Infrared spectra of complex organic molecules in astronomically …linnartz/files/191-2018... · 2020. 8. 17. · Infrared spectra of complex organic molecules in astronomically relevant

AstronomyampAstrophysics

AampA 611 A35 (2018)httpsdoiorg1010510004-6361201731998copy ESO 2018

Infrared spectra of complex organic molecules inastronomically relevant ice matrices

I Acetaldehyde ethanol and dimethyl ether

J Terwisscha van Scheltinga12 N F W Ligterink12 A C A Boogert34E F van Dishoeck25 and H Linnartz1

1 Raymond and Beverly Sackler Laboratory for Astrophysics Leiden Observatory Leiden University PO Box 9513 2300 RA LeidenThe Netherlandse-mail [jeroentvslinnartz]strwleidenunivnl

2 Leiden Observatory Leiden University PO Box 9513 2300 RA Leiden The Netherlands3 Universities Space Research Association Stratospheric Observatory for Infrared Astronomy NASA Ames Research Center Moffett

Field CA 94035 USA4 Institute for Astronomy University of Hawaii 2680 Woodlawn Dr Honolulu HI 98622 USA5 Max-Planck Institut fuumlr Extraterrestrische Physik (MPE) Giessenbackstr 1 85748 Garching Germany

Received 26 September 2017 Accepted 27 November 2017

ABSTRACT

Context The number of identified complex organic molecules (COMs) in inter- and circumstellar gas-phase environments is steadilyincreasing Recent laboratory studies show that many such species form on icy dust grains At present only smaller molecular specieshave been directly identified in space in the solid state Accurate spectroscopic laboratory data of frozen COMs embedded in icematrices containing ingredients related to their formation scheme are still largely lackingAims This work provides infrared reference spectra of acetaldehyde (CH3CHO) ethanol (CH3CH2OH) and dimethyl ether(CH3OCH3) recorded in a variety of ice environments and for astronomically relevant temperatures as needed to guide or inter-pret astronomical observations specifically for upcoming James Webb Space Telescope observationsMethods Fourier transform transmission spectroscopy (500ndash4000 cmminus120ndash25 microm 10 cmminus1 resolution) was used to investigate solidacetaldehyde ethanol and dimethyl ether pure or mixed with water CO methanol or COmethanol These species were deposited ona cryogenically cooled infrared transmissive window at 15 K A heating ramp was applied during which IR spectra were recorded untilall ice constituents were thermally desorbedResults We present a large number of reference spectra that can be compared with astronomical data Accurate band positions andband widths are provided for the studied ice mixtures and temperatures Special efforts have been put into those bands of each moleculethat are best suited for identification For acetaldehyde the 7427 and 5803 microm bands are recommended for ethanol the 1136 and7240 microm bands are good candidates and for dimethyl ether bands at 9141 and 8011 microm can be used All spectra are publicly availablein the Leiden Database for Ice

Key words astrochemistry ndash methods laboratory molecular ndash techniques spectroscopic ndash molecular processes

1 IntroductionWater was the first molecule to be detected in the solid statein the interstellar medium (Gillett amp Forrest 1973) Since thenmore than 10 other molecules have been identified in icy form(ie CO CO2 CH4 NH3 and CH3OH) and it has become clearthat icy dust grains play a key role in the formation of boththese small molecules and more complex organic molecules(COMs) such as glycolaldehyde (HOCH2CHO) and ethyleneglycol (HOCH2CH2OH) The combined outcome of astronom-ical observations specifically space based missions such as theInfrared Space Observatory (ISO) and Spitzer Space Telescope(Kessler et al 1996 Werner et al 2004) laboratory and astro-chemical modelling studies have resulted in a detailed picture ofthe composition and structure of ice mantles on interstellar dustgrains and the chemical processes taking place (see reviews byGibb et al 2000 Herbst amp van Dishoeck 2009 Oumlberg et al 2011 Both authors contributed equally to this work

Caselli amp Ceccarelli 2012 Tielens 2013 Boogert et al 2015Linnartz et al 2015 Oumlberg 2016) It is generally accepted thatinterstellar ices form on the surface of dust grains in cold darkclouds through accretion in two distinct layers a polar H2O-richand an apolar CO-rich layer Water together with NH3 CO2and CH4 forms through atom addition reactions in lower densityenvironments (Hiraoka et al 1995 Miyauchi et al 2008 Obaet al 20092012 Dulieu et al 2010 Hidaka et al 2011 Linnartzet al 2011 Lamberts et al 2013 2014 Fedoseev et al 2015) Atlater stages when densities increase and temperatures decreasealong with the ongoing cloud collapse CO freeze-out from thegas phase occurs forming a CO coating on top of the water richlayer (Tielens et al 1991 Pontoppidan 2006) Subsequent hydro-genation processes transform CO to H2CO and H2CO to CH3OH(Watanabe amp Kouchi 2002 Fuchs et al 2009) resulting in COice intimately mixed with methanol (Cuppen et al 2011 Pen-teado et al 2015) Radical recombination processes in variousstarting mixtures triggered by energetic (ie UV photons or

Article published by EDP Sciences A35 page 1 of 43

AampA 611 A35 (2018)

Fig 1 Spectra of pure acetaldehyde (blue) ethanol (red) dimethyl ether (green) water (orange) methanol (purple) and CO (dark green) normalizedto one in the range of 25ndash200 microm The bands investigated in this work are indicated with an asterisk ()

cosmic rays) or non-energetic (ie atom additions) were shownto provide pathways towards the formation of more complexmolecules (see reviews of Linnartz et al 2015 Oumlberg 2016)

The molecules H2O CO CO2 CH4 NH3 and CH3OH makeup the bulk of interstellar ice (Ehrenfreund amp Charnley 2000Oumlberg et al 2011) but less abundant species have been observedas well These include species such as OCS and OCNminus (Palumboet al 1995 van Broekhuizen et al 2004) A number of COMssuch as formic acid (HCOOH) acetaldehyde (CH3CHO) andethanol (CH3CH2OH) have been tentatively detected based onspectroscopic features at 72 and 74 microm (Schutte et al 1999Oumlberg et al 2011) Several other spectroscopic features such asthe 60 and 68 microm bands remain only partly identified (Schutteet al 1996 Boudin et al 1998 Gibb amp Whittet 2002 Boogertet al 2008) Limited astronomical detection sensitivity com-bined with a lack of high resolution laboratory data have thusfar prohibited secure solid state identifications of COMs otherthan methanol but their presence in interstellar ices is generallyaccepted and also further supported by the recent detection of anumber of COMs on comet 67PChuryumovndashGerasimenko andin its coma (Goesmann et al 2015 Altwegg et al 2017)

With the upcoming launch of the James Webb Space Tele-scope (JWST) in 2019 new instruments such as MIRI (MidInfraRed Instrument Wright et al 2015) and NIRSpec (NearInfraRed Spectrograph Posselt et al 2004) will become avail-able to record telluric free spectra of interstellar ices at higherspectral and spatial resolution and with higher sensitivity thanpossible so far This opens up new possibilities to search for andstudy the level of molecular complexity in interstellar ices Toaid in the search for larger molecules in the solid state highresolution IR laboratory spectra are required The ice matrixenvironment and its temperature have to be taken into account

since these influence the spectral appearance of vibrationalbands

In this work we present the infrared spectra of acetalde-hyde ethanol and dimethyl ether respectively CH3CHOCH3CH2OH and CH3OCH3 The choice for these three speciesan aldehyde an alcohol and an ether is motivated by previoustentative identifications (Boudin et al 1998 Schutte et al 1999Oumlberg et al 2011) their astronomical gas-phase identificationand high abundance (eg Turner 1991 Gibb et al 2000 Cazauxet al 2003 Bisschop et al 2007b Taquet et al 2015 Muumllleret al 2016) and their common formation scheme upon UVirradiation of methanol ice (Oumlberg et al 2009) Formation ofthese molecules is seen in energetic processing experiments ofmethanol ice (Gerakines et al 1996 Bennett et al 2007 Oumlberget al 2009 Boamah et al 2014) and starts with cleavage of theCH3OH bonds This results in a reservoir of radicals that can beused for their formation as follows

bullCH3 +bullCHOrarr CH3CHO

bullCH3 +bullCH2OHrarr CH3CH2OH

bullCH3 +bullOCH3 rarr CH3OCH3

Formation of dimethyl ether and ethanol has also been stud-ied by radical recombination reactions starting from CH4H2Omixtures (Bergantini et al 2017) Besides energetic radicalrecombination reactions other formation pathways and linksbetween the three molecules exist as well For exampleacetaldehyde has been proposed as a solid state precursor ofethanol A hydrogen atom addition experiment showed thatacetaldehyde can at least partially (gt20) be transformedinto ethanol (Bisschop et al 2007a) Acetaldehyde itself hasbeen proposed to form as a spin-off in the well-studied

A35 page 2 of 43

J Terwisscha van Scheltinga et al Infrared spectra of complex organic molecules

CO + HrarrHCOrarrH2COrarrH3COrarrCH3OH chain (Charnley2004) HCO may directly interact with a C-atom to form HCCOthat upon hydrogenation yields CH3CHO (Charnley amp Rodgers2005)

This work presents a detailed study of the IR spectral char-acteristics of CH3CHO CH3CH2OH and CH3OCH3 in pureform and mixed in the interstellar relevant ice matrices H2O COCH3OH and COCH3OH Section 2 contains the experimentaldetails and measurement protocols The results of the measure-ments are presented and discussed in Sect 3 In Sect 4 theastronomical relevance of the new data is illustrated The con-clusions are summarized in Sect 5 A complete overview withall data obtained in this study is available from the appendices

2 Experimental

21 Set-up

The ice spectra are recorded in a high-vacuum (HV) set-upwhich is described in detail by Bossa et al (2015) A centralstainless steel chamber is evacuated by a 300 l sminus1 turbomolecu-lar pump backed by a double stage rotary vane pump (8 m3 hminus1)This allows a base pressure of sim10minus7 mbar at room tempera-ture The pressure is monitored by an Agilent FRG-720 fullrange gauge Ices are grown on an infrared transmissive ZnSewindow that is cryogenically cooled to a lowest temperatureof 12 K by a closed cycle helium cryostat The temperatureof the window is monitored by a LakeShore 330 temperaturecontroller which regulates a feedback loop between a resistiveheating wire and a silicon diode temperature sensor An abso-lute temperature accuracy of plusmn2 K and a relative accuracy ofplusmn1 K is acquired with this diode The IR beam of a FourierTransform InfraRed Spectrometer (FTIRS Varian 670-IR) isaligned through the window to obtain IR spectra of the samplesThe spectrometer covers a range of 4000ndash500 cmminus1 (25ndash20 microm)at spectral resolutions as high as 01 cmminus1 Samples are exter-nally prepared in a 2L glass bulb using a separate multi-linegas mixing system The gas mixing line is turbomolecularlypumped to pressures lt1times 10minus4 mbar Gas mixtures are madeby sequential addition of its components Two gas independentgauges covering various pressure ranges ensure that accuratemixing ratios are obtained with a maximum error of lt10 Theliquids and gases used in these experiments are acetaldehyde(SigmandashAldrich 995) ethanol (Biosolve 999) dimethylether (SigmandashAldrich 999) water (Milli-Q Type I) carbonmonoxide (Linde gas 99997) and methanol (SigmandashAldrich999) Liquid samples are purified with freeze-pump-thawcycles before use

22 Measurement protocol

Pure or premixed gases are background deposited onto the 15 Kcold sample via an inlet valve A standard pressure of 20 mbar inthe glass bulb is used to prevent a decreasing inlet pressure gra-dient during deposition Bi-mixed gases are prepared in a 120ratio and tri-mixed gases in a 12020 ratio where the smallestfraction is the COM under investigation These dilution fac-tors ensure that the COM mainly interacts with the surroundingmatrix resulting in matrix shifted IR vibrational bands Icesare grown at 15 K to a column density of sim4500 ML (1 mono-layer is 1times 1015 mol cmminus2) on the window This coverage ensuresthat any influence of background contamination mainly waterdepositing at a rate of less than 30 ML hminus1 can be neglectedDuring deposition IR spectra are recorded at 1 cmminus1 resolution

(05 cmminus1 step size) and averaged over 61 scans (equals 2 min)to trace the ice growth and determine when the ice is sim4500 MLthick From the integration of the IR band absorption the col-umn density of the species Nspecies is determined according to

Nspecies = ln(10)

intband log10

(I0(ν)I(ν)

)dν

Aprime (1)

whereint

band log10

(I0(ν)I(ν)

)dν is the integrated absorbance of the

band and I0(ν) and I(ν) are the flux received and transmitted bythe sample respectively and Aprime is the apparent band strength incm molminus1 It is important to realize that strongly absorbing bandsmay get saturated at high coverages resulting in unreliable col-umn density measurements In the experiments conducted theCO band at 2135 cmminus1 reaches saturation at high coverage asdo certain bands of pure acetaldehyde and dimethyl ether Forthese species bands with a lower band strength or isotopologuescan be used The measured column densities give an indicationwhether the mixed ice composition still matches the gas-phasemixing ratio and whether the COMs are sufficiently diluted inthe matrix Small variations in the composition of the gas mix-ture and matrix interactions complicate accurate ice mixing ratiodeterminations The apparent band strengths are listed in Table 1and taken from literature for acetaldehyde and ethanol For thedimethyl ether bands at 923 1095 and 1164 cmminus1 the bandstrength value is approximated from a methanedimethyl ethermixture prepared at a one to one ratio in the gas phase Assum-ing this ratio is maintained in the ice and matrix interactions arenegligible the apparent band strength is determined from a com-parison with the methane 1302 cmminus1 band area and its knownapparent band strength of 80times 10minus18 cm molminus1 (Bouilloud et al2015)

After deposition the sample is linearly heated at a rate of25 K hminus1 until it is fully desorbed from the window The lowtemperature ramp ensures that the ice has sufficient time toundergo structural changes particularly from the amorphous tothe crystalline phase During heating IR spectra are continuouslyrecorded and averaged over 256 scans to trace spectral changesvs temperature

23 Analysis protocol

Owing to the very large amount of spectra that are recorded dur-ing the experiments we only present samples of representativeIR spectra for temperatures at which significant spectral changesoccur These spectra are baseline subtracted and the peak posi-tion and band width at full width at half maximum (FWHM)are determined for selected spectral features When the band ofa COM overlaps with a spectral feature of a matrix moleculealso the matrix feature is subtracted where possible In the caseof band splitting the least intense component is only taken intoaccount when its peak position is clearly distinguishable In a fewcases splitted peaks rival in intensity and are heavily overlappingand it is not possible to fit a FWHM for the individual compo-nents Here the FWHM of the combined peaks is determinedPeaks are selected for analysis mainly based on their intensityand potential as an ice tracer ie selecting wavelengths for whichno strong overlap with known interstellar features exist

Identification of vibrational modes of the three speciesstudied here is realized by comparison with available spectrafrom liquid and solid state literature (Plyler 1952 Evans ampBernstein 1956 Barnes amp Hallam 1970 Allan et al 1971

A35 page 3 of 43

AampA 611 A35 (2018)

Table 1 Selected bands of acetaldehyde ethanol and dimethyl ether

Species Formula Mode Peak position Aprime

cmminus1 microm cm molminus1

Acetaldehyde CH3CHO

CH3 rock + CC stretch 11223 8909+ CCO bendCH3 s-deform + CH wag 13462 7427CH3 deform 14294 6995CO stretch 17230 5803 13times 10minus17 a

Ethanol CH3CH2OH

CC stretch 8798 1136 324times 10minus18 b

CO stretch 10510 9514 141times 10minus17 b

CH3 rock 10905 9170 735times 10minus18 b

CH2 tors 12752 7842OH deform 13302 7518CH3 s-deform 13813 7240

Dimethyl ether CH3OCH3

COC stretch 9213 1085 50times 10minus18 c

COC stretch + CH3 rock 10939 9141 92times 10minus18 c

COC stretch + CH3 rock 11638 8592 98times 10minus18 c

CH3 rock 12482 8011

Notes (lowast) Peak position of the pure molecule at 15 K Note that throughout literature there seems to be disagreement in the assignment of certainmodes particularly for ethanolReferences (a) Schutte et al (1999) (b) Hudson (2017) (c) This work

Hollenstein amp Guumlnthard 1971 Mikawa et al 1971 Coussan et al1998) Optical effects such as longitudinal opticalndashtransverseoptical (LO-TO) splitting and particle shape effects are notexplicitly taken into account Since spectra are recorded atnormal incidence with unpolarized light only the TO modesare recorded However certain combinations of polarized lightand angles of incidence can result in the LO phonon modeshowing up (Baratta et al 2000 Palumbo et al 2006) Alsoparticle shape effects can shift transition bands with respect torecorded laboratory spectra (Baratta amp Palumbo 1998) Sucheffects affect only the spectra of more abundant species suchas CO or CO2 and are not considered to be relevant forCOMs

3 Results and discussion

In this section selected results of the acetaldehyde ethanol anddimethyl ether experiments are presented These are represen-tative for the much larger data set given in the appendix Allthe selected spectra used in this work are publicly availablefrom the Leiden Database for Ice1 spectra recorded for othertemperatures are available on request Figure 1 shows the IRspectra of pure acetaldehyde ethanol and dimethyl ether iceat 15 K the bands that are fully analysed are indicated withan asterisk () and spectra of pure water CO and methanolice Figures of the spectra of COMs mixed in water COmethanol and COmethanol at 15 K are shown in Appendix AIn Table 1 these selected bands are listed together with theirpeak positions and if available apparent band strength in pureices at 15 K Appendix B presents the results of the analy-sis of the selected bands listing peak positions FWHMs and

1 httpicedbstrwleidenunivnl

integrated absorbance ratios at various temperatures and for dif-ferent ice matrices A representative example of the tables listedin the appendix is shown in Table 2 for the acetaldehyde CH3s-deformation + CH wagging mode at 13466 cmminus1 at 15 K

For easier interpretation the results are represented in a num-ber of plots see Figs 2ndash4 for examples Each plot covers thedata of one band In all plots the top panels show spectroscopicchanges of the band under thermal processing in pure and mixedices The bottom left panels plot peak position vs FWHM show-ing trends in the band The bottom right panels give an indicationof how the band strengths change relative to each other in variousmatrices Assuming that the ice column density is roughly thesame for each experiment and that the gas mixing ratio is closeto the ice mixing ratio the mixed ices are corrected for theirdilution factor Owing to various uncertainties in this methodthis results in relatively large error bars for the band strengthsThis is unfortunate as this would allow us to interpret the spec-troscopic identifications ndash the primary goal of this work ndash alsoin terms of accurate column densities The remaining figures ofother bands can be found in Appendix C

A few general statements can be made Most peaks displaypeak narrowing under thermal processing which is due to theice changing to a crystalline phase with increasing temperatureMixed ice in CO and COCH3OH are exceptions due to thevolatility of CO and its removal from the ice at relatively lowtemperatures Above 30 K the desorption temperature of CO(Oumlberg et al 2005) these ices are often seen to display peakbroadening

Peak splitting especially at high temperatures is anothereffect that is generally seen This can be caused by two or moremodes contributing to a single feature at low temperatures andbecoming visible as the peaks begin to narrow at higher tem-peratures Alternatively the matrix can play a role and a peak is

A35 page 4 of 43

J Terwisscha van Scheltinga et al Infrared spectra of complex organic molecules

Fig 2 Top from left to right the acetaldehyde 7427 microm band pure (black) and in water (red) CO (blue) methanol (purple) and COCH3OH(green) at various temperatures Bottom left peak position vs FWHM plot using the same colour coding Bottom right the relative band strengthfor the 7427 microm band at 15 K in various matrices

split owing to different interactions of a functional group with itssurroundings For example an ice can segregate under thermalprocessing and have part of the COM still intimately mixed withthe matrix molecule while another part is forming COM clus-ters Segregation is an effect most clearly seen in the COMCOice mixtures

Integrated absorbance ratios are given for the bands underinvestigation in each ice mixture These ratios can provide a toolto estimate the likelihood of observing other bands upon detec-tion of a specific transition They can also be used as conversionfactors to determine band strengths from known band strengths

The bands are normalized on the band with highest integratedabsorbance at 15 K unless this band is suspected to be in satu-ration or when the data set is incomplete over the investigatedtemperature range

31 Acetaldehyde

Acetaldehyde hosts four significant features in the 55ndash125 micromregion (see Fig 1) Some smaller features are also visible suchas the CC stretching + CH3 rocking mode close to 11 microm how-ever its intensity is very small compared to the other bands Two

A35 page 5 of 43

AampA 611 A35 (2018)

Table 2 Peak position and FWHM of the acetaldehyde CH3 s-deformation + CH wagging mode at 15 K in various matrices

Mixture Temperature λpeakminusbaseline λpeakminusmatrix FWHM(K) (cmminus1) (microm) (cmminus1) (microm) (cmminus1) (microm)

CH3CHO

15

13466 74264 ndash ndash 135 00744CH3CHOH2O 13499 74078 13499 74078 92 00502CH3CHOCO 13494 74104 ndash ndash 48 00262CH3CHOCH3OH 13475 74211 ndash ndash 130 00714CH3CHOCOCH3OH 13494 74105 ndash ndash 126 00691

Notes Excerpt from Table B1 (lowast) FWHM result of two or more blended peaks

characteristic vibrational modes of acetaldehyde at 6995 and8909 microm coincide with methanol CH3 rocking and deformationmodes and are likely obscured in interstellar spectra A solidstate identification of acetaldehyde based on these vibrationalmodes is unlikely The CO stretching mode is the most prominentband in this spectrum However its location at 58 microm coincideswith the CO stretching mode of many other molecules suchas formaldehyde (H2CO) formic acid (HCOOH) or formamide(NH2CHO) which are expected to be present in interstellar icemaking it likely that this band is blended The fourth band isthe CH3 s-deformation + CH wagging mode around 7427 micromwhich is found to have no substantial overlap with abundant bulkinterstellar ice components and therefore is most suited for asuccessful solid state identification of this molecule

Figure 2 shows the results of the analysed data of the CH3s-deformation + CH wagging band Under thermal processingthe band widths are generally seen to decrease this is causedby crystallization in the ice Peak positions shift as well withclear blue shifting trends visible for the COCH3OH and watermixtures In the case of the COCH3OH mixture this is likelybecause of the loss of CO from the matrix while for the watermixture the interaction between acetaldehyde and crystallinewater is more likely the cause In some cases at high temper-ature CH3CHO undergoes peak splitting making identificationthrough FWHM challenging However this can also be used asa tool to determine the ice temperature The comparison of peakposition makes it in general easy to distinguish between pureacetaldehyde acetaldehyde mixed in CH3OH and COCH3OHacetaldehyde mixed in CO and acetaldehyde mixed in waterThe 7427 microm band shows a substantial decrease in band strengthby about 40 when acetaldehyde is surrounded by matrixmolecules

The acetaldehyde CO stretching band underlines the abovefindings given it is clearly observed (see Fig C3) Especially atlow ice temperatures of 15 and 30 K clear peak shifts are visiblebetween the COCH3OH matrix at 584 microm the water matrix at5825 microm and the pure matrix or in a CH3OH matrix at around5805 microm

32 Ethanol

The spectrum of pure ethanol in Fig 1 shows a strong CC stretch-ing band at 1136 microm CO stretching mode at 9514 microm and CH3rocking mode at 9170 microm A number of weaker modes are seenbetween 65 and 85 microm specifically the CH2 torsion mode at7842 microm the OH deformation mode at 7518 microm and the CH3symmetric deformation mode at 7240 microm Overlap with spec-tral features of bulk interstellar ice species such as water andmethanol is an issue for the three strongest bands coincidingwith either the water libration mode or CO stretching and CH3rocking modes of methanol Also the prominent broad silicate

feature is present at sim97 microm Although the other ethanol modesare substantially weaker they fall within a spectral region that isgenerally clean of strong transitions

The ethanol 7240 microm band is a possible candidate for iden-tification Figure 3 shows the data of this band Ethanol mixedin water can be distinguished from other features by a sim3 cmminus1

peak shift from other mixtures In general it is found thatthe ethanolwater mixture is relatively easy to distinguish butthe other mixtures display much overlap in peak position andFWHM The CH2 torsion OH deformation mode and CH3 sym-metric deformation mode are hard to identify in the ethanolCOmixture owing to the appearance of many more modes Bandareas and relative band strengths of these modes are thereforenot considered The band strength is seen to vary substantiallyfor the various bands but does not show a clear trend

33 Dimethyl ether

Three strong bands of dimethyl ether are found at 1085 9141and 8592 microm for the COC stretching and two COC stretch-ing + CH3 rocking modes respectively A much weaker CH3rocking mode is found at 8011 microm The first two overlap withknown interstellar ice features of methanol water and silicatesand are therefore less suited for an identification while the thirdlikely falls in the wing of such features and may still be usedEven though it is a weak mode the 8011 microm band falls in arelatively empty region of interstellar ice spectra This featurecould therefore be most suited for a dimethyl ether identificationsee Fig 4

For the 8011 microm band clear differences are seen dependingon the matrix The spectra of pure and methanol mixture are dis-tinguishable from those of the water and COCH3OH mixturesby a sim2 cmminus1 peak shift of the low temperature spectra at 15 and30 K In water this band displays a narrower peak compared tothe other ices The other bands also show many clear differencesin peak position and FWHM between the various ice mixturesA characteristic peak splitting structure at low temperatures isseen for the 1085 microm band when mixed in water methanolor COCH3OH Interestingly the relative band strength showsa substantial increase in the CH3OH and COCH3OH mixturesfor the 8011 microm band Other modes do not show such clear dif-ferences Also it is interesting to note the fact that the COCstretching mode has the largest band area when mixed in waterwhile in the other mixtures this is always the CH3 rocking modeat 8592 microm (see Appendix B6)

4 COM ice features in W33A

Our extensive measurements of frozen COMs are needed inthe analysis of the many spectra of dense clouds embedded

A35 page 6 of 43

J Terwisscha van Scheltinga et al Infrared spectra of complex organic molecules

Fig 3 Top from left to right the ethanol 7240 microm band pure (black) and in water (red) CO (blue) methanol (purple) and COCH3OH (green)at various temperatures Bottom left peak position vs FWHM plot using the same colour coding Bottom right the relative band strength for the7240 microm band at 15 K in various matrices

protostars and inclined protoplanetary disks that will beobtained with the upcoming JWST mission at high sensitivityand medium spectral resolution (R of up to 3500) Here wedemonstrate their use by a reanalysis of a spectrum of themassive protostar W33A obtained with the Infrared SpaceObservatoryrsquos Short Wavelength Spectrometer (AstronomicalObservation Template 1 R = 800) This is one of the few sourcesfor which a high quality mid-IR spectrum is available (Gibbet al 2000) In the 7 to 8 microm region three prominent featuresat 725 741 and 767 microm have been described previously in

the literature The 725 microm feature has been attributed to bothCH3CH2OH and HCOOH (Schutte et al 1999 Oumlberg et al2011) the 741 microm feature has been attributed to HCOOminus andCH3CHO (Schutte et al 1999) and the 767 microm band has beenidentified as solid methane with potentially contributions of SO2(Boogert et al 1996)

In this work we make use of the water and silicate subtractedspectrum of W33A shown in Fig 5 with a straight line localcontinuum subtraction The aforementioned features are visiblealthough the 741 microm feature seems to have two contributions at

A35 page 7 of 43

AampA 611 A35 (2018)

Fig 4 Top from left to right the dimethyl ether 8011 microm band pure (black) and in water (red) CO (blue) methanol (purple) and COCH3OH(green) at various temperatures Bottom left peak position vs FWHM plot using the same colour coding Bottom right the relative band strengthfor the 8011 microm band at 15 K in various matrices

747 and 740 microm and the 725 microm feature is found at 722 micromThe spectra of ethanol and acetaldehyde mixed in COCH3OHand H2O are plotted in the same figure The peak position ofthe 740 microm feature can be reproduced well by the acetaldehydeCH3 s-deformation mode in both mixtures However the band inCOmethanol mixture seems to be too broad to justly reproducethe W33A 740 microm feature and also this band covers the 747 micromfeature next to it The other two features at 722 and 747 micromcould be the result of the CH3 s-deformation and OH defor-mation modes of ethanol Particularly the CH3CH2OHH2Omixture coincides with the peak locations of the 722 and 747 microm

features in the W33A spectrum While the identification ofacetaldehyde and ethanol are plausible detection of additionalfeatures would strengthen the assignment We checked and foundthat none of the other CH3CHO and CH3CH2OH bands have ananti-coincidence with the W33A spectrum

Upper limits to the ice column densities of ethanol andacetaldehyde can be given based on the integrated optical depthof their potential features Schutte et al (1999) give integratedτ values of 20plusmn 03 and 16plusmn 05 cmminus1 respectively Bandstrength values of ethanol and acetaldehyde are taken from theliterature and used to calculate the column densities of the two

A35 page 8 of 43

J Terwisscha van Scheltinga et al Infrared spectra of complex organic molecules

Fig 5 Continuum and water and silicate subtracted spectrum of W33Aplotted together with ice spectra of ethanol and acetaldehyde at 15 Kmixed in COCH3OH and H2O Features in the W33A spectrum areindicated with dashed lines at 722 740 and 747 microm The large spectralfeature at 767 microm is due to CH4 and SO2

features The ethanol band strength of the CO stretch mode at9514 microm has been determined to be 141times 10minus17 cm molminus1 by(Hudson 2017) Using the integrated absorbance ratio CH3 s-defCO str = 020 at 15 K from Table B16 the band strength ofthe CH3 stretch mode is determined to be 28times 10minus18 cm molminus1The effect of the matrix on the relative band strength is smallfor both the ethanol CO stretch and CH3 s-deformation modesas can be seen from Figs 3 and C5 and therefore assumed tobe negligible Assuming the entire 722 microm feature is caused byethanol this results in a column density of 71plusmn 02times 1017 cmminus2

In Schutte et al (1999) the acetaldehyde band strength isgiven as 13times 10minus17 cm molminus1 for the CO stretch mode basedon data from Wexler (1967) The integrated absorbance ratioof CO stretchingCH3 s-deforming + CH wagging = 432 in pureacetaldehyde at 15 K in laboratory experiments As the COstretching mode is likely saturated the ratio may thus behigher Using this ratio the band strength of the CH3 s-deformation mode is found to be 30times 10minus18 cm molminus1 As canbe seen in Fig 2 the relative band strength of this modedecreases substantially in mixtures by about 40 The band

Table 3 Ice upper limits and gas-phase abundances of ethanol andacetaldehyde towards W33A

Species Ice Gas phasec

N(H2O)a N(CH3OH)b N(CH3OH)

CH3CH2OH le19 le42 24CH3CHO le23 le52 le02

Notes Abundances given in References (a) Keane et al (2001) (b) Dartois et al (1999) (c) Bisschopet al (2007b)

strength of the CH3 s-deformation mode in mixed ices is thus18times 10minus18 cm molminus1 If the entire 740 microm feature is attributedto CH3CHO the resulting column density is 89plusmn 3times 1017 cmminus2

In all likelihood the 722 and 740 microm features contain con-tributions of other molecules mainly HCOOH and HCOOminus andthe reported values should therefore be seen as upper limitsUsing solid water and methanol column densities of 38times 1019

and 17times 1018 cmminus2 respectively towards W33A (Dartois et al1999 Keane et al 2001) the upper limit abundance ratiosof ethanol and acetaldehyde can be determined The abun-dance ratio N(COM)N(H2O) is found to be le19 and le23while N(COM)N(CH3OH) is le42 and le52 for ethanol andacetaldehyde respectively The abundances with respect to waterare in good agreement with previously reported values of le4and le36 for ethanol and acetaldehyde respectively (Boudinet al 1998 Schutte et al 1999)

The N(COM)N(CH3OH) upper limit ice abundance can becompared with known gas-phase abundances towards W33AThese are given as N(CH3CH2OH)N(CH3OH) = 24 andN(CH3CHO)N(CH3OH) lt 02 (Bisschop et al 2007b) andare substantially lower than the ice upper limits Interferometricobservations with the Atacama Large MillimetersubmillimeterArray are needed to spatially resolve these molecules and deter-mine more accurate abundances Beside being upper limits thisdifference may be linked to the process that transfers solid statespecies into the gas phase causing molecules to fragment or toother destruction of species in the gas phase An overview of theCOM abundances in ice and in the gas phase towards W33A isgiven in Table 3

The spectroscopic data presented in this paper combinedwith the improvements in terms of sensitivity and resolution ofJWST will aid in confirming these detections and distinguishother potential contributors to these features More observa-tions particularly towards low-mass sources will give additionalinformation about the carriers of these features

5 Conclusions

This paper adds to and extends on data of three important inter-stellar ice candidates acetaldehyde ethanol and dimethyl etherA number of selected bands are fully characterized in FWHMand peak positions and show clear changes in various matricesOur conclusions are summarized as follows

1 The most promising bands to identify the COMs studied herein interstellar ice spectra are the 7427 and 588 microm bands ofacetaldehyde the 7240 and 1136 microm bands of ethanol andthe 8011 and 8592 microm bands of dimethyl ether

2 Matrix characteristic shifts in peak position and FWHM areseen for several bands The acetaldehyde CH3 deformation

A35 page 9 of 43

AampA 611 A35 (2018)

and CO stretching mode can be distinguished in the H2OCO CH3OH and COCH3OH matrices Ethanol shows gen-erally less distinctive shifts and only bands in the watermatrix are unique At low temperatures matrix specificdimethyl ether band shifts can be identified specifically forthe CH3 rocking mode at 8011 microm

3 Given the higher complexity of the involved spectra unam-biguous identifications need to involve different bands thatreflect bandwidths and intensity ratios as found in thelaboratory studies The dependence on matrix environmentand temperature provides a tool to use these transitions as aremote diagnostic instrument

4 Analysis of the ISO W33A spectrum in the 7 microm regionshows a number of features that can be assigned to theCOMs studied in this work The 740 microm feature matchesthe position of the CH3 s-deformation mode of acetalde-hyde and the 722 microm feature is plausibly caused by theCH3 s-deformation mode of ethanol It is likely that 722 micromband is specifically caused by ethanol mixed in water Abun-dances of both molecules with respect to water ice aredetermined to be le23 and le34 for acetaldehyde andethanol respectively

Acknowledgements The authors thank ME Palumbo for useful discussions onband profile changes due to grain shape differences We also thank S Ioppolofor many discussions This research was funded through a VICI grant of NWOthe Netherlands Organization for Scientific Research Astrochemistry in Lei-den is supported by the European Union A-ERC grant 291141 CHEMPLANby the Netherlands Research School for Astronomy (NOVA) and by a RoyalNetherlands Academy of Arts and Sciences (KNAW) professor prize

ReferencesAllan A McKean D C Perchard J-P amp Josien M-L 1971 Spectrochim

Acta A Mol Spectr 27 1409Altwegg K Balsiger H Berthelier J J et al 2017 MNRAS 469 S130Baratta G A amp Palumbo M E 1998 J Opt Soc Am A 15 3076Baratta G A Palumbo M E amp Strazzulla G 2000 AampA 357 1045Barnes A J amp Hallam H E 1970 Trans Faraday Soc 66 1932Bennett C J Chen S-H Sun B-J Chang A H H amp Kaiser R I 2007

ApJ 660 1588Bergantini A Maksyutenko P amp Kaiser R I 2017 ApJ 841 96Bisschop S E Fuchs G W van Dishoeck E F amp Linnartz H 2007a AampA

474 1061Bisschop S E Joslashrgensen J K van Dishoeck E F amp de Wachter E B M

2007b AampA 465 913Boamah M D Sullivan K K Shulenberger K E et al 2014 Faraday

Discuss 168 249Boogert A C A Schutte W A Tielens A G G M et al 1996 AampA 315

L377Boogert A C A Pontoppidan K M Knez C et al 2008 ApJ 678 985Boogert A C A Gerakines P A amp Whittet D C B 2015 ARAampA 53

541Bossa J-B Mateacute B Fransen C et al 2015 ApJ 814 47Boudin N Schutte W A amp Greenberg J M 1998 AampA 331 749Bouilloud M Fray N Beacutenilan Y et al 2015 MNRAS 451 2145Caselli P amp Ceccarelli C 2012 AampARv 20 56Cazaux S Tielens A G G M Ceccarelli C et al 2003 ApJ 593 L51Charnley S B 2004 Adv Space Res 33 23

Charnley S B amp Rodgers S D 2005 in Astrochemistry Recent Successes andCurrent Challenges eds D C Lis G A Blake amp E Herbst IAU Symp 231237

Coussan S Bouteiller Y amp Perchard JP 1998 Spectrochim Acta A MolSpectr 102 5789

Cuppen H M Penteado E M Isokoski K van der Marel N amp Linnartz H2011 MNRAS 417 2809

Dartois E Schutte W Geballe T R et al 1999 AampA 342 L32Dulieu F Amiaud L Congiu E et al 2010 AampA 512 A30Ehrenfreund P amp Charnley S B 2000 ARAampA 38 427Evans J amp Bernstein H 1956 Can J Chem 34Fedoseev G Ioppolo S Zhao D Lamberts T amp Linnartz H 2015 MNRAS

446 439Fuchs G W Cuppen H M Ioppolo S et al 2009 AampA 505 629Gerakines P A Schutte W A amp Ehrenfreund P 1996 AampA 312 289Gibb E L amp Whittet D C B 2002 ApJ 566 L113Gibb E L Whittet D C B Schutte W A et al 2000 ApJ 536 347Gillett F C amp Forrest W J 1973 ApJ 179 483Goesmann F Rosenbauer H Bredehoumlft J H et al 2015 Science 349Herbst E amp van Dishoeck E F 2009 ARAampA 47 427Hidaka H Watanabe M Kouchi A amp Watanabe N 2011 Phys Chem

Chem Phys 13 15798Hiraoka K Yamashita A Yachi Y et al 1995 ApJ 443 363Hollenstein H amp Guumlnthard H H 1971 Spectrochim Acta A Mol Spectr 27

2027Hudson R L 2017 Spectrochim Acta A Mol Spectr 187 82Keane J V Tielens A G G M Boogert A C A Schutte W A amp Whittet

D C B 2001 AampA 376 254Kessler M F Steinz J A Anderegg M E et al 1996 AampA 315 L27Lamberts T Cuppen H M Ioppolo S amp Linnartz H 2013 Phys Chem

Chem Phys 15 8287Lamberts T Cuppen H M Fedoseev G et al 2014 AampA 570 A57Linnartz H Bossa J-B Bouwman J et al 2011 in The Molecular Universe

IAU Symp 280 390Linnartz H Ioppolo S amp Fedoseev G 2015 Int Rev Phys Chem 34 205Mikawa Y Brasch J amp Jakobsen R 1971 Spectrochim Acta A Mol Spectr

27 529Miyauchi N Hidaka H Chigai T et al 2008 Chem Phys Lett 456 27Muumlller H S P Belloche A Xu L-H et al 2016 AampA 587 A92Oba Y Miyauchi N Hidaka H et al 2009 ApJ 701 464Oba Y Watanabe N Hama T et al 2012 ApJ 749 67Oumlberg K I 2016 Chem Rev 116 9631Oumlberg K I van Broekhuizen F Fraser H J et al 2005 ApJ 621 L33Oumlberg K I Garrod R T van Dishoeck E F amp Linnartz H 2009 AampA 504

891Oumlberg K I Boogert A C A Pontoppidan K M et al 2011 ApJ 740 109Palumbo M E Tielens A G G M amp Tokunaga A T 1995 ApJ 449 674Palumbo M E Baratta G A Collings M P amp McCoustra M R S 2006

Phys Chem Chem Phys 8 279Penteado E M Boogert A C A Pontoppidan K M et al 2015 MNRAS

454 531Plyler E 1952 J Res Natl Bur Stand 48 281Pontoppidan K M 2006 AampA 453 L47Posselt W Holota W Kulinyak E et al 2004 in Optical Infrared and

Millimeter Space Telescopes ed J C Mather Proc SPIE 5487 688Schutte W A Tielens A G G M Whittet D C B et al 1996 AampA 315

L333Schutte W A Boogert A C A Tielens A G G M et al 1999 AampA 343

966Taquet V Loacutepez-Sepulcre A Ceccarelli C et al 2015 ApJ 804 81Tielens A G G M 2013 Rev Mod Phys 85 1021Tielens A G G M Tokunaga A T Geballe T R amp Baas F 1991 ApJ 381

181Turner B E 1991 ApJS 76 617van Broekhuizen F A Keane J V amp Schutte W A 2004 AampA 415 425Watanabe N amp Kouchi A 2002 ApJ 571 L173Werner M W Roellig T L Low F J et al 2004 ApJS 154 1Wexler A 1967 Appl Spectr Rev 1 29Wright G S Wright D Goodson G B et al 2015 PASP 127 595

A35 page 10 of 43

J Terwisscha van Scheltinga et al Infrared spectra of complex organic molecules

Appendix A Spectra

The following figures show the spectra of acetaldehydeethanol and dimethylether mixed in water CO methanol andCOmethanol in the range of 25ndash200 microm All spectra are takenat 15 K

Fig A1 Spectra of acetaldehyde (blue) ethanol (red) and dimethyl ether (green) mixed in water at 15 K in the range of 25ndash200 microm

Fig A2 Spectra of acetaldehyde (blue) ethanol (red) and dimethyl ether (green) mixed in CO at 15 K in the range of 25ndash200 microm

A35 page 11 of 43