Mid- and far-infrared absorption spectroscopy of Titan's aerosols analogues
1 2 3
4 5
Author List
6
Alexandre Giuliani3,4, Cyril Szopa1 , Carrie M. Anderson5, Jean- Jacques Correia1,
7
Paul Dumas3 and Guy Cernogora1
8
1
9
d'Alembert 78280 Guyancourt, France
Thomas Gautier1, Nathalie Carrasco1, Ahmed Mahjoub1, Sandrine Vinatier2,
LATMOS, Université Versailles St Quentin, UPMC Univ. Paris 06, CNRS, 11 Bvd
10
2
LESIA, Observatoire de Paris, 5 place Jules Janssen, 92195, Meudon, France
11
3
Synchrotron SOLEIL, L’orme des Merisiers, BP 48, Saint Aubin, F-91192 Gif sur
12
Yvette, France
13
4
INRA, U1008 CEPIA, Rue de la Géraudière, F-44316 Nantes, France
14
5
NASA Goddard Space Flight Center, Solar System Exploration Division, Greenbelt,
15
MD 20771, United States
16
Editorial correspondence to:
17 18 19 20 21 22 23 24 25 26
Mr. Thomas Gautier LATMOS - University of Versailles St Quentin 11 Boulevard d'Alembert 78280 Guyancourt France Phone: (0033) 1 80 28 52 77 Fax: (0033) 1 80 28 52 90 E-mail:
[email protected] Published in Icarus 221 (1) : 320-327
Page 1
27
28
Abstract
29
In this work we present mid- and far-Infrared absorption spectra of Titan's aerosol
30
analogues produced in the PAMPRE experimental setup. The evolution of the linear
31
absorption coefficient ε (cm-1) is given as a function of the wavenumber.
32
We provide a complete dataset regarding the influence that the concentration of
33
methane vapor in the gas mixture has on the tholin spectra. Among other effects, the
34
intensity of the 2900 cm-1 (3.4 µm) pattern (attributed to methyl stretching modes)
35
increases when the methane concentration increases. More generally, tholins
36
produced with low methane concentrations seem to be more amine based polymers,
37
whereas tholins produced with higher methane concentrations contains more
38
aliphatic carbon based structures.
39
Moreover, it is shown that the position of the bands around 2900 cm -1 depends on
40
the chemical environment of the methyl functional group. We conclude that the
41
presence of these absorption bands in Titan's atmosphere, as measured with the
42
VIMS instrument onboard Cassini is in agreement with an aerosol contribution.
43
We also compare the far-infrared spectrum of tholin to spectra of Titan's aerosols
44
derived from recent Cassini-CIRS observations displaying many similarities,
45
particularly with absorption bands at 325 cm -1, 515 cm-1, and the methyl attributed
46
1380 cm-1 and 1450 cm-1 bands.
47
Page 2
48
1 Introduction
49
The atmosphere of Titan is mainly comprised of N2 and CH4. Organic chemical
50
reactions are induced by solar irradiation and electrically charged particles
51
accelerated in Saturn's magnetosphere (Sittler Jr et al. 2009). These reactions lead
52
to the production of an opaque layer of organic solid aerosols in the atmosphere.
53
These aerosols have a major impact on several parameters of Titan's atmosphere,
54
such as the greenhouse effect (McKay et al. 1991), or condensation (Lavvas et al.
55
2011). Studying Titan's aerosols is then of primary importance in order to understand
56
both the physics and the chemistry of Titan's atmosphere.
57
One possible and achievable method to study Titan’s aerosol is to produce and study
58
laboratory analogues, coined "tholins". A discussion on the different experimental
59
setups designed for such a purpose can be found in Cable et al. (2011). The
60
properties of the produced tholins allow a better analysis and understanding of
61
observational data of the atmosphere of Titan.
62
However, it should be noted that, even if Titan's aerosol spectrum is better
63
characterized at present by observations in the infrared spectral range thanks to the
64
Cassini mission (Bellucci et al. 2009; Rannou et al. 2010; Vinatier et al. 2010;
65
Anderson and Samuelson 2011; Vinatier et al. 2012) very few laboratory studies
66
provide data on tholin in the extended far- to mid-infrared spectral domain. Up to
67
now, the complex refractive indices of tholin published in Khare et al. (1984) were
68
often used since they cover a large spectral range from 0.02 µm to 920 µm but
69
present some discrepancies with observations done by both by Cassini Composite
70
Infrared Spectrometer (CIRS) and Visible and Infrared Mapping Spectrometer (VIMS)
71
instruments.
Page 3
72
Quirico et al. (2008) and Imanaka et al. (2004) have reported different absorption
73
spectra on laboratory tholin, but only down to 500 cm-1. Very recently, Imanaka et al.
74
(2012) derived the optical constants in the 400 - 4000 cm-1 range at three different
75
pressures and with a given gas mixture (10% of CH4).
76
Here we report new optical data, from the far-IR ranges down to 100 cm-1, a spectra
77
region that was not studied since the Khare et al. (1984) work, to the mid-IR up to
78
4000 cm-1. We performed experiments with four different initial gas mixtures. The
79
tholins spectra are compared with those of Titan's aerosols recently acquired by the
80
CIRS and VIMS instruments
81
2 Experimental setup and protocol
82
2.1 PAMPRE experiment
83
Tholins were produced with the PAMPRE device (Szopa et al. 2006). PAMPRE is a
84
low pressure radiofrequency capacitively coupled cold plasma (RF CCP) at
85
13.56 MHz. RF CCP discharges are well-known for producing thin films on substrates
86
and solid particles in the volume of reactive gas mixtures (Bouchoule 1999). The
87
plasma discharge is generated from a gaseous mixture between a polarized
88
electrode and a cylindrical grid grounded electrode confining the plasma (Alcouffe et
89
al. 2010). The gas mixture is continuously injected through the meshed polarized
90
electrode and pumped through a rotary valve vacuum pump. This discharge design
91
carries out a uniform gas flow in the confining box where tholins are produced. The
92
reactive gas mixture is at room temperature and can be adjusted with gas flow
93
controller in order to introduce from 0% to 10% of CH4 in N2. Some chemical and
94
physical properties of tholins produced in PAMPRE have already been investigated
95
(Carrasco et al. 2009; Hadamcik et al. 2009; Pernot et al. 2010) as well as the gas
Page 4
96
phase chemistry leading to tholins formation (Sciamma-O'Brien et al. 2010; Gautier
97
et al. 2011)
98
2.2 Sample production
99
The production was performed with a continuous 55 standard cubic centimeter
100
(sccm) gas flow rate. In this work, samples were prepared with various N 2:CH4 gas
101
mixtures including: 1%, 2%, 5% and 10% of methane. For the production of tholin,
102
the power injected in the reactor was 30 W and the gas pressure 100 Pa. Substrates
103
for thin film deposition were placed on the grounded electrode.
104
Two types of substrates were used for tholin thin film deposition, MirrIRTM and silicon
105
wafers.
106
Tholin films on MirrIRTM substrates (Low-e microscope slides from Kevley
107
Technologies, dim. 25mm x 75mm x 2mm) were used for acquiring data in the mid-IR
108
frequency region, from 700 cm-1 up to 4000 cm-1 (14.2 µm to 2.5 µm).
109
For the far-IR analyses, the thin films were deposited onto circular silicon wafers
110
(1 cm diameter, 0.5 mm thickness). Thin films on silicon substrates allow for
111
obtaining the spectra from 700 cm-1 down to 100 cm-1 (14.2 µm to 100 µm).
112
Deposition of thin films on MirrIRTM substrates were achieved after 2 hour long
113
experiments. Since the absorption of tholin is weak in the far-IR, productions of thin
114
film on silicon were three hours long in order to increase the thickness of the sample.
115
Three Si substrates were used in order to ensure film properties in the far-IR. For the
116
mid-IR, only one MirrIRTM substrate was used, but as the substrate is large enough,
117
three different areas were studied.
Page 5
118
2.3 Measurement of film thickness
119
The thin film thicknesses were determined by spectroscopic ellipsometry. This
120
technique is based on the measurement of the change of the light polarization upon
121
light reflection on a sample (Fujiwara 2007). The reflected light from the sample
122
surface is elliptically polarized.
123
coefficients rp and rs, respectively for the light polarized perpendicularly and parallel
124
to the incident plane is related to and parameters defined as:
, the ratio between the complex reflection
125
()
126
The ellipsometry parameters, and, are related to the complex refractive indices of
127
the substrate and of the different layer and to the thickness of the different sample
128
layers. A model taking into account the optical properties and thickness of the
129
different layers must established in order to deduce both the optical constants and
130
the layer thickness. The method is described in details in Sciamma-O’Brien et al.
131
(2012) and Mahjoub et al. (2012).
132
We used an M-2000V spectroscopic ellipsometer from J.A. Woollam Co. The M-
133
2000V is a rotating compensator ellipsometer with a CCD detector that measures all
134
wavelengths simultaneously across the spectral range 370 nm to 1000 nm.
135
The layer modeling was performed using the Complete-EASETM software (Complete
136
EASETM Data Analysis Manual By J.A Woollam Co Inc. June 15, 2008).
137
thicknesses measured for the different studied samples are presented in Table 1.
138
Uncertainties on each sample are due to film non uniformities calculated by the
139
Complete-EASETM software.
Page 6
(1)
The
140
2.4 Far- and Mid-Infrared spectroscopy
141
Spectra were recorded at the SMIS (Spectroscopy and Microscopy in the Infrared
142
using Synchrotron) beamline of SOLEIL synchrotron radiation facility in France
143
(Dumas et al. 2006). A NicPlan microscope was used coupled to a Nicolet Magna
144
System 560 Fourier Transform Infrared (FTIR) spectrometer. The IR sources utilized
145
for the present work were the synchrotron radiation for mid-IR and the internal Globar
146
source for the far-IR. The detectors were either the Mercury-Cadmnium-Telluride
147
(MCT) detector of the microscope (mid-IR), or a silicon doped bolometer from
148
Infrared Laboratories, cooled down to 4.2 K with liquid helium (far-IR). Analyses were
149
performed in transmission mode.
150
To ensure the repeatability of the measurements in the far-IR, thin film analyses were
151
performed on the three samples, three times each, i.e. 9 sample spectra were taken
152
in the far-IR range. We also collected three points of reference on two tholin-free
153
silicon wafers to be used as the substrate reference spectra.
154
In the mid-IR, 6 spectra were collected on the thin film deposited on MirrIRTM
155
substrates and blanks were performed on a tholin free substrate.
156
Spectra were recorded at a spectral resolution of 4 cm-1 after co-adding 512 scans at
157
a Michelson mirror velocity of 1.26 cm.s-1.
158
3 Results
159
All the spectra presented hereafter correspond to average spectra. Error bars are the
160
standard deviations of measurements performed on samples produced in the same
161
conditions.
Page 7
162
3.1 Thickness calibration and linear absorption coefficient determination
163
Figure 1 presents an absorbance uncalibrated spectrum of tholin (produced with 5%
164
of methane) from 100 cm-1 to 3500 cm-1. The spectrum is made of two parts:
165
1) below 700 cm-1 corresponding to the measurement of tholin film deposited onto
166
silicon,
167
2) above 700 cm-1 corresponding to the measurement of film deposited on MirrIRTM
168
substrates.
169
The non continuity of the spectrum at 700 cm-1 is due to the difference of thickness of
170
the two samples.
171 172
Figure 1
173
In order to consider the effect of the film thickness, the linear absorption coefficient, ,
174
should be used. The calculation of this coefficient requires the sample thickness, d.
Page 8
175
The reflected signal was measured and found negligible compared to the transmitted
176
contribution. The linear absorption coefficient, (cm-1), can thus be defined using the
177
Beer-Lambert law: (
178
)
(2)
179
where d is the sample thickness, It is the intensity of the transmitted signal, I0 the
180
incident intensity.
181
Note that the absorption coefficient is a function of the absorption cross-section, σ,
182
and the density of absorbent molecules, [n], not measurable here: (3)
183 184
Absorbance, A, being defined as:
185
(
186
The average thickness of thin films on silicon wafers was measured to be
187
1300 ±45 nm.
188
The multilayer structure of MirrIRTM substrates does not allow ellipsometric
189
measurements. The thickness of the film was thus estimated by fitting both parts of
190
the spectrum depicted in Fig.1 in order to get continuity of ε on the whole spectral
191
range. Using this method, the estimated thickness of the tholin films on MirrIRTM
192
substrates was 580 ± 50 nm. In order to validate this thickness estimation, tholin films
193
have been deposited on CaF2 substrates of the same thickness that MirrIRTM
194
substrates and during the same duration. Measurement on the film deposited on pure
195
CaF2 gives a thickness of 550 ± 25 nm. This value is in agreement with the thickness
196
estimated by fitting the spectra. We thus consider that the thicknesses of films on
197
pure CaF2 provides a good estimation of the thickness of the film on MirrIR TM
198
substrates, and we used this method to infer the thickness of other films deposited on
199
MirrIRTM as shown in Table 1.
Page 9
)
(4)
Sample
Injected % CH4
Deposit duration
Thickness (nm) 1250 ± 40
Si n°1 3h
Si n°2
1340 ± 40
5% Si n°3
1300 ± 40
CaF2
550 ± 25
CaF2
1%
420 ± 10 2h
CaF2
2%
490 ± 15
CaF2
10%
580 ± 25
200 201
Table 1: Experimental conditions and thickness of the different studied samples
202
The difference of thickness between films on silicon and on CaF2 comes from both
203
their respective production times and the dielectric properties of the substrates, as
204
discussed in Mahjoub et al. (2012). Note that since both Si and MirrIRTM substrates
205
are reflective mirrors, the optical path length during the absorption is equal to two
206
times the film thickness.
207
3.2 Band assignment
208
In the following, absorption spectra are measured from tholins produced with 5% CH 4
209
in N2 as the gas mixture. It has been shown that in the PAMPRE experiment, 5% of
210
methane injected corresponds to approximately 2% in steady state conditions
211
(Sciamma-O'Brien et al. 2010). The wavenumber dependence of the linear
212
absorption coefficient is shown in Fig. 2.
Page 10
213 214
Figure 2
215
The red curve corresponds to the average spectrum, and the pink envelope includes
216
both the film thickness uncertainty and the statistical 1 σ standard deviation errors
217
from the different measurements.
218
The larger uncertainty observed at shorter wavenumbers (around 700 cm-1) is due to
219
both the limitations of the MirrIRTM substrates and the detection efficiency of the MCT
220
detector.
221
Silicon has a strong absorption band in the 590 cm-1 - 620 cm-1 range which prevents
222
the determination of the linear absorption coefficient in this range when the spectra is
223
divided by blank spectra. The spectrum is therefore interpolated between 590 and
224
620 cm-1.
225
The spectrum shown in Fig. 2 depicts common tholin absorption features in the mid-
226
IR, described for example in Coll et al. (1999), Imanaka et al. (2004), Quirico et al.
227
(2008) or Imanaka et al. (2012). Broad and intense bands at 3200 cm-1 (3.13 µm)
228
and 3330 cm-1 (3.00 µm) are due to primary -NH and secondary amines -NH2. The
229
2880 cm-1 (3.47 µm) band is attributed to –CH3 symetric stretching. The 2930 cm-1 Page 11
230
(3.41 µm) and the 2960 cm-1 (3.38 µm) bands are attributed to -CH2 asymmetric
231
stretching and –CH3 asymmetric stretching.
232
The three bands observed at 2140 cm-1 (4.67 µm), 2175 cm-1 (4.60 µm) and
233
2250 cm-1 (4.44 µm) can be attributed to nitriles -C≡N, isocyanides -N≡C or
234
carbodiimide –N=C=N- stretching modes.
235
The 1560 cm-1 and 1630 cm-1 bands cannot be assigned unambiguously. They can
236
correspond to several possible functional groups, such as aromatic or aliphatic –NH2,
237
C=N double bonds, C=C double bonds, aromatics or heteroaromatics (bearing
238
nitrogen). This precludes a precise assignment of these bands.
239
At lower wavenumbers, we observe two bands at 1380 cm-1 (7.2 µm) and 1450 cm-1
240
(6.9 µm). As suggested in Vinatier et al. 2012, the 1450 cm -1 band is probably a
241
contribution of asymmetric C-H bending of CH3 and scissor in plane bending of C-H
242
in CH2, and the 1380 cm-1 band the symmetric bending of C-H in CH3.
243
Several absorption features are also visible in the far-IR frequency range, as shown
244
in Fig. 7. First a broad feature extends from 400 to 600 cm-1, which might be due to
245
amorphous carbon nitride (Rodil et al. 2001; Quirico et al. 2008). But as said in
246
Quirico et al. (2008) "it cannot be simply interpreted in terms of wagging or skeletal
247
modes of simple molecules, but rather as lattice vibrations within the covalent solids".
248
Thus, this band is left unassigned.
249
The band observed at 690 cm-1 (14.49 µm), could possibly be attributed to C-H
250
bending out of plane in carbon-carbon double bonds (usually 665 to 730 cm -1) or to
251
ring out of plane deformation vibration in aromatics (usually 670-720 cm-1).
252
Absorption due to aromatic rings deformations are known to be more intense than
Page 12
253
absorption due to C-H out of plane bending mode. Further analysis on the aliphatic
254
and aromatic content in tholin could confirm a preferential assignement of theses
255
bands to aromatics ring deformations.
256
Bands are also visible at 515 cm-1 (19.42 µm), 324 cm-1 (30.86 µm), 255 cm-1
257
(39.22 µm) and 170 cm-1 (58.82 µm). However, their assignements are not obvious
258
because of a lack of litterature data for these weak bands.
259
3.3 Impact of the percentage of methane on tholin spectra
260
We studied the influence of the methane concentration of the reactive gas mixture in
261
the PAMPRE experiment on tholin spectra. Figure 3 presents ε as a function of the
262
wavenumber for tholins produced with 1%, 2%, 5% and 10% of methane in the gas
263
mixture injected. Sciamma-O'Brien et al. (2010) have shown that these injected ratios
264
respectively correspond to ~0.2%, ~0.5%, ~2%, ~5% of methane in the steady state
265
conditions of the reactive plasma. 3500 3000
-1 (cm )
2500
1% 2% 5% 10%
2000 1500 1000 500 0
266
1500
2000
2500
3000
3500
(cm-1)
267
Figure 3
268
When comparing the spectra presented in Fig. 3, the most striking point is a growth in
269
intensity of the amine bands at 3200 cm-1 and 3330 cm-1 when the concentration of
270
methane decreases in the initial gas mixture. These bands are obvious when 1% of Page 13
271
methane is injected, whereas they are quite low in the spectrum when 10% of
272
methane is used. This is consistent with the elemental analysis presented in
273
Sciamma-O'Brien et al. (2010), showing an increase of the nitrogen content in tholin
274
for the lower methane concentrations. Nevertheless, in spite of the fact that amines
275
were recently confirmed to be functional groups present in our tholin with solid
276
nuclear magnetic resonance (Derenne et al. 2012), their predominance was not
277
expected considering the unsaturation supported by nitrogen reported in previous
278
studies performed both on the gas phase (Coll P et al. 1999; Gautier et al. 2011) and
279
on tholin (Somogyi et al. 2005; Carrasco et al. 2009; Pernot et al. 2010).
280
This increase of the amine bands is correlated to the decrease of the intensity of the
281
2900 cm-1 pattern, attributed to aliphatic methyl. For the lowest methane
282
concentration (1%) the 2900 cm-1 pattern is very weak, whereas for the highest
283
methane concentration studied (10%) it is one of the strongest absorption bands in
284
the spectra. The increase in the strength of these bands suggests that the amount of
285
aliphatic methyl increases in tholin when they are produced with a higher methane
286
percentage.
287
Another variation observed on the spectra presented in Fig. 3 is the progressive
288
vanishing of the band at 2140 cm-1 compared to the 2175 cm-1 band. As said in
289
section 3.2, these bands and the one at 2250 cm -1 are due to stretching modes of
290
several kinds of carbon-nitrogen triple bonds. Their exact attribution is still unclear,
291
but the 2140 cm-1 band decreases with respect to the 2175 cm-1 band when the
292
methane percentage increases. This is consistent with the increase of the saturated
293
aliphatic carbons, also observed on the 2900 cm -1 pattern. The 2140 cm-1 and the
294
2175 cm-1 bands could respectively be attributed to unsaturated and saturated
295
nitriles. Page 14
296
Another visible effect visible in Fig. 3 is the plummeting of the band at 1560 cm-1
297
respective to the 1630 cm-1 band. Indeed for low methane concentrations, the
298
1560 cm-1 band is as intense as the one at 1630 cm-1, whereas it is almost
299
undetectable in the spectra of tholin produced with 10% methane. This band is not
300
clearly identified (cf. part 3.2), but may involve nitrogen, such as cyclic or aliphatic
301
amines (c-NHx or n-NHx), C=N or heteroaromatics, since this band intensity
302
decreases with the decrease of the N2 concentration in the gas mixture introduced in
303
the experiment.
304
Finally, the relative increase of the absorption bands at 1450 cm-1 and 1380 cm-1 with
305
increasing CH4 percentage is also noticeable in Fig.3, as is the growth of broad low
306
intensity bands centered at 1230 cm-1 and 1140 cm-1. These bands are possibly due
307
to carbon-carbon or carbon-hydrogen bonds.
308
All these band variations enforce the notion that the amount of molecules with
309
hydrocarbon skeleton in tholin increases with the methane concentration in the
310
reactive gas mixture, whereas tholins produced with low methane concentrations
311
seem to be based on amine rich polymers.
312
4 Comparison with Cassini CIRS and VIMS observations
313
Figure 4 shows tholin absorption spectrum in the far- and mid-IR spectral range
314
(black curve). Also plotted in this figure is the tholin spectrum of Khare et al. (1984),
315
in blue. The CIRS and VIMS spectra of Titan's aerosols in the mid- and far-IR are
316
plotted in red (Anderson and Samuelson 2011; Vinatier et al. 2012; Bellucci et al.
317
2009; Kim et al. 2011).
Page 15
318
In order to compare the tholin absorption spectra obtained with different conditions, a
319
normalization coefficient was applied to the spectra. The Khare et al. (1984)
320
spectrum was normalized to the maximum intensity of the entire tholin spectrum (i.e.
321
the 1560 cm-1 band). VIMS and CIRS spectra were normalized to the maximum of
322
our spectrum (i.e. the CIRS spectrum was normalized at the 1450 cm -1 band, and the
323
VIMS spectrum at the 2930 cm-1 band).
324
Figure 4
325 326
327
The analysis of Fig. 4 is discussed in the mid- and far-IR spectral regimes: a) Mid-Infrared
328
The 2900 cm-1 (3.4µm) pattern, emphasized in Fig. 5, is in general agreement with
329
the feature observed by VIMS (Bellucci et al. 2009; Rannou et al. 2010). This pattern
330
is not present in the spectrum presented by Khare et al. (1984), produced with a
331
10% initial methane concentration. But in our set up, the intensities of these bands
332
are maximum with 10% methane introduced in the experiment (cf. Fig. 3), and almost
333
as intense as the feature at 1500 cm-1.
Page 16
1 Tholins 5% (this work) VIMS spectrum (adapted from Kim et al. 2011) Data from Khare et al. 1984 (adapted from Quirico et al. 2008)
Absorbance (a.u.)
0.8
0.6
0.4
0.2
0 2700
2800
2900
3000
3100
3200
3300
3400
3500
334
(cm-1)
335
Figure 5
336
A possible explanation is as follows. Despite the same initial methane concentration
337
between our work and Khare et al. (1984), the effective consumption of methane
338
might be different. Indeed, the residual methane concentration in a steady state in
339
the plasma is quite different than from the initial one as shown in Sciamma-O'Brien et
340
al. (2010). And yet, as shown in section 3.3, the intensity of the 2900 cm-1 pattern is
341
highly correlated with the methane concentration. However, it must be underlined that
342
other plasma parameters can influence the chemical composition of tholins in
343
addition to the injected amount of CH4, such as the working pressure as shown in
344
Imanaka et al. (2012).
345
Figure 5 compares the shape of the 2900 cm-1 feature between tholin spectrum and
346
the derived aerosol spectrum from VIMS observations (extracted from Kim et al.
347
2011).
348
Among the five main bands observed in the VIMS data (2885 cm-1, 2930 cm-1,
349
2965cm-1, 3000 cm-1 and 3030 cm-1) two are not found in our tholin spectrum: the
350
3000 cm-1 band, attributed to CH3CN ice by Kim et al. (2011), and the weak 3030 cm-
351
1
band. The spectral position of the 2930 cm-1 band (attributed to n-CH2 assymetric
Page 17
352
stretching) in our data corresponds exactly to one of the absorption bands detected
353
with VIMS in Titan's atmosphere. For the 2885 cm -1 and 2965 cm-1 bands, we
354
assume they are actually the same as the bands detected in tholin respectively at
355
2880 cm-1 and 2960 cm-1 and attributed to CH3 stretching modes. Indeed infrared
356
spectra of solids can present slight frequency shifts due to the chemical environment
357
for the functional group generating the absorption band (see Fig.6) ("Infrared
358
Spectra" by NIST Mass Spec Data Center, S.E. Stein, director). Particularly, a shift of
359
a few wavenumbers toward higher wavenumbers may arise from the influence of
360
aromatics on the methyl vibration. In the same way, the chemical environment of the
361
methyl group responsible for this pattern has a major impact on the presence of the
362
bands shown in Fig. 6. In this figure the impact of the insertion of an heteroatom in
363
the cycle of toluene on the 2900 cm-1 pattern is observed.
364
365 366
Figure 6
367
The red line represents the spectrum of a methyl group bonded to a pure carbon
368
aromatic ring (toluene), whereas green and blue lines are spectra with the methyl
Page 18
369
group bonded to an heteroaromatic ring (here pyridine) in two different positions. It is
370
clear from Fig. 6 that the composition of the chemical environment of the methyl
371
group (and even the position of the nitrogen in the cycle comparing blue and green
372
spectra) has a major influence on the shape of the spectra, and it can induce both
373
band intensity changing and band position shifting.
374
The fact that our tholin bands do not perfectly fit the bands seen with VIMS still does
375
not imply that these bands in Titan's atmosphere cannot come from aerosols
376
absorption. But this might suggest that the chemical environment of the methyl
377
functional groups in Titan's aerosols are, even slightly, different from the obsevred
378
features in our tholin.
379
In Kim et al. (2011), the authors proposed that the 2900 cm -1 absorption pattern
380
observed by VIMS could be due to a mixing of different ices (C2H6, CH4, CH3CN,
381
C5H12, C6H12). As shown in this work, tholins present an intense absorption at this
382
wavenumber. We thus suggest that aerosol absorption can explain the VIMS
383
observations better than ices, and in agreement with Rannou et al. (2010) and
384
Bellucci et al. (2009). Moreover, the condensation of these ices seems unlikely
385
considering the SVP curves of these compounds (Linstrom P.J. and W.G. Mallard,
386
Eds., NIST Chemistry WebBook, NIST Standard Reference Database Number 69) at
387
the altitude of the observation since Fulchignoni et al. (2005) measured a
388
temperature of about 180 K at 300 km, with a pressure of about 10 Pa.
389
b) Far Infrared
390
A close up of Fig. 4 in the 100 cm-1 – 1500 cm-1 wavenumber range (100 µm- 6.6 µm
391
wavelength range) is given in Fig. 7. In this range, our tholin spectrum is compared
392
with the absorption spectra attributed to Titan's aerosols derived from CIRS
Page 19
393
observations (Vinatier et al. 2010; Anderson and Samuelson 2011; Vinatier et al.
394
2012). 1
Absorbance (a.u.)
0.8
This work CIRS spectrum (adapted from Anderson & Samuelson 2011) CIRS spectrum (adapted from Vinatier et al. 2011) Data from Khare et al. 1984 (adapted from Quirico et al. 2008)
1450
1380
0.6
0.4
0.2
0 0
170
515
255
690
324
500
1000
1500
-1
395
(cm )
396
Figure 7
397
In this frequency range, two absorption bands are clearly present in both the CIRS
398
and the tholin spectrum, the band at 1450 cm -1 (6.9 µm), slightly left-shifted in tholin
399
spectrum compared to observations, and the band at 1380 cm-1 (7.2 µm) which is
400
exactly co-located in both spectra. A third band is visible arround 1320 cm -1 (7.6 µm)
401
in the CIRS spectrum but not in our tholin spectrum (or really weak) and may be due
402
to an overtone of the C-H bending in ≡C-H group signature observed at 630 cm -1
403
from CIRS. This absence could mean that tholins produced with PAMPRE contain
404
less alkynes than the Titan aerosols.
405
Furthermore, as illustrated in Fig. 7, the broad band visible around 600 cm -1 in our
406
spectrum is also present in other laboratory tholins spectra (e.g. Imanaka et al. 2004;
407
Quirico et al. 2008; Imanaka et al. 2012), but is not present in Cassini-CIRS spectra.
408
As discussed in part 3.2, this feature is speculated to arise from the lattice vibration in
409
solids (Rodil et al. 2001). This would mean that this band could only originate from
410
large scale solid material such as laboratory tholin film, whereas this band could not
Page 20
411
be observed on suspended single particles (free and not assembled), such as
412
aerosols in Titan's atmosphere. We suggest that obtaining the spectra of tholins in
413
suspension would be a good way to validate this hypothesis.
414
Finally, spectrum of our tholin exhibit a few features in the far-IR below 700 cm-1,
415
(14.3 µm). Some of these features, especially the bands at 325 cm -1 (30.77 µm) and
416
513 cm-1 (19.49 µm) are also visible in CIRS spectra (Anderson and Samuelson
417
2011). Up to now these bands are attributed to noise in CIRS spectrum.
418
The fact that these two bands are also present in our tholin spectrum could tend to
419
confirm that these weak absorption bands visible in CIRS spectrum were not noise
420
but were due to aerosols. Further investigation on this spectral range of observations
421
of Titan's atmosphere could confirm such hypothesis.
422
5 Conclusion
423
In this work we present a study of the infrared absorption properties of tholins
424
produced with the PAMPRE plasma device for several gas mixtures.
425
This work provides the wavenumber dependence of the linear absorption coefficient
426
ε, from the far-IR (100 cm-1) to the mid-IR (4000 cm-1) range.
427
We also show the way in which the percentage of methane in the experiment directly
428
impacts the spectrum of tholins in the mid-IR. This influence is maximum on the
429
amine bands at 3330 cm-1 and 3200 cm-1, and also visible on the 2900 cm-1 pattern
430
due to aliphatic methyl, and in the 2200 cm-1 - 2400 cm-1 and 1300 cm-1 - 1650 cm-1
431
ranges.
Page 21
432
Furthermore, some comparisons are performed between different tholins material
433
and data derived from Cassini CIRS and VIMS observations. The 2900 cm -1 pattern
434
is in relatively good agreement with the Cassini-VIMS spectra obtained in Titan
435
atmosphere and possibly attributed to aerosols.
436
The high intensity of these bands in PAMPRE tholin spectra supports the hypothesis
437
that aerosols are the main contributors to the 2900 cm-1 (3.4 µm) absorption pattern
438
in Titan's atmosphere.
439
Comparing tholin spectrum with spectra obtained with Cassini-CIRS in the far-IR
440
frequency range, we see that the two spectra are in good agreement with many
441
absorption bands, especially at 1450 cm-1, 1380 cm-1, 515 cm-1 and 325 cm-1,
442
reinforcing the detection of these bands from CIRS.
443
Furthermore, the mid- and far-IR data provided in this work could also be used for
444
comparison to other astronomical environment where tholin material is supposed to
445
be relevant, such as cometary ices, Triton and TNO surfaces or diffuse interstellar
446
medium (Gradie and Veverka 1980; Pendleton and Allamandola 2002; Dotto et al.
447
2003).
448
6 Acknowledgments
449
This work was financially supported by CNRS (PNP, ANR-09-JCJC-0038 contract).
450
All the PAMPRE team gratefully thanks the SMIS beam line team for their help and
451
contribution and SOLEIL synchrotron facility for accepting and supporting the
452
project.n° 20100103.
Page 22
453
7 Figure Caption
454
Figure 1: Reconstructed experimental spectrum of tholin produced in PAMPRE with
455
5% of methane before calibration of the sample thickness.
456
Figure 2:
457
PAMPRE with 5% of methane. Average spectrum is given by the red curve. Pink
458
envelopes indicate this standard deviation (1 σ) of the spectra, representing the
459
variation of the spectrum from one measure to another on the same sample. Possible
460
attributions are given for major bands of the spectrum. n-X and c-X mean that the
461
functional group X is attached to respectively an aliphatic or an aromatic skeleton.
462
Figure 3: Absorption spectra of tholins in the mid-Infrared with different initial CH4
463
concentrations in the gas mixture (1% in blue, 2% in green, 5% in red and 10% in
464
cyan).
465
Figure 4: Absorbance spectrum from far-IR to mid-IR of PAMPRE tholin made with
466
5% of CH4 (black curve) compared to Quirico et al. 2008 tholin spectra (extracted
467
from
468
Anderson and Samuelson 2011 and Vinatier et al. 2011) and Cassini-VIMS spectra
469
(red dots, from Kim et al. 2011).
470
Figure 5: Close up of Fig. 3 on the 2900 cm -1 pattern. Black line represents our tholin
471
5% spectrum. Red dots are the spectrum derived from observations with Cassini-
472
VIMS attributed to Titan's aerosols (extracted from Kim et al. 2011). The blue line
473
represent tholin spectra reconstituted from Khare et al. 1984 data (from Quirico et al.
474
2008). Black vertical dashed lines represent bands seen by VIMS, respectively from
475
left to right: 2885 cm-1, 2930cm-1, 2965 cm-1, 3000 cm-1, 3030 cm-1.
Evolution of from the far- to the mid-Infrared for tholin produced in
Khare et al. 1984 tholin, blue), Cassini-CIRS observations (red line, from
Page 23
476
Figure 6: Infrared spectra of toluene (red), 2-methylpyridine (blue) and 3-
477
methylpyridine (green) extracted from the NIST database.
478
Figure 7: Close up of Fig.3 in the far-IR range. Black line represents our tholin 5%
479
spectrum. Titan's aerosols spectra derived from observations with Cassini CIRS
480
(Anderson and Samuelson 2011; Vinatier et al. 2011) are plotted in red. The blue line
481
represents tholin spectra reconstituted from Khare et al. 1984 data (from Quirico et
482
al. 2008).
Page 24
483
References
484
Alcouffe, G., M. Cavarroc, G. Cernogora, F. Ouni, A. Jolly, L. Boufendi and C. Szopa
485
(2010), "Capacitively coupled plasma used to simulate Titan's atmospheric
486
chemistry." Plasma Sources Science and Technology 19(1): 015008.
487 488
Anderson, C. M. and R. E. Samuelson (2011), "Titan's aerosol and stratospheric ice
489
opacities between 18 and 500 µm: Vertical and spectral characteristics from
490
Cassini CIRS." Icarus 212(2): 762-778.
491 492
Bellucci, A., B. Sicardy, P. Drossart, P. Rannou, P. D. Nicholson, M. Hedman, K. H.
493
Baines and B. Burrati (2009), "Titan solar occultation observed by
494
Cassini/VIMS: Gas absorption and constraints on aerosol composition." Icar
495
201(1): 198-216.
496 497 498
Bouchoule, A. (1999), Dusty Plasmas: Physics, Chemistry, and Technological Impact in Plasma Processing, Ed. Wiley, 418p.
499 500
Cable, M. L., S. M. Hörst, R. Hodyss, P. M. Beauchamp, M. A. Smith and P. A. Willis
501
(2011), "Titan Tholins: Simulating Titan Organic Chemistry in the Cassini-
502
Huygens Era." Chemical Reviews, Article ASAP.
503 504
Carrasco, N., I. Schmitz-Afonso, J.-Y. Bonnet, E. Quirico, R. Thissen, O. Dutuit, A.
505
Bagag, O. Laprévote, A. Buch, A. Giulani, G. Adandé, F. Ouni, E. Hadamcik,
506
C. Szopa and G. Cernogora (2009), "Characterization of Titan's Tholins:
Page 25
507
Solubility, Morphology and Molecular Structure Revisited." Journal of Physical
508
Chemistry A 113: 11195-11203.
509 510
Coll, P., D. Coscia, D. Smith, M.-C. Gazeau, S. I. Ramirez, G. Cernogora, G. Israel,
511
and F. Raulin (1999), "Experimental laboratory simulation of Titan's
512
atmosphere: aerosols and gas phase." Planet. Space Sci. 47(10-11): 1331.
513 514
Derenne S., C. Coelho, C. Anquetil, C. Szopa, A.S. Rahman, P.F. McMillan, F. Corà,
515
C.J. Pickard, E. Quirico, C. Bonhomme (2012), "New insights into the structure
516
and chemistry of Titan's tholins via
517
resonance spectroscopy", Icarus (in press).
13
C and
15
N solid state nuclear magnetic
518 519
Dotto, E., M. A. Barucci and C. de Bergh (2003), "Surface composition of TNOs and
520
Centaurs: visible and near-infrared spectroscopy." Comptes Rendus Physique
521
4(7): 775-782.
522 523
Dumas, P., F. Polack, B. Lagarde, O. Chubar, J. L. Giorgetta and S. Lefrançois
524
(2006), "Synchrotron infrared microscopy at the French Synchrotron Facility
525
SOLEIL." Infrared Physics & Technology 49(1-2): 152-160.
526 527 528
Fujiwara, H. (2007), Spectroscopic ellipsometry: principles and applications, Ed. Marusen, Tokyo, 388p.
529 530
Fulchignoni, M., F. Ferri, F. Angrilli, A. J. Ball, A. Bar-Nun, M. A. Barucci, C. Bettanini,
531
G. Bianchini, W. Borucki, G. Colombatti, M. Coradini, A. Coustenis, S. Debei,
Page 26
532
P. Falkner, G. Fanti, E. Flamini, V. Gaborit, R. Grard, M. Hamelin, A. M. Harri,
533
B. Hathi, I. Jernej, M. R. Leese, A. Lehto, P. F. Lion Stoppato, J. J. Lopez-
534
Moreno, T. Makinen, J. A. M. McDonnell, C. P. McKay, G. Molina-Cuberos, F.
535
M. Neubauer, V. Pirronello, R. Rodrigo, B. Saggin, K. Schwingenschuh, A.
536
Seiff, F. Simoes, H. Svedhem, T. Tokano, M. C. Towner, R. Trautner, P.
537
Withers and J. C. Zarnecki (2005), "In situ measurements of the physical
538
characteristics of Titan's environment." Nature 438(7069): 785-791.
539 540 541
Gautier, T., N. Carrasco, A. Buch, C. Szopa, E. Sciamma-O'Brien and G. Cernogora (2011), "Nitrile gas chemistry in Titan atmosphere." Icarus 213(2): 625-635.
542 543 544
Gradie, J. and J. Veverka (1980), "The composition of the Trojan asteroids." Nature 283(5750): 840-842.
545 546
Hadamcik, E., J.-B. Renard, G. Alcouffe, A. C. Levasseur-Regourd and C. Szopa
547
(2009), "Laboratory light-scattering measurements with Titan's aerosols
548
analogues produced by a dusty plasma." Planetary and Space Science
549
57(13): 1631-1641.
550 551
Imanaka, H., D. P. Cruikshank, B. N. Khare and C. P. McKay (2012), "Optical
552
constants of Titan tholins at mid-infrared wavelengths (2.5–25μm) and the
553
possible chemical nature of Titan’s haze particles." Icarus 218(1): 247-261.
554 555
Imanaka, H., B. N. Khare, J. E. Elsila, E. L. O. Bakes, C. P. McKay, D. P. Cruikshank,
556
S. Sugita, T. Matsui and R. N. Zare (2004), "Laboratory experiments of Titan
Page 27
557
tholin formed in cold plasma at various pressures: implications for nitrogen-
558
containing polycyclic aromatic compounds in Titan haze." Icarus 168: 344-
559
366.
560 561
Khare, B. N., C. Sagan, E. T. Arakawa, F. Suits, T. A. Callcott and M. W. Williams
562
(1984), "Optical constants of organic tholins produced in a simulated Titanian
563
atmosphere: From soft x-ray to microwave frequencies." Icarus 60 (1): 127-
564
137.
565 566
Kim, S. J., A. Jung, C. K. Sim, R. Courtin, A. Bellucci, B. Sicardy, I. O. Song and Y.
567
C. Minh (2011), "Retrieval and tentative indentification of the 3μm spectral
568
feature in Titan's haze." Planetary and Space Science 59(8): 699-704.
569 570 571
Lavvas, P., C. A. Griffith and R. V. Yelle (2011), "Condensation in Titan’s atmosphere at the Huygens landing site." Icarus 215(2): 732:750.
572 573
Linstrom P.J. and W.G. Mallard, Eds., NIST Chemistry WebBook, NIST Standard
574
Reference Database Number 69, National Institute of Standards and
575
Technology, Gaithersburg MD, 20899, http://webbook.nist.gov
576 577
Mahjoub, A., P. R. Dahoo, T. Gautier, N. Carrasco, C. Szopa and G. Cernogora
578
(2012), "Influence of methane concentration on the optical indices of Titan's
579
aerosols analogues." Icarus (accepted).
580
Page 28
581 582
McKay, C. P., J. B. Pollack and R. Courtin (1991), "The greenhouse and antigreenhouse effects on Titan." Science 253(5024): 1118-1121.
583 584
Pendleton, Y. J. and L. J. Allamandola (2002), "The Organic Refractory Material in
585
the Diffuse Interstellar Medium: Mid-Infrared Spectroscopic Constraints." The
586
Astrophysical Journal Supplement Series 138(1): 75.
587 588
Pernot, P., N. Carrasco, R. Thissen and I. Schmitz-Afonso (2010), "Tholinomics-
589
Chemical Analysis of Nitrogen Rich Polymers." Analytical Chemistry 82(4):
590
1371–1380.
591 592
Quirico, E., G. Montagnac, V. Lees, P. F. McMillan, C. Szopa, G. Cernogora, J.-N.
593
Rouzaud, P. Simon, J.-M. Bernard, P. Coll, N. Fray, R. D. Minard, F. Raulin, B.
594
Reynard and B. Schmitt (2008), "New experimental constraints on the
595
composition and structure of tholins." Icarus 198(1): 218-231.
596 597
Rannou, P., T. Cours, S. Le Mouélic, S. Rodriguez, C. Sotin, P. Drossart and R.
598
Brown (2010), "Titan haze distribution and optical properties retrieved from
599
recent observations." Icarus 208(2): 850-867.
600 601
Rodil, S. E., A. C. Ferrari, J. Robertson and W. I. Milne (2001), "Raman and infrared
602
modes of hydrogenated amorphous carbon nitride." Journal of Applied Physics
603
89(10): 5425-5430.
604
Page 29
605
Sciamma-O'Brien, E., N. Carrasco, C. Szopa, A. Buch and G. Cernogora (2010),
606
"Titan's atmosphere: an optimal gas mixture for aerosol production?" Icarus
607
209(2): 704-714.
608 609
Sciamma-O’Brien, E., P. R. Dahoo, E. Hadamcik, N. Carrasco, E. Quirico, C. Szopa
610
and G. Cernogora (2012), "Optical constants from 370 nm to 900 nm of Titan
611
tholins produced in a low pressure RF plasma discharge." Icarus 218 (1) 356-
612
363.
613 614
Sittler Jr, E. C., A. Ali, J. F. Cooper, R. E. Hartle, R. E. Johnson, A. J. Coates and D.
615
T. Young (2009), "Heavy ion formation in Titan's ionosphere: Magnetospheric
616
introduction of free oxygen and a source of Titan's aerosols?" Planetary and
617
Space Science 57(13): 1547-1557.
618 619
Somogyi, A., C.-H. Oh, M. A. Smith and J. I. Lunine (2005), "Organic Environments
620
on Saturn's Moon, Titan: Simulating Chemical Reactions and Analyzing
621
Products by FT-ICR and Ion-Trap Mass Spectrometry." Journal of the
622
American Society for Mass Spectrometry 16(6): 850-859.
623 624
Szopa, C., G. Cernogora, L. Boufendi, J. J. Correia and P. Coll (2006), "PAMPRE: A
625
dusty plasma experiment for Titan's tholins production and study." Planetary
626
and Space Science 54: 394-404.
627 628
Vinatier, S., B. Bézard, R. de Kok, C. M. Anderson, R. E. Samuelson, C. A. Nixon, A.
629
Mamoutkine, R. C. Carlson, D. E. Jennings, E. A. Guandique, G. L. Bjoraker,
Page 30
630
F. Michael Flasar and V. G. Kunde (2010), "Analysis of Cassini/CIRS limb
631
spectra of Titan acquired during the nominal mission II: Aerosol extinction
632
profiles in the 600-1420 cm-1 spectral range." Icarus 210(2): 852-866.
633 634
Vinatier, S., P. Rannou, C. M. Anderson, R. de Kok and R. E. Samuelson (2012),
635
"Optical constans of Titan's stratospheric aerosols in the 70-1500 cm-1 spectra
636
range constrained from Cassini/CIRS observations." Icarus 219(1): 5-12.
Page 31