηW (kη)G�(ln η)

+ ln

�1 +

12

�� ηmax

ηmin

ηX(kη)G�(ln η)

�2�

   Constant  curvature  for  modes  outside  the  horizon.                   Well  controlled  at  large  values  of  the  source:  percent  level  errors.            (Mme  integrals  are  small  even  if  the  source  is  not  small  everywhere).  

   Simple  to  relate  to  the  Inflaton  PotenMal:  G� ≈ 3�

V,φ

V

�2

− 2�

V,φφ

V

Page 11: GeneralizedSlowRoll* and CMB*Constraintson*the*Inflaton*Poten8al · 2011. 12. 15. · Ordinary!slow!roll!(brief!background)!!Technique!for!compuMng!the!iniMal!curvature! power!spectrum!for!inflaMonary!models!where!

Second  order  Generalized  Slow  Roll:  Well  controlled  

11  

TEMPERATURE   POLARIZATION  

M.Mortonson,  C.  Dvorkin,  H.V.Peiris,  W.Hu,  PRD  (2009)  

PolarizaHon  can  test  the  feature  hypothesis      at  2.5  sigma  with  Planck,  5-­‐6  sigma  with  CMBpol.  

Page 12: GeneralizedSlowRoll* and CMB*Constraintson*the*Inflaton*Poten8al · 2011. 12. 15. · Ordinary!slow!roll!(brief!background)!!Technique!for!compuMng!the!iniMal!curvature! power!spectrum!for!inflaMonary!models!where!

Second  order  Generalized  Slow  Roll:  Well  controlled  

12  

TEMPERATURE   POLARIZATION  

C.  Dvorkin,  W.  Hu,  PRD  (2009)  

Page 13: GeneralizedSlowRoll* and CMB*Constraintson*the*Inflaton*Poten8al · 2011. 12. 15. · Ordinary!slow!roll!(brief!background)!!Technique!for!compuMng!the!iniMal!curvature! power!spectrum!for!inflaMonary!models!where!

13  

We  can  use  features  in  the  power  spectrum    to  directly  constrain  features  in  the  potenHal.  

   Power  spectrum  

   Source  

   PotenMal  

Page 14: GeneralizedSlowRoll* and CMB*Constraintson*the*Inflaton*Poten8al · 2011. 12. 15. · Ordinary!slow!roll!(brief!background)!!Technique!for!compuMng!the!iniMal!curvature! power!spectrum!for!inflaMonary!models!where!

Model  independent  constraints  (PC’s)  (Work  in  progress)  

 Principal  components:  Basis  for  a  complete  representaMon  of  observable  properMes  of  the  source  funcMon.  

14  

GSR  approximaHon  allows  us  to  go    beyond  specific  infla8onary  models    and  think  of  the  data  as  directly    constraining  the  source  funcHon.  

C.  Dvorkin,  W.  Hu  (2010,  in  prepara8on)  

G� = 1− ns +N�

a=1

maSa(ln η)

Page 15: GeneralizedSlowRoll* and CMB*Constraintson*the*Inflaton*Poten8al · 2011. 12. 15. · Ordinary!slow!roll!(brief!background)!!Technique!for!compuMng!the!iniMal!curvature! power!spectrum!for!inflaMonary!models!where!

WMAP7  constraints  from  MCMC’s  (Work  in  progress)  

15  

C.  Dvorkin,  W.  Hu    (2010,  in  prepara8on)  

As  a  by-­‐product:    We  parallelized  the    

WMAP  likelihood  evaluaMon    for  mulM-­‐core  systems    

and  we  improved  its  speed  by  ~  5*Ncore  

Page 16: GeneralizedSlowRoll* and CMB*Constraintson*the*Inflaton*Poten8al · 2011. 12. 15. · Ordinary!slow!roll!(brief!background)!!Technique!for!compuMng!the!iniMal!curvature! power!spectrum!for!inflaMonary!models!where!

Summary  

•   The  Generalized  Slow  Roll  approximaMon  is  accurate  at  the  percent  level  for  order  unity  deviaMons  in  the  CMB  temperature  and  polarizaMon  power  spectra.  

•   There  is  a  single  source  funcMon  responsible  for  observed  features  and  it  is  simply  related  to  the  local  slope  and  curvature  of  the  Inflaton  PotenMal.  

•   ApplicaMons  (work  in  progress):      Use  this  technique  for  model  independent  constraints  on  the  

Inflaton  PotenMal:  think  of  the  data  as  directly  constraining  the  source  funcMon.  

16