18h29m49.6s49.8s50.0s

Right Ascension (J2000)

+1�1501600

1800

2000

2200

2400

Dec

linat

ion

(J20

00)

400AU

1.5

3.0

4.5

6.0

7.5

9.010.5

Pol

ariz

edflux

den

sity

(mJy

bea

m�

1)

Serpens SMM1

Hull+2016, in prep. See also Hull+2016a

ALMA

18h29m48.5s48.6s48.7s48.8s48.9s49.0s

Right Ascension (J2000)

+1�1605200

5400

5600

5800

Dec

linat

ion

(J20

00)

400AU

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.45

0.50

Pol

ariz

edflux

den

sity

(mJy

bea

m�

1)

Ser-emb 8(N)

Hull+2016, in prep.

ALMA

18h29m47.5s48.0s48.5s49.0s49.5s

Right Ascension (J2000)

+1�1603500

4000

4500

5000

5500

1700000

0500

Dec

linat

ion

(J20

00)

Ser-emb 8, 8(N)

8

8(N)

(b)

B-field

Ser-emb 8, 8(N)

2000 AU

Hull+2014

Ser-emb 8, 8(N) CARMA 1mm

Will ALMA see toroidally wrapped B-fields (or mm-wave dust scattering?!) at ~25 AU disk scales in these sources?

Stay tuned for ALMA Cycle 3 data!

Proposal 2015.1.00768.S PI: Hull, high priority

0.06ʺ (~25 AU) resolution dust polarization (@ Band 7)

Class 0 protostellar dust polarization

HH 211

REFERENCES Hull et al. 2014, ApJ, 768, 159 (outflows vs. B-fields) Hull et al. 2013, ApJS, in press (data release)

ABSTRACT Magnetic fields are thought to play an important role in the formation of stars. However, that importance has been called into question by previous observations showing misalignment between protostellar outflows and magnetic fields (B-fields), as well as inconsistency in B-field morphology between 10,000 AU and 1,000 AU scales. To investigate these inconsistencies, we used the 1.3 millimeter full-Stokes polarimeter at CARMA to map dust polarization with ~2.5" resolution toward 30 star-forming cores and 8 high-mass star-forming regions as part of the TADPOL survey.

Our main findings are: (1) A subset of the sources have consistent magnetic

field orientations between large (~20") and small (~2.5") scales. Those same sources also tend to have higher fractional polarizations than the sources with inconsistent large-to-small-scale fields. We interpret this to mean that in at least some cases B-fields play a role in regulating the infall of material all the way down to the ~1000AU scales of protostellar envelopes.

(2) Outflows appear to be randomly aligned with B-fields; although, in sources with low polarization fractions there is a hint that outflows are preferentially perpendicular to small-scale B-fields, which suggests that in these sources the B-fields have been wrapped up by envelope rotation.

Bipolar outflows are randomly aligned with magnetic fields at the 1000 AU

scales of protostellar envelopes

CD

F

𝜽outflow – 𝜽B-field (deg)

0 1

0º 90º

Hull et al. 2014, ApJ, 768, 159

Photo credit: Chat Hull

with ALMA chat.hull@cfa.harvard.edu — Harvard/NRAO — Jansky Fellow — Chat Hull

More info — chathull.com References

•  Hull et al. 2016, ApJL, 823, 27 •  Hull et al. 2015, JAI, 450005 •  Hull et al. 2014, ApJS, 213, 13 •  Hull et al. 2013, ApJ, 768, 159 •  Machida et al. 2006, ApJ, 645, 1227 •  Yang et al. 2016, MNRAS, 456, 2794

Hull+2016, in prep.

an angle of !B ’ 90!. On the other hand, the rotation axis hardlychanges its direction from the initial state and remains directedalong the z-axis. The disk normal is also oriented along thez-axis (i.e., the rotation axis). From Figure 8b, it can be seen thata disk forms by the effect of the rotation and the disk normaldirection coincides with the rotation axis.

Figure 8c shows the central region at the core formation epoch(nc ¼ 2:3 ; 1011 cm#3). It can be seen from this figure that anonaxisymmetric structure has formed and the central core haschanged its shape from a circular disk (bottom panels of Figs. 8aand 8b) to a bar (Fig. 8c, bottom). The magnetic field lines runlaterally, i.e., jBrj, jB"j3 jBzj, in the adiabatic phase (Figs. 8cand 8d ). Figure 8d shows an adiabatic core when the centraldensity has reached nc ¼ 6:9 ; 1014 cm#3. A spiral structure isseen in this figure, which indicates that a nonaxisymmetricpattern has formed, even if no explicit nonaxisymmetric densityperturbation is assumed at the initial stage. (Although the non-axisymmetric patterns also appear in some models of Papers I,II, and III, it should be noted that these patterns are due to anonaxisymmetric perturbation added to the density and magneticfield at the initial stage.) The magnetic field lines are considerablytwisted in Figure 8d . It should be noted that in this model, theinclined magnetic field induces nonaxisymmetric perturbations,on behalf of the initial explicit perturbation.

Figure 11 shows the magnetic field lines, the shape of thecore, and the velocity vectors on the z ¼ 0 plane in the adiabaticphase for model C00. This figure shows that a ring is formed, asfound in Paper III, without any growth of a nonaxisymmetricpattern. In Papers I, II, and III, we assumed a cylindrical cloud inhydrostatic equilibrium, in which the magnetic field and angularvelocity are functions of the radius r in cylindrical coordinates.On the other hand, the cloud is assumed to be spherical with auniform magnetic field and angular velocity at the initial stagein model C00. In spite of these differences, a similar ring struc-ture appears in both models C00 and CS of Paper II. Figure 9(bottom) plots the evolution of the axis ratio against the centraldensity for group C. The axis ratios for models C30, C45, andC60 begin to grow after a thin disk is formed (nc k 5 ; 106 cm#3)and reach "ar ’ 0:5 at the core formation epoch. The axis ratio

grows to "ar ’ 1 at nc ¼ 1012 cm#3 in models C30, C45, andC60, while no nonaxisymmetric pattern appears in model C00.This shows that the nonaxisymmetric pattern arises from theanisotropy of the Lorentz force around the rotation axis. A barstructure is formed by the nonaxisymmetric force exerted by theinclined magnetic field, as shown in Figures 8b–8d . This isconfirmed by the fact that the short axis of the density distribu-tion on the z ¼ 0 plane (the disk midplane) and the bar patternrotate together with the magnetic field lines. The axis ratio (thenonaxisymmetry) grows in proportion to #1=6 (107 cm#3 P nc P1010 cm#3 in Fig. 9, bottom), as found by Hanawa &Matsumoto(1999). Since the lateral component of the magnetic field(jBjsin !0) is large (Fig. 9, bottom), the axis ratio grows more inmodels with large !0.

The evolution of the angles !B, !!, !P, and "B for group C isplotted against the central density in Figure 12. The angle be-tween the magnetic field and z-axis becomes !B ’ 90! even inthe early phase of isothermal collapse for all the models C30,C45, and C60. The rotation axis and the disk normal maintaintheir angles !!; !P ’ 0!. Figures 4 and 12 show that in bothmagnetic- and rotation-dominant models the directions of themagnetic field, rotation axis, and disk normal are qualitativelythe same for models with the same $ and !, irrespective of !0 inthe range 30! $ !0 $ 60!.

Figure 13 shows the magnetic field lines, velocity vectors,and density distribution for the epoch t ¼ 1:52 ; 106 yr (nc ¼1:5 ; 109 cm#3) for model C30. Note that the box scale andlevel of grid are different for each panel. The spatial scale ofeach successive panel is different by a factor of 4, and thus thescale between Figures 13a and 13d is different by a factor of 64.The magnetic field has an angle !0B % 30! in Figure 13a, where!0B is defined as the angle between the volume average magnetic

Fig. 11.—Same as Fig. 1, but for C00 at the core formation epoch.

Fig. 10.—Same as Fig. 5 (left), but for model C45. The lower left inset is anenlarged view of the center.

CLOUD EVOLUTION WITH OBLIQUE FIELD 1237No. 2, 2006

Future

?

8H

.Yan

get

al.

Figu

re5.

Mod

elfo

rthe

pola

rizat

ion

ofH

LTa

udi

sk.P

lotte

dar

eth

epo

lar-

ized

inte

nsity

(col

orm

ap,i

nun

itsof

2⌃2 0

abs

sca⌫

2 kT0/

c2 )and

pola

rizat

ion

vect

ors(

line

segm

ents

,with

leng

thpr

opor

tiona

lto

the

pola

rizat

ion

frac

tion

ofth

eto

tali

nten

sity

)for

i=

45� .

Ach

arac

teris

ticsc

atte

ring

optic

alde

pth

�⌧ s,c⇡

0.07

isne

eded

tobr

ing

the

high

pola

rizat

ion

frac

tion

ofte

nsof

perc

entr

elat

ive

toth

esc

atte

red

inte

nsity

(sho

wn

inth

elo

wer

-left

pane

lof

Fig.

2)do

wn

toth

eob

serv

edle

velo

f⇠1%

rela

tive

toth

eto

tali

nten

sity

.

the

low

pola

rized

inte

nsity

near

the

edge

.Thi

stre

ndw

aspr

esen

tin

our

mod

elw

hen

only

the

scat

tere

dra

diat

ion

was

cons

ider

ed(s

eeFi

g.2

and

3),b

utw

asw

ashe

dou

tby

the

tota

lint

ensi

tyin

Fig.

5.Th

ese

mi-a

naly

ticth

eory

that

we

have

deve

lope

dso

faru

nder

the

assu

mpt

ion

ofge

omet

rical

lyth

indi

sk,o

ptic

ally

thin

emis

sion

,an

don

lyR

ayle

igh

scat

terin

gis

inde

pend

ento

fthe

deta

iled

prop

er-

ties

ofdu

stgr

ains

.Thi

sin

depe

nden

cem

akes

the

broa

dag

reem

ent

betw

een

ourm

odel

and

the

mai

npo

lariz

atio

nfe

atur

esob

serv

edin

HL

Tau

rath

erro

bust

.In

the

next

sect

ion,

we

will

tryto

putc

on-

stra

ints

onth

egr

ain

size

dist

ribut

ion

inth

eH

LTa

udi

sk,w

hich

ism

uch

mor

eun

certa

in.

4IM

PLIC

ATIO

NS

ON

GR

AIN

SIZE

4.1

Scat

teri

ngO

paci

tyan

dth

eN

eed

for

Larg

eG

rain

s

The

mai

nfr

eepa

ram

eter

ofth

efin

alm

odel

for

the

HL

Tau

disk

pola

rizat

ion

disc

usse

din

§3

isth

ech

arac

teris

ticsc

atte

ring

op-

tical

dept

h�⌧ s,c

,w

hich

cont

rols

the

pola

rizat

ion

frac

tion.

Inor

-de

rto

prod

uce

the

obse

rved

pola

rizat

ion

frac

tion

of⇠

1%,

ava

lue

of�⌧ s,c⇡

0.07

isre

quire

d.Th

isva

lue

yiel

dsa

scat

terin

gop

acity s

ca⇡

10�3

cm2

g�1

(cro

ssse

ctio

npe

run

itto

tal,

rath

erth

andu

st,m

ass)

at1.

3m

m(th

ew

avel

engt

hof

the

HL

Tau

disk

pola

rizat

ion

obse

rvat

ion)

,usi

ngth

ech

arac

teris

ticco

lum

nde

nsity

⌃0=

68g

cm�2

from

the

best

-fitd

isk

mas

sof

0.13

M�

ofK

won

etal

.(20

11).

This

scat

terin

gop

acity

can

putc

onst

rain

tson

the

grai

nsi

zedi

strib

utio

n,al

thou

ghth

eyde

pend

onth

edu

stco

mpo

sitio

n,w

hich

isun

certa

in.A

san

illus

tratio

n,w

eco

nsid

erth

em

odel

ofdu

stgr

ains

adop

ted

byK

atao

kaet

al.(

2015

),w

hich

are

sphe

res

with

am

ixtu

reof

silic

ate

(8%

),w

ater

ice

(62%

)and

orga

nics

(30%

).A

llfr

actio

nala

bund

ance

sar

ein

volu

me

and

are

take

nfr

omPo

llack

etal

.(19

94).

We

assu

me

aca

noni

calg

as-to

-dus

tmas

sra

tioof

100,

and

use

the

Mie

theo

ryto

calc

ulat

eth

eab

sorp

tion

and

scat

terin

gop

aciti

es(B

ohre

n&

Hu↵

man

1983

).Th

ein

ferr

edsc

atte

ring

opac

-ity

corr

espo

nds

toa

grai

nra

dius

a=

37µ

mfo

rgr

ains

ofa

sing

le

size

.For

theM

RN

-type

pow

er-la

wdi

strib

utio

nn(

a)/

a�3.

5(M

athi

set

al.1

977)

,we

obta

ina

max

imum

grai

nsi

zeof

a max=

108µ

m.

The

incr

ease

ofth

ism

axim

umov

erth

esi

ngle

size

case

com

esfr

omav

erag

ing

over

the

grai

nsi

ze.I

nbo

thca

ses,

the

dim

ensi

onle

sspa

-ra

met

erx=

2⇡a/�⌧

1,so

that

the

Ray

leig

hlim

itus

edfo

rtre

atin

gth

esc

atte

ring

inth

epr

evio

usse

ctio

nsis

self-

cons

iste

nt(s

eeFi

g.6

belo

w). In

sum

mar

y,to

repr

oduc

eth

e⇠

1%po

lariz

atio

nfr

actio

nob

-se

rved

inth

edi

skof

HL

Tau

thro

ugh

dust

scat

terin

g,th

egr

ains

mus

thav

egr

own

tote

nsof

mic

rons

(the

exac

tval

uede

pend

son

the

assu

med

grai

nsi

zedi

strib

utio

nan

dco

mpo

sitio

n).H

owev

er,t

his

pict

ure

isco

mpl

icat

edby

the

opac

itysp

ectra

lind

ex�

infe

rred

for

HL

Tau,

asw

edi

scus

sne

xt.

4.2

Opa

city

Spec

tral

Inde

x�

and

the

Nee

dfo

rLa

rger

Gra

ins?

Kw

onet

al.(

2011

)ob

tain

eda

best

-fitv

alue

0.73

for

the

spec

tral

inde

x�

ofth

edu

stop

acity a

bs/⌫�

for

the

HL

Tau

disk

base

dm

ostly

onC

AR

MA

obse

rvat

ions

at1.

3an

d2.

7m

m.I

tis

inag

ree-

men

twith

the

spat

ially

aver

aged

valu

eob

tain

edfr

omA

LMA

ob-

serv

atio

nsfr

om0.

87to

2.9

mm

(ALM

APa

rtner

ship

:B

roga

net

al.

2015

).Th

isva

lue

issi

gnifi

cant

lylo

wer

than

the

typi

cal

ISM

valu

eof�⇠

1.5�

2.Th

edi↵

eren

ceis

usua

llyta

ken

asev

iden

cefo

rgr

ain

grow

thto

mill

imet

ersi

zeor

larg

er(T

esti

etal

.20

14),

alth

ough

othe

rin

terp

reta

tions

are

poss

ible

.For

exam

ple,

Ric

ciet

al.(

2012

)sh

owed

that

ava

lue

of�⇠

1or

low

erca

nbe

obta

ined

with

outm

m/c

msi

zed

grai

nsif

part

ofth

edi

skis

optic

ally

thic

k.So

me

supp

ort

for

this

poss

ibili

tyis

prov

ided

byth

esp

atia

llyre

-so

lved

dist

ribut

ion

of�

deriv

edfr

omth

eA

LMA

data

,whi

chsh

ows

�⇠

0in

dica

tive

ofop

tical

lyth

ick

emis

sion

atth

ece

ntra

lcon

tin-

uum

peak

and

two

rings

(B1

and

B6,

ALM

APa

rtner

ship

:Bro

gan

etal

.201

5,se

eth

eir

Fig.

3).A

noth

erpo

ssib

ility

isth

atth

ein

dex

�is

sens

itive

tono

tonl

yth

esi

zebu

tals

oth

esh

ape

ofth

egr

ains

.In

deed

,Ver

hoe↵

etal

.(20

11)

was

able

tore

prod

uce

the

spec

tral

ener

gydi

strib

utio

n(S

ED)

ofth

edi

skof

HD

1425

27(w

ith�⇠

1in

the

mill

imet

erre

gim

e)w

ithirr

egul

argr

ains

ofsi

zes

upto

only

2.5µ

m;t

hegr

ain

shap

ew

astre

ated

with

the

dist

ribut

ion

ofho

llow

sphe

res

Min

etal

.(20

05).

The

grai

nsin

ferr

edin

ourm

odel

ofdu

stsc

atte

ring-

indu

ced

pola

rizat

ion

fort

heH

LTa

udi

skha

vea

sign

if-ic

antly

larg

erm

axim

umsi

ze(o

ford

erte

nsof

mic

rons

).Th

eym

ayst

illbe

able

tore

prod

uce

the

obse

rved

(ave

rage

d)op

acity

spec

tral

inde

xof�⇠

0.73

ifth

egr

ains

are

irreg

ular

and/

orpa

rtof

the

disk

isop

tical

lyth

ick.

Det

aile

dex

plor

atio

nof

this

poss

ibili

tyis

beyo

ndth

esc

ope

ofth

epr

esen

twor

k.If

larg

e,m

m/c

msi

zed,

grai

nsar

ere

spon

sibl

efo

rthe

rela

tivel

ylo

wva

lue

of�

obse

rved

inth

eH

LTa

udi

sk,i

tis

natu

ral

toas

kw

heth

erth

eyca

npr

oduc

ea

pola

rizat

ion

patte

rnth

atm

atch

esth

eob

serv

edon

eth

roug

hsc

atte

ring.

Itis

unlik

ely,

beca

use

the

key

topr

oduc

ing

the

obse

rved

patte

rnis

the

pola

rizat

ion

degr

eeof

the

scat

tere

dlig

htpe

akin

gne

ar90�

(as

inth

eR

ayle

igh

limit)

,and

this

requ

irem

enti

sno

tsat

isfie

dfo

rmm/c

msi

zed

grai

ns.F

orex

ampl

e,fo

rthe

grai

nm

odel

adop

ted

byK

atao

kaet

al.(

2015

),th

epo

lariz

a-tio

nde

gree

(defi

ned

asth

era

tioof

the

two

elem

ents

inth

esc

at-

terin

gm

atrix

,�Z 1

2/Z 1

1,w

hich

ises

sent

ially

the

pola

rizat

ion

frac

-tio

nbu

tcan

beei

ther

posi

tive

orne

gativ

e)is

near

lyze

roat

0.87

mm

fora

llsc

atte

ring

angl

esex

cept

arou

nd13

5�,w

here

itre

ache

sa

(neg

ativ

e)“p

eak”

valu

eof⇠�0.2

for

a max=

1m

man

d1

cm(s

eeth

erig

htpa

nelo

fth

eir

Fig.

2).T

hene

gativ

eva

lue

iskn

own

asth

epo

lariz

atio

nre

vers

al(e

.g.,

Mur

akaw

a20

10;K

irchs

chla

ger&

Wol

f201

4)w

hich

,tog

ethe

rwith

the

shift

ofth

epo

lariz

atio

n“p

eak”

MN

RA

S00

0,1–

12(2

015)

Yang+2016

?

x [AU]

y[AU]

shock

shock

shock

shock

shock

0 0.5 1 1.5 2 2.5 3x 104

0

0.5

1

1.5

2

2.5

3

x 104

log10(ρ)[g

cm−3]

−20.5

−20

−19.5

−19

−18.5

−18

−17.5

−17

−16.5

Hull, Mocz, Burkhart, Hernquist+2016, in prep.

2000 AU

10000 AU

Compression of B-fields along

walls of outflow cavities?

18h29m47.9s48.0s48.1s48.2s48.3s

Right Ascension (J2000)

+1�1604000

4200

4400

4600

Dec

linat

ion

(J20

00)

400AU 0.08

0.16

0.24

0.32

0.40

0.48

0.56

0.64

Pol

ariz

edflux

den

sity

(mJy

bea

m�

1)

Ser-emb 8

Hull+2016, in prep.

Accretion along magnetized filaments?

ALMA

outflows

magnetic fields

850 micron dust emission

sources

ALMA 850 micron dust emission

AREPO simulation

ALMA figure panels (top-left, top-right, bottom-right): 850 micron observations of dust polarization toward Serpens SMM1, Ser-emb 8(N), and Ser-emb 8, three Class 0 young stellar objects in the Serpens star-forming region (d = 415 pc). Yellow line segments are magnetic field orientations, rotated by 90° from the polarization orientations. Grayscale is polarized thermal dust emission. Black contours are total intensity (Stokes I) dust emission, starting at 3𝜎 and increasing logarithmically (SMM1 has a gray 3𝜎 contour to highlight a “hole” in the dust in the SE). Blue and red arrows indicate outflow orientations. The black ellipse in the lower-left is the synthesized beam (resolution element), which is ~0.3ʺ (or ~100 AU). Data are from ALMA Cycle 2 project 2013.1.00726.S. Figures were produced using APLpy.CARMA/ALMA/simulation comparison (bottom-left): comparison of CARMA data (right), ALMA data (left), and an AREPO simulation (bottom). CARMA: black and gray line segments are inferred magnetic field orientation; contours are 1mm dust emission; arrows indicate outflows. ALMA: colorscale is total intensity 850 micron dust emission at the same scale as the CARMA map. Simulation: a slice from a simulation of a star-forming core produced using AREPO (credit: Phil Mocz); colorscale is gas density; white line segments are magnetic fields; magenta vectors are velocity vectors; white line is a shock front created by colliding flows.

ALMA & CARMA Same sources!