Doppler limited rotational transitions of OH and SH radicals measured by continuous-wave terahertz photomixing

Journal of Molecular Structure 1006 (2011) 13–19

Contents lists available at SciVerse ScienceDirect

Journal of Molecular Structure journal homepage: www.elsevier.com/locate/molstruc

Doppler limited rotational transitions of OH and SH radicals measured by continuous-wave terahertz photomixing Sophie Eliet a,b, Marie-Aline Martin-Drumel c,d, Mickaël Guinet a,b, Francis Hindle a,b, Gaël Mouret a,b, Robin Bocquet a,b, Arnaud Cuisset a,b,⇑ a

Laboratoire de Physico-Chimie de l’Atmosphère, EA 4493, Université du Littoral – Côte d’Opale, F-59140 Dunkerque, France Université Lille Nord de France, F-59000, France Institut des Sciences Moléculaires d’Orsay, CNRS, Bât. 210, Université Paris–Sud, 91405 Orsay Cedex, France d Synchrotron SOLEIL, L’Orme des Merisiers, Saint-Aubin – BP48, 91192 Gif-sur-Yvette, France b c

a r t i c l e

i n f o

Article history: Available online 19 August 2011 Keywords: Rotational spectroscopy THz photomixing Doppler-limited THz spectroscopy Hydroxyl radical Mercapto radical Fine and hyperfine structure

a b s t r a c t A continuous-wave terahertz (CW-THz) source generated by photomixing has been employed to detect and quantify radicals produced in a cold plasma probing their spin-rotation transitions. Due to their dual interest for both atmospherists and astrophysicists, the hydroxyl OH and the mercapto SH radicals have been chosen. The photomixing technique which can access the largest range of THz frequencies of any known coherent source, allowed to resolve the Doppler-limited hyperfine transitions of OH in the 2.5 THz frequency region. Line profile analysis of the hyperfine components demonstrated that OH radicals have been detected in this region at a ppm level at a temperature close to 490 K. The hyperfine structure of SH has been resolved for the first time above 1 THz. Ten new frequency transitions have been measured in the 1.3–2.6 THz frequency range using the CW-THz synthesizer based on a frequency comb. With relative uncertainties better than 107, the CW-THz frequencies measured in this study are now competitive with those measured by other instruments such as frequency multiplication chains or FTFIR spectrometers and are now capable to improve the predictions of the complete high-resolution spectra of these radicals collected in the atmospheric and astrophysical spectroscopic databases. versioncorrigeeAC 2011-07-18 17:32 Ó 2011 Arnaud Cuisset. Ó 2011 Elsevier B.V. All rights reserved.

1. Introduction Molecules have been detected throughout the electromagnetic spectrum in the microwave, submillimeter-wave, infrared and UV–Vis regions. Approximately 20 years ago, technological opening of the terahertz (THz) region to laboratory, atmospheric as well as interstellar spectroscopy, allowed great scientific advances to be performed, providing particularly highly accurate spectral line THz frequencies of higher rotational transitions of medium heavy molecules (CH3OH, CH3Cl, OCS, etc.) or low J transitions of lighter molecules, notably the light hydrides molecules (OH, CH, NH, SH, etc.) [1]. The forthcoming Far-IR high resolution interferometric astrophysical observations (e.g. with the HIFI instrument on board of the HERSCHEL Space Observatory, or the ALMA telescope) will provide new reliable spectroscopic informations on chemical composition of the interstellar medium whose understanding will rely on highly accurate experimental data [2]. Although the most relevant hydride radicals were widely studied by high resolution ⇑ Corresponding author at: Laboratoire de Physico-Chimie de l’Atmosphère, EA 4493, Université du Littoral – Côte d’Opale, F-59140 Dunkerque, France. E-mail address: [email protected] (A. Cuisset). 0022-2860/$ - see front matter Ó 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.molstruc.2011.07.055

spectroscopy, detailed knowledges of line parameters (line frequencies and intensities, line-widths and line-broadening parameters) are relatively sparse and require the association of submillimeter data measured with pure electronic sources and far-infrared data measured with pure optical sources. Using an optoelectronic generation of THz radiation, we propose to demonstrate in this article that the photomixing technique is a powerful tool to provide very accurate spectroscopic data on hydride radicals on a large spectral range from the submillimeter to the FarIR regions. The hydroxyl radical (OH) and the mercapto radical (SH) have been chosen for this study. OH plays a key role in chemistry of the Earth’s atmosphere, combustion processes and the interstellar medium. This radical is directly involved in a catalytic cycle which is the main process for controlling ozone destruction in the upper stratosphere and mesosphere. Moreover OH is one of the major chemical oxidant in the lower atmosphere, reacting with volatile organic compounds and inorganic species trace gas cycles [3,4]. In 1963, OH became the first free radical to be identified by radioastronomy in the interstellar medium [5]. Since then, OH is known to be present in a variety of sources from its K-doubling transitions at cmwavelengths, transitions which are very difficult to interpret in

14

S. Eliet et al. / Journal of Molecular Structure 1006 (2011) 13–19

terms of molecular abundances. THz lines of OH are well-known diagnostics of interstellar molecular shocks and are also a valuable probe of oxygen-rich circumstellar envelopes. Chemistry of sulfur-containing species in space is also especially interesting because of their chemical activity and relatively high abundance. Among the 17 sulfur-containing molecules detected in the interstellar medium, only the SH radical has not been detected by pure rotational spectroscopy [6]. Only its cation SH+ has been recently detected in the submillimeter range [7]. The identification of SH was performed in 2000 by Yamamura et al. from the Dv = 1 rovibrational lines using FTIR spectroscopy [8]. Pure rotational transitions in the THz region are more suitable for searches in interstellar space due to larger intensities and smaller Doppler-limited line-widths. Nevertheless, the THz lines measured in laboratory [9,10] did not allow a detection of SH in space which may contribute to the understanding of sulfur chemistry, especially of the production and reaction mechanism of SHn [9]. The mercapto radical (SH) is also an important species in the atmospheric process. For instance, it is a reaction intermediate involved in numerous phenomena such as acid rains. In the troposphere, H2S is removed mostly through its reaction with the OH radical, which produces the SH radical. With a lifetime shorter than 1 s in the atmosphere, SH is mainly oxidized by O3 and NO2 leading eventually to the formation of SO2 [11]. Both for OH and SH, the most intense rotational transitions are observed in the 1.5–4.5 THz frequency region [12]. In Fig. 1, the low-lying rotational energy levels of OH and SH radicals involved in the most intense rotational transitions in the 2P3/2 component of the vibrational ground state are described in the Hund’s case [b]. The quantum number N results from the coupling between the electronic orbital angular momentum L and the rotational angular momentum R. The coupling of N with the electronic spin S produces the total angular momentum J describing the fine structure of the radicals with the 2P1/2 component lying 3.8 THz and 11.0 THz above the 2P3/2 component, respectively for OH and SH. However, deviations from this idealized case occur as a consequence of the interaction of the molecular rotation with the electronic motion. In particular, for SH and OH with a weak spin–orbit interaction, L and S are uncoupled by the electrostatic field of the nuclei that yields to the well-known K-doubling which splits the energy levels from several hundreds of MHz for SH to

several GHz for OH. The K-doubling splitting may be resolved by most of sources commonly used in THz high-resolution spectroscopy. Finally, a hyperfine rotational structure is also present in OH and SH due to the coupling of nuclear spin magnetic moment of the hydrogen nuclei with the total angular momenta. For both OH and SH, resolution of the hyperfine structure requires to work in the Doppler-broadened regime using THz sources with a relevant line-width better than 10 MHz. For the OH radical, the Far-IR laser side-band spectrometer developed by the JPL allowed to resolve for the first time the hyperfine structure of the strongest rotational transitions in the 1.8–3.0 THz frequency range [13]. This study provided a set of molecular constants accurate enough to predict the pure rotational spectrum of OH to better than 10 MHz throughout the Far-IR and the Mid-IR regions. With a precision estimated to 500 kHz, these measurements are actually used as reference in the spectroscopic databases [6,12]. For the SH radical, the hyperfine structure has been resolved below 0.9 THz by Klisch et al. [10] with the Cologne THz spectrometer based essentially on a frequency- and phase-stabilized high-frequency, broadband tunable backward waved oscillators. Pure rotational transitions of SH have been measured at higher frequencies by Morino and Kawaguchi using a Fourier transform spectrometer with a resolution of 225 MHz insufficient to observe the hyperfine splitting [9]. The present study revisits the pure rotational spectroscopy of OH and SH radicals using for the first time a continuous-wave (CW) THz radiation produced by photomixing. After a short description of the experimental set-up, we will demonstrate the ability of CW-THz spectroscopy by photomixing to detect and quantify radicals such as OH and SH from strongly intense rotational transitions and to resolve hyperfine splitting up to 3 THz. The frequencies measured with the CW-THz synthesizer are now capable to complete the atmospheric and astrophysical spectroscopic databases and to improve the predictions of the complete high-resolution spectra of these radicals. 2. Experimental details 2.1. CW-THz spectrometer The scheme of the CW-THz spectrometer used for the detection and quantification of OH and SH radicals is presented in Fig. 2. The

Fig. 1. Low-lying rotational energy levels of OH and SH radicals (energy levels are labeled in the Hund’s case [b]). For the sake of clarity K-doubling and hyperfine splittings are magnified and 2P1/2 levels of SH radical are not drawn. Observed transitions are represented by bold arrows.

15

S. Eliet et al. / Journal of Molecular Structure 1006 (2011) 13–19 Table 1 Experimental conditions used for the OH and SH transitions measurements. Radical

OH

SH

Precursor Pressure (mbar) RF power (W) Cell windows Spectral range (THz) Frequency stabilization Frequency tuning Resolution (MHz) Frequency accuracy (MHz)a

H2O 0.8 20–80 Polypropylene 1.8–2.5 Rb transition iScan 1 100

H2S 0.6 50–60 Teflon 1.3–2.6 Freq. comb Synthesizer 0.1 0.01

a Frequency accuracies correspond to the ultimate accuracy accessible with the frequency stabilization technique. For OH and SH line position measurements, these accuracies depend also on the S/N ratio observed on the spectra.

Fig. 2. CW-THz spectrometer used for OH and SH studies.

operation of a photomixing source is based on the heterodyning of two lasers with the optical beatnote being converted to free space THz radiation by an ultrafast LTG-GaAs device known as a photomixer [14,15]. Here two extended cavity diode lasers (ECDL) operating around 780 nm were detuned so that the difference frequency of the lasers corresponds to the desired THz frequency. The lasers were spatially overlapped and the resulting beatnote were used to illuminate the interdigitated electrode array located at the center of the photomixer. The spiral antenna on the photomixer and a hemispherical Si lens allows the THz beam to be extracted from the GaAs substrate. The highly divergent beam was collimated and propagated through a sample chamber by means of a pair of off-axis parabolic mirrors before being focused onto either a Si or InSb helium cooled detector. The ECDL used to display a frequency stability of the order of 5 MHz over a period of 1 s, which is directly transferred to the generated THz frequency. To undertake the spectroscopy of Doppler limited transitions a stability of 1 MHz or better is required. This may be realized by implementing a frequency stabilization scheme on the ECDLs, in this study two different schemes were used. Firstly a system composed of commercially available components was used for the measurements of OH [16]. The frequency of one of the ECDL was locked to a saturated absorption feature of the Rubidium D2-line (CoSy, TEM Messtechnik), whereas the second ECDL was stabilized to a low contrast Fabry–Perot interferometer (iScan, TEM Messtechnik). This system allowed the THz frequency with a stability of 1 MHz to be obtained and tuned over several GHz. The absolute value of the THz frequency was however established by a standard wavelength-meter which has an accuracy of around 100 MHz. To overcome the shortcomings of the previous scheme a frequency stabilization scheme based on a frequency comb was developed and used for the SH measurements. A frequency doubled erbium doped femtosecond fiber laser was used to generate a frequency comb (FC). Each mode of the FC is separated by the laser repetition rate which is known with an excellent accuracy. Two phase locked loops are used to lock the two ECDLs to the FC accurately fixing the difference frequency [17,18]. In this case a spectral purity of 100 kHz was achieved and the transition line centers can be determined to within absolute accuracy.

2.2. Sample environment The THz radiation probes a plasma created by an electrode-less radiofrequency (RF) discharge in a Pyrex cell (diameter: 5 cm, length: 120 cm). The discharge cell has been previously described in Ref. [19]. In this study, the Pyrex cell was equipped with two windows transparent in the THz region (polypropylene or Teflon).

While the complete set-up generating the THz radiation was at atmospheric pressure, the discharge cell was pumped out by means of a primary pump. The coil of the transmitter of the RF power supply was wrapped around the cell and placed co-linearly to the axis of the coil of the RF transmitter. In our experiments, precursor vapor flews through the cell at a pressure below 1 mbar. The length of the plasma generated by the discharge was adjusted between 50 cm and 70 cm depending on the precursor flow and the applied RF power. The RF power which is not transferred to the plasma, radiates into the laboratory hindering the efficiency of the locking scheme and may affect the useful linewidth of the THz spectrometer. So no more than 80 W is used to prevent any interaction with all parts of electronics of the set-up. The experimental conditions used in the measurements of OH and SH radicals are summarized in Table 1. 3. OH observed spectra and discussion 3.1. Line frequency measurements Strong transitions of OH involving the lowest electronic states 2

P1/2 and 2P3/2 are observed in the 1.8 THz and 2.5 THz regions, respectively. Due to K-doubling and hyperfine splitting, two

groups of three individual lines may be recorded in these regions. To underline the importance of this frequency domain, it can be noted that recently several airborne instruments have been developed to detect the stratospheric OH by probing these specific rotational transitions [20]. Up to now, Fourier Transform spectrometers, Fabry–Perot interferometers or Far-IR heterodyne techniques were used to determine the concentration profiles of OH. In this article, we propose the CW-THz by photomixing as an alternative technique for the detection and the quantification of OH. The frequencies of the rotational transitions of OH recorded in this study are listed in Table 2. The experimental conditions of pressure and RF power have been optimized in the 1.8 THz and 2.5 THz regions aiming to resolve the hyperfine structures of the 5/2 3/2 and the 3/2 1/2 of the 2P1/2 and the 2P3/2 electronic states, respectively. Unlike the 2P3/2 state, the 3/2 1/2 transitions of the 2P1/2 state show weaker intensities and hyperfine splitting. For the transitions in the 1.8 THz region, the two K-doubling components were not recorded in the Doppler-broadened regime and the hyperfine structures were not resolved. At 2.5 THz, the transition intensities are stronger and the Doppler limit was reached with a working pressure of 0.8 mbar and a RF power of 20 W. In these conditions, the hyperfine splitting was resolved. As example, the Fig. 3 shows the two DF = 1 hyperfine components of the 5/2+ 3/2 transition separated by only 33 MHz in a good agreement with the 39 MHz frequency difference measured at JPL [13]. Due to their weak intensities, hyperfine components with DF = O were never observed. In Table 2, we can notice that the

16

S. Eliet et al. / Journal of Molecular Structure 1006 (2011) 13–19

Table 2 Transition frequencies measured for the OH radical. All frequencies are given in GHz units. Experimental frequency uncertainties on the last digit are given in parentheses. Quantum numbers for the upper state are denoted by primes and those for the lower state by double primes. The total angular momentum, the K-doubling and the nuclear hyperfine components are labelled respectively with the J, p and F quantum numbers. Electronic state 2

P1/2

2

P1/2

2

P3/2

2

P3/2

J0

p0

F0

J00

p00

F00

3/2 3/2 3/2 3/2 3/2 3/2 5/2 5/2 5/2 5/2 5/2 5/2

   + + + + + +   

1 2 1 1 2 1 2 3 2 2 3 2

1/2 1/2 1/2 1/2 1/2 1/2 3/2 3/2 3/2 3/2 3/2 3/2

+ + +       + + +

1 1 0 1 1 0 2 2 1 2 2 1

Experimental Gaussian fit

1.04

Transmission a.u.

1834.79(5)

1838.03(5)

2509.96(5) 2509.99(5) 2514.22(5) 2514.25(5)

A IðTÞL

mthis work  mRef.[3] 0.043

0.213

0.011 0.015 0.097 0.104

ð1Þ

1.00

m

0.98

IðTÞ ¼ Ið300Þ

0.96 0.94 0.92 0.90 0.88 0.02 0.00 -0.02

mexp (Ref. [13]) 1834.7355(5) 1834.7473(5) 1834.7504(5) 1837.7466(5) 1837.8168(5) 1837.8370(5) 2509.9353(5) 2509.9493(5) 2509.9885(5) 2514.2987(5) 2514.3171(5) 2514.3539(5)

Although line intensities have been evaluated at 300 K, values of I at other temperature can also be obtained from Eq. (2) [12].

1.02

Residual a.u.

N¼

mexp (this work)

2509.92

2509.94

2509.96

2509.98

2510.00

Frequency (GHz) Fig. 3. The 5=2þ3 3=22 and the 5=2þ2 3=21 hyperfine components of OH radical in the 2P3/2 state (P = 0.8 mbar, RF power = 20 W). The Doppler line-widths of the two lines have been fitted to 9.0 ± 0.7 MHz. Respectively for the stronger and the weaker components, 2.7 ± 0.2 MHz and 1.7 ± 0.1 MHz have been obtained for the integrated intensities.

absolute line positions of the measured transitions with a commercial wavelength-meter may differ by more than 100 MHz from the frequencies listed in Ref. [13]. The experimental frequencies of OH measured by the JPL group were calibrated using carbon monoxide as reference gas and an average accuracy of 0.5 MHz was estimated. No calibration with reference gases was done in this study. In the part Section 4, we will demonstrate how a THz synthesizer based on a frequency comb allowed to improve the accuracy on the measured THz frequencies by several orders of magnitude without any standard of calibration. 3.2. Line intensity measurements In addition to perform an unambiguous detection, an absolute quantification of OH is also possible from the integrated intensity measurements of the THz pure rotational transitions. Several previous studies demonstrated the ability of CW-THz spectroscopy by photomixing to quantify stable species in gas phase without any calibration [15,21]. From the Eq. (1), the concentration N of absorbing molecules in molecules/cm3 unit is directly obtained from the integrated area A of the THz line in cm1, the line intensity at a given temperature I(T) in cm1/(molecule/cm2) and the absorption path length L in cm:

Q sr ð300Þ ð1  ekT Þ ð 1  1 ÞE00 e kT 300k m Q sr ðTÞ ð1  e300k Þ

ð2Þ

In this study, we used the lower state rotational energy E00 and the ambient temperature line intensities I(300) of OH listed in the JPL catalog [12]. In a simple approximation [22], the temperature dependence of the spin-rotation partition function is given by:
n Q sr ð300Þ ¼ 300 with n = 1 or 3/2, respectively for linear or asymmetT Q sr ðTÞ ric molecules. In this study, the main difficulty consisted in the determination of the rotational temperature T. In principle, T can be determined by comparing the measured absorption intensities of two rotational lines involving different energy levels. But this approach was not possible with the rotational lines of OH listed in Table 2 because the lines were not recorded in the same experimental conditions and the probed energy levels were too close to determine the temperature variation. Because the 2P3/2 transitions were recorded at the Doppler limit, an easier solution consisted in the determination of the thermodynamical temperature from the Doppler line-width using the relation (3) (see Ref. [23]):

T ¼ 7:798  1012 M

 2 Dm

m

ð3Þ

In Eq. (3), M is the molar weight of OH (17 g mol1). Dm/m was evaluated to (1.80 ± 0.15)  106 using the fitted HWHM linewidths and line frequencies of the 3/2 5/2 Doppler-limited hyperfine components (see Fig. 3). Finally, a thermodynamic temperature of 490K ± 78K has been estimated taking into account the uncertainties of the fitted line-widths. The estimated temperature indicates that the use of a weak power for the RF discharge allows to detect OH in a cold plasma. Taking into our experimental conditions, especially the weak RF power used, rotational and thermodynamic temperatures have been assumed equal (this is generally not the case in discharge plasma [24]). Therefore, using a temperature of 490 K and the fitted integrated intensities (Fig. 3), the Eqs. (2) and (1) yielded to a OH concentration of (2.9 ± 0.7)  1011 molecules/cm3. Taking into account the total number of molecules in has been estithe absorption cell, the efficiency of the discharge NNOH tot mated to (1.7 ± 0.7)  105 ensuring a detection of OH at the ppm level for the CW-THz photomixing technique. Another way to estimate the discharge efficiency consists in the direct comparison of the rotational intensities of the radical and its precursor. Contribution from moisture in the laboratory environment at atmospheric pressure did not allow to quantify water

17

S. Eliet et al. / Journal of Molecular Structure 1006 (2011) 13–19

This difference may be explained taking into account the fact that different RF powers have been used in these experiments (typically, 80 W and 20 W for the OD and OH experiments respectively).

4. SH observed spectra and analysis 4.1. Line frequency measurements

Fig. 4. The 5/2+ 3/2 transitions of the OD radical in the 2P3/2 state (P = 1.1 mbar, RF power = 80 W).

precursor from its THz rotational lines. Nevertheless, by use of deutered water in order to produce the OD radical, a direct comparison between OD and D2O rotational lines has been performed allowing an estimation of the discharge efficiency. The Fig. 4 shows the 5/ 2+ 3/2 transition of OD recorded in the 1.39 THz region at a pressure of 1.1 mbar with a RF power of 80 W. In the same frequency region, the 147,7 138,6 rotational transitions of D2O has been recorded using the same experimental conditions. Compared to OH, the OD rotational transitions shows a weaker hyperfine splitting and the different hyperfine components could not be resolved. Therefore, the thermodynamic temperature could not be deduced from the Doppler linewidth as it was done with OH. Nevertheless, assuming a weak variation of temperature, the approximation in Eq. (4) may be considered:

NOD AOD ID2 O ðTÞ AOD ID2 O ð300Þ ¼ ’ ND2 O AD2 O IOD ðTÞ AD2 O IOD ð300Þ

ð4Þ

Using tabulated line intensities at 300 K of JPL catalog [12] and the integrated intensities of OD and D2O determined respectively to 26.6 ± 2.0 kHz and 14.3 ± 3.9 kHz, Eq. (4) gave a ratio of (34.7 ± 12.0)  105. This last value is around 20 times superior to the discharge efficiency determined with OH measurements.

SH radical has been studied with the recently developed CWTHz synthesizer based on a frequency comb [17,18]. Compared to OH measurements, the frequency stabilization scheme allowed to obtain a spectral purity and an accuracy on line position improved respectively by one and four order of magnitude (see Table 1). The recorded transitions of SH are summarized in Table 3. Unlike OH, new THz measurements have been performed for the SH radical: for the first time, hyperfine splitting has been observed for frequency transitions superior to 1 THz. The four hyperfine components (DF = 1) of the 3/2 1/2 have been resolved in the 1.38 THz frequency region. The measured frequency transitions are in very good agreement with the predicted frequencies in the CDMS catalog [6] with frequency differences from 0.2 MHz to 1.8 MHz. Below 1.5 THz, rotational transitions of SH have been recorded with the InSb detector using both amplitude and frequency modulations. Above 1.5 THz, the InSb detector has been replaced by a Si bolometer more sensitive at higher THz frequencies but with a slower response time excluding frequency modulated THz measurements. In Fig. 5, the 5/2 3/2+ rotational line of SH recorded in the 1.455 THz region is presented. The frequency modulation permitted to improve the S/N ratio by a factor 5 providing therefore a better sensibility. This last affirmation agrees with the first observation of a HDS rotational line up to 0.8 THz. Indeed, in addition to the SH transition centred at 1,455,101 MHz, an other THz absorption line centred at 1,455,062 MHz is observed on the spectrum shown in Fig. 5. This absorption signal was observed only in presence of the H2S precursor in the cell with or without the application of the RF discharge. Referring to the JPL database [12], the observed line was assigned to the 70,7 60,6 rotational transition of HDS predicted at 1,455,054 MHz. The 7/2 5/2+ transitions have been recorded for both 2P3/2 and 2P1/2 electronic states in the 1.935 THz and 2.026 THz frequency region, respectively. For the 2P3/2 state, a direct comparison has be done with the previous measurements of Morino et al. using a FT-FIR spectrometer [9]. The difference mexp  mcalc of

Table 3 Transitions frequencies measured for the SH radical. All frequencies are given in MHz units. Experimental frequency uncertainties on the last digit are given in parentheses. Quantum numbers for the upper state are denoted by primes and those for the lower state by double primes. The total angular momentum, the K-doubling and the nuclear hyperfine components are labelled respectively with the J, p and F quantum numbers. Electronic state 2

P3/2

2

P3/2

2

P1/2

2

P3/2

2

P1/2

2

P1/2

J0

p0

F0

J00

p00

F00

3/2 3/2 3/2 3/2 3/2 3/2 5/2 5/2 5/2 7/2 7/2 7/2 7/2 7/2 7/2 9/2 9/2 9/2

+ + +             + + +

2 3 2 2 3 2 2 3 2 3 4 3 4 3 3 4 4 4

1/2 1/2 1/2 1/2 1/2 1/2 3/2 3/2 3/2 5/2 5/2 5/2 5/2 5/2 5/2 7/2 7/2 7/2

   + + + + + + + + + + + +   

2 2 1 2 2 1 2 2 1 3 3 2 3 2 3 3 4 4

mexp (Ref. [9])

mexp(This work) 1382909.9(4) 1382916.4(4) 1383240.1(4) 1383246.0(4) 1455101.4(4)

1,935,206(5)

1935200.0(4) 2025950.2(4) 2025950.2(4)

2,603,301(5)

2603299.1(6) 2603299.1(6)

mcalc (CDMS)

Dmcalc

mthis work  mCDMS

1382905.6 1382910.1 1382916.8 1383236.5 1383241.2 1383247.8 1455073.4 1455100.4 1455104.0 1935200.2 1935200.9 1935204.8 2025950.1 2025950.4 2025952.4 2603300.0 2603300.2 2603302.4

 0.9 0.9  0.9 0.9 0.1 0.1 0.1 1.2 1.2 1.2 0.3 0.3 0.3 0.6 0.6 0.6

 0.2 0.4  1.2 1.8  1.0   0.9  0.1 0.2  0.9 1.1 

18

S. Eliet et al. / Journal of Molecular Structure 1006 (2011) 13–19

efficiency (NSH/Ntot) of (1.1 ± 0.8)  105 have been determined. These results are very close to those obtained for OH even if the uncertainties are larger due to a lower S/N ratio level. Nevertheless, compared to OH, the THz line intensities of SH are weaker highlighting the improved sensitivity of the CW-THz synthesizer based on frequency comb. 5. Conclusion

Fig. 5. The 70,7 60,6 rotational transition of HDS and the 5/2 3/2+ SH transitions in the 2P1/2 state. Amplitude (top) and frequency modulated (bottom) CW-THz signal (P = 0.6 mbar, RF power = 60 W).

the CW-THz measurement is reduced by a factor 5 compared to the previous FT-FIR measurement. Both in the 2P3/2 and 2P1/2 electronic states, the difference mexp  mcalc obtained in the present study is inferior to the uncertainties Dmcalc of the predicted frequencies. Finally, the 9/2+ 7/2 transitions of SH in the 2P1/2 state have been recorded up to 2.6 THz with an accuracy of 1 MHz. The frequencies measured by the CW-THz synthesizer listed in Table 3 may be integrated in the spectroscopic databases for two reasons: the accuracy of the measurements are now competitive with the accuracies of the frequencies below 1 THz obtained with pure electronic sources [10]; the addition of these new CW-THz measurements should have a significant influence of the molecule fitted parameters of the SH radical. A more complete line frequency list of SH including CW-THz and FT-FIR measurements using a synchrotron source will be published providing a better fit of the ground state spin-rotation parameters of SH.

4.2. Line intensity measurements As for OH, line intensity measurements have been performed for SH in order to quantify the radical production. The integrated area of the 3/22 1/2+1 Doppler-limited hyperfine component have been measured yielding to an estimation of the thermodynamic temperature of 328 K ± 50 K close to the ambient temperature (see Eq. (3)). Using Eqs. (2) and (1), a molecular concentration of (1.4 ± 1.1)  1011 molecules/cm3 and a discharge

To the best of our knowledge, optoelectronic THz sources have never be used for the detection and quantification of radicals except few THz-Time Domain Spectroscopy studies where rotational signatures of radicals such as NO or CH were recorded at a GHz resolution [25,26]. In this study, for the first time, a high-resolution spectroscopic analysis of radicals is performed by means of photomixing CW-THz spectroscopy. The hyperfine structure of OH has been resolved in the 2.5 THz frequency region where strongly intense 5/2 3/2 transitions of the 2P3/2 state are observed. At low pressure, the Doppler line-widths and the integrated area of the hyperfine components allowed to estimate a ppm level detection of OH at a temperature of 470 K. The SH radical was also detected at the ppm level from its Doppler-limited spin-rotation THz transitions. For the SH radical study, the CW-THz synthesizer based on a frequency comb has been used to provide accurate line positions above 1 THz. Ten new frequency transitions of SH have been measured in the 1.3–2.6 THz frequency range with a subMHz accuracy. This frequency list should be used as reference data for the spectroscopic databases. Particularly, for the first time, the hyperfine structure of SH has been resolved above 1 THz. Many radicals and reactive molecules (such as OH, CH, SH, HCO, NO, NH, and NO2) are spectroscopically active in the THz region [12]. This study establishes the capability of CW-THz spectroscopy by photomixing to perform high-resolution spectroscopic measurements on these species and to provide accurate data for the community of atmospherists and astrophysicists. Photomixing technique is particularly flexible being able to cover all frequencies from 0.3 THz to 3.3 THz. The large spectral range delivered by such a source introduces a new feature that is the simultaneous spectroscopic recognition and evolution of different species. Time-resolved quantitative measurements of stable and unstable species aiming to determine the kinetic of chemical reactions should be now considered in the THz frequency domain. Acknowledgments The authors thank Olivier Pirali and Michel Vervloet for their assistance. This work was supported by both IRENI (Institut Régional en ENvironnement Industriel) and the European Commission within the Interreg IVA (CleanTech project) programme. References [1] G. Winnewisser, J. Mol. Struct. 408 (1997) 1–10. [2] T. De Graauw et al., Astron. Astrophys. 518 (2010) 1–7. [3] M. Carlotti, P. Ade, B. Carli, M. Chipper1eld, P.A. Hamilton, M. Mencaraglia, I. Nolt, M. Ridol, J. Atmos. Sol. – Terr. Phys. 63 (2001) 1509–1518. [4] D.E. Heard, M.J. Pilling, Chem. Rev. 103 (2003) 5163–5198. [5] S. Weinreb, A.H. Barret, M.L. Meeks, J.C. Henry, Nature 200 (1963) 829–831. [6] H.S.P. Müller, F. Schlöder, J. Stutzki, G. Winnewisser, J. Mol. Spectrosc. 742 (2005) 215–227. [7] K.M. Menten, F. Wyrowski, A. Belloche, R. Güsten, L. Dedes, H.S.P. Müller, Astron. Astrophys. 525 (2011). [8] I. Yamamura, K. Kawaguchi, S.T. Ridgway, Astrophys. J. 528 (2000) 33–36. [9] I. Morino, K. Kawaguchi, J. Mol. Spectrosc. 170 (1995) 172–177. [10] E. Klisch, T. Klaus, S.P. Belov, A. Dolgner, R. Schieder, G. Winnewisser, E. Herbst, Astrophys. J. 473 (1996) 1118–1124. [11] S.C. Herndon, K.D. Froyd, E.R. Lovejoy, A.R. Ravishankara, J. Phys. Chem. A 103 (1999) 6778–6785. [12] H.M. Pickett, R.L. Poynter, E.A. Cohen, M.L. Delitsky, J.C. Pearson, H.S.P. Müller, J. Quant. Spectrosc. & Rad. Transfer 60 (1998) 883–890.

S. Eliet et al. / Journal of Molecular Structure 1006 (2011) 13–19 [13] J. Farhoomand, G.A. Blake, H.M. Pickett, APJ 291 (1985) 19–22. [14] F. Hindle, A. Cuisset, R. Bocquet, G. Mouret, CR Acad. Sci. 9 (2008) 262–275. [15] F. Hindle, C. Yang, G. Mouret, A. Cuisset, R. Bocquet, J.F. Lampin, K. Blary, E. Peytavit, T. Akalin, G. Ducournau, Sensors 9 (2009) 9039–9057. [16] C. Yang, J. Buldyreva, I.E. Gordon, F. Rohart, A. Cuisset, G. Mouret, R. Bocquet, F. Hindle, J. Quant. Spectrosc. & Rad. Transfer 109 (2008) 2857–2868. [17] G. Mouret, F. Hindle, A. Cuisset, C. Yang, R. Bocquet, M. Lours, D. Rovera, Opt. Expr. 17 (2009) 22031–22040. [18] F. Hindle, G. Mouret, C. Yang, A. Cuisset, R. Bocquet, M. Lours, D. Rovera, Proc. SPIE 7763 (2010). [19] O. Pirali, M. Vervloet, Chem. Phys. Lett. 423 (2006) 376–381. [20] R. Titz, D. Hausamann, R. Nitsche, F. Schreier, J. Urban, H. Kulmann, H.P. Röser, Infrared Phys. Technol. 36 (1995) 883–891.

19

[21] D. Bigourd, A. Cuisset, F. Hindle, S. Matton, E. Fertein, R. Bocquet, G. Mouret, Opt. Lett. 31 (2006) 2356–2358. [22] C.L. Verdes, A. von Engeln, S. Buehler, A. Perrin, J. Quant. Spectrosc. & Rad. Transfer 90 (2004) 217–238. [23] C.H. Townes, A.L. Schawlow, Microwave Spectroscopy, Dover, New-York, 1955 (for Doppler linewidths see Chapter 13.2 on p. 337). [24] D. Collet, J.L. Destombes, I. Hadj Bachir, T.R. Huet, Chem. Phys. Lett. 286 (1998) 311–316. [25] S. Le Boiteux, N. Del Fatti, I. Manek-Honinger, L. Sarger, CR Acad. Sci. 2 (2001) 983–988. [26] R.A. Cheville, D. Grischkowsky, Opt. Lett. 20 (1995) 1646–1648.