22.12.2013 Views

Investigations of Faraday Rotation Maps of Extended Radio Sources ...

Investigations of Faraday Rotation Maps of Extended Radio Sources ...

Investigations of Faraday Rotation Maps of Extended Radio Sources ...

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

3.3. THE METHOD 47<br />

(e.g. Subramanian 1999) where the longitudinal, normal, and helical autocorrelation<br />

functions, M L (r), M N (r), and M H (r), respectively, only depend on the distance, not<br />

on the direction. The condition ∇ ⃗ · ⃗B = 0 leads to ∂/∂r i M ij (⃗r) = 0 (here and<br />

below the sum convention is used). This allows to connect the non-helical correlation<br />

functions by<br />

d<br />

dr (r2 M L (r)) (3.10)<br />

M N (r) = 1 2r<br />

(Subramanian 1999). The zz-component <strong>of</strong> the magnetic autocorrelation tensor depends<br />

only on the longitudinal and normal correlations, and not on the helical part:<br />

M zz (⃗r) = M L (r) r2 z<br />

r 2 + M N(r) r2 ⊥<br />

r 2 with ⃗r = (⃗r ⊥, r z ) , (3.11)<br />

which implies that <strong>Faraday</strong> rotation is insensitive to magnetic helicity. It is also useful<br />

to introduce the magnetic autocorrelation function<br />

w(⃗r) = 〈 ⃗ B(⃗x) · ⃗B(⃗x + ⃗r)〉 = M ii (⃗r) , (3.12)<br />

which is the trace <strong>of</strong> the autocorrelation tensor, and depends only on r (in the case<br />

<strong>of</strong> a statistically isotropic magnetic field distribution, in the following called briefly<br />

isotropic turbulence):<br />

w(⃗r) = w(r) = 2M N (r) + M L (r) = 1 r 2 d<br />

dr (r3 M L (r)) . (3.13)<br />

In the last step, Eq. (3.10) was used. Since the average magnetic energy density is given<br />

by 〈ε B 〉 = w(0)/(8π) the magnetic field strength can be determined by measuring the<br />

zero-point <strong>of</strong> w(r). This can be done by <strong>Faraday</strong> rotation measurements: The RM<br />

autocorrelation can be written as<br />

C ⊥ (r ⊥ ) = 1 2<br />

∫ ∞<br />

−∞<br />

√<br />

dr z w( r⊥ 2 + r2 z) =<br />

∫ ∞<br />

r ⊥<br />

dr<br />

r w(r)<br />

√ . (3.14)<br />

r 2 − r⊥<br />

2<br />

and is therefore just a line-<strong>of</strong>-sight projection <strong>of</strong> the magnetic autocorrelations. Thus,<br />

the magnetic autocorrelations w(r) can be derived from C ⊥ (r ⊥ ) by inverting an Abel<br />

integral equation:<br />

w(r) = − 2<br />

π r<br />

= − 2 π<br />

d<br />

dr<br />

∫ ∞<br />

r<br />

∫ ∞<br />

r<br />

dy<br />

dy y C ⊥(y)<br />

√<br />

y 2 − r 2 (3.15)<br />

C′ ⊥ (y) √<br />

y 2 − r 2 , (3.16)<br />

where the prime denotes a derivative. For the second equation, it was used that w(r)<br />

stays bounded for r → ∞.<br />

Now, an observational program to measure magnetic fields is obvious: From a<br />

high quality <strong>Faraday</strong> rotation map <strong>of</strong> a homogeneous, (hopefully) isotropic medium <strong>of</strong><br />

known geometry and electron density (e.g. derived from X-ray maps) the RM autocorrelation<br />

has to be calculated (Eq. (3.8)). From this an Abel integration (Eq. (3.15) or<br />

(3.16)) leads to the magnetic autocorrelation function, which gives 〈B 2 〉 at its origin.<br />

Formally,<br />

〈B 2 〉 = w(0) = − 2 ∫ ∞<br />

dy C′ ⊥ (y) , (3.17)<br />

π y<br />

0

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!