InstrumentalWidthOfEIS
Back to current versionRestore this version

I'm getting first level data from zero level after usual reduction by 'eis_prep'. However, when i am fitting the lines by xcfit, i'm getting the spectrum slightly displaced from the wavelength values given in the data window. What error this will make in the calculations?. Secondly, can we get the accurate width variation ? , and can we use these datas for line-width variation studies ?

--A.K. Srivastava, 02-July-2007


In response to your first query - do you get the same shift (and what is it) if you got with a simple gauss_fit?

In response to your second query we would need more information. Various work has been carried on line widths looking at the differences at different stages or positions of loops by Hara et al., and Doschek et al (both been submitted), so clearly line widths can be measured.What variation are you referring to?

--Hiro Hara, 09-Jul-2007


About the instrumental width, Doschek et al 2007 proposed a FWHM of 2.5 pixels (i.e. 56 mA) in orbit given that the FWHM is 1.956 pixels in laboratory and Brown et al. 2008 give us between 54mA and 57mA depending on the wave band and the wavelength.

Is there any instrumental profile available?

Is it a Gaussian profile? What is the width commonly used ?

--Celine Boutry, 16-Jan-2009


Hi Céline,

I think the answer may be in your question, actually.

The Doschek et al. article states a single number, but that is based on comparisons with older data. However, it is still consistent with the numbers derived from comparing the pre-launch laboratory EIS calibration data with the on-orbit EIS data by Brown et al. (2008). The Brown et al. numbers are inferred widths, but the reasoning is pretty logical; the fact that the Doschek et al. (2007) number falls within that range of 0.054nbsp;— 0.057nbsp;Å is comforting.

In summary, Brown et al. (2008) assume a Gaussian instrumental line profile, and they infer an instrumental width of 0.054 Å in the short-wavelength channel (170 — 210 Å) and 0.057 Å for the long-wavelength channel (250 — 290 Å).

NOTE! However, the instrumental width that is discussed by Brown et al. is not the FWHM of the instrumental width, but rather the (1/e)-1/2 half-width σ that naturally falls out of the Gaussian function:

NOTE! As Céline Boutry points out, this is the FWHM of the instrumental width, not the Gaussian width σ as previously incorrectly stated here.


Hi,

Thank you for your answer. I have to say that I'm puzzled about which kind of width is discussed in the papers.

About the Brown et al 2008 paper, they wrote : "The thermal Doppler FWHM is given by ΔD=7.162× 10e(-7)λ(T/M)1/2". The 7.162× 10e(-7) coefficient is : 2 × (2 ln(2) )1/2 × *(kB/u) 1/2×c

where u is the Atomic Mass Unit in kg, kB is the Boltzmann constant and c is the Speed of light. So ΔD is well a FWHM.

and then they wrote : "Δ=(Δinst2NT2D2)1/2"

So Δinst needs to be a FWHM to be homogeneous.

The value given just then is Δinst=0.055=(0.0562-0.0092)1/2 which is consistent with the Doschek et al value if it is the FWHM but not if it is the gaussian σ.


So, I don't understand your comment that this value in Brown et al is the gaussian σ and not the FWHM.

--Celine Boutry, 21-Jan-2009

Hi, Céline.

Vous avez raison! :-)

I looked at my original notes, and you're quite right: the σ-to-FWHM factor is included.

I've amended the above to show this now. Thanks.

--Dave Williams, 22-Jan-2009