SPACE TELESCOPE SCIENCE INSTITUTE Operated for NASA by AURA TIPS / COS Update The NUV Gratings 18 October 2007 Last COS TIPS presentation: 21 December.

SPACETELESCOPESCIENCEINSTITUTE

Operated for NASA by AURA

TIPS / COS UpdateThe NUV Gratings

18 October 2007

Last COS TIPS presentation: 21 December 2006

Keyes – 18 October 2007Slide 2 of 19

NUV grating characteristics

COS NUV grating blanks coated with Au and Cr Original COS NUV gratings (G185M, G225M, G285M, and

G230L) also were MgF2 coated over Al to protect reflective surface

Routine design process shifts Wood’s anomaly out of spectral region of interest to shorter wavelengths– Anomalous distribution of energy in diffracted light; highly polarization

dependent; sensitive to groove spacing and depth comparable to wavelength diffracted

– this is a resonance effect – it will not show up in bulk reflectivities, such as from flat-mirror witness samples, but as a complicated wavelength and polarization-dependent grating efficiency modulation.

Keyes – 18 October 2007Slide 3 of 19

NUV Gratings

For COS NUV gratings the addition of the last layer of coating (MgF2) apparently shifts Wood’s anomaly directly into bandpass of interest decreasing throughput substantially – Solution 1: use MgF2-coated “longer-wavelength” grating in shorter

wavelength region where throughput is nominal> Consequence: resolution degraded as does not change> Shifting G225M for use in G185M region works well, G285M to G225M does not

– Solution 2: do not apply MgF2 coating to Al surface; anomaly stays at design location

> This is employed for G225M and G285M gratings> Consequence: Al oxidizes within days of deposition; forms thin coating; literature

indicates oxide layer reaches maximum depth of ~5 nm.

Keyes – 18 January 2007Slide 4 of 19

Variances in measured COS sensitivities in vacuum 2006/2003

Bare Al

Bare Al

Grating Efficiency Issues:

2003 T/V data compared to 2006 T/V data- this data shows significant variations in efficiency between 2003 and 2006, including indications that some channels at some wavelengths improved in efficiency by as much as 40% and others diminished by ~25%- the majority, but not all, of these discrepancies can be explained by the polarization sensitivity of COS, and the difference in polarization generated by the 2003 Calibration delivery system and the 2006 calibration delivery system - this polarization sensitivity has been demonstrated on the flight spares, and in the flight optics

COS polarization data

Polarization and COS

0.700

0.800

0.900

1.000

1.100

1.200

1.300

1.400

1.500

1.600

1.700

1.800

1200 1700 2200 2700 3200

Wavelength (Anstroms)

Ve

rtic

al/

Ho

rizo

nta

l

Scrambled, CDS+RAS/Cal Unscrambled, CDS Only

Grating Efficiency Monitoring

Grating efficiency monitoring tests of the NUV gratings (no monitoring is done of the FUV gratings)

– a sequence of exposures are performed on the NUV gratings at various wavelength settings using the internal lamps, and the count rates are recorded – ratios of count rates in different channels at the same wavelength (when possible) are also calculated – this process should remove the possible effects of lamp variability

These tests indicate that the lamp is slowly losing brightness (not unexpected and at an acceptable rate) – and that G225M and G285M are experiencing steady and continuing degradation

Keyes – 18 October 2007Slide 9 of 14

C B A C B A

PtNeWavecal

ExternalScience

NUVMAMA

COS Spectral Layout Internal Wavecals and Science Spectra

– Obtain (continuous or flashed) internal PtNe spectra at same time as science exposure

– Track internal PtNe lines and apply shifts to science spectrum (all events time-tagged) in COS data pipeline

Keyes – 18 October 2007Slide 10 of 19

G185M / G225M

G225M / G285M

Left: G185M / G230L Both MgF2

Right: G225M / G230L

G285M / G230L

COS NUV Performance Changes

Compare G285M and G225M to G230L, TA1, and Compare G285M and G225M to G230L, TA1, and to witness coupons (dashed)to witness coupons (dashed)

0.7

0.75

0.8

0.85

0.9

0.95

1

1.05

1.1

170 190 210 230 250 270 290 310

TV

06

/TV

03

Wavelength

TV06/TV03 Performance ratios with coupon ratios (dashed)

TA1 Raw Ratio

G225 Ratio

G285 Ratio

G185 Ratio

G230 Ratio

'COS Aft' Al/MgF2

CW-9 Al/MgF2

CW-14 Al/MgF2

225F1 Bare Al

285D1 Bare Al

225F2 Bare Al

NUV Performance Summary

The degradation of the G225M and G285M gratings are real – at the rate of 1.4%/year and 4.4%/year respectively. To date no process has been identified that can explain the observed behavior

Over the past 3 years measurements of the flight spare gratings have shown very similar degradation rates (1% and 4.5%)

Measurements of the witness sample mirror coupons have shown no degradation over the past 5 years

Summary of observations

Little loss of efficiency in FUV channels Anomalous efficiency loss in G225M and G285M

(high density, bare aluminum) – External and internal calibration data give similar results

GSFC analysis shows gratings and coupons are very clean

Possible explanations for loss of efficiency considered to date

Is efficiency loss due to simple hydrocarbon contamination?

– This is not consistent with FUV, coupon or TA1 data – we would expect more loss in the FUV channels and in short wavelength coupon data than we see

– wash/analysis of witness samples inside instrument revealed no significant contamination

Is loss due to migration of Au substrate into Al?

– No gold was observed in GSFC analysis, not consistent with Al/MgF2 optic stability, and we would expect to see a similar reflectivity loss in the bare aluminum witness coupon data

Possible explanations for loss of efficiency considered to date (cont)

Is the ‘loss’ due to change in test setups – change in polarization of calibration signal?– Polarization testing of COS indicates that this cannot explain all of

the loss in apparent efficiency– NUV efficiency monitoring indicates continuing, steady

degradation Could in-air NUV testing be polymerizing hydrocarbons onto mirrors

and gratings, with stronger effect on high line density optics than on low density or purely reflective optics?– GSFC testing rules this out

Possible explanations for loss of efficiency considered to date (cont)

Could pin-holes be forming in Al coating or could mixing of Au through Cr layer into Al be degrading reflectivity?– pin-holing and layer mixing cannot be ruled out on flight gratings without

examination of flight units ; however no indications of either in detailed special tests of spare gratings at GSFC

JY modeling of gratings with a thick (9nm) Al2O3 layer is consistent with our observed performance.– This thickness is twice what is suggested likely by literature– Continuing effort at GSFC to determine actual thickness of oxide

layer – possible update at Oct COS MSR next week

Current Status Summary – Oct 2007

Degradation is real for G225M and G285M; no effect seen for MgF2 coated NUV or FUV optics

COS gratings show definite polarization sensitivity which will be evaluated on orbit; most polarized astronomical sources <5% polarized

Current degradation rate projects to COS ≥ 4x STIS efficiencies per exposure at launch (for likely COS target flux-levels)

Contamination has been ruled out Observational uncertainty and systematics have been ruled out

Current Status Summary – Oct 2007(cont)

JY modeling with a thick (9nm) Al2O3 layer is consistent with the observed G225M/G285M performance; literature suggests that this is thicker than we should expect to see by ~2x; effort continues at GSFC to determine oxide layer thickness

Will it continue on orbit? – unknown at present Spare gratings have been fabricated and coated; are in

testing at GSFC at present No grating swap is likely at this point

Keyes – 18 October 2007Slide 18 of 19

Background Information

Keyes – 18 October 2007Slide 19 of 19

G285M flight and spare grating performance changes

Estimated G285M-D (flight) vs G258M-E (spare)

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

250 260 270 280 290 300 310 320

Wavelength

Ab

solu

te E

ffic

ien

cy (

Un

po

lari

zed

)

G285M-E 2/1/02 (G285M spare) Bare Al

G285M-E 2/9/07 (G285M spare) Bare Al

G285M-D 1/16/02 (G285M Flight) Bare Al

G285M-D 12/06 est. (G285M flight) Bare Al

G285M-D 2006 modeled by multiplying 2002 grating only values by the 2003-2006 change in COS G285M efficiency, and removoing estimated TA1 contribution.

Keyes – 18 October 2007Slide 21 of 19

The Degradation Projection Scenario

Linear extrapolation of observed degradation to 9/11/2008 Linear extrapolation of observed degradation to 9/11/2008 launch date at ~ 0.5% per month yields additional ~11% loss of launch date at ~ 0.5% per month yields additional ~11% loss of sensitivity before COS gets to orbit; compared to TV I values:sensitivity before COS gets to orbit; compared to TV I values:– G285M will have lost ~25% throughput waiting to flyG285M will have lost ~25% throughput waiting to fly

– G225M will have lost ~20% throughput waiting to flyG225M will have lost ~20% throughput waiting to fly

In comparisons to follow, we have assumed 25% degradation In comparisons to follow, we have assumed 25% degradation (i.e., launch throughput = 0.75 x TV I throughput)(i.e., launch throughput = 0.75 x TV I throughput)

Keyes – 18 January 2007Slide 22 of 19

HST orbits required to reach S/N=10 at 2500 Ǻ (R~20,000 (0.12 Ǻ) binning)

-16.5

-16.0

-15.5

-15.0

-14.5

-14.0

-13.5

-13.0

0 10 20 30 40 50 60 70 80 90 100HST Orbits

log

Flu

x (e

rg c

m-2

sec

-1 Ǻ

-1)

COS G225M; 3x7 binningno degradation; ground dark

STIS E230M + 0.2x0.2 aperture3x7 binning

COS G225M; 3x7 binning25% degradation; worst-case dark

COS G225M; 3x7 binning25% degradation; ground dark

COS G225M; 3x7 binningno degradation; worst-case dark

Observing Efficiency Comparison: COS vs STIS – darks included

Keyes – 18 October 2007Slide 23 of 19

Impacts: Ratios of exposure to achieve same S/N with degraded sensitivity

Object Flux

COS Exposure ratio to reach S/N = 1025%-degraded Sλ vs no-loss COS Sλ

with COS ground dark (with worst-case on-orbit dark)

COS/STIS Exposure ratio to reach S/N = 10

25%-degraded COS Sλ vs STIS Sλ

with COS ground dark (with worst-case on-orbit dark)

1.e-13 1.33 (1.34) 0.43 (0.44)

1.e-14 1.35 (1.41) 0.29 (0.36)

1.e-15 1.45 (1.64) 0.09 (0.24)

5.e-16 1.50 (1.69) 0.07 (0.23)

2.e-16 1.62 (1.74) [>40 orbits] 0.05 (0.22) [>40 orbits]

1.e-16 1.68 (1.76) [>40 orbits] 0.04 (0.21) [>40 orbits]

Keyes – 18 October 2007Slide 24 of 19

Impacts – Single COS grating setting used

Single COS grating setting used:Single COS grating setting used:– Bright limit (ignore background): simply increase science exposures to Bright limit (ignore background): simply increase science exposures to

compensate for sensitivity loss: 1.25-1.3x to achieve same S/Ncompensate for sensitivity loss: 1.25-1.3x to achieve same S/N> In bright limit: STIS/COS(no-loss) exposure ratio ~3x; so COS SIn bright limit: STIS/COS(no-loss) exposure ratio ~3x; so COS Sλλ must must

degrade to 0.33 TV I level for COS=STIS efficiencydegrade to 0.33 TV I level for COS=STIS efficiency– Faint “limit”: Faint “limit”:

> For a 40-orbit COS observation to achieve S/N=10 at 2500 For a 40-orbit COS observation to achieve S/N=10 at 2500 ǺǺ.:.:• No-loss case with ground dark: FNo-loss case with ground dark: Fλλ==2.0 e-162.0 e-16• No-loss case with worst-case dark: FNo-loss case with worst-case dark: Fλλ=4=4.0 e-16.0 e-16• With 25% degradation and ground dark: FWith 25% degradation and ground dark: Fλλ=2.5 e-16=2.5 e-16 ; however, this ; however, this

flux requires 26 orbits in no-loss case (1.5x longer with degradation)flux requires 26 orbits in no-loss case (1.5x longer with degradation)• With 25% degradation and worst-case dark: FWith 25% degradation and worst-case dark: Fλλ=5.5 e-16 ; however, this =5.5 e-16 ; however, this

flux requires 24 orbits in no-loss case (1.7x longer with degradation) flux requires 24 orbits in no-loss case (1.7x longer with degradation) – the limiting flux for a STIS 40-orbit observation is 1.2e-15the limiting flux for a STIS 40-orbit observation is 1.2e-15

Keyes – 18 October 2007Slide 25 of 19

Impacts: Summary and Questions

For observing the fainter targets with degraded SFor observing the fainter targets with degraded Sλλ, two considerations , two considerations important:important:– Brighter limiting flux for observation at a particular S/NBrighter limiting flux for observation at a particular S/N– Increase of exposure time to reach a target at a specific S/NIncrease of exposure time to reach a target at a specific S/N

For most STIS targets the modest difference in COS limiting flux due to For most STIS targets the modest difference in COS limiting flux due to the degradation does not appear to be an important consideration the degradation does not appear to be an important consideration

Targets with the faintest fluxes attempted by STIS (~1.e-15) in 40 orbits Targets with the faintest fluxes attempted by STIS (~1.e-15) in 40 orbits are important, but with 25% degradation and worst-case background would are important, but with 25% degradation and worst-case background would require ~12-13 orbits require ~12-13 orbits for a single grating settingfor a single grating setting with COS (or ~7-8 orbits with COS (or ~7-8 orbits if no degradation). if no degradation). – Question: is 8 versus 12 orbits significant for science at this flux level?Question: is 8 versus 12 orbits significant for science at this flux level?

> Depends on number of targets and/or COS grating settings requiredDepends on number of targets and/or COS grating settings required> In most cases where multiple COS grating settings are needed; COS In most cases where multiple COS grating settings are needed; COS

probably would not be chosen as the SI of choiceprobably would not be chosen as the SI of choice The difference in limiting flux between 25% degradation and no The difference in limiting flux between 25% degradation and no

degradation (for S/N=10 at 2500 degradation (for S/N=10 at 2500 ÅÅ and worst-case background) is 6. e-16 and worst-case background) is 6. e-16 versus 4. e-16 (or for best-case dark, 2. e-16)versus 4. e-16 (or for best-case dark, 2. e-16)