! $Id: 5488,v 5.1 1994/08/03 17:33:41 pepsa Exp $ coverpage: title_1: MEASUREMENT OF THE COSMOLOGICAL DEUTERIUM TO HYDROGEN title_2: ABUNDANCE RATIO (CYCLE 4, HIGH) sci_cat: QUASARS & AGN sci_subcat: QUASAR ABSORPTION proposal_for: GO pi_fname: DAVID pi_mi: R. pi_lname: TYTLER pi_inst: UNIVERSITY OF CALIFORNIA AT SAN DIEGO pi_country: USA hours_pri: 4.83 num_pri: 1 hrs: Y funds_amount: 62000 funds_length: 12 ! end of coverpage abstract: line_1: We propose to directly measure the primordial Deuterium/Hydrogen (D/H) line_2: abundance ratio in one QSO absorption system by observing the Lyman line_3: Alpha line with the GHRS (G270M). The D/H ratio is in principle the line_4: best measure of the cosmological baryon to photon ratio (ETA), which line_5: leads to the baryon mass density, a critical parameter which tests the line_6: Big Bang model. Solar system and local interstellar measurements of D/H line_7: are highly uncertain (factor of 10) because of chemical fractionation, line_8: line blending and destruction of D in stars. Selected QSO absorption line_9: systems should be free of all these uncertainties, giving a direct, line_10: high precision measurement. During the past decade we have been line_11: searching for suitable systems, which must have high HI column density line_12: N(HI) and extraordinarily low velocity dispersion (b). The one line_13: absorption system which we propose to observe is the only one that we line_14: know of. It is a Lyman-Alpha forest cloud (primordial gas) with line_15: log N(HI)=17.17 +/- 5% towards Q1718+481. If primordial D/H=10(-4), line_16: the best guess, our error on D/H would be 30% (+/- 15% on ETA). If D is line_17: not detected, our 95% confidence upper limit would be 1.9x10(-5) line_18: (barely consistent with corrected local measurements). ! ! end of abstract general_form_proposers: lname: TYTLER fname: DAVID title: PI mi: R. inst: UNIVERSITY OF CALIFORNIA-SAN DIEGO country: USA esa: N ! lname: FAN fname: XIAOMING inst: UNIVERSITY OF CALIFORNIA-SAN DIEGO country: USA esa: N ! lname: LANZETTA fname: KENNETH mi: M. inst: UNIVERSITY OF CALIFORNIA-SAN DIEGO country: USA esa: N ! lname: LU fname: LIMIN inst: UNIVERSITY OF WISCONSIN-MADISON country: USA esa: N ! lname: SAVAGE fname: BLAIR mi: D. inst: UNIVERSITY OF WISCONSIN-MADISON country: USA esa: N ! lname: WOLFE fname: ARTHUR mi: M. inst: UNIVERSITY OF CALIFORNIA-SAN DIEGO country: USA esa: N ! ! end of general_form_proposers block general_form_text: ! question: 3 section: 1 line_1: The GHRS spectrum of the Lyman-Alpha line at absorption system line_2: zabs=0.703 towards Quasar Q1718+4807 will be carried out with G270M at line_3: 2043-2120 Angstroms. The Small Science Aperture (SSA) will be used line_4: to guarantee the resolution (FWHM=15 km/s) needed to resolve line_5: the DI and HI lines. line_7: VALIDATION ERROR line_8: We choose to use G270M at 2082A central wavelength. This is line_9: unusual, because it is below its usual minimum of 2200A, and line_10: this causes a validation error. We asked Dennis Ebbets about line_11: this validation error. He does not know why it ocurrs, but he does line_12: known that there is no problem taking data at 2082 with G270H, and line_13: that the spectrum does fall on the detector. line_15: G270M IS MORE EFFICIENT THAN G00M AT 2082A. line_16: We choose G270M rather than G200M because we believe that G270M gives line_17: about 2 times the flux. There are two reasons why be believe this: line_19: 1. Blaire Savage has GTO spectra of star U-Columbae with both G200M line_20: and G270M, both covering 1950-2200A. These spectra show that G270M gives line_21: twice the signal at 2082A. We do not believe that this is an aperture line_22: centering problem, though this needs to be checked. ! question: 3 section: 2 line_1: line_2: 2. The pre-launch calibration of the gratings (e.g. HRS Inst Handbook line_3: Oct 1985, p.38) gave the same factor of two advantage to G270M at 2082A. line_5: However we are aware that other data contradict this result, and suggest line_6: that G200M is better. The calibrations in the current instrument line_7: handbook, are taken from the SV report. On p.6-9 of the ``Final Report line_8: of the Sci. Verification Program, 24 Feb 1992'' there are efficiency line_9: curves which give: line_11: item G200M G270M line_12: efficiency E from p.6-9 9.6e11 9.7e11 line_13: Ang/diode at 2082 0.0766 0.0966 line_14: efficicy per Angstrom 1.27e13 0.95e13 line_16: We asked Steve Hulbert (STScI 410-338-4911) about these, and he line_17: gave similar numbers (9.7e11 for G200, and 9,2e11 for G270M) on line_18: Sept 28 1993. Blaire Savage (608 262 2395), Dennis Ebbets line_19: (Ball Aerospace, Boulder 303-939-4000x5964) and Rich Robinson line_20: (GHRS, Goddard) have looked into this and find that we are line_21: correct, G270 is 2 times faster, and the SV numbers are wrong. line_22: We will report this to Steve Hulbert. ! question: 3 section: 3 line_1: line_2: A second difference between G200M and G270M is resolving power. line_3: At 2082A, G270M has a slightly lower resolution of R=20,000, line_4: compared to 24,000 with G200M. (the Sci. Verification Manual line_5: page 7-9). line_7: We MAY change to the LSA if we change to the G200M. Depends on the line_8: resolution and how well we will know the psf and wavelength scale. ! question: 3 section: 4 line_1: Our IUE spectrum of Q1718+4807 shows a flux F=1.5 x 10^(-14) line_2: ergs/s/cm^2/A at the wavelength of the Lyman-Alpha line (2070A). The line_3: sensitivity of G270M with the LSA is twice that of G200M at that line_4: wavelength, E=2x9.5x10^(11) cts/sec/diode at 2070A. Taking into account line_5: the factor of 2/3 deduction in sensitivity when the SSA is used, the line_6: signal count rate of 0.020 cts/sec/diode is two times the background line_7: of 0.01 cts/sec/diode. Thus 5.02 hours is needed to reach the desired line_8: S/N=15 per diode. line_10: All measurements will be carried out with a substep pattern of 2 line_11: samples per diode. Standard FP-SPLIT will be used in order to reduce line_12: detector window and photocathode fixed pattern noise. The total line_13: spacecraft time is estimated to be 6.88 hours using the phase I line_14: resource estimator. line_16: We will follow recommended observation procedures to use LSA with a line_17: 5x5 spiral search area (search-size=5) for initial onboard acquisition line_18: of our target. After initial acquisition, we will use ACQ/PEAKUP for line_19: both LSA and SSA respectively to more precisely center the object in line_20: the aperture. Then we take an IMAGE of the field to confirm that line_21: the telescope has centered our target properly in the dark-taking line_22: aperture before our final real exposure. ! question: 4 section: 1 line_1: HST observations are required because the redshift of the target line_2: absorption system places the Deuterium and Hydrogen Lyman lines in line_3: the UV. We must observe low redshift absorbers for the following line_4: reasons: line_6: (1) Blending is a major problem at high z. This is clearly shown in line_7: Figure 2, where we show three REAL spectra, each with the same line_8: resolution and velocity coverage (as seen from a Lyman-Alpha line in line_9: the middle of each spectrum). Note that blending remains serious at line_10: HST wavelengths around 2800A (middle panel), but is much less a line_11: problem at our chosen wavelengths (bottom panel). This marked change line_13: is partially due to the evolution of the Lyman-Alpha forest. It is line_14: also due to the number of metal lines. A metal line systems can have line_15: many lines in the UV, so the total number of lines in a spectrum line_16: scales with the integrated number of systems between us and the QSO, line_17: which is about (1+zem)^2. line_19: (2) The worst sort of blending is that with components of the very line_20: system under study. Systems at low z <= 1 are simpler, with fewer line_21: velocity components, and (or) lower Doppler parameter b values, than line_22: those at moderate to high z (Boisse etal 1992, Steidel & Sargent 1992). ! question: 4 section: 2 line_1: (3) Our goal is to measure primordial D/H much more precisely than has line_2: been deduced from local ISM data, and the most accurate and reliable line_3: D/H measurements must come from the lowest z systems, for the above line_4: two reasons. line_6: (4) We have been searching for suitable ground based targets absorption line_7: systems for three years now, with no success. We frequently find line_8: systems which look promising at the start, but when we obtain more, line_9: higher resolution data, we invariably find that they have more velocity line_10: components, and so are much less likely to yield a clean D/H line_11: measurement. We now all agree that it is unlikely that a believable line_12: result will be obtained with a 4-m telescope. The Keck should be able line_13: to measure D/H because it can (i) get much higher S/N, and (ii) do many line_14: more targets, which will be fainter because we have ruled out nearly line_15: all known QSOs with magnitude V <= 18. But we stress that the two HST line_16: targets are much better than any known Keck targets. line_18: D/H measurement is of sufficient cosmological significance that line_19: consistent measurements in many systems will ultimately be required line_20: to give a reliable, high precision value. We expect that this effort line_21: will be continued for years to come (c.f. Hubble constant, microwave line_22: background). There are many reasons why an individual system could line_23: give a misleading D/H (line blends, unobserved velocity substructure, ! question: 4 section: 3 line_1: non-gaussian velocities, stellar processing etc.). Candidates for line_2: D/H measurement are sufficiently rare, that measurements should be line_3: attempted in all apparently favorable cases. ! question: 5 section: 1 line_1: No request for special scheduling or real-time observations. line_2: Note that we do request an image with GHRS prior to the exposure, line_3: to confirm that the target is in the aperture. ! question: 6 section: 1 line_1: We request a PT-LAMP exposure be obtained immediately after the line_2: science observation. This is necessary because the proposed line_3: scientific analysis requires comparison of absorption velocities line_4: between different ion species. Therefore, accurate wavelength line_5: calibration is crucial to our final result. ! ! ! ! ! question: 8 section: 1 line_1: Our observations use the G270M grating at wavelength around 2070 A line_2: (which is not generally recommended). We have direct evidence that line_3: the G270M can be used at 2000