日本天文学会欧文報告 ＰＡＳＪ 48, 601(1996)
加藤賢一・渡部義弥（大阪市立科学館）、定金晃三（大阪教育大学）
ＣＣＤの登場により近赤外線領域の分光データが容易に得られるようになったため、Ｆ型分光標準星プロキオン（α CMi、こいぬ座）の観測を行い、近赤外線領域に検出しやすい吸収線を示す軽元素の定量を行った。可視光領域にはこれらの元素の線は少ないので、これまであまり調べられていなかった。カリウムＫとリンＰについてはこれが最初の報告である。観測は岡山天体物理観測所の太陽望遠鏡で行った。
結果は、炭素、酸素、カリウムが太陽より 0.2 〜 0.4 dex 多く、非熱平衡状態 non-LTE で吸収線が形成されたものと推察された。酸素の三重線（7771.94Å、7774.17Å、7775.39Å）にこの効果が見られることはさまざまな研究者が指摘していたことであり、実際、プロキオンの上層大気で酸素の三重線は形成されたと考えられる。また、カリウム線（ 7698.96Å）は大変強い共鳴線であり、同様のnon-LTE 効果が期待される。この点については、竹田洋一氏（東京大学）が大変厳密な定量的扱いを行い、その結果を下記論文で発表しているので、それを参照していただきたい。なお、これには併せて、プロキオンの大気モデルに関する問題も追究してある。
Takeda Y., Kato K., Watanabe Y., Sadakane 1996, PASJ 48, 511(1996)
その他、N、 Mg、 Al、 Si、 P、鉄などの量も求められたが、これらはほぼ太陽値で、これまでの可視域スペクトルの解析と矛盾しない。
なお、下に掲載したのは上記論文の元になったもので、上記論文とはやや異なっている。
Ken-ichi KATO and Yoshiya WATANABE
Science Museum of Osaka, Nakanoshima, Kita-ku, Osaka 530
E-mail(KK) kato@sci-museum.kita.osaka.jp
and
Kozo SADAKANE
Astronomical Institute, Osaka Kyoiku University, Asahigaoka, Kashiwara, Osaka 582
Abstract
High dispersion spectra in the near-infrared region of the F-type star Procyon (α CMi, F5IV-V) are obtained and absorption lines of eight light elements including phosphorus P and potassium K are analyzed with ATLAS9 model atmospheres. The LTE abundance is found to be solar for N, Mg, Al, Si, and P, which is consistent with previous studies in the visual region. However, two C I lines, triplet lines of O I at λ7771-5Å, and the resonance line of K I at λ7699・ are fairly strong, and their estimated LTE abundances are greater than the solar values by more than 0.2dex. These discrepancies are discussed in terms of the non-LTE line formation.
Key words:Stars:abundances - light elements - Stars:invidual(Procyon)
1. Introduction
This paper presents the abundance analysis of light metals (with atomic number z smaller than 20) in the atmosphere of the F-type spectroscopic standard star Procyon (α CMi = HR 2943 = HD 61421). This work was initiated in an attempt to extend our previous abundance analyses of Procyon (Kato, Sadakane 1982, 1986; Kato 1987) into the near-infrared region and to detect absorption lines of light elements unexplored so far in this star.
Most of the deduced abundances of light elements in the atmosphere of Procyon have values very close to the Sun. Exceptions are light elements such as Li and Be (e.g., Conti, Danziger 1966; Boesgaard 1976). Extensive abundance analyses made by Griffin(1971), Kato and Sadakane(1982), Steffen(1985), and Lane and Lester(1987) include such elements as C, N, O, Na, Mg, Al, Si, and S. Tomkin and Lambert(1978) and Altas(1987) studied abundances of C, N, and O. They found that their abundances are very close to solar. In their elaborate work for 189 F and G disk dwarfs, Edvardsson et al.(1993) obtained abundances of 13 elements whose deviations from solar values are within 0.11dex. Takeda(1992, 1994)ﾕs recent detailed investigations of non-LTE effects for neutral lines of C, N, and O in this star demonstrate fairly large deviations from LTE amounting to 0.36dex for stronger lines.
2. Observations and Data Reduction
Spectroscopic observations of Procyon in the near-infrared region ranging from 6700・ to 9800・ were carried out during a nine nights observing run in February, 1994, with the 0.65m solar coude telescope at the Okayama Astrophysical Observatory, a branch of the National Astronomical Observatory of Japan. The spectrograph was used with the 1200 grooves mm-1 grating and spectra were recorded with a LN2 cooled CCD camera of Osaka Kyoiku University. The CCD (EEV88200) has an effective imaging area of 1152 × 790 pixels and the pixel size is 22.5mm. This configuration yielded a linear dispersion of 0.017 pixel-1 = 0.72Å mm-1. Shutter durations of individual stellar exposure were fixed at 10 minutes. The spectral resolution, defined as the FWHM of thorium-argon arc lines, was close to 45mÅ, corresponding to 2.7 pixels on the CCD chip.
Data reduction was carried out using the software IRAF following the standard procedure including bias subtraction, trimming, flat-fielding, and extracting stellar and comparison spectra into one dimensional data. The wavelength calibration was carried out with Th-Ar comparison lines. Individual stellar spectra were adjusted in the wavelength scale and then co-added in order to increase the S/N ratio. The measured S/N ratio in our final spectrum is about 150. This is certainly valid when we could achieve a complete flat-fielding. In some cases, weak wavy interference patterns, which are probably produced in the thin surface layer of the CCD chip or in the cover glass of the dewer, remained on final spectra. Unfortunately, we have failed in removing completely the wavy pattern in the region of O I triplet lines near 7774Å.
3. Abundance Analysis
Equivalent widths of absorption lines for light metals were measured using an IRAF routine splot. We also measured Fe I lines to utilize them for checking the result of abundance computations.
We adopt a set of model parameters of (Teff, log g, ξt ) = ( 6500K, 4.0, 1.8 km s-1). The surface gravity log g and microturbulent velocity ξt are the same as used in Kato and Sadakane(1982), while the effective temperature Teff is lower by 150K. In table 1, we summarize effective temperatures of Procyon derived by various methods within these 20 years. The simply averaged effective temperature and RMS scatter are 6512K and 97.0K, respectively. We can see that the temperatures deduced from flux data point toward Teff 〜 6500K, while those from (b-y) index indicates a higher value. Kato and Sadakane(1982) found the effective temperature of Procyon determined from flux data is cooler by 100K 〜 400K than that deduced from the ionization balance of iron group elements. This discrepancy in the effective temperature was also found by Steffen(1985). At the final stage of abundance analysis, Kato and Sadakane(1982) and Steffen(1985) have chosen Teff = 6650K and 6750K, respectively, to minimize abundance differences between those derived from neutral and from ionized species. However, the stellar effective temperature should be determined from flux data following its definition. Therefore we adopt Teff = 6500K in this study. We can find further discussion on the effective temperature of Procyon in Lane and Lester(1984, 1987) and Drake and Martin Laming(1995).
Most of the gf values for each line are taken from literature. When reliable data are not available, solar gf values are empirically deduced from equivalent widths measured on the Solar Flux Atlas from 296 to 1300nm (Kurucz et al. 1984).
To calculate elemental abundances, fully line-blanketed ATLAS9 model atmospheres(Kurucz 1993) were used, where they were computed by adopting the convection of mixing length l/Hp =1.25 (Hp represents the pressure scale height) and solar metallicity. Elemental abundances were computed with the program WIDTH9, a companion to the program ATLAS9 assuming the LTE line formation.
Table 2 contains the analysis of light elemental lines in Procyon. Consecutive columns give the wavelengths (in
Å), lower excitation potentials (in eV), log gf values, sources of gf values, the measured equivalent widths (in mÅ, the lines with lower accuracy are marked with a colon), derived LTE abundances(in the usual scale where log ε(H)=12.00), and the solar abundance value. All the solar abundances are taken from Anders and Grevesse(1989) except for iron. The solar iron abundance is from Holweger et al.(1995). For recent investigations of solar iron, see, for instance, Biemont et al.(1991a), Holweger et al.(1991), Blackwell et al.(1995), and Holweger et al.(1995).
4. Abundance Results
4.1. C I
Two fairly strong neutral carbon lines are analyzed. The van der Waals interaction with hydrogen atoms is important for these lines, and an enhancement of Δlog C6 = +1.0 is applied to the collisional damping term in the WIDTH9 program. This correction factor is empirically determined to bring the abundances derived from weak lines and those from strong lines into agreement for the Sun. The solar line data for neutral carbon are taken from Biemont et al.(1993). The resulting carbon abundance in Procyon is higher by 0.24dex with respect to the Sun, contradicting with previous results.
From the analysis of near-infrared C I lines, Tomkin and Lambert(1978) deduced a solar abundance for carbon in Procyon. Studies of neutral carbon lines in the visual region also show the solar abundance. According to Steffen(1985), the logarithmic abundance relative to the Sun is -0.03. Altas(1987) and Kato(1988) concluded that the carbon in Procyon is slightly underabundant, but the dicrepancy disappears after applying new gf values of Biemont et al.(1993). From an analysis of the forbidden line at l 8727.13Å, Andersson and Edvardsson(1994) gives the carbon abundance [e(C)] = -0.01 with respect to the Sun. Based on the observation of 7100Å C I lines, Tomkin et al.(1995) obtained a value [ε(C)] = -0.22 by adopting model parameters of (Teff, log g) = (6704, 4.03). The abundance will increase to [ε(C)] = -0.10, when we calculate employing the same line data and model parameters (Teff, log g) = (6500, 4.0) as those in the present analysis.
The departure from LTE for C I lines in Procyon is investigated by Takeda(1994). His non-LTE computations have found no significant departures from LTE for weak C I lines (Wλ<52mÅ). However, his recent computations(Takeda 1995, private communication) including strong C I lines show that non-LTE effects are fairly large for the two lines in table 2, where corrections are 0.23 and 0.38dex for 8335.13 and 9111.80Å lines, respectively. After the correction for the non-LTE effects noted above, the final carbon abundance would be solar, which is consistent with previous studies.
4.2. N I
We use a weak neutral nitrogen line at 8216.35Å Its equivalent width (31mÅ) is slightly larger than the measurement (26mÅ) given in Tomkin and Lambert (1978). The derived abundance of nitrogen seems to be somewhat higher than that in the Sun. According to Takeda (1994), the non-LTE correction for this line is 0.11dex in Procyon. His elaborate work using a very detailed model atom of nitrogen shows that the non-LTE effect in the atmosphere of Procyon reduces the abundance systematically with respect to the LTE abundance from 0.09 to 0.13 dex even for weak lines (Wλ = 12 〜 38mÅ). Applying the non-LTE correction to the value in table 2, an agreement with the solar nitrogen abundance is achieved.
4.3. O I
The oxygen abundance has been derived from strong triplet lines at 7771-5Å. We used the radiative damping log γgrad = 8.04, the Stark broadening log C4 = -14.46, and the van der Waals broadening log C6 = -30.87 following Baschek et al.(1977). Spectral data in this region have suffered from interference features produced in the thin surface layer of CCD chip. Even if we take the uncertainty due to the noise into account, the abundances derived from these lines are evidently higher than the solar value. The differences range from +0.27 to +0.44dex, corresponding to the estimation of [O(7774)/O(6300)] = 0.35dex (normalized to the Sun) by Nissen and Edvardsson (1992). It is well known that the O I triplet lines are fairly strong in early-type stars, especially in supergiants, so that they have been used as a luminosity indicator. The unusual strength is now explained as an effect of departures from local thermodynamic equilibrium (e.g., Baschek et al. 1977; Eriksson, Toft 1979; Faraggiana et al. 1988; Kiselman 1991, 1993; Takeda 1992; Nissen, Edvardsson 1992). Takeda(1992, 1994) performed detailed non-LTE computations of the O I triplet lines for Procyon by adopting a realistic model atom of oxygen (86 terms and 294 transitions) and found that LTE abundances deduced from the triplet were overestimated by 0.3 to 0.4dex. Previous LTE studies in the visual region (Griffin 1971; Tomkin, Lambert 1978; Steffen 1985; Altas 1987; Kato 1988) show no significant departure from the solar composition. The apparent overabundance of oxygen in table 2 can be interpreted as the effect of non-LTE. The correction necessary to the LTE abundance is around 〜0.3 dex, which is much less than the results of, for example, Baschek et al.(1977), and is very close to those of Takeda(1992, 1994).
4.4. Mg I, Al I, and Si I
Solar gf values for Mg I and Si I lines are obtained from equivalent widths(151 and 41 mÅ for Mg I and Si I lines, respectively) measured on the Solar Flux Atlas(Kurucz et al. 1984) by adopting the ATLAS9 solar model atmosphere, where the microturbulence of 0.85 km s-1, an enhancement of ΔlogC6 = +1.0, and the solar composition given by Anders and Grevesse (1989) are used. For the two Al I lines at 7835.39Å and at 7836.15Å, their solar gf values are also derived by the same way. Final abundances for these elements are consistent with those obtained in the visual region and coincide with the solar abundances.
4.5. P I
Two lines of P I are identified and used in the present analysis. The weaker line at 9750.75Å is located on a noisy continuum so that its equivalent width may be less reliable. The derived abundance of phosphorus appears to be slightly underabundant when compared to the Sun. More analyses based on higher quality data are necessary to confirm the conclusion.
4.6. K I
The LTE abundance of potassium has been determined from the strong resonance line of K I at 7698.96Å by applying ΔlogC6 = +1.0 to the van der Waals damping term. The resulting abundance log ε(K)=5.51 is fairly larger than the solar value of 5.12. Bedford et al.(1994) measured weak magnetic fields of Procyon from the observation of circular polarization in wings of this potassium line. The measured magnetic fields are so weak that no significant enhancement of the line due to the Zeeman splitting is expected(e.g., Takeda 1993). It is unlikely that pottasium is heavily overabundant in the atmosphere of Procyon, because no other metallic elements shows such a large anomaly. It is reasonable to assume that the overabundace is due to non-LTE effect appeared in the strong resonance line.Takeda's recent non-LTE computation demonstrates that the line at 7698.96Å is very sensitive to the radiation field in the atmosphere of Procyon, and its correction to LTE abundance amounts to 〜0.7 dex ! Therefore, we can conclude the overabundance of pottasium in table 2 is superficial due to the non-LTE effect. Details of the non-LTE analysis of the K I line is discussed in a separate paper(Takeda et al. 1996).
4.7. Fe I
The gf values for the Fe I lines in table 2 were taken from May et al.(1974), Cayrel et al.(1985, solar gf values), and Kurucz(1994). The averaged iron abundance derived after applying Δlog C6 = +1.0 is 7.51, which is in good agreement with previous analyses and identical to the current solar value.
5. Conclusions
Final LTE abundances of eight light elements in Procyon with respect to the Sun are summarized in table 3 together with results of previous studies. Fairly good agreements are found except for C and O which probably reflect the non-LTE effect. The iron abundance is also solar, which indicates that our analysis is consistent with previous studies. Remaining some abundance differences are partly due to the choice of model parameters and due to the atmospheric structure corresponding to these parameters. A difference of 500K in effective temperature changes the abundance from 0.01 to 0.38dex depending on elements. Particularly, neutral nitrogen, pottasium, and iron are very sensitive to the change of effective temperature. Taking into account ambiguities involved in abundance analysis, we conclude the final abundances including newly derived phosphorus and potassium are solar. Apparent overabundances of carbon, oxygen and potassium deduced from stronger lines are caused by the departure from LTE for these lines.
We thank Professor H. Maehara, director of the Okayama Astrophysical Observatory, for his kind help and arrangements. Technical supports of Mr. Y. Norimoto and K. Koyano are gratefully acknowledged. The authors express their hearty thanks to Dr. R. L. Kurucz for kindly sending the CD-ROM set. Thanks are also due to Dr. Y. Takeda for helpful discussions and for providing his unpublished results of non-LTE computations. This work was supported in part by the Scientific Research Fund of the Ministry of Education, Science, and Culture under grant No.05916028 to K.K.
References
Altas L. 1987, Astrophys. Space Science 134, 85
Anders E., Grevesse N. 1989, Geochim. Cosmochim. Acta 53, 197
Andersson H., Edvardsson B. 1994 A&A 290, 590
Baschek B., Scholz M., Sedlmayr E. 1977 A&A 55, 375
Bedford D.K., Chaplin W.J., Innis J.L., Isaak G.R., McLeod C.P., Speake C.C. 1994, MNRAS 269, 639
Beeckmans F. 1977, A&A 60, 1
Biemont E., Baudoux M., Kurucz R.L., Ansbacher W., Pinnington E.H. 1991a, A&A 249, 539
Biemont E., Hibbert A., Godefroid M., Vaeck N., Fawcett B.C. 1991b, ApJ 375, 818
Biemont E., Hibbert A., Godefroid M., Vaeck N. 1993, ApJ 412, 431
Biemont E., Martin F., Quinet P., Zeippen C.J. 1994, A&A 283, 339
Blackwell D.E., Lynas-Gray A.E., Smith G. 1995, A&A 296, 217
Blackwell D.E., Petford A. D., Shallis M. J. 1980, A&A 82, 249
Blackwell D.E., Shallis M. J. 1977, MNRAS 180, 177
Boesgaard A.M. 1976, ApJ 210, 466
Bohm-Vitense E. 1982, ApJ 255, 191
Burkhart C., Coupry M.F. 1989, A&A 220, 197
Cayrel R., Cayrel de Strobel G., Campbell B. 1985, A&A 146, 249
Code A.D., Davis J., Bless R.C., Hanbury Brown R. 1976 ApJ 203, 417
Conti P.S., Danziger I.J. 1966, ApJ 146, 383
Drake J. J., Martin Laming J. 1995, Observatory 115, 118
Edvardsson B., Andersen J., Gustafsson B., Lambert D.L., Nissen P.E., Tomkin J. 1993, A&A 275, 101
Eriksson K., Toft S.C. 1979, A&A 71, 178
Faraggiana R., Gerbaldi M., van 't Veer C., Floquet M. 1988 A&A 201, 259
Griffin R. 1971, MNRAS 155, 139
Hibbert A., Biemont E., Godefroid M., Vaeck N. 1991, A&AS 88, 505
Holweger H., Bard A., Kock A., Kock M. 1991, A&A 249, 545
Holweger H., Kock M., Bard A. 1995, A&A 296, 233
Kato K. 1987, PASJ 39, 517
Kato K. 1988, Thesis, Tohoku University
Kato K., Kuroda T. 1992, Pub. Science Museum of Osaka 2, 61
Kato K., Sadakane K. 1982, A&A 113, 135
Kato K., Sadakane K. 1986, A&A 167, 111
Kiselman D. 1991, A&A 245, L9
Kiselman D. 1993, A&A 275, 269
Kurucz R.L. 1979, ApJS 40, 1
Kurucz R.L. 1993, ATLAS9 Stellar Atmosphere Programs and 2km/s Grid (Kurucz CD-ROM No.13)
Kurucz R.L. 1994, Atomic Data for Fe and Ni (Kurucz CD-ROM No.22)
Kurucz R.L., Furenlid I., Brault J., Testerman L. 1984, Solar Flux Atlas from 296 to 1300nm (National Solar Observatory, Sunspot)
Lane M.C., Lester J.B. 1984, ApJ 281, 723
Lane M.C., Lester J.B. 1987, ApJS 65, 137
May M., Richter J., Wichelmann J. 1974, A&AS 18, 405
Moon T. T, Dworetsky M. M. 1985, MNRAS 217, 305
Morton D.C. 1991, ApJS 77, 119
Nissen P.E., Edvardsson B. 1992, A&A 261, 255
Saxner M., Hammarback G. 1985, A&A 151, 372
Smalley B., Dworetsky M.M. 1995, A&A 293, 446
Sokolov N. A. 1995, A&AS 110, 553
Steffen M. 1985, A&AS 59, 403
Takeda Y. 1992, PASJ 44, 309
Takeda Y. 1993, PASJ 45, 453
Takeda Y. 1994, PASJ 46, 53
Takeda Y., Kato K., Watanabe Y., Sadakane 1996, PASJ 48, in press
Theodossiou, E. 1985, MNRAS 214, 327
Tomkin J., Lambert D.L. 1978, ApJ 223, 937
Tomkin J., Woolf V.M., Lambert D.L. 1995, AJ 109, 2204
Table 1. Effective temperatures derived from various methods.
The number in parentheses indicates references.
------------------------------------------------------
Integrated flux 6510( 1) 6520( 2) 6427( 9) 6420( 9) 6560(16)
IR flux method 6359( 3) 6421( 5) 6601(10)
Bolometric Corr. 6450( 9)
cont. flux vis. 6500( 6) 6650( 9) 6550( 9) 6560(13)
cont. flux vis-UV 6400( 8) 6500(11)
cont. flux UV 6340(15)
four color (b-y) 6600( 4) 6650( 6) 6650(12) 6704(14)
UBVRIJKL color 6470( 9)
13 color
6500( 9)
UV color 6550( 7) 6370(13)
R-I color 6500(13)
------------------------------------------------------
Notes: Teff References
1. Code et al.(1976) 2. Beeckmans(1977)
3. Blackwell and Shallis(1977) 4. Tomkin and Lambert(1978)
5. Blackwell et al.(1980) 6. Kato and Sadakane(1982)
7. Bohm-Vitense(1982) 8. Lane and Lester(1984)
9. Steffen(1985) 10. Saxner and Hammarback(1985)
11. Theodossiou(1985) 12. Moon and Dworetsky(1985)
13. Kato and Kuroda(1992) 14. Edvardsson et al.(1993)
15. Sokolov(1995) 16. Smalley and Dworetsky(1995)
Table 2. Measured equivalent widths and derived LTE abundances.
-----------------------------------------------
Species λ(Å) χ(eV) log gf Ref Wλ(mÅ) logε logε(Sun)
-----------------------------------------------
C I 8.60
8335.13 7.68 -0.46 BHGV 203: 8.81:
9111.80 7.49 -0.32 BHGV 270: 8.86:
N I 8.00
8216.35 10.34 0.147 HBGV 31 8.13
O I 8.93
7771.94 9.15 0.35 BHGVF 174: 9.37:
7774.17 9.15 0.21 BHGVF 155: 9.32:
7775.39 9.15 -0.02 BHGVF 124: 9.19:
Mg I 7.58
8213.03 5.75 -0.67 sol 100 7.54
Al I 6.47
6695.97 3.14 -1.60 BC 21 6.48
6698.42 3.14 -1.90 BC 12 6.52
7835.39 4.02 -0.68 sol 26 6.37
7836.15 4.02 -0.54 sol 38 6.46
Si I 7.55
8215.15 6.26 -0.90 sol 27 7.45
P I 5.45
9750.75 6.95 -0.17 BMQZ 9: 5.29:
9796.91 6.99 0.26 BMQZ 23 5.37
K I 5.12
7698.96 0.00 -0.168 M 138 5.51
Fe I 7.51
6703.56 2.76 -3.060 MRW 15 7.51
6705.10 4.61 -1.073 CCC 28 7.47
7832.24 4.43 0.018 K 98 7.37
8220.44 4.32 0.249 K 124 7.32
9800.36 5.09 -0.394 K 83 7.88
-----------------------------------------------
Notes: gf values references
BC : Burkhart and Coupry(1989) HBGV: Hibbert et al.(1991)
BHGV : Biemont et al.(1993) M : Morton(1991)
BHGVF: Biemont et al.(1991b) MRW : May et al.(1974)
BMQZ : Biemont et al.(1994) K : Kurucz(1994)
CCC : Cayrel et al.(1985) sol : solar gf-values
Table 3. Comaprison of derived abundances relative to the Sun.
N is the number of lines used in this analysis.
----------------------------------------------------------
Species N This study
TL78 KS82 S85
A87 LL87 E93
Teff (K) 6500
6600 6650 6750 6500
6400 6704
log g
4.0 4.0
4.0 4.04 4.0
3.95
4.03
ξt
1.8 2.0
1.8 2.1 2.1
1.80 2.4
----------------------------------------------------------
C I 2
+0.22 +0.07 --- -0.03
-0.16 --- ---
N I 1
+0.13 +0.07 --- +0.07
-0.02 --- ---
O I 3
+0.36 +0.07 --- -0.08
-0.13 --- -0.05
Mg I 1
-0.04 --- -0.04 +0.07
--- -0.24 +0.07
Al I 4
-0.01 --- -0.05
-0.01 --- +0.40 0.00
Si I 1
-0.10 --- 0.00
+0.12 --- -0.07 +0.01
P I 2
-0.12 ---
--- --- ---
--- ---
K I 1
+0.39 ---
--- --- ---
--- ---
Fe I 5
0.00 -0.14 -0.01
-0.02 --- -0.11 -0.02
----------------------------------------------------------
Notes:
TL78: Tomkin and Lambert(1978) KS82: Kato and Sadakane(1982)
S85 : Steffen(1985)
A87 : Altas(1987)
LL87: Lane and Lester(1987) E93 : Edvardsson et al.(1993)