FIFI-LS: The Field-Imaging Far-Infrared Line Spectrometer on SOFIA

Christian Fischer1,12, Simon Beckmann1, Aaron Bryant2, Sebastian Colditz1, Fabio Fumi2,Norbert Geis3, Mourad Hamidouche4, Thomas Henning5, Rainer H€onle2, Christof Iserlohe1,Randolf Klein6, Alfred Krabbe1,2, Leslie Looney7, Albrecht Poglitsch3, Walfried Raab8,

Felix Rebell2, Dirk Rosenthal9, Maureen Savage10,6, Mario Schweitzer11,Christopher Trinh6 and William Vacca6

1Deutsches SOFIA InstitutUniversity of Stuttgart, 70569 Stuttgart, Germany

2Institute of Space SystemsUniversity of Stuttgart, 70569 Stuttgart, Germany3Max Planck Institute for Extraterrestrial Physics

85748 Garching, Germany4Remote Sensing Technology Institute

DLR (German Aerospace Center)82234 Wessling, Germany

5Max Planck Institute for Astronomy69117 Heidelberg, Germany

6SOFIA-USRA, NASA Ames Research CenterMo®ett Field, CA, 94035, USA

7Department of Astronomy, University of IllinoisUrbana, IL, 61801, USA

8ESTEC, European Space Agency, 69117, Nordwijk2201 AZ, The Netherlands

9OSRAM GmbH, Industriestraße 20B85072 Eichst€att, Germany

10UC Observatories, 1156 High StreetSanta Cruz, CA, 95064, USA

11OHB System AG, 82234 Wessling, Germany12fischer@dsi.uni-stuttgart.de

Received April 14, 2018; Accepted June 18, 2018; Published July 20, 2018

We describe the design of the Field-Imaging Far-Infrared Line Spectrometer (FIFI-LS), operated as aFacility-Class instrument on the Stratospheric Observatory for Infrared Astronomy (SOFIA). FIFI-LS isan imaging spectrometer for medium resolution spectroscopy. Since being commissioned in 2014, it hasperformed over 50 SOFIA commissioning and science °ights. After operating as a principal investigatorinstrument in 2014 and early 2015, it was accepted as a Facility Science Instrument in 2015. In addition tothe description of the design, we report on the in-°ight performance and the concept of operation. We alsoprovide an overview of the science opportunities with FIFI-LS and describe how FIFI-LS observationscomplement and complete observations with the PACS instrument on the Herschel observatory.

Keywords: Integral ¯eld spectroscopy, spectrometer, far-Infrared, FIFI-LS, SOFIA.

12Corresponding author.This is an Open Access article published by World Scienti¯c Publishing Company. It is distributed under the terms of the CreativeCommons Attribution 4.0 (CC-BY) License. Further distribution of this work is permitted, provided the original work is properlycited.

Journal of Astronomical Instrumentation, Vol. 7, No. 4 (2018) 1840003 (14 pages)#.c The Author(s)DOI: 10.1142/S2251171718400032

1840003-1

J. A

stro

n. I

nstr

um. 2

018.

07. D

ownl

oade

d fr

om w

ww

.wor

ldsc

ient

ific

.com

by G

ER

MA

N A

ER

OSP

AC

E C

EN

TE

R (

DL

R)

on 0

1/09

/19.

Re-

use

and

dist

ribu

tion

is s

tric

tly n

ot p

erm

itted

, exc

ept f

or O

pen

Acc

ess

artic

les.

1. Introduction

The Field-Imaging Far-Infrared Line Spectrometer(FIFI-LS) is a far-infrared integral ¯eld spectro-graph designed for observations on the StratosphericObservatory for Infrared Astronomy (SOFIA;Erickson & Davidson, 1993). The ¯rst observationswith FIFI-LS on SOFIA were obtained in March2014 and to date the instrument has been operatedon more than 50 SOFIA science and commissioning°ights. The FIFI-LS instrument features two par-allel spectral channels with wavelength ranges of51–125�m (\blue channel") and 115–203�m (\redchannel"), respectively, that can be operated si-multaneously. Each channel has a ¯eld of view(FOV) consisting of 5 by 5 spatial pixels with aplate scale of 6 arc sec/pixel in the blue channel and12 arc sec/pixel in the red channel. The spectralresolution of the instrument is in the range of R �500 to � 2000 with an instantaneous coverage be-tween 800 km/s and 3000 km/s.The combination ofwavelength range and coverage, spectral resolutionand imaging capabilities implies that FIFI-LS isparticularly useful for investigations of the inter-stellar medium in our own and other galaxies.

Development of FIFI-LS began in 1997 at theMaxPlanck Institute for Extraterrestrial Physics (MPE)in Garching, Germany. It was originally designed tobe a Principal Investigator (PI)-instrument. In 2010,the development was terminated at MPE due to ashift in scienti¯c objectives. However, reviews byboth NASA and DLR emphasized the scienti¯cpotential of FIFI-LS and in 2012 the instrumentwas transferred to the Institute of Space Systems(IRS) in Stuttgart to ¯nish the development and

characterization. The instrument was thencommissioned aboard SOFIA in the spring of 2014and was accepted by the SOFIA program as a Fa-cility Science Instrument (FSI) in 2015.

FIFI-LS has a design and capabilities similar tothose of the PACS instrument (Poglitsch et al.,2010) onboard the Herschel satellite. FIFI-LScomplements and builds on PACS observationswhile also enabling new science investigations withadditional wavelength coverage and with thesuperior mapping capabilities of SOFIA.

In Fig. 1, a line °ux map for the [OI] line at145.5�m in NGC2024 is shown together with thecorresponding continuum °ux map as an examplefor data obtained with FIFI-LS.

2. Science Opportunities

FIFI-LS on SOFIA plays a major role in investiga-tions of star formation (primarily massive) and theinterstellar medium in both our own galaxy andexternal galaxies. With its sensitivity, the lattertopic is a prime capability of FIFI-LS on SOFIA.The spatial resolution of FIFI-LS is a huge im-provement over available data from ISO (Kessleret al., 1996) and allows to spatially resolve extra-galactic sources, such as M82, better. The key fea-tures that FIFI-LS observes are the bright far-Infrared ¯ne structure lines, e.g. [OI] 63�m and146�m, [OIII] 52�m and 88�m, [NII] 122�m, [NIII]57�m, [CII] 158�m, and molecular lines like COlines between J ¼ 13–12 and J ¼ 38–37. An over-view of lines observed with FIFI-LS to date is givenin Table 1.

Fig. 1. [OI] at 145.525�m in NGC 2024 as an example for FIFI-LS/SOFIA data. Left: Integrated line °ux obtained with aGaussian ¯t in Wm�2 Sr�1. Right: Continuum °ux obtained with a ¯rst-order polynomial ¯t in Jy Sr�1.

C. Fischer et al.

1840003-2

J. A

stro

n. I

nstr

um. 2

018.

07. D

ownl

oade

d fr

om w

ww

.wor

ldsc

ient

ific

.com

by G

ER

MA

N A

ER

OSP

AC

E C

EN

TE

R (

DL

R)

on 0

1/09

/19.

Re-

use

and

dist

ribu

tion

is s

tric

tly n

ot p

erm

itted

, exc

ept f

or O

pen

Acc

ess

artic

les.

2.1. Typical use cases

One typical application of FIFI-LS is obtainingmaps of [CII] emission in nearby galaxies. Largemaps of [CII] emission in nearby galaxies like M51and NGC6946 have been obtained in 10–15 h ofobserving time. An area of 60 arcmin2 can be inte-grated to reach a noise level of 10�17 W/m2 perspatial pixel (\spaxel", 12� 12 arc sec2) in about15 h of wall clock time. For smaller sources that ¯tinto 1� 1 arcmin2 FOV of the red channel, deepintegrations are easily possible. [CII] has beendetected in sources with red shifts of up to 0.04 inwall clock times below two hours.

The oxygen lines [OI] at 63 �m and 146 �mare also frequently observed simultaneously withFIFI-LS as they play a major role in the analysisof photo-dissociation regions (PDRs). The line°ux ratio provides constraints on the density andthe UV °ux in the region. In addition, FIFI-LSobservations of the relatively weak [OI] line at146 �m (typically a few times weaker than the [OI]63 �m line) are essential as a diagnostic of theconditions within the PDR since the often usedline at 63 �m can be optically thick for manysources. When mapping extended sources withmultiple FOVs in the blue channel (30�30 arc sec2Þ, the red channel gets four times theintegration time due to the overlap created by itslarger size (1� 1 arcmin2Þ. This helps to detectthe fainter line in the red channel. So mappingboth [OI] lines simultaneously with FIFI-LS isfrequently carried out.

The line pair of [OIII] at 52�m and 88�m isvery valuable for the analysis of HII regions sincethe ratio probes the electron density in the region.With the addition of the also available [NIII] line at57�m, it is also possible to probe the metallicity(Nagao et al., 2011). FIFI-LS observations of the

[OIII] line at 52�m and the [NIII] line at 57�m arefrequently obtained to fully utilize existing data setsof other spectral lines obtained with PACS at sim-ilar spatial resolution.

Using the quickly movable SOFIA telescope,large maps of bright galactic regions can beobtained very e±ciently. With the FIFI-LS \brightobject mode", a 32.5 arcmin2 map of the Orion BN/KL region in the [CII] 158�m and [OI] 63�m lineswas completed in 76min. In the region around theCircumnuclear Ring (CNR) at the galactic center(Iserlohe et al., 2018), a deep map of about136 arcmin2 was executed in a total time of about5 hours.

Another unique capability of FIFI-LS is thee±cient mapping of high J CO lines. Such mapshave been successfully produced for the BN/KL(Looney et al., 2018) as well as the CNR regions.The sensitivity of FIFI-LS enables the study offainter sources like protostars as well.

3. Instrument Description

3.1. Imaging spectroscopy

FIFI-LS is an imaging spectrometer. As such, itallows the observation of spectra at every spatialpixel in a two-dimensional (2D) FOV in a singleobservation. Therefore, imaging spectroscopy isoften referred to as 3D spectroscopy. An overviewof how imaging spectroscopy is carried out withFIFI-LS is given in Fig. 2. After the light fromthe telescope is separated into the red andblue channels by a dichroic, the 5� 5 spaxels arerearranged into a pseudo slit of 25 spaxels, withfour gaps, that later help to separate the spaxelson the detector. The light is then dispersed on thetwo gratings and projected onto the 25� 16 pixeldetectors.

Table 1. Spectral lines already observed with FIFI-LS.

Line � [�m] Line � [�m] Line � [�m] Line � [�m]

[OIII] 51.8 CH4 80.1 CO 104.4 CO 162.8OH 55.9 CO 84.4 CO 118.6 OH 163.3[NIII] 57.3 OH 84.5 [NII] 122 HCN 169.4[OI] 63.2 CO2 (ice) 86 CO 124.2 CO 174C2H2 68.6 CO 87.2 CO 130.4 CO 186CO 69.1 CH4 87.3 [OI] 145.5 CO 200.3C2H2 69.7 [OIII] 88.4 CO2 (ice) 146CO 70.9 OH 96.3 CO 153.3CO 77.1 CO 96.8 [CII] 157.7

FIFI-LS: The Field-Imaging Far-Infrared Line Spectrometer on SOFIA

1840003-3

J. A

stro

n. I

nstr

um. 2

018.

07. D

ownl

oade

d fr

om w

ww

.wor

ldsc

ient

ific

.com

by G

ER

MA

N A

ER

OSP

AC

E C

EN

TE

R (

DL

R)

on 0

1/09

/19.

Re-

use

and

dist

ribu

tion

is s

tric

tly n

ot p

erm

itted

, exc

ept f

or O

pen

Acc

ess

artic

les.

3.2. Instrument design

A picture of the instrument installed on the SOFIAtelescope, as well as a drawing of the cryostatinstalled at the telescope °ange, are provided inFig. 3. The cryostat (gray) sits on top of a cradle(also gray). A pressure coupler (orange) operates asa pressure barrier for the aircraft cabin, sincestratospheric pressure prevails inside the telescopetub during °ight.

The instrument operates at four temperaturelevels. The bore sight box, which houses the warmdichroic is at room temperature and stratosphericpressure. Inside the vacuum vessel there arethree temperature levels: \nitrogen" (red), \helium"

(blue) and \super°uid helium" (magenta), realizedwith three cryogen tanks. For each tank there is aneck assembly with redundant vent paths to ensuresafe venting of the boiled-o® cryogens during nor-mal operation, as well as in case of a loss-of-vacuumevent in the cryostat.

A schematic overview of the complete opticalpath of FIFI-LS is presented in Fig. 4. A detailedoverview over the optical design of the instrument isavailable in Looney et al. (2000) and Raab et al.(2003). The light coming from the telescope throughthe Nasmyth tube is re°ected into the FIFI-LScryostat at the \Warm Dichroic". Currently, a fullyre°ective mirror is installed at that location. The

Fig. 2. Schematic of the FIFI-LS concept for imaging spectroscopy. The blue and red channel observations overlap on the sky, splitinternally by the entrance optics. The FOV of 5� 5 spaxels in each channel is rearranged into a pseudo slit of 25 spaxels (and 4 gap-spaxels). This slit is dispersed with the gratings onto the 25 by 16 pixel detectors. One of the 5 spaxels in each slice is highlighted toshow how the light is distributed on the detector.

Fig. 3. Left: FIFI-LS and its counter weight rack (CWR) installed at the SOFIA telescope. The telescope is at the lower end of its20–60� elevation range. Right: A drawing of FIFI-LS with some main components labeled.

C. Fischer et al.

1840003-4

J. A

stro

n. I

nstr

um. 2

018.

07. D

ownl

oade

d fr

om w

ww

.wor

ldsc

ient

ific

.com

by G

ER

MA

N A

ER

OSP

AC

E C

EN

TE

R (

DL

R)

on 0

1/09

/19.

Re-

use

and

dist

ribu

tion

is s

tric

tly n

ot p

erm

itted

, exc

ept f

or O

pen

Acc

ess

artic

les.

optional instrument guider optics plus guidingcamera originally planned for that location havenot been implemented since the guiding providedby the telescope has proven su±cient for main-taining the FIFI-LS image quality close to thedi®raction limit. The fully re°ective mirror restson a stage that permits rotation in two axes. Sincethe mirror is located close to a focus in the opticalpath, it is used for pupil alignment of the instru-ment. Pupil alignment was performed on theSOFIA facilities Telescope Assembly AlignmentSimulator (Thompson, 2011). The incoming lightis then transmitted through the entrance windowthat forms the pressure boundary of the cryostat.The entrance optics are located on the nitrogenplate. A K-mirror is used to rotate the quadraticfootprint on the sky to optimize the coverage forsources as well as to keep the ¯eld orientation ¯xedon the detectors throughout long integrationswhile the sky rotates relative to the telescope.More details can be found in Colditz et al. (2014).A °ip mirror is used to switch the optical pathbetween the sky and a calibration source. Thecalibration source itself is thermally connected tothe helium optical bench, but located outside ofthe helium shield. It contains an electrically heatedradiator, an integrating sphere and a chopper togenerate signal levels comparable to that fromthe sky (Schweitzer et al., 2008; Rebell, 2018). It isused to generate spectral and spatial °at ¯elds.The LN2 plate also houses several motors for thedi®raction grating coarse drives and the dichroic¯lter changer on the LHe plate. Transfer of themotion through the helium shield is e®ected viamagnetic coupling.

The light enters the helium cooled optical platethrough the Lyot stop and the entrance ¯lter as-sembly. The ¯lter has a short wavelength cut o® at�40�m. On the LHe plate the light is ¯rst split upinto the blue and red channels by the cold dichroic.This is a mesh dichroic that transmits the longerwavelength for the red channel and re°ects the lightfor the blue channel. Since the transition betweentransmission and re°ection is not a perfect stepfunction at a single wavelength but extends overseveral microns, two interchangeable dichroics areavailable to maintain full wavelength coverage inboth channels. The dichroics also provide some°exibility in the channel in which a speci¯c wave-length is to be observed. For example, the CO lineat �119�m can be observed in either the blue or thered channel.

The integral ¯eld units (IFU) rearrange the 5� 5spaxel FOV into the 25 spaxel pseudo slit (seeSec. 3.1) that is projected onto the di®ractiongratings via the collimator optics. The instanta-neous wavelength range covered by the detector isset by rotating the gratings with a multi-stage drivein a nested control loop. More details on the gratingsystems can be found in Rebell et al. (2014) andRebell (2018). The beam passes through the colli-mator optics for the second time on its way to thedetectors. In the blue channel, observations areperformed in both ¯rst and second order and aninterchangeable ¯lter is located in that channel onthe LHe plate before the light enters the detectorassembly. The motor running that ¯lter changer isthe only motor operating on the LHe plate. In op-eration, the plate typically reaches a temperature ofabout 6� K.

Fig. 4. Schematic of the FIFI-LS optical path from the telescope to the detectors of both channels.

FIFI-LS: The Field-Imaging Far-Infrared Line Spectrometer on SOFIA

1840003-5

J. A

stro

n. I

nstr

um. 2

018.

07. D

ownl

oade

d fr

om w

ww

.wor

ldsc

ient

ific

.com

by G

ER

MA

N A

ER

OSP

AC

E C

EN

TE

R (

DL

R)

on 0

1/09

/19.

Re-

use

and

dist

ribu

tion

is s

tric

tly n

ot p

erm

itted

, exc

ept f

or O

pen

Acc

ess

artic

les.

The detector assembly is cooled by heliumpumped down to about 20mbar to operate at atemperature of about 1.8�K. FIFI-LS uses gallium-doped germanium pixels in both channels. Thespectral sensitivity of gallium-doped germaniumpixels is high between �40�m and �120�m, idealfor the blue channel. For the spectral range of thered channel the detector elements are compressed intheir mounting structures in order to shift the areaof good spectral sensitivity into the wavelengthrange of �120�m to �210�m (Rosenthal et al.,2000). The pixels in the blue channel have a size of6� 6 arc sec2 on the sky and in the red channel theprojected size is 12� 12 arc sec2; the plate scaleswere chosen to roughly match the di®raction-limit-ed point spread functions at the mean wavelengthsof the two channels. Having two detectors operatingin parallel also enables FIFI-LS to map two linessimultaneously, which is a multiplex advantage onan airborne observatory operating a limited numberof science hours per year.

3.3. Electronics

Thearrangementof the electronics forFIFI-LS re°ectsthe arrangement of the instrument at the telescopein the aircraft. An overview showing the signal con-nections between the locations is provided in Fig. 5.

Instrument operation is performed at the PIrack, which holds three computers and two term-inals to allow two operators to control and monitorthe instrument in parallel. The MASTER computerruns the main user interface for the instrumentcontrol. It controls the observation process based onscripted parameters, by distributing commands anddata to the other machines in the system andmonitoring their feedback. The MIDDLEMANcomputer provides the interface to the SOFIAtelescope, receiving the commands from MASTERand translating them into commands that are usedto control the SOFIA telescope. It also monitors thetelescope's status and relays any information im-portant for an observation, e.g. telescope errors,to MASTER. The MAKER computer records thedigitized observational data and forwards it viathe on-board computer network to a storage disk,where it may be reduced, viewed and analyzed bythe scientists on board the SOFIA and used to makereal-time observation planning decisions.

The counter weight rack (CWR) is located atthe telescope assembly close to the instrument°ange (see Fig. 3). It contains electronics that mustbe close to the instrument to ensure good signalquality. The MOTHER computer controls themechanisms in the cryostat and the detector read-out timing by generating a set of high level clock

Fig. 5. Overview over the electronics layout of FIFI-LS.

C. Fischer et al.

1840003-6

J. A

stro

n. I

nstr

um. 2

018.

07. D

ownl

oade

d fr

om w

ww

.wor

ldsc

ient

ific

.com

by G

ER

MA

N A

ER

OSP

AC

E C

EN

TE

R (

DL

R)

on 0

1/09

/19.

Re-

use

and

dist

ribu

tion

is s

tric

tly n

ot p

erm

itted

, exc

ept f

or O

pen

Acc

ess

artic

les.

signals. In the EGG-Box two PC104 Stacks (\red"and \blue" EMBRYO) control the coarse drives forthe gratings and command the driving voltages(which are ampli¯ed by the HV-Ampli¯er) of thepiezo stacks of the ¯ne drive. The grating positionsare controlled using digital PID feedback loopsrunning on the EMBRYO machines to reduceoscillations of the gratings. The MISC-Box containsthe low-level controllers for the mechanisms, ex-cluding the grating drives. A temperature displayfor the di®erent zones in the cryostat is available inthe MATCH-Box. Electronics that need to be lo-cated very close to the cryostat in order to avoidnoise on analogue data lines were built into theelectronics BACKPACK. The BACKPACK is di-rectly attached to the rear of the cryostat and isbuilt symmetrically, having the same set of elec-tronics boards for the blue and for the red channel.Tasks performed here include the digitization andrelay of the detector and grating position readouts,and providing power and clock signals for thedetectors, as well as bias voltages.

The cryostat contains the detector readoutelectronics, operating at �1:8�K. The cold readoutelectronics (CREs) were developed for the PACSinstrument. The CREs amplify the voltage mea-surements and relay the ampli¯ed signal to a set ofwarm pre-ampli¯ers located on the inner wall of thecryostat shell. After ampli¯cation, the signals arerelayed to the BACKPACK for digitization. Thecryostat also contains the motors and actuators forthe mechanisms, the respective end switches of thesemechanisms, as well as the Inductosynr rotarytransducers used to measure the positions of the redand blue gratings.

3.4. Software

The instrument uses information stored in script¯les called \scan descriptions" to command theinternal mechanisms and the telescope during anobservation. Each scan description de¯nes a shortobservation and contains information like theposition of the observed target, the integrationtimes for the detectors, the chopping frequency forthe secondary mirror and the wavelengths to beobserved. A set of multiple scan descriptions isnecessary to de¯ne and execute a complete astro-nomical observation including nods. This sequenceof scan descriptions is, in turn, stored in a sched-ule ¯le, which is completed sequentially bythe instrument software. More details on the

generation of the scan descriptions can be found inBryant et al. (2014).

Multiple processes then run the instrument,with \¯¯main" being the major one. This processreads the current scan description according to theschedule ¯le. The information is then evaluated anddistributed to the other processes in the systemusing the inter-process communication protocolkosma io, which is part of the KOSMA-controlsoftware package also used by the GREAT instru-ment (Guesten et al., 2003). More details onthe processes can be found in Klein et al. (2004).Communication with the telescope is maintainedvia the KOSMA translator (Stutzki, 2004), whichconverts the telescope commands from the FIFI-LSinstrument (conforming to the KOSMA-controlsoftware) into commands that may be executed bythe telescope's Mission Controls and Communica-tions System (MCCS).

The astronomical data are processed by a ded-icated pipeline that reduces the data obtainedduring an entire observation, beginning with thetime-resolved raw values, into a single cube (twospatial dimensions and one spectral dimension)comprising a map that can be used for astronomicalanalysis. The pipeline has been implemented in IDLin the observatory's \Redux" environment, whichuni¯es the pipeline development e®orts acrossSOFIA's suite of instruments as well as with theinstrument team (Clarke et al., 2015). The instru-ment records the cumulative signal detected every1/250 s. The sequence of cumulative signal values asa function of exposure time constitutes a \ramp".The typical duration of a ramp is 32/250 s (or about1/8 s), during which the telescope and instrumentparameters remain ¯xed. The assignment to gratingpositions, chop phases, nod phases and map positionis provided by information in the ¯le header of theraw data. First, the pipeline splits up the gratingpositions and chop phases recorded in a single ¯le,which represents only one map position and nodphase. The ramps are then ¯tted with linear slopesand are averaged for a single grating position andchop phase. Chop and nod subtractions are per-formed, and each of the 400 pixels is as signed withits spatial position and a wavelength. Flat ¯elds arethen applied. The °at ¯elds are created from \skydips", where the sky background is observed at twodi®erent telescope elevations to generate a signalwithout telescope and instrument background. Fordata obtained in Cycle 5 and later, °at ¯elds taken

FIFI-LS: The Field-Imaging Far-Infrared Line Spectrometer on SOFIA

1840003-7

J. A

stro

n. I

nstr

um. 2

018.

07. D

ownl

oade

d fr

om w

ww

.wor

ldsc

ient

ific

.com

by G

ER

MA

N A

ER

OSP

AC

E C

EN

TE

R (

DL

R)

on 0

1/09

/19.

Re-

use

and

dist

ribu

tion

is s

tric

tly n

ot p

erm

itted

, exc

ept f

or O

pen

Acc

ess

artic

les.

with the internal calibration source are available;these o®er superior spectral coverage.The full spec-tral scans necessary to generate these are timeconsuming, but can be easily obtained in the lab,while time in-°ight is very limited.

After combining the di®erent grating positionswithin a ¯le into a single spectrum, a telluric cor-rection is also performed.With the relevant informa-tion from the ¯le headers (°ight altitude, telescopeelevation) an ATRAN (Lord, 1992) transmissioncurve is calculated, convolved with the spectralresolution of the instrument and then used to cor-rect the data. Since the data are not yet combinedinto a map, this allows one to account for changing°ight altitudes and telescope elevations. However,this technique can be problematic near narrow at-mospheric features since the convolution with theinstrumental resolution signi¯cantly changes theshape of those features whereas the astronomicalsignal interacts with the unconvolved transmissionpro¯le of the atmosphere. Nevertheless, for manydata sets this method has proven su±cient. If thetransmission pro¯le in the relevant spectral regiondoes not allow for this method to be applied, telluriccorrections can still be applied to the completedmap using the following procedure.The astronomi-cal signal is assumed to consist of an emission line atop a continuum that varies linearly with wave-length. A telluric transmission pro¯le is applied tothis signal and the result is convolved with the in-strumental spectral resolution. This model is then¯tted to the observed spectrum. Here the para-meters of altitude and elevation have to be assumedconstant throughout the map, or the map has to besplit into smaller entities during which this as-sumption holds. If there is a su±ciently strongcontinuum source available in the mapping area anda telluric water line is covered within the wave-length range of the observations, this method can beused to estimate the water vapor content of theatmosphere at the time of the observations. Thismethod has been successfully used for the telluriccorrection applied to data obtained from the ga-lactic center (Iserlohe et al., 2018).

To overcome the unavailability of a calibratedwater vapor monitor, the FIFI-LS team has suc-cessfully tested a method to measure the watervapor with the instrument itself. In intervals ofabout 20min the astronomical observations werepaused, and several water lines close to the observedwavelength were measured in an unchopped

manner. By choosing the right water lines, it waspossible to create a setting sensitive to the watervapor content in the atmosphere. This techniquecreates an additional overhead of about 10% wallclock time and is suggested by the team for thoseobservations where a priori investigation with anATRAN modeled atmosphere shows the transmis-sion to be sensitive to water vapor.

Both the telluric corrected as well as theuncorrected data cubes are available as pipelineproducts. The next reduction steps are the °uxcalibration and the bary-centric velocity correction.Since the data are not combined into a single cubeyet, it is still possible to use di®erent calibrationcurves (e.g. for di®erent ¯lter and dichroic con¯g-urations). Then the data are resampled onto a reg-ular wavelength grid. Finally, the pipeline resampleseach wavelength plane onto a regular spatial grid.This step combines the spatial information from allinput nod-combined dither positions into a singleoutput map with a smooth local polynomial surface¯t. An example for data reduced with the FIFI-LSpipeline is shown in Fig. 1.

In-°ight the pipeline is also used to create quick-look results of the observed data to enable real-timedecisions on the observing strategy.

4. Performance

4.1. Spectral resolution

To verify the delivered spectral resolution of FIFI-LS on-sky, observations of the width of ¯ne struc-ture lines were taken toward the planetary nebulaeNGC 7027 and NGC 6543 as well as the carbon starIRCþ10216. The full width at half maximum in-tensity (FWHM) of those lines was then comparedto the values obtained with the FIFI-LS telescopesimulator in the laboratory (Colditz, 2017). Obser-vations of calibration targets were taken on multiple°ight series between April 2014 and March 2017.

The planetary nebulae proved ideal, since theyfeature bright narrow line emission and relativelylow continuum °ux levels. In these cases, the re-duced data cube without telluric correction can beused for the analysis. The CO emission lines fromIRCþ 10216 o®er access to even more wavelengths.Due to the strong underlying continuum for somewavelengths (�69�m and 77�m), the telluric cor-rected cube generated by the pipeline had to be usedto obtain a ¯rst-order polynomial ¯t for the base-line. Each line is ¯tted with a Gaussian pro¯le to

C. Fischer et al.

1840003-8

J. A

stro

n. I

nstr

um. 2

018.

07. D

ownl

oade

d fr

om w

ww

.wor

ldsc

ient

ific

.com

by G

ER

MA

N A

ER

OSP

AC

E C

EN

TE

R (

DL

R)

on 0

1/09

/19.

Re-

use

and

dist

ribu

tion

is s

tric

tly n

ot p

erm

itted

, exc

ept f

or O

pen

Acc

ess

artic

les.

determine its FWHM. Pixels with an integrated line°ux below a threshold of 10–50% of the peak inte-grated line °ux are cut from the analysis since theyare classi¯ed o®-source. The median of the remain-ing values is used to create a data point. If a source/line was observed multiple times during a °ight se-ries, the data sets are evaluated separately. Theresults are shown in Fig. 6. Some scatter is presentand it should be noted that the signal-to-noise ratio(SNR) is much higher for the gas cell measurementsin the lab compared with those from the astronom-ical data, especially at 200�m where the detectorssensitivity is quite low. Generally, the measurementscon¯rm that FIFI-LS delivers the same spectralresolution in-°ight as it does in the lab.

4.2. Sensitivity

The sensitivity of the instrument was veri¯ed in-°ight as a by product to full range spectral scans of

Mars that were primarily used to obtain the °uxcalibration of FIFI-LS. To obtain values for thesensitivity, the noise on a single ramp was evaluatedwithout any chop or nod subtraction. The typicalastronomical signal for FIFI-LS is usually below 1%of the background. This makes the chop and nodphase insigni¯cant for the noise analysis. A singleramp represents an integration time of about 1/8 s.Then, the standard deviation of that ramp com-pared to the linear ¯t applied in the FIFI-LS datareduction is the noise in the signal.

The results are presented and compared topredictions from Klein et al. (2014) in Fig. 7. Shownare the results for both channels (red and blue)with the two orders (second order, 50 to 70�m, and¯rst order, 70 to 120�m) in the blue channel.The left graph shows the minimum detectablecontinuum °ux (MDCF) in Jy, and the right graphshows the minimum detectable line °ux (MDLF)in 10�17 W/m2. Both are scaled to 4� in 15min

Fig. 6. Spectral resolution of the instrument in-°ight compared to prediction from laboratory measurements. The solid black linesshow the results from the laboratory measurements (Colditz et al., 2018). The symbols show the veri¯cation measurements per-formed on sky.

Fig. 7. Continuum (left) and line sensitivity (right) of FIFI-LS. Both are scaled to 4� in 15min on-source.

FIFI-LS: The Field-Imaging Far-Infrared Line Spectrometer on SOFIA

1840003-9

J. A

stro

n. I

nstr

um. 2

018.

07. D

ownl

oade

d fr

om w

ww

.wor

ldsc

ient

ific

.com

by G

ER

MA

N A

ER

OSP

AC

E C

EN

TE

R (

DL

R)

on 0

1/09

/19.

Re-

use

and

dist

ribu

tion

is s

tric

tly n

ot p

erm

itted

, exc

ept f

or O

pen

Acc

ess

artic

les.

on-source integration times. For most of the spectralcoverage, data is shown for both dichroic beamsplitters with a solid line for the D105 (cuto® at�105�m) and a dashed line for the D130 (cuto® at�130�m). Values are also scaled to the size of theFIFI-LS spaxels (6� 6 arc sec2 in the blue channeland 12� 12 arc sec2 in the red).

For the continuum sensitivity, it is assumedthat the signal is averaged over the whole spectralcoverage of the detector. The noise on a single pixelis divided by the square root of the number of res-olution elements available within the spectral cov-erage of the detector. The number of resolutionelements is calculated from the spectral resolution.For the line sensitivity, the noise on a single pixel isscaled to the FWHM of a narrow line. The predic-tions are based on noise estimates based on thetemperatures of the optics of the telescope andthe instrument, as well as the predicted spectralresolution.

The steep slopes of the sensitivity (both MDCFand MDLF) at the edges of the spectral coverage arecaused by the rapidly decreasing transmission of the¯lters in the instrument. In the blue channel secondorder (50–70�m) the continuum sensitivity mat-ches the prediction for � >�54�m with the D105dichroic and � >�59�m for the D130 dichroic. Thehigher wavelength cuto® of the D130 dichroic isnot an issue for the observation setup, since it isprimarily used to access longer wavelengths� >�90�m in the blue channel. In the ¯rst order,both dichroics show a similar performance up untilabout 90�m where the ¯lter curve of the D105starts to drastically reduce transmission. The per-formance is signi¯cantly better than predictedcaused by a change of the instantaneous spectralcoverage of the blue channel late in development ofFIFI-LS. For the red channel with the D105 di-chroic, the performance meets or exceeds the pre-diction roughly between 125�m and 160�m with adecreased sensitivity for the D130 dichroic. Thedecreasing sensitivity for wavelengths about 160�mis mainly caused by the decreasing grating e±ciencyand the rapidly decreasing detector sensitivity afterabove �190�m.

The line sensitivity for the second order is abouta factor of two lower than the prediction. This iscaused by the spectral resolution being about afactor of two lower than originally anticipated(Colditz et al., 2018). The important [OIII] lineat 52�m is located on the slope with rapidly

decreasing sensitivity. To address this issue, a ¯lterupgrade has been performed that doubles thetransmission for the [OIII] line and improves thetransmission in the whole second order by about30%. The ¯lters have been installed, the perfor-mance has been veri¯ed in the lab, and it a waits¯nal in-°ight veri¯cation. The line sensitivity in the¯rst order of the blue channel (�70–120�m) isclose to the prediction with a dip around 85�m.Since the spectral resolution for the red channel iswithin about 10% of the prediction, line and con-tinuum sensitivity behave very similar relative tothe prediction. This guarantees excellent perfor-mance for the two main red channel lines of [OI] at146�m and [CII] at 158�m.

4.3. Flux calibration

The °ux calibration of FIFI-LS was computed usingobservations of asteroids, planets and planetarymoons. They are used, along with time-dependentmodels of such objects, to derive instrumental re-sponse curves. To date, observations of Mars havebeen used as the primary °ux calibration source.Predicted total °uxes for Mars across the FIFI-LSpassband at the speci¯c UT dates of the observa-tions have been generated using the model fromLellouch & Amri (2006).

In order to check the calibration, °uxes mea-sured with FIFI-LS are compared with ISO LWSmeasurements from Liu et al. (2001). To account forthe uncertainty of telluric correction, 50% is addedto or subtracted from the standard zenith values ofwater vapor used by ATRAN (e.g. 7.3�m precipi-table water vapor at 41,000 ft). For the two [OI]lines at 63�m and 146�m, the telluric correctiontool described above (Iserlohe et al., 2018) has beenused. For the other lines, which are farther awayfrom telluric features, the line °ux was ¯tted to theuncorrected result cubes and then divided by thetransmission at the line location. The results forthat comparison are shown in Table 2.

General agreement between FIFI-LS and ISOLWS is very good. We strongly suspect the [NIII]data taken on 2014 April 24 were a®ected by watervapor and þ50% higher than the assumed value.The data showed a low SNR compared to the other[NIII] data from the same source with similar inte-gration time. The 12 h weather updated °ight planpredicted about 34�m of zenith precipitable watervapor compared to the 11�m used as maximum

C. Fischer et al.

1840003-10

J. A

stro

n. I

nstr

um. 2

018.

07. D

ownl

oade

d fr

om w

ww

.wor

ldsc

ient

ific

.com

by G

ER

MA

N A

ER

OSP

AC

E C

EN

TE

R (

DL

R)

on 0

1/09

/19.

Re-

use

and

dist

ribu

tion

is s

tric

tly n

ot p

erm

itted

, exc

ept f

or O

pen

Acc

ess

artic

les.

value above. Using the predicted value to calculatean atmospheric transmission for the line, the maxi-mum °ux relative to ISO LWS increases to 1.25,indicating that this is the most likely explanationfor the low measured °ux for that °ight.

Under the somewhat arbitrary assumption ofþ/� 50% for the minimum and maximum watervapor values, all but one measured °ux values de-viate from the ISO LWS values by less than 15%, aresult which indicates good agreement between theFIFI-LS measured °uxes with those measured fromspace with ISO-LWS, and therefore con¯rms theoverall °ux calibration of the instrument.

4.4. FIFI-LS and PACS

FIFI-LS has a very similar design to the PACSinstrument onboard the Herschel satellite. SincePACS is no longer operational, FIFI-LS is an idealtool for follow-up observations. Since SOFIA is nota space telescope, FIFI-LS does not o®er the sensi-tivity of PACS, but it does o®er several features tobridge the sensitivity gap and enable observationsnot feasible with PACS:

. FIFI-LS employs two complete, and completelyindependent, spectrometer channels, whereasPACS used only one over the entire wavelengthrange. This di®erence enables simultaneousobservations of two individually selectable spec-tral lines with FIFI-LS, which increases FIFI-LS'sobserving e±ciency signi¯cantly. In the PACSspectrometer, the wavelengths in the two chan-nels are interdependent — bound by the commongrating — and are not individually selectable.

. Due to the two independent grating channels inFIFI-LS, the useful wavelength range of FIFI-LS

(51–200�m) exceeds that of PACS (55–200�m).FIFI-LS can observe additional diagnostic lines,especially the [OIII] line at 51.815�m.

. The pixel size in the PACS spectrometer, as set bythe one IFU, is 9.4 00, which is optimum in the longwavelength part of the instrument, but noticeablyunder-samples the telescope PSF at shorterwavelengths. As FIFI-LS uses two independentIFUs, one for each channel, the projected pixelsizes are 12 00 in the red channel and 6 00 in theblue channel, providing FIFI-LS with a betterangular resolution in the short wavelength range(51�m�120�m) compared with PACS, eventhough Herschel provides a larger e®ective tele-scope aperture.

. The spectral resolution of the two instruments iscomparable, but the execution time per observa-tion and the sensitivities di®er. The \raw" sensi-tivity (point source, unresolved spectral line) ofHerschel/PACS is about eight times better thanthat of SOFIA/FIFI-LS, owing to the lower tele-scope temperature and emissivity, the lack of at-mospheric emission and absorption, and thelarger telescope diameter. The \atomic buildingblock" of a PACS observation executes in�7min, after which time the noise in a line-de-tection is �3� 10�18 W/m2. For deep observa-tions, this is a useful building block, of which anumber of repetitions can be executed to reachthe desired depth (SNR). However, for (fullysampled) mapping of su±ciently bright, largeobjects (e.g. nearby galaxies), this leads to verylong integration times and highly \over-integrated" line-detections (e.g. SNR values >100), which are then limited by systematic e®ectsand calibration precision. Here SOFIA/FIFI-LS

Table 2. Overview over cross-calibration with FIFI-LS; the last two columns show the minimum and maximum °ux relative to theISO LWS values. The values are calculated using standard ATRAN water vapor �50% (maximum) and þ50% (minimum).

� [�m] Line Date [dd.mm.yyyy] Source Altitude [ft]Av. elevation

angle [�]Min. °ux relative

to ISO LWSMax. °ux relative

to ISO LWS

51.815 [OIII] 26.04.2014 NGC 7027 41,000 23 1.02 1.2951.815 [OIII] 14.10.2015 NGC 6543 41,000 40 1.05 1.1251.815 [OIII] 26.10.2015 NGC 6543 41,000 43 1.07 1.1457.317 [NIII] 22.04.2014 NGC 7027 41,000 45 0.73 0.8657.317 [NIII] 24.04.2014 NGC 7027 41,000 44 0.50 0.5857.317 [NIII] 26.04.2014 NGC 7027 41,000 21 0.74 1.0063.184 [OI] 20.10.2015 NGC 7027 43,000 39 0.87 1.3663.184 [OI] 27.10.2015 NGC 7027 37,000 47 0.80 2.9088.356 [OIII] 02.03.2016 NGC 6543 43,000 48 0.95 0.98145.525 [OI] 26.04.2014 NGC 7027 41,000 23 0.60 0.96

FIFI-LS: The Field-Imaging Far-Infrared Line Spectrometer on SOFIA

1840003-11

J. A

stro

n. I

nstr

um. 2

018.

07. D

ownl

oade

d fr

om w

ww

.wor

ldsc

ient

ific

.com

by G

ER

MA

N A

ER

OSP

AC

E C

EN

TE

R (

DL

R)

on 0

1/09

/19.

Re-

use

and

dist

ribu

tion

is s

tric

tly n

ot p

erm

itted

, exc

ept f

or O

pen

Acc

ess

artic

les.

has an advantage, since the telescope time tomove from pointing to pointing on the sky is onlyabout 8 s on average. This enables fast mapping ofrelatively bright sources. In the symmetric chopmode the minimum on-source integration timetypically used with FIFI-LS is 30 s, with a totalwall-clock time of about 75 s. FIFI-LS has alreadysuccessfully carried out mapping projects of M51(noise �1:6W/m2 in �15 h total wall clock time)and NGC6946 (noise �2:3W/m2 in �10 h wallclock time).

. For mapping extended targets (e.g. nearby gal-axies) in several spectral lines, the capability toobserve two lines simultaneously with the fasterand more e±cient observing scheme providesFIFI-LS with a multiplex advantage of more thana factor of two. FIFI has done this successfullyfor galactic and extragalactic sources. For exam-ple, FIFI-LS has created maps of the [OI] linesat 63.184�m and 145.525�m, covering an area of32.5 arc min2 in 130 tiled ¯elds in about 75min ofwall-clock time.

5. Observing with FIFI-LS

5.1. Observing modes

To subtract the high background present in far-infrared observations, FIFI-LS uses chopping of thetelescope secondary for all standard observingmodes. The typical chop frequency for FIFI-LS is�2Hz. Only about 3.1% of the integration time isa®ected by the chopper transition at this frequency,which has proven to be su±ciently fast for thequickly changing weather conditions and climbs ofthe aircraft during °ights. Since the two chop beamsuse di®erent portions of the primary mirror, nod-ding is used to subtract the residual background.FIFI-LS o®ers three di®erent standard observingmodes to general observers:

. Symmetric chop,

. Asymmetric chop,

. Bright object mode.

The most commonly-used mode is the symmetricchopping mode. It is also the most e±cient one,since in symmetric chopping, also known as beamswitching, the source is present in one chop phasefor both nod positions (A and B). Overheads arecreated by the time spent in o® chop phase and bythe time the telescope takes to switch betweennod beams and/or map positions. For symmetric

chopping FIFI-LS uses an on-source nod time of30 s, which results in a nod e±ciency of about 80% ifone also accounts for the average telescope movetime of 8 s. The e±ciency degrades quickly forshorter nod dwells and improves only very slowly forlonger nod dwell times (Fischer et al., 2016). Sym-metric chopping does create a coma e®ect thatlimits the total chop-throw to 5–8 arcmin in theFIFI-LS wavelength range and requires an area freeof astronomical signal at the chop angle as well as inthe opposite direction. If symmetric chop is notpossible, an asymmetric chop can be consideredallowing for a total chop throw of up to 10 arcminand requiring an emission free area only in the di-rection of the chop, since the nod can be performedin areas much further away in the sky. The down-side of that mode is that only 25% of the total in-tegration time is spent on the source cutting theobserving e±ciency in half to about 20%.

Some of that loss can be recovered by using thebright object mode, where an identical grating setupis used on all integrations, thus limiting the spectralwidth available to roughly the instantaneous spec-tral coverage (Klein et al., 2014). Here one \o® nod"(B) is used for multiple \on nods" (A) in an ABA(30 s on nod) or AABAA (20 s on nod) scheme. Thelatter being the fastest mapping mode availablewith FIFI-LS. Such a strategy pushes the e±ciencycloser to the symmetric chopped mode. This mode isalso the fastest available for very bright sources,since it allows shorter integration times per mapposition (down to 10 s on-source) compared tosymmetric chopping (30 s on-source). If the ob-served spectral line of a weaker source does not re-quire a wide spectral scan, this mode can also beused for deep integrations.

The feasibility of an unchopped \total power"mode has also been investigated using an on-sourceintegration time of 10 s in an ABA scheme. With ane±ciency of 37% this mode almost reaches the e±-ciency of the symmetric chopped mode (39%). It isalso superior to the asymmetric modes with theire±ciency of 20–29%. For comparison, the same maphas been acquired with the bright object mode onthe same °ight with the same on-source integrationtime. The analysis of the test data showed a com-parable but lower data quality for the total powermode. There is also a higher risk associated withthose observations since background °uctuationsare subtracted about 100 times slower comparedto chopped observations. Also the aircraft cannot

C. Fischer et al.

1840003-12

J. A

stro

n. I

nstr

um. 2

018.

07. D

ownl

oade

d fr

om w

ww

.wor

ldsc

ient

ific

.com

by G

ER

MA

N A

ER

OSP

AC

E C

EN

TE

R (

DL

R)

on 0

1/09

/19.

Re-

use

and

dist

ribu

tion

is s

tric

tly n

ot p

erm

itted

, exc

ept f

or O

pen

Acc

ess

artic

les.

climb during those observations. Still the mode isavailable and feasible for those observations other-wise showing a high risk of self-chopping in crowded¯elds on the sky. Also, existing chopped observa-tions have been reduced \total power style" whenthere were indications of self-chopping. This didresult in some loss of SNR, but still enabled scien-ti¯c interpretation of the data.

For observation planning with FIFI-LS, theSOFIA USPOT tool is available to set up and vi-sualize the observations.(a) This includes the spec-tral setup, chop and map/dither. Some instrumentspeci¯c observation parameters and the observingstrategy are then later developed by the SOFIAsupport scientist together with the observer, oncethe proposal has been accepted. FIFI-LS has mul-tiple recommended map/dither schemes that can befound in Fischer et al. (2016). An observation timeestimator based on the noise performance and theoverheads mentioned above is available online.(b)

5.2. Data analysis tools

Multiple tools to analyze the data are available forthe community online.(c) Available are tools forcube exploration (SOSPEX) line and continuum¯tting (FLUXER) and a ¯tting tool for telluriccorrection. FLUXER was used to extract the lineand continuum °ux shown in Fig. 1.

6. Conclusion and Outlook

With FIFI-LS, a unique and versatile instrumenthas been added to the SOFIA observatory. It canvery quickly map bright and large galactic line andcontinuum sources of bright galactic sources andprovides the sensitivity to augment existing datasets of, e.g., the PACS instrument with furtherspectral transitions or to explore new extragalactictargets. Since it has been o®ered to the communityin 2015, there has been a constantly high demandfor FIFI-LS with about 30 proposals for everyobserving cycle.

All aspects of the instrument as well as the wayto utilize it are constantly under development. Thewater vapor measurement is the latest improvement

to the operation of FIFI-LS. Both the instrumentcontrol software as well as the pipeline are con-stantly updated with bug ¯xes and other improve-ments as well.

For Cycle 6 in 2018, the hardware of FIFI-LSwas updated for the ¯rst time with a new entrance¯lter and new order sorter ¯lters for the bluechannel to improve the sensitivity for numerouslines in the blue channel. Currently, there is also amajor upgrade under investigation for FIFI-LS toincrease the spatial and instantaneous spectralcoverage of the instrument and potentially widenthe spectral range to continue to improve the e±-ciency of FIFI-LS on SOFIA.

Acknowledgments

SOFIA, the \Stratospheric Observatory for InfraredAstronomy" is a joint project of the DeutschesZen-trumfürLuft- und Raumfahrte.V. (DLR; GermanAerospace Center, grant: 50OK0901) and theNational Aeronautics and Space Administration(NASA). It is funded on behalf of DLR by theFederal Ministry of Economics and Technologybased on legislation by the German Parliament, thestate of Baden-Württemberg and the Universit€atStuttgart. Scienti¯c operation for Germany iscoordinated by the German SOFIA Institute (DSI)of the Universit€at Stuttgart, in the USA by theUniversities Space Research Association (USRA).

The team would like to thank the SOFIAobservatory including NASA, USRA, DLR andDSI. The eight FIFI-LS °ight series conducted up tothe time of writing were successful only due to theexcellent cooperation of all participants.

References

Bryant, A., Fischer, C., H€onle, R. et al. [2014] \FIFI-LS ob-servation planning and data reduction," Proc. SPIE 9147,91474C.

Clarke, M., Vacca, W. D. & Shuping, R. Y. [2015] \Redux: ACommon Interface for SOFIA Data Reduction Pipelines,"ADASS, ASP Conference Series, Vol. 495, 355.

Colditz, S., Klein, R., Beckmann, S. et al. [2014] \BoresightCalibration of FIFI-LS: in theory, in the lab and on sky,"Proc. SPIE 9147, 91474S.

Colditz, S. [2017] \FIFI-LS – A Field-Imaging Far-Infrared LineSpectrometer for SOFIA: Completion of the Instrument,Laboratory and In-°ight Calibration and Characterization,"(Dr. Hut Verlag), Dissertation, University of Stuttgart,Germany.

Colditz, S., Beckmann, S., Bryant, A. et al. [2018] \Spectral andspatial characterization and calibration of FIFI-LS — The

a https://dcs.arc.nasa.gov/observationPlanning/install-USPOT/uspotDownload.jsp.bhttps://dcs.arc.nasa.gov/.chttps://www.so¯a.usra.edu/science/proposing-and-observ-ing/data-products/data-resources.

FIFI-LS: The Field-Imaging Far-Infrared Line Spectrometer on SOFIA

1840003-13

J. A

stro

n. I

nstr

um. 2

018.

07. D

ownl

oade

d fr

om w

ww

.wor

ldsc

ient

ific

.com

by G

ER

MA

N A

ER

OSP

AC

E C

EN

TE

R (

DL

R)

on 0

1/09

/19.

Re-

use

and

dist

ribu

tion

is s

tric

tly n

ot p

erm

itted

, exc

ept f

or O

pen

Acc

ess

artic

les.

¯eld imaging spectrometer on SOFIA," J. Astron. Instrum7, 1840004.

Erickson, E. F. & Davidson, J. A. [1993] \SOFIA: StratosphericObservatory for Infrared Astronomy," Adv. Space Res.13(12), 549–556.

Fischer, C., Bryant, A., Beckmann, S. et al. [2016] \Observingwith FIFI-LS on SOFIA: Time estimates and strategies to usea ¯eld imaging spectrometer on an airborne observatory,"Proc. SPIE 9910, 991027.

Guesten, R., Camara, I., Hartogh, P. et al. [2003] \GREAT:The German Receiver for Astronomy at TerahertzFrequencies," Proc. SPIE 4857, doi: 10.1117/12.458820.

Iserlohe, C., Bryant, A., Krabbe, A. et al. [2018] \FIFI-LSobservations of the Circumnuclear Ring, Probing the high-density phase of the PDR," to appear in ApJ.

Kessler, M. F., Steinz, J. A., Anderegg, M. E. et al. [1996] \TheInfrared Space Observatory (ISO) mission," A&A 315, L27–L31.

Klein, R., Poglitsch, A., Fumi, F. et al. [2004] \Real-time op-eration without a real-time operating system for instrumentcontrol and data acquisition," Proc. SPIE 5496, doi:10.1117/12.551111.

Klein, R., Beckmann, S., Bryant, A. et al. [2014] \FIFI-LS: thefacility far-infrared spectrometer for SOFIA," Proc. SPIE9147, 91472X.

Lellouch, E. & Amri, H. [2006] Mars brightness model, http://www.lesia.obspm.fr/perso/emmanuel-lellouch/mars/index.php.

Liu, X.-W., Barlow, M. J., Cohen, M. et al. [2001] \ISO LWSobservations of planetary nebula ¯ne-structure lines,"MNRAS 323, 343.

Looney, L. W., Geis, N., Genzel, R. et al. [2000] \Realizing 3Dspectral imaging in the far-infrared: FIFI LS," Proc. SPIE4014, doi: 10.1117/12.389112.

Looney, L. W., Klein, R. et al. [2018] First Look at Orion-KLwith FIFI-LS onboard SOFIA, to appear.

Lord, S. D. [1992] NASA Technical Memorandum 103957.Nagao, T., Maiolino, R., Marconi, A., Matsuhara, H., [2011]

\Metallicity diagnostics with infrared ¯ne-structure lines,"A&A 526, A149.

Poglitsch, A., Waelkens, C., Geis, N. et al. [2010] \The Pho-todetector Array Camera and Spectrometer (PACS) on theHerschel Space Observatory," A&A 518, L2.

Raab, W., Looney, L. W., Poglitsch, A. et al. [2003] \FIFI LS:the optical design and di®raction analysis," Proc. SPIE 4857,91474G.

Rebell, F., Raab, W., Colditz, S. et al. [2014] \Precise angularpositioning at 6K: the FIFI-LS grating assembly," Proc.SPIE 9147, 914735.

Rebell, F., [2018] \Laboratory characterisation of FIFI-LScryogenic subsystems and their in-°ight performanceonboard SOFIA," to appear.

Rosenthal, D., Beeman, J. W., Geis, N. et al. [2000] \16 x 25 Ge:Ga detector arrays for FIFI LS," Proc. SPIE 4014, doi:10.1117/12.389096.

Schweitzer, M. [2008] \Optical Alignment and Characterizationof FIFI-LS — The Field-Imaging Far-Infrared Line Spec-trometer," Dissertation of the faculty of physics of theLudwig-Maximilian-University Munich.

Stutzki, J. [2004] \Sub-mm-instrumentation: Current projectsat KOSMA/Universit€at zu K€oln," Proc. SPIE 5498, doi:10.1117/12.555951.

Thompson, R. [2011] User Manual for the Telescope AssemblyAlignment Simulator, SOFIA Program, SCI-AR-MAN-OP02-2068.

C. Fischer et al.

1840003-14

J. A

stro

n. I

nstr

um. 2

018.

07. D

ownl

oade

d fr

om w

ww

.wor

ldsc

ient

ific

.com

by G

ER

MA

N A

ER

OSP

AC

E C

EN

TE

R (

DL

R)

on 0

1/09

/19.

Re-

use

and

dist

ribu

tion

is s

tric

tly n

ot p

erm

itted

, exc

ept f

or O

pen

Acc

ess

artic

les.