20.07.2015 Views

Spatio-spectral encoding of fringes in optical long ... - GSU Astronomy

Spatio-spectral encoding of fringes in optical long ... - GSU Astronomy

Spatio-spectral encoding of fringes in optical long ... - GSU Astronomy

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.

D. Mourard et al.: <strong>Spatio</strong>-<strong>spectral</strong> <strong>encod<strong>in</strong>g</strong> <strong>of</strong> <strong>fr<strong>in</strong>ges</strong> <strong>in</strong> <strong>optical</strong> <strong>long</strong>-basel<strong>in</strong>e <strong>in</strong>terferometry60Differential Phase (degree)Differential Closure Phase (degree)40200-20-40-60720 725 730 735 740 745 7506040200-20-40-60720 725 730 735 740 745 750Wavelength (nm)Fig. 8. Top: plots represent the differential phases <strong>in</strong> a narrow <strong>spectral</strong>band <strong>of</strong> 0.6 nm with respect to the wavelength (basel<strong>in</strong>e E1E2 <strong>in</strong> black,E2W2 <strong>in</strong> red, E1W2 <strong>in</strong> green). Bottom: plot represents the differentialclosure phase deduced from the three <strong>in</strong>dividual differential phases.Each estimate corresponds to 400 s <strong>of</strong> observation <strong>of</strong> calibrator starHD 55185 us<strong>in</strong>g VEGA with three telescopes (E1E2W2) at medium<strong>spectral</strong> resolution and fr<strong>in</strong>ge track<strong>in</strong>g done with CLIMB. Individualerror bars are not plotted for clarity. On basel<strong>in</strong>es E1E2 and E2W2, theerrors are at the level <strong>of</strong> 0.5 ◦ , whereas on the basel<strong>in</strong>e E1W2 they arecloser to 5 ◦ .Eqs. (23)<strong>in</strong>to(22)givesΔφ 12,i = θ λ1 ,i − θ λ2 ,i + 2πχ i( 1λ 1− 1 λ 2)+ φ disp,12 , (24)where the dispersive differential phase is( δχi (λ 1 )φ disp,12 = 2π − δχ )i(λ 2 )· (25)λ 1 λ 2Hence, if the <strong>in</strong>terferometer operates <strong>in</strong> vacuum and if the OPDis servo<strong>in</strong>g to zero, the third and fourth terms <strong>in</strong> Eq. (24) disappearand the differential phase depends only on the phase <strong>of</strong>the object. In the case <strong>of</strong> VEGA, the OPDs <strong>of</strong> each basel<strong>in</strong>eare servo<strong>in</strong>g to a non-zero value and the delay l<strong>in</strong>es operate <strong>in</strong>air. Therefore, the differential phases are corrupted by both thestatic OPD and the dispersive differential phase φ disp,12 ,whichexpla<strong>in</strong>s the l<strong>in</strong>ear and/or quadratic trends <strong>of</strong> the curves presented<strong>in</strong> Fig. 8.Extraction <strong>of</strong> the astrophysical signal requires the removal<strong>of</strong> these l<strong>in</strong>ear and/or quadratic trends. This is usually doneby fitt<strong>in</strong>g and subtract<strong>in</strong>g a low-order polynomial function ora model <strong>of</strong> the dispersive differential phase to the differentialphase (Matter et al. 2010). The ma<strong>in</strong> drawback <strong>of</strong> this methodis the loss <strong>of</strong> both the l<strong>in</strong>ear and/or quadratic terms <strong>in</strong> the astrophysicalsignal. For studies <strong>of</strong> circumstellar discs for example,it is not a major problem because the astrophysical signal ismore complex and conf<strong>in</strong>ed only to <strong>spectral</strong> l<strong>in</strong>es. So the neighbor<strong>in</strong>gcont<strong>in</strong>uum could be used to remove the l<strong>in</strong>ear and/orquadratic trends. However, other astrophysical studies, such asb<strong>in</strong>ary stars, require us to measure signatures <strong>in</strong> the differentialphase that are very close to l<strong>in</strong>ear or quadratic trends. In thatcase, the differential phase could not be used directly. We def<strong>in</strong>eda new <strong>in</strong>terferometric observable, which is the differentialclosure phaseΔφ (3)12 =Δφ 12,1 +Δφ 12,2 − Δφ 12,3 . (26)Fig. 9. Sky coverage <strong>of</strong> potential calibrator stars well adapted toVEGA/CHARA <strong>in</strong> 3T/4T mode with CLIMB as external IR group delaytrack<strong>in</strong>g. All these stars have a <strong>spectral</strong> type rang<strong>in</strong>g from O (cross),B (star), A (diamond), to F (triangle).It can easily be shown that this quantity does not depend on theOPD but is the difference <strong>in</strong> the classical closure phase <strong>of</strong> each<strong>spectral</strong> bandΔφ (3)12 =Δφ 1 − Δφ 2 , (27)where Δφ 1 and Δφ 2 are respectively the classical closure phase<strong>of</strong> the reference channel and <strong>of</strong> the narrow channel. Figure 8(bottom plot) illustrates that the l<strong>in</strong>ear and/or quadratic trendsdue to the OPD have disappeared <strong>in</strong> the differential closurephase. The mean <strong>of</strong> the differential closure phase <strong>in</strong> one narrowchannel is Δφ (3)12 = 0.2 ± 11 ◦ all over the <strong>spectral</strong> band <strong>of</strong> 30 nm.4. Discussion <strong>of</strong> possible VEGA/3T-4T sciencecases4.1. 3T-4T operation issuesEven if progress <strong>in</strong> fr<strong>in</strong>ge track<strong>in</strong>g allows fa<strong>in</strong>ter objects to beobserved and/or lower visibilities to be recorded, this requiresf<strong>in</strong>d<strong>in</strong>g calibrators that are bright enough for both the scientific<strong>in</strong>strument and the fr<strong>in</strong>ge tracker. Moreover, us<strong>in</strong>g very <strong>long</strong>basel<strong>in</strong>es makes the choice <strong>of</strong> calibrators much more difficults<strong>in</strong>ce lots <strong>of</strong> them are partially resolved. The calibrators mustalso be close to the scientific targets <strong>in</strong> terms <strong>of</strong> both distanceand <strong>spectral</strong> type to avoid any bias <strong>in</strong> the calibrated <strong>in</strong>terferometricobservables. As an example, we selected calibrators amongthe 9500 stars (decl<strong>in</strong>ation larger than −30 ◦ and magnitude Vbrighter than 7.0) <strong>in</strong> the VEGA <strong>in</strong>ternal catalog by apply<strong>in</strong>g filtersto parameters such as magnitude and diameter so as to fit theVEGA capabilities. Angular diameters were calculated us<strong>in</strong>g aspectrophotometric polynomial law based on the V magnitudeand the (V − K) color <strong>in</strong>dex (Bonneau et al. 2006b). Accord<strong>in</strong>gto the current limit<strong>in</strong>g magnitude, we considered only the starsbrighter than 6.5<strong>in</strong>theV band and brighter than 7 <strong>in</strong> the K band(for CLIMB that performs the group delay track<strong>in</strong>g). We identified1866 potential calibrators <strong>in</strong> all parts <strong>of</strong> the sky with anangular diameter smaller than 0.4 mas provid<strong>in</strong>g a high visibilitymeasurement even on the largest basel<strong>in</strong>es. All <strong>of</strong> them areOBAF stars (Fig. 9). Figure 10 shows the histogram <strong>of</strong> the number<strong>of</strong> calibrators found per regions <strong>of</strong> sky <strong>of</strong> 1h <strong>in</strong> right ascensionand 5 ◦ <strong>in</strong> decl<strong>in</strong>ation.A110, page 7 <strong>of</strong> 9

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

Saved successfully!

Ooh no, something went wrong!