ISSN 10248560, Atmospheric and Oceanic Optics, 2013, Vol. 26, No. 4, pp. 329–336. © Pleiades Publishing, Ltd., 2013. Original Russian Text © V.I. Serdyukov, L.N. Sinitsa, S.S. Vasil’chenko, B.A. Voronin, 2013, published in Optica Atmosfery i Okeana.
OPTICAL INSTRUMENTATION
HighSensitive FourierTransform Spectroscopy with ShortBase Multipass Absorption Cells V. I. Serdyukov, L. N. Sinitsa, S. S. Vasil’chenko, and B. A. Voronin V.E. Zuev Institute of Atmospheric Optics, Siberian Branch, Russian Academy of Sciences, pl. Akademika Zueva 1, Tomsk, 634021 Russia Received December 14, 2012
Abstract—A highsensitive spectrometer operating in the range 20000–9000 cm–1 with an absorption sensi tivity of 1 × 10–8 cm–1 and a spectral resolution of 0.05 cm–1, based on the Bruker IFS125M Fourier spec trometer with a short multipass cell, is described. The high sensitivity of the spectrometer was gained through the use of the multipass absorption cell (a base length of 60 cm) with a high transmission (the ratio of the col lecting mirror diameter to the base length is 1 : 4) and a high intensity light source. The comparison of the recorded spectra with experimental and theoretical water vapor spectra shows that the spectrometer reliably detects absorption lines of natural isotopomers of water. The threshold sensitivity of the spectrometer was found from the signal/noise ratio and recorded water lines of minimal intensities. DOI: 10.1134/S1024856013040131
INTRODUCTION At present, the problem of recording absorption spectra in the visible and UV spectral ranges is urgent, first of all, due to the importance of these ranges in the balance of incoming solar radiation and the complex ity of the theoretical calculation of spectra [1]. The main absorbing component in this spectral range is water vapor. In addition, there are O2 and CO2 absorption spectra in this range, as well as rather strong wideband O3 and O4 absorption, clearly visible in spectra of solar radiation propagated through the Earth’s atmosphere [2]. To record such weak (from 10–5 to 10–8 cm–1), but very dense spectra, different methods are used, providing for a high sensitivity in the recording of the absorption coefficient. Mostly, these are laser methods with the sensitivity at a level of 10–7–10–10 cm–1 [3]. Standard diffraction spectrometers with multipass cells (MPC) fail to provide high resolution and high accuracy in the measurement of the parameters of spectral lines recorded. Therefore, the most optimal in this case is Fourier spectrometry, having significantly higher transmission even at a high resolution as com pared to the standard diffraction spectrometry [4]. This is particularly important in the visible and UV spectral ranges, where the magnitudes of line intensi ties are of the order of 10–25–10–28 cm/mol; and the spectrometer’s sensitivity should reach 10–6–10–8 cm–1 in order to record them. To gain such a high sensitivity, Fourier spectrome ters are used with long multipass vacuum cells with a base length of more than 20 m; for example, the base of spectrometers in Reims (France) is 50 m [5], in
Tomsk (Russia), 30 m [6]. These cells are unique and quite expensive devices, requiring a large quantity of the gas studied (about 10000 liters). The large sizes of the cells hinder their thermal stabilization, necessary during longterm accumulation of measurement results, and make it difficult to achieve a high sig nal/noise ratio. Therefore, the questions arise as to whether these unique constructions are necessary, and whether or not it is possible to gain high sensitivity with shorter cells. The goal of this work is to show the capabilities of highsensitive Fouriertransform spectroscopy with small multipass cells. We have designed a simple and reliable multipass cell with a high transmission (the ratio of the collective mirror diameter to the base length is 1 : 4) of the vertical type. Multiple passes were provided by a threemirror White scheme with a base length L0 = 0.6 m in the Bernstein and Herzberg mod ification [7] and a radiation source of high brightness. With the help of this cell, water vapor absorption spec tra were measured in the region 20000–9000 cm–1 at absorption sensitivity of up to 10–8 cm–1. LIMITATION OF THE THRESHOLD SENSITIVITY The spectrometer’s threshold sensitivity Kthr is defined by the formula
K thr = (1 L)x(Δ I I )thr ,
(1)
where L is the length of the absorption layer; I is the intensity of the transmitted radiation recorded by the photodetector [3].
329
330
SERDYUKOV et al.
value in order to decrease the threshold absorption coefficient. Earlier, we have shown that optimal number of reflections Nopt (neglecting the sensitivity function of the photodetector) after reaching the threshold value (ΔI/I)thr is determined by the equality of the relative increase in the optical path length and relative decrease in the intensity [8]: 1 – R = 1/Nopt. (4)
Kthr 1 4
3
0.1 2
It immediately follows that the optimal number of reflections is Nopt = 1/(1 – R). (5)
1 0.01 0
10
20
30
40
50
N
Fig. 1. Kthr as a function of N for different R: 1 (1), 0.99 (2) 0.95 (3), and 0.9 (4).
Therefore, the threshold sensitivity depends on two experimental parameters: the path length of the absorbing layer Li and the capability of the recording system to measure small variations in the signal (ΔI/I)thr. Thus, for example, Kthr = 10–5 cm–1at L = 1 m and (ΔI/I)thr = 0.001. When using the multipass cell, the path length of the absorbing layer L becomes the product of the cell base length L0 and the number of passes N (the num ber of reflections from mirrors). When N increases, on the one hand, the length of the absorbing layer increases, and on the other hand, the radiation inten sity falls and, hence, the ratio (ΔI/I)thr changes. The intensity of a light beam passing through the cell depends on reflection losses (without accounting for the medium absorption): (2) I = I 0R N , where I0 is the intensity of radiation incident on the cell; and R is the reflection coefficient of cell mirrors. In this case, Kthr takes the form
(
)
K thr = ( 1 NL0 ) Δ I (I 0 R N ) .
(3)
Figure 1 shows the dependence of Kthr on N for dif ferent R (1, 0.99, 0.95, and 0.9). It follows from Fig. 1 that Kthr begins to differ noticeably from Kthr of an ideal cell with 100%mirrors as the number of reflections increases. The lower the reflection coefficient of mirrors, the stronger and more distinct the worsening of the threshold sensitiv ity. Starting from a certain number of reflections, a break point appears in the curve that describes the dependence of spectrometer sensitivity on the number of reflections. The lower the reflection coefficient of mirrors, the sharper and the more distinct the break, and it is observed at lower N. Consequently, at the given reflectivity of the mirrors and radiation intensity, the increase in N is reasonable only up to a definite
Thus, the threshold sensitivity strongly depends on the length of the absorbing layer and the radiation intensity recorded by the spectral system’s photode tector. There are three ways to decrease (ΔI/I)thr: to increase the reflection coefficient of the mirrors, to increase the accumulation time of signals (in this case the signal/noise ratio is proportional to t ), and to increase the radiation source brightness accompanied by a decrease in its noise characteristics. Whereas the first two approaches have reached their optimal limit, the use of light sources of high brightness has a certain reserve. EXPERIMENT. MULTIPASS OPTICAL CELL The construction designed is similar to the well known multipass optical cells of vertical type for the study of absorption spectra of gases. It is based on the White threemirror scheme in the Bernstein and Herzberg modification [7]. One collecting and two objective mirrors with a radius of curvature of 60 cm are installed on two foun dations, rigidly fastened by three steel bars 16 mm in diameter. The adjustment of mirrors allows from 3 to 60 reflections of the halogen lamp through the cell. Thus, the absorbing layer length in the cell is changed from 2.4 to 36 m. The system of spherical mirrors is hermetically covered with a steel cap with an inside diameter of 200 mm and a height of 700 mm. The working volume of the cell is 22 liters. The cap is fas tened to the cell bottom (heavy steel flange with quartz windows and the collecting mirror installed on the alignment unit). The cell is installed on the massive table with legs of adjustable height. The gas is filled and evacuated by a vacuum system. The cell pressure is controlled by an AIR20M pressure converter in the pressure range 0– 100 kPa with an error of 0.1%. Water vapor (of distilled water) is input into the cell through the vacuum inlet. In the MPC, we used wide band silvercoated mirrors with the SiO2 and Al2O3 protecting layers, which provided a reflection coeffi cient R = 96–98% up to 0.4 μm. In the absence of pro tecting layers, the silver coating was quickly oxidized
ATMOSPHERIC AND OCEANIC OPTICS
Vol. 26
No. 4
2013
ATMOSPHERIC AND OCEANIC OPTICS
10834.051
–
10834.256
12
13
11
10834.943
10835.039
10835.096
10835.227
11
11
11
11
No. 4
2013
10834.057 10834.088 10834.201
3.1 × 10–26 7.0 × 10–27 1.4 × 10–27
10835.201
6.7 × 10–27
10835.030
5.1 × 10–27 –
10834.947
4.8 × 10–26
–
10834.780
–
8.8 × 10–27
–
10834.347
10834.007
1.0 × 10–26
2.7 × 10–25
Frequency, cm–1 [10]
Int. [10], cm/mol
10835.202
10835.191
10835.032
10834.947
–
–
10834.349
10834.211
–
–
10834.016
BT2**, cm–1
Notes: * Isotopologue: 11 – H216O, 12 – H218O, 13 – H217O [1]. ** Frequency from the BT2 calculation [18] improved by data [19], *** Estimated intensity.
10834.782
10834.609
12
Vol. 26
?
10834.341
–
11
11
Our mea surements, cm–1
Isotopo logue*
Water vapor absorption lines within the 10834–10835.2 cm–1 range
4.6 × 10–27
–
3.1 × 10–27
4.2 × 10–26
–
3.0 × 10–28***
2.2 × 10–25
1.0 × 10–27
–
–
8.2 × 10–27
Int. [18], cm/mol
201
121
013
022
003
–
003
013
300
003
013
v1 v2 v3
14 4 10
11 7 4
331
744
616
–
515
330
101
524
515
J Ka Kc
000
000
010
000
000
–
000
010
000
000
010
v1 v 2 v3
13 4 9
11 5 7
432
633
717
–
532
431
110
625
616
J K a Kc
–
–
–
–
+
–
+
–
+
+
–
[1, 12]
+
–
+
+
+
–
+
–
–
+
–
[15]
+
–
+
+
+
–
+
–
+
+
+
[11]
HIGHSENSITIVE FOURIERTRANSFORM SPECTROSCOPY 331
332
SERDYUKOV et al.
Recorded signal, rel. units
1.0 0.8 0.6 0.4 0.2 0 9500
10000
10500 11000 Wavenumber, cm–1
11500
Fig. 2. Water vapor transmission spectrum at P = 19 Torr, neglecting the spectral sensitivity function of the photode tector.
and became impractical. A ratio of the collecting mir ror diameter to the base length was 1 : 4. The leakage into the cell was less than 0.1 Torr per 5 days. All measurements were conducted with the Bruker IFS125M Fourier spectrometer. The recording con ditions are given below. Spectral range Photodetector Divider Resolution Apodization function Scanning rate Aperture diameter Total number of scans Total time of recording
9000–20000 cm–1 silicon photodiode quartz 0.05 cm–1 triangular 10 kHz 1.1 mm 23328 216 h
The longterm measurement of spectra (over 9 days) was provided by temperature stabilization in the measurement room (volume of 75 m3) with the help of a Midea MSE24HR conditioner with an error better than 1 K.
In our experiment, we used a halogen lamp (OSRAM GmbH) with a power of 20 W as an emitter in the overheating mode. The GPR30600 power source was used, providing an instability in voltage of 1 mV and a current less than 3 mA. The lamp emission in such a mode more than doubled the signal ampli tude in the visible region. The maximal signal does not lead to the saturation of the photodetector and pro vides for the required signal/noise ratio, which was achieved at L = 36 m. Note that the lamps worked over 6.5 days in such rigorous conditions. Radiation from the Fourier spectrometer cell was input through the emission inlet. It was necessary to ensure conditions under which a light beam would remain parallel after leaving the cell. This could be achieved via placing a spherical telescope before the rotating reflecting mirror of the spectrometer. How ever, we proceeded as follows. We placed a correcting lens between the rotating mirror and the spectrometer diaphragm, which did not distort the spectrometer geometry and, conse quently, allowed us to avoid the correction of the spec trometer input diaphragm. As a result, the intensity of the incoming radiation into the spectrometer increased by ∼3–4 times, which ensured the same increase in the signal/noise ratio. The recordings of the oxygen absorption spectra at a pressure of 4 Torr in the 13000–13200 cm–1 range with the installed dia phragm for a resolution of 0.01 cm–1 (the width of the Doppler profile of oxygen lines in this range is 0.0285 cm–1 at a temperature of 297 K) have shown that the recording of spectra with the correcting lens does not lead to a deterioration of the spectrometer resolution. As the result, we achieved a signal/noise ratio of 5 × 104 at a beam length of 36 m in the cell during 9 days of recording. Assuming that the useful signal of the absorbing line should be at least twice as higher as that of the noise, we found that (ΔI/I)thr = 4 × 10–5, which corresponded to a threshold sensitivity of 1 × 10–8 cm–1. This allowed us to state that the recorded spectra were comparable in sensitivity with those measured in long multipass cells with the beam path length L = 650−864 m with daily recording.
As follows from (5), Nopt = 50 for mirrors with R = 0.98, which in our cell provides for the absorbing layer length L = 30.6 m.
EXPERIMENT. DETERMINATION OF THE THRESHOLD SENSITIVITY FROM A RECORDED SPECTRUM
However, accounting for the radiation intensity and the spectral photodetector sensitivity function can significantly change the situation. In our case, the radiation intensity is much higher than that at which (ΔI/I)thr is achieved; consequently, the number of passes in the cell can be increased by N0 when (ΔI/I) becomes equal to (ΔI/I)thr, and then Nopt (Eq. (5)) is added to N0.
The panoramic spectrum of water recorded in the 9500–12000 cm–1 range at a pressure of 19 Torr and a temperature of 298 K is shown in Fig. 2. The absorp tion is sufficiently high only in the band center, while in its wings it is equal to a fraction of one percent. To illustrate the capabilities of the recording method, the narrow 10820–10840 cm–1 range with relatively weak lines (earlier recorded in [9–15]) was
ATMOSPHERIC AND OCEANIC OPTICS
Vol. 26
No. 4
2013
HIGHSENSITIVE FOURIERTRANSFORM SPECTROSCOPY
333
(a) 0.009
0.006
Absorption, rel. units
0.003
0 0
(b)
–0.003
–0.006
–0.009 10820
10825
10830 Wavenumber, cm–1
10835
10840
Fig. 3. Water vapor absorption spectrum in the 10820–10 840 cm–1 range (in the plot, the absorption coefficient is increased by 1000 times): (a) absorption spectrum, measured in this work, (b) absorption spectrum, calculated by data from [11] (for illustra tive purposes, the absorption spectrum is multiplied by –1).
considered in more detail. The spectrum measured included more than 130 lines. There are only 59 lines of three isotopomers: H216O(32), H217O(15), and H218O(12) in the 10820– 10840 cm–1 range in the HITRAN database [1]. At Kitt Peak, a Fourier spectrum was recorded along a path of 433 m under a pressure of 1.97 × 10–3 atm [9]; in the range of our interest, 29 transitions were recorded. As reported in a later work [12], 27 transi tions were recorded with the Fourier spectrometer at the same site along the same path within a pressure range 2–16 Torr. Frequencies of 50 transitions were measured in [13] along a path length of 600 m at a pressure between 4.5 and 13.5 Torr. The spectrum recorded with the Bruker 120M Fourier spectrometer at a temperature of 3000 K, is presented in [14]; 17 transitions are described there in the range of our interest. In [11], parameters of weak lines are improved on the basis of earlier recorded spectra; and parameters of 57 transitions of the basic isotopologue are presented for the 10820–10840 cm–1 range. In [10], a narrow range was studied using the highsensi tive Cavity Ring Down Spectroscopy (CRDS). Figure 3 shows a part of the absorption spectrum A = 1 – I(ν)/I0(ν) for the 10820–10840 cm–1 range calculated using the database [16, 17]. ATMOSPHERIC AND OCEANIC OPTICS
Vol. 26
The comparison of the recorded spectrum with data [17] shows that a good agreement is observed in frequencies of transitions for strong lines, while cen ters of weak lines are often shifted relative to each other in both calculated and experimental spectra, because the correction of energy levels to experimental data is lacking in the calculation linelist [17]. At present, the most accurate H2O linelist is BT2 [18]. To use it in our range, we have applied the proce dure of calculated frequencies refinement through the replacement of calculated top and bottom energy lev els by corrected ones with the help of works [19, 20]. This range includes transitions of bands of the first decade H216O [19] (300)–(000), (003)–(000), (220)– (000), (022)–(000), (140)–(000), (041)–(000), (102)–(000), (201)–(000), (121)–(000) and hot tran sitions of bands of the second decade (013)–(010), (032)–(010), (112)–(010), (211)–(010). As a result of the refinement of centers, the spectrometer frequency scale was calibrated. We considered a narrower range 10834– 10835.3 cm–1 in order to find the threshold sensitivity of the spectrometer. Within this range, the HIT RAN database [1] includes four water vapor lines. One line (10 834.347 cm–1 with an intensity of 1.94 × 10–25 cm/mol) belongs to the basic isotopo
No. 4
2013
334
SERDYUKOV et al.
logue H216O and dominates in the range. In addition, the HITRAN database includes two transitions of H218O (10834.057 cm–1; 2.7 × 10–26 cm/mol, –1 –27 10834.780 cm ; 7.8 × 10 cm/mol) and a transition of H217O (10834.099 cm–1; 4.8 × 10–27 cm/mol). Measurements of this spectral range of water vapor were performed using CRDS, which is the most sensi tive method at the present time [10]. Table 2 shows transitions we recorded with a radiation source of high brightness, as well as calculation data [18, 19] and results of the measurements in a long multipass cell with a beam path of 434 m at Kitt Peak [1] and with a beam path of 600 m at Brussels [15]. In [10], when studying water vapor with natural content of isotopomers, nine lines were recorded in the above range: five transitions of H216O, two of H218O, one transition of H217O, and one transition free of identification (10834.201 cm–1 with an intensity of 1.39 × 10–27 cm/mol). As seen from the table, mea surement data represent the most complete set of lines recorded to date. There are only four lines in the HITRAN [1]; six and eight lines have been recorded in works [15] and [11], respectively. It may also be seen that the threshold sensitivity of our spectrometer exceeds that of the Fourier spectrometer at Kitt Peak with a path length of 434 m (it is represented in HITRAN by results on water vapor spectra in the 10000 cm–1 range [1, 12]) and of Reims spectrometer with a path length of 600 m [15]; it is comparable with the sensitivity of CRDS spectrometer [10]. We recorded nine lines of 3 × 10–28–1 × 10–26 cm/mol in intensity, including two lines, lacking in [10]. Pressures of 4.1 and 7.8 Torr were used in [10] and of 19 Torr in this work for weak lines, hence, lines 10834.007 and 10834.088 cm–1 belong to the strong line wing and are not resolved in our spectrum. In BT2 calculation [18], there is the H216O transi tion (013)[330]–(010)[431] (10834.156 cm–1 with an intensity of 1.0 × 10–27 cm/mol). On refining the top and bottom energy levels of this transition by [19], its frequency was 10834.211 cm–1. This transition is marked in [10] as the weakest among all recorded tran sitions. A satisfactory agreement of lines in frequency and intensity with the calculation [18] and indepen dent experimental data [10, 19] allows us to state that a new H216O absorption line (013)[330] – (010)[431] is identified. We note also a new line 10835.096 cm–1. We assume that this line is generated by the transition (121) [11 7 4] – (000) [11 5 7]. The level (121) [11 7 4] with an energy of 12820.880 cm–1 is found for the first time. In addition to coincidence in the frequency and intensity with the given transition, there is also agreement with the transition (BT2) (121) [11 7 4] – (000) [10 7 3] (10766.51 cm–1 with an intensity of 2.4 × 10–27 cm/mol) according to the combination
rule. The transition to the vibrationrotation level (121) [11 7 4] is observed in the recorded spectrum as well, though it is impossible to determine the line cen ter with a high accuracy because of its weakness and the presence of a stronger neighbor line in the wing. For the transition 10835.201 cm–1 with an inten sity of 6.7 × 10–27 cm/mol, the identification is given, which differs from that in [10], but agrees with [17]. It remains to elucidate the situation with the line 10834.605 cm–1 with an intensity estimated approxi mately as 3 × 10–28cm/mol. It is impossible to deter mine whether this is the water vapor line or not within this work. Let us estimate the spectrometer threshold sensi tivity with the help of the weakest of the recorded lines. The threshold sensitivity of spectra, recorded in the 10100–11200 cm–1 range with the Fourier spectrometer, and determined from the 10834.211 cm–1 –1 lines with intensities of 1.4 × 10–27 and 10835.032 cm 5.1 × 10–27 cm/mol, are 0.6 × 10–8 and 2 × 10–8 cm–1, respectively, at a spectral line halfwidth of 0.05 cm–1. This value agrees well with the threshold sensitivity found from the signal/noise ratio of the recorded spec trum. This threshold sensitivity was attained from a very large number of scans over 216 hours. This shows a good longterm stability of the spectrometer. As a rule, spectra are recorded with Fourier spectrometers with a high resolution over no more than 24 hours. In this case, the signal/noise ratio decreases by n times (n is the ratio of the number of scans during 9day measure ment to those for daily measurement), i.e., by three times, and will be 17000, which corresponds to a spec trometer threshold sensitivity of 3 × 10–8 cm–1. Such a threshold sensitivity is below the best values of CRDS (5 × 10–11 cm–1 [21]) and intraresonator (10–9 cm–1 [22]) laser spectrometers; however, it allows reliable recording of weak spectra of molecules with a small number of atoms, due to transitions to highly excited vibration states. It is interesting to compare the capabilities of the given system with commercial cells, in particular, with a Bruker multipass cell, parameters of which are close to those of our cell. The Bruker cell base is 80 cm, mir rors are goldcoated, which provides for a beam path length of up to 41.6 m [23]. The gold reflects radiation well in the IR range (R = 0.98), it decreases to R = 0.95 in the 1μm range and falls to 0.29 in the 0.42μm range with a further shift to the visible range. It follows from Fig. 1 and Eq. (5) that Nopt = 50 for our cell (R = 0.98) in the 1μm range, while Nopt = 20 for the Bruker cell (R = 0.95), which provides for path lengths of 30 and 16 m, respectively. In order to com pensate losses for lacking 20 reflections in the Bruker cell and to obtain the same output intensity as in our cell (i.e., the same sensitivity with an equal number of
ATMOSPHERIC AND OCEANIC OPTICS
Vol. 26
No. 4
2013
HIGHSENSITIVE FOURIERTRANSFORM SPECTROSCOPY
scans), it is necessary to increase the source intensity by five orders of magnitude. The situation becomes more complicated when shifting to the visible range, where R of gold falls sharply, and the use of this com mercial cell in our range of interest, with the aim of attaining a high sensitivity (10–7–10–8 cm–1), becomes practically impossible. CONCLUSIONS We have recorded absorption spectra in the visible and nearIR ranges using a Fourier spectrometer with a path length of 36 m. The sensitivity of the spectrom eter (1 × 10–8 cm–1 ) is comparable with that attained in laser methods even with the use of incandescent lamps as emitters. The use of a highintensity light emitting diode is capable of increasing the sensitivity of Fourier spectrometers by an order of magnitude. The emitting range of modern lightemitting diodes is not high (500–800 cm–1). In this case, one of the main advantages of the Fouriertransform spectroscopy is lost, i.e., its capability of a panoramic recording of spectra. Therefore, the use of highintensity light emitting diodes in the IR and UV ranges will allow absorption spectra measurements at the level k(ν) ∼ 10–8−10–9 cm–1 within comparatively small ranges, including individual rovibrational molecular bands. ACKNOWLEDGMENTS The authors are grateful to Yu.G. Borkov for useful consultations. This work was partly supported by the Russian Foundation for Basic Research, the Russian Academy of Sciences (Program 3.9), and the Ministry of Educa tion and Science of the Russian Federation (Contract 11.519.11.5009). REFERENCES 1. L. S. Rothman, I. E. Gordon, A. Barbe, D. C. Benner, P. F. Bernath, M. Birk, V. Boudon, L. R. Brown, A. Campargue, J.P. Champion, K. Chance, L. H. Coudert, V. Dana, V. M. Devi, S. Fally, J.M. Flaud, R. R. Gam ache, A. Goldman, D. Jacquemart, I. Kleiner, N. Lacome, W. J. Lafferty, J.Y. Mandin, S. T. Massie, S. N. Mikhailenko, C. E. Miller, N. MoazzenAhmadi, O. Naumenko, A. V. Nikitin, J. Orphal, V. I. Perevalov, A. Perrin, A. PredoiCross, C. P. Rinsland, M. Rotger, M. Simecková, M. A. H. Smith, K. Sung, S. A. Tash kun, J. Tennyson, R. A. Toth, A. C. Vandaele, and Auw era J. Vander, “The HITRAN 2008 Molecular Spectro scopic Database,” J. Quant. Spectrosc. and Radiat. Transfer 110 (9–10), 533–572 (2009). 2. S. S. Vasil’chenko, V. I. Serdyukov, and L. N. Sinitsa, “Spectral System for Measuring Gaseous Atmospheric Components with a FiberOptic Tracking System, and Certain Analysis Results of Atmospheric Spectra,” Atmos. Ocean. Opt. 26 (3), 227–231 (2013). ATMOSPHERIC AND OCEANIC OPTICS
Vol. 26
335
3. L. N. Sinitsa, HighResolution Spectroscopy Techniques (Publishing House of Tomsk State University, Tomsk, 2006) [in Russian]. 4. M. V. Tonkov, “FourierTransform Spectroscopy— Maximum Information for Minimal Time,” Soros. Obrazovat. Zh. 7 (1), 83–88 (2001). 5. C. Oudot, Le. Wang, and X. Thomas, P. Von Der Hey den, L. Daumont, and L. Régalia, “Intensity Measure ments of H216O Lines in the Spectral Region 8000– 9350 cm–1,” J. Mol. Spectrosc. 262 (1), 22–29 (2010). 6. Yu. N. Ponomarev, T. M. Petrova, A. M. Solodov, A. A. Solodov, and S. A. Sulakshin, “A FourierSpec trometer with a 30m BaseLength Multipass Cell for the Study of Weak Absorption Spectra of Atmospheric Gases,”. Atmos. Ocean. Opt. 24 (6), 593–595 (2011). 7. H. J. Bernstein and G. Herzberg, “RotationVibration Spectra of Diatomic and Simple Polyatomic Molecules with Long Absorbing Paths,” J. Chem. Phys. 16 (1), 30–38 (1948). 8. Yu. A. Poplavskii and V. I. Serdyukov, “LightEmitting Diode FourierTransform Spectrometry of Gases in the Visible Range,” in Proc. of the 14th Intern. Symposium “Atmospheric and Oceanic Optics. Atmospheric Physics”, Tomsk, 2009 (Tomsk, 2009) [in Russian]. 9. J.P. Chevillard, J.Y. Mandin, J.M. Flaud, and C. CamyPeyret, “H2O: Line Positions and Intensities Between 9500 and 11500 cm–1. The Interacting Vibra tional States (041), (220), (121), (022), (300), (201), (102), and (003),” Can. J. Phys. 67 (11) 1065–1084 (1989). 10. D. Lisak and J. T. Hodges, “LowUncertainty H2O Line Intensities for the 930nm Region,” J. Mol. Spectrosc. 249 (1), 6–13 (2008). 11. R. Tolchenov and J. Tennyson, “Water Line Parameters from Refitted Spectra Constrained by Empirical Upper State Levels: Study of the 9500–14500 cm–1 Region,” J. Quant. Spectrosc. and Radiat. Transfer 109 (8), 559– 568 (2008). 12. L. R. Brown, R. A. Toth, and M. Dulick, “Empirical Line Parameters of H216O near 0.94 µm: Positions, Intensities and AirBroadening Coefficients,” J. Mol. Spectrosc. 212 (1), 57–82 (2002). 13. R. N. Tolchenov, O. Naumenko, N. F. Zobov, S. V. Shi rin, O. L. Polyansky, J. Tennyson, M. Carleer, P.F. Coheur, S. Fally, A. Jenouvrier, and A. C. Vandael, “Water Vapour Line Assignments in the 9250–26000 cm–1 Frequency Range,” J. Mol. Spectrosc. 233 (1), 68–76 (2005). 14. N. F. Zobov, S. V. Shirin, R. I. Ovsyannikov, O. L. Poly ansky, R. J. Barber, J. Tennyson, P. F. Bernath, M. Car leer, R. Colin, and P.F. Coheur, “Spectrum of Hot Water in the 4750–13000 cm–1 Wavenumber Range (0.769–2.1 µm),” Mon. Not. Roy. Astron. Soc. 387 (3), 1093–1098 (2008). 15. M. F. Merienne, A. Jenouvrier, C. Hermans, A. C. Van daele, M. Carleer, C. Clerbaux, P. F. Coheur, R. Colin, S. Fally, and M. Bach, “Water Vapor Line Parameters in the 13000–9250 cm–1 Region,” J. Quant. Spectrosc. and Radiat. Transfer 82 (1–4), 99–117 (2003). 16. H. Partridge and D. W. Schwenke, “The Determination of an Accurate Isotope Dependent Potential Energy Surface for Water from Extensive Ab Initio Calculations
No. 4
2013
336
SERDYUKOV et al.
and Experimental Data,” J. Chem. Phys. 106 (11), 4618–4639 (1997). 17. http://spectra.iao.ru/1280x866/ru/mol/ 18. R. J. Barber, J. Tennyson, G. J. Harris, and R. N. Tolchenov, “A High Accuracy Computed Water Line List—BT2,” Mon. Not. Roy. Astron. Soc. 368 (3), 1087–1094 (2006). 19. J. Tennyson, P. F. Bernath, L. R. Brown, A. Campar gue, A. G. Cseszar, L. Daumont, R. R. Gamache, J. T. Hodges, O. V. Naumenko, O. L. Polyansky, S. Rothman, A. C. Vandaele, N. F. Zobov, A. R. Derzi, C. F. Csaba, A. Z. Fazliev, T. Furtenbacher, I. E. Gor don, L. Lodi, and I. I. Misus, “IUPAC Critical Evalu ation of the RotationalVibrational Spectra of Water Vapor. Part III. Energy Levels and Transition Wave numbers for H216O,” J. Quant. Spectrosc. and Radiat. Transfer 117, 29–58 (2013).
20. L. P. Vorob’eva, B. A. Voronin, and O. V. Naumenko, “Assignment of 9250–13000 cm–1 Water Vapor Absorp tion Spectrum,” Atmos. Ocean. Opt. 6 (12), 961–966 (2003). 21. A. Campargue, Le Wang, D. Mondelain, S. Kassi, B. Bezard, E. Lellouch, A. Coustenis, C. de Bergh, M. Hirtzig, and P. Drossart, “An Empirical Line List for Methane in the 1.26–1.71 μm Region for Planetary Investigations (T = 80–300 K). Application to Titan,” Icarus 219 (1), 110–128 (2012). 22. S. Beguier, S. Mikhailenko, and A. Campargue, “The Absorption Spectrum of Water between 13540 and 14070 cm–1: ICLAS Detection of Weak Lines and a Complete Line List,” J. Mol. Spectrosc. 265 (2) 106– 109 (2011). 23. http://www.bruker.de, Manual WhiteCell.doc. Translated by S. B. Ponomareva
ATMOSPHERIC AND OCEANIC OPTICS
Vol. 26
No. 4
2013