27

9

Mini-reviewReceived: 2 September 2008 Revised: 19 October 2008 Accepted: 22 October 2008 Published online in Wiley Interscience: 9 January 2009

(www.interscience.wiley.com) DOI 10.1002/pi.2524

Control of stereochemical structures ofsilicon-containing polymeric systems†

Yusuke Kawakami,∗ Yuriko Kakihana, Osamu Ooi, Motoi Oishi,Keigo Suzuki, Satoshi Shinke and Kazuya Uenishi

Abstract

Various optically active silicon compounds have been synthesized or separated, and used to synthesize silicon-containingpolymers with well-controlled stereochemical structures. Hydrosilylation, anionic ring-opening polymerization and cross-coupling reactions have been used to synthesize optically active and/or stereoregular silicon-containing polymers.c© 2009 Society of Chemical Industry

Keywords: stereochemistry; silicon; structure; optically active

INTRODUCTIONThe properties of polymeric systems in general strongly depend ontheir chemical structures. Molecular weight and its distribution arethe primary factors to be controlled. Living anionic polymerization,established by Szwarc and co-workers1 – 3 in 1956, has beenutilized to control these factors in many polymers. Stereochemicalstructure is another very important factor in the control of polymerproperties. Such control is usually achieved by enantioselectivepolymerization of unsaturated carbon–carbon bonds, typicallyseen in stereospecific olefin polymerization.

The properties of silicon-containing polymeric systems verymuch depend on the type and number of atoms attached to thesilicon atoms in the molecular structure. Stereochemical aspectsof silicon centers, as well as molecular weight of the polymers, arevery important in the control of the properties of these systems.One of the characteristic stereochemical aspects of the bondingof silicon compounds is that sp2 or sp configurations are not wellstabilized and sp3 is the most stable configuration, a fact thatlimits enantioselective polymerization of silicon–silicon doublebond compounds. Optically active silicon compounds with sp3

configuration must be used as the monomer, and the stereo-chemistry of the polymerization reaction must be controlled toobtain polymeric systems with stereochemically controlled struc-ture. However, since there are limitations in the variety of reactionsof silicon compounds, we have to develop new reactions with con-trolled stereochemistry. Since silicon derivatives are non-naturalcompounds, such optically active compounds must be separatedor synthesized to obtain monomers. The most common start-ing material is (methyl)phenylnaphthylmenthoxysilane, shown inFig. 1.

The stereochemistry of the reduction to (methyl)phenylnaph-thylsilane was established to be a complete retention process.Chlorination of silane function is also a retention process.4 – 6 Nu-cleophilic substitutions at halogen–silicon bonds are basicallyinversion processes, as shown in Scheme 1, but the stereochem-istry strongly depends on the reaction conditions.

Some silicon compounds can be separated using achromatograph.7 In this article, the development of new reac-

tions and control of the stereochemistry of silicon compounds arepursued in terms of the synthesis of silicon-containing polymerswith controlled stereochemical structures.

OPTICALLY ACTIVE SILICON COMPOUNDSAND STEREOREGULAR AND/OR OPTICALLYACTIVE SILICON-BASED POLYMERSIsotactic polycarbosilane was first synthesized by polyaddition viahydrosilylation reaction. The starting optically active allylsilane wassynthesized from methylphenyldi[(−)-bornyloxy]silane, anotheroptically pure starting material, and allyllithium, followed byreduction using lithium aluminium hydride to give a colorlessoil: [α]25

D = −16.0 (c 0.50, pentane). The reaction scheme for thesynthesis of the polymer is shown in Scheme 2, and the 750 MHz1H NMR spectrum of the polymer is shown in Fig. 2.8,9

In the spectrum, the SiCH3 signal is split into three singletsat 0.120, 0.125 and 0.131 ppm according to the triad tacticity.In spectrum in Fig. 2(b) of the optically active monomer, thesignal at 0.120 ppm becomes relatively stronger and that at0.131 ppm relatively weaker compared with those of the polymerfrom racemic monomer. The signals at 0.120 and 0.131 ppm areassigned to the isotactic and syndiotactic triad, respectively, andthat at 0.125 to the heterotactic triad. The calculated concentrationof each triad starting from the optically active monomer with60.8% enantiomer excess (ee) assuming complete retention ofsilicon stereochemistry in the reduction and in the polymerizationis S : H : I = 1.0 : 2.0 : 3.3 (0.16 : 0.32 : 0.52). The actual concentrationwas 1.0 : 2.0 : 2.3 (0.19 : 0.37 : 0.44).

∗ Correspondence to: Yusuke Kawakami, School of Materials Science, JapanAdvanced Institute of Science and Technology (JAIST), Asahidai 1-1, Nomi,Ishikawa 923-1292, Japan. E-mail: kawakami@jaist.ac.jp

† Dedicated to the occasion of the retirement of Prof. Dick Jones.

School of Materials Science, Japan Advanced Institute of Science andTechnology (JAIST), Asahidai 1-1, Nomi, Ishikawa 923-1292, Japan

Polym Int 2009; 58: 279–284 www.soci.org c© 2009 Society of Chemical Industry

28

0

www.soci.org Y Kawakami et al.

Si

OMen

Me

LiAlH4

retention

(S)

Si

H

Me

(R)

Cl2Si

Cl

Me

(S)

S, R configuration of chiral silicon center is determined by the absoluteconfigurations of chiral centers of menthol (1S,2R,5S) and (1R,2S,5R)

retention

Scheme 1. Stereochemical aspects of the basic transformations of optically active silicon compounds.

Si*CH2CH=CH2

CH3Ph

H

SiCH2

CH3Ph

CH2

SiCH2

CH3PhH2C

CH2

H2C

CH2

SiCH2

CH3Phcatalyst

hydrosilylation

optically active silane isotactic poly(carbosilane)

Scheme 2. Synthesis of optically active allylsilane and isotactic poly[(phenylmethylsilylene)trimethylene].

Figure 1. ORTEP structure of (S)-(methyl)naphthylphenylmenthoxysilane.

Si*CH3

Si

CH3

CH2

n

BuLi

optically active

Scheme 3. Optically active polycarbosilane by anionic ring-opening poly-merization of optically pure 1-methyl-1-(naphthyl)-2,3-benzosilacyclobut-2-ene.

Recently we found that ring opening of an optically pure1-methyl-1-(naphthyl)-2,3-benzosilacyclobut-2-ene using butyl-lithium (BuLi) gave an optically active polymer as shown inScheme 3.10 However, the polymer did not have a well-controlledstereochemical structure.

Figure 2. 1H NMR (750 MHz) SiCH3 signal of isotactic poly[(phenylmethyl-silylene)trimethylene].

The 1H NMR signals are shown in Fig. 3. The signals at −0.34,−0.40 and −0.48 ppm reasonably assignable to the isotactic (I),heterotactic (H) and syndiotactic (S) triad sequence, respectively,are quite similar to those of the polymer obtained from racemicmonomer (Fig. 3(b)).

BuLi attacked the silicon atom in the reaction of 1-methyl-1-(naphthyl)-2,3-benzosilacyclobut-2-ene. Phenyllithium and ben-zyllithium also attacked the silicon atom regioselectively to cleavethe silicon–methylene carbon bond. The bulky silyl anions wereconcluded to attack at the methylene carbon rather than thesilicon leading to the regeneration of silyl anions.11

The initial dimer, 1-[{butylmethyl(naphthyl)}silyl]-2-[{methyl(o-methylphenyl)-(naphthyl)}silyl]methylbenzene, shown in Fig. 4,was separated using SEC on polystyrene gel followed by HPLC onCHIRALCEL OD, as shown in Fig. 5.

In the 1H NMR spectrum, methylene signals were observedas apparently eight peaks at 2.819, 2.811, 2.789, 2.781, 2.777,2.766, 2.747 and 2.737 ppm. These signals were assignableto each diastereoisomer of the dimer having two asymmetric

www.interscience.wiley.com/journal/pi c© 2009 Society of Chemical Industry Polym Int 2009; 58: 279–284

28

1

Control of stereochemical structures www.soci.org

by catalytic ROPof racemic monomer

by anionic ROPof optically pure monomer

by catalytic ROPof optically pure monomer

(b)

(a)

(c)

Figure 3. Stereoregularity of polymers obtained from 1-methyl-1-(naphthyl)-2,3-benzosilacyclobut-2-ene: (a) racemic using platinum–1,3-divinyl-1,1,3,3-tetramethyldisiloxane complex; (b) optically pure usingBuLi; (c) optically pure using a platinum catalyst. (ROP, ring-opening poly-merization).

Si2Bu

Si1 CH3**

CH3 NpNpH3C

Figure 4. Structure of 1-[{butylmethyl(naphthyl)}silyl]-2-[{methyl(o-met-hylphenyl)-(naphthyl)}silyl]methylbenzene.

silicon centers. The first (minus sign) and the second (plus sign)eluted isomers showed well-separated quartets at 2.804 and 2.752

ppm (J = 14.9 Hz) and 2.796 and 2.762 ppm (J = 14.9 Hz), re-spectively. The first and second fractions have opposite signsin optical activity, and the third and fourth fractions have thesame chemical shifts and coupling constants as well as thesame optical sign as the first and second fractions, respectively.It might be reasonable to assign, from the fact that polymerrich in isotacticity showed minus sign in optical activity, andby considering generally accepted inversion stereochemistry ofthe nucleophilic attack on silicon atoms, the first fraction withminus sign as (S,S′)- or (R,R′)-1-[{butylmethyl(naphthyl)}silyl]-2-[{methyl(o-methylphenyl)(naphthyl)}silyl]methylbenzene. The(S,S′)- and (R,R′)-isomers should have similar 1H NMR signals,and opposite sign in optical activity to each other. The fourthfraction, which showed the same 1H NMR and opposite sign ofoptical rotation to that of the first fraction, was concluded to bethe (R,R′)-isomer, if the first fraction was the (S,S′)-isomer, andvice versa. The second and third fractions can be assigned to(S,R′)- or (R,S′)-isomers. The ratio of the four isomers was estimatedfrom 1H NMR spectra to be [(first + fourth fractions), i.e. (R,R′)-+ (S,S′)-]:[(second + third fractions), i.e. (R,S′)- + (S,R′)-] = 60:40.Chromatographic data give the ratio first : second : third : fourth =36 : 26 : 23 : 15. The isomer ratios of the dimer calculated by as-suming the stereoselectivity of the ring-opening reaction of (+)-1-methyl-1-(naphthyl)-2,3-benzosilacyclobut-2-ene by BuLi (67 : 33)as initiation reaction and by benzyllithium (60 : 40) as model prop-agation reaction are (S,S′)- or (R,R′)- = (0.67 × 0.60:0.33 × 0.40),(S,R′)- or (R,S′)- = (0.67 × 0.40:0.33 × 0.60) = 40:13 : 27 : 20. Thesefacts strongly suggest that the stereoselectivities in the ring open-

Figure 5. SEC and liquid chromatograms of 1-[{butylmethyl(naphthyl)}silyl]-2-[{methyl(o-methylphenyl)(naphthyl)}silyl]methylbenzenes.

ing of (+)-1-methyl-1-(naphthyl)-2,3-benzosilacyclobut-2-ene byBuLi and ring opening by benzyl species in propagation wereboth low.

Ring-opening polymerization of 1-methyl-1-(naphthyl)-2,3-benzosilacyclobut-2-ene with platinum–1,3-divinyl-1,1,3,3-tetra-methyldisiloxane (PDT) complex as catalyst (0.1 mol%) gave apolymer with Mn = 1 210 000, dispersity index (D = Mw/Mn) of1.77, and the cyclic dimer gave 1-triethylsilyl-2-{methyl(naphthyl)silylmethyl}benzene, 1-{methyl(naphthyl)(2′-triethylsilylphenyl-methyl)silyl}-2-{methyl(naphthyl)silylmethyl}benzene and 1-{me-thyl(naphthyl)-(2′-triethylsilylphenylmethyl)silyl}-2-[methyl(nap-hthyl)[2′-{methyl-(naphthyl)silylmethylphenyl}]silylmethyl]benzene (the structures of which are shown in Fig. 6) in 54, 18and 11% yield, respectively, through regioselective ring-openingreaction, followed by the σ -bond metathesis process.12

Polym Int 2009; 58: 279–284 c© 2009 Society of Chemical Industry www.interscience.wiley.com/journal/pi

28

2

www.soci.org Y Kawakami et al.

Et3Si Si

Np CH3

Si

Np CH3

H

Si

H Np

CH3

SiEt3

SiH

CH3Np

Et3Si Si

Np CH3

Si

Np CH3

Figure 6. Low molecular weight products from 1-methyl-1-(naphthyl)-2,3-benzosilacyclobut-2-ene using a platinum catalyst in the presence of Et3SiH.

Si

Cl

Me

(S)

Me3SiLi

Inv.Si

SnMe3

MeMeLi

Ret.Si

Li

Me

(R) (S)

Scheme 4. Synthesis of optically active silyllithium from optically active chlorosilane.

When optically pure (+)-1-methyl-1-(naphthyl)-2,3-benzosila-cyclobut-2-ene was used, 1-triethylsilyl-2-{methyl(naphthyl)silyl-methyl}benzene and 1-{methyl(naphthyl)(2′-triethylsilylphenyl-methyl)silyl}-2-{methyl(naphthyl)silylmethyl}benzene have ee oroptical purity (op) higher than 99%. The subsequent propagationalso seems to proceed regio- and stereoselectively; thus, platinum-catalyzed ring-opening polymerization of optically pure (+)-1-methyl-1-(naphthyl)-2,3-benzosilacyclobut-2-ene provided anisotactic and optically pure polymer (as shown in Fig. 3(c)).

Optically active silyllithium derivatives can be versatilesynthetic intermediates to obtain various optically active siliconcompounds. Such optically active silyllithium derivatives make itpossible to synthesize silicon–carbon bonds in which asymmetryis introduced at both silicon and carbon centers, which cannotbe achieved by the substitution reaction by a carbanion at anasymmetric silicon center. Silyllithium derivatives are reported toform by the cleavage reaction of silicon-containing inter-elementlinkages with lithium metal or alkyllithium.13 – 26 The disilane andsilylstannane derivatives were obtained by the reaction of chlorosi-lane with silyllithium or stannyllithium, as shown in Scheme 4. Theformation of silyllithium is a retention process. The optical puritywas kept higher than 85% ee at −78 ◦C even after 5 h.27

The reaction of (S)-chloro-, (S)-bromo- and (R)-fluoro[methyl(naphthyl)phenyl]silane (>99% ee) with the optically activesilyllithium gave almost quantitative yields of the expecteddisilane, 1-methyl(naphthyl)phenyl-2-dimethyl(4-methoxynaph-thyl)disilane. Interestingly enough, the fluoride with the oppositeconfiguration to the chloride also gave the same (R)-antipode of thedisilane as the major product at room temperature, not only in pen-tane and Et2O, but also even in tetrahydrofuran (THF), indicatingretention of configuration at the chiral silicon center. Surprisingly,in contrast to the reaction at room temperature, the opposite(S)-antipode was produced as 97% inverted product at −78 ◦C inpentane. Figure 7 shows the change of percentage of (S)-antipodeas a function of temperature, with or without additives such asHMPA and LiBr, in THF. At −78 ◦C, (S)-1-methyl(naphthyl)phenyl-2-dimethyl(4-methoxynaphthyl)disilane was obtained as 93% in-verted product from (R)-fluoro[methyl(naphthyl)phenyl]silane.

By increasing the temperature, at around −20 ◦C the stere-ochemistry crossover point was observed where (R)- and (S)-antipodes are equal. The reaction proceeded with 90% retentionat 60 ◦C.

Very interestingly, the presence of 3 equiv. HMPA selectivelyafforded the (R)-antipode (100% ee) with completely oppositeconfiguration compared with that without HMPA. At room

3 equiv HMPA

1 equiv HMPA

1 equiv HMPA

inversion

retention

3 equiv LiBr

Figure 7. Change in stereochemistry with reaction conditions in nucle-ophilic substitution of an optically active fluorosilane.

temperature, 1 equiv. HMPA also afforded the (R)-antipode with100% ee.28

Optically active disilanes with two chiral centers, (1R,2R)-1,2-dimethyl-1,2-di(naphthyl)-1,2-diphenyldisilane and (1S,2S)-1,2-di(4-methoxynaphthyl)-1,2-dimethyl-1,2-diphenyldisilane, wereobtained with high optical purity by the reactions of thecorresponding optically active halogenosilanes (chloro or flu-oro) with optically active silyllithiums. Although silicon–siliconbonds, as well as silicon–naphthyl bonds, of the naphthylderivative were cleaved without selectivity on bromination,the silicon–(4-methoxynaphthyl) bond of (R)-1,2,2-trimethyl-2-(4-methoxynaphthyl)-1-(naphthyl)-1-phenyldisilane was regiospecif-ically cleaved, followed by the stereoselective cleavage of theremaining chiral silicon–naphthyl bond (94% inversion).

Although the silicon–(4-methoxynaphthyl) bond of (1S,2S)-1,2-di(4-methoxynaphthyl)-1,2-dimethyl-1,2-diphenyldisilane (>99%ee) was regioselectively cleaved without silicon–silicon bondscission, unfortunately a marked racemization could not beavoided during the one-pot reaction, as shown in Scheme 5.29

Cleavage by TfOH gave completely racemized products at roomtemperature.

CATALYTIC CROSS-COUPLINGPOLYMERIZATIONMost of the methods employed for the synthesis of polysiloxanesinvolve the self-polycondensation of silanol functional groups.

www.interscience.wiley.com/journal/pi c© 2009 Society of Chemical Industry Polym Int 2009; 58: 279–284

28

3

Control of stereochemical structures www.soci.org

CHCl3,-64°C Et2O,0°C

LiAlH4Si

MePh

MeONp Si

NpOMe

PhMe

Br2 SiH Si

HPh Me

Me Ph

SiBr Si

BrPh Me

Me Ph

>99% ee

Scheme 5. Regioselective but non-stereoselective cleavage of optically active disilane with two methoxynaphthyl groups.

HOSi

O

CH3

SiOH

H3C

SiO

CH3

Si

H3C

On

HSi

O

CH3

SiH

H3C Rh

Scheme 6. Synthesis of stereoregular polymethylphenylsiloxane by cross-coupling.

(a) (b)

Figure 8. 13C NMR spectra of polymethylphenylsiloxane: (a) random;(b) syndiotactic rich.

Meanwhile, it has been reported that many compounds readilycatalyze (i) the hydrolysis of SiH to produce silanols (SiOH) or(ii) the dehydrocoupling of SiH/SiOH to form siloxane (SiOSi)linkages. Tris(pentafluorophenyl)borane, B(C6F5)3, was also foundto be very effective for the formation of siloxane bonds by cross-condensation between silane and silanol or alkoxysilane.30 – 32

Cross-dehydrocoupling polymerization of 1,3-dimethyl-1,3-diphenyl-1,3-disiloxanediol with 1,3-dihydro-1,3-dimethyl-1,3-diphenyl-1,3-disiloxane in the presence of [RhCl(cod)]2 gavepolymethylphenylsiloxane, as shown in Scheme 6.

Assignment of the triad signals of the polymer was made by 13CNMR spectroscopy of ipso carbons of phenyl groups (S = 136.7,H = 136.9 and I = 137.1 ppm), as shown in Fig. 8.33,34

Although the reaction of optically pure (S,S)-1,3-dimethyl-1,3-diphenyl-1,3-disiloxanediol with 1,3-dihydro-1,3-dimethyl-1,3-diphenyl-1,3-disiloxane [(S,S) : (S,R) : (R,R) = 84 : 16 : 0], obtained in80% yield by the reduction of the disiloxanediol [((S,S) + (R,R)):(S,R)= 91:9] using LiAlH4, gave a polymethylphenylsiloxane of ratherlow molecular weight, its triad tacticity was found to be rich insyndiotacticity (S : H : I = 60 : 32 : 8). Some racemization seems tohave occurred for the silane derivative.

Stereochemistry in silylation reactions of silanols and methoxysi-lanes with optically pure (R)-methyl(naphthyl)phenylsilane in thepresence of B(C6F5)3 were studied, as shown in Scheme 7.

The reaction with (R)-(methoxy)methyl(naphthyl)phenylsilane(88% ee) gave (R,R)-1,3-dimethyl-1,3-di(naphthyl)-1,3-diphenyl-disiloxane [(R,R):(R,S):(S,S) = 87:12 : 0.5]. The stereochemistrywas proved to be almost completely inversion and retention for thechiral silicon centers of the silane and methoxysilane, respectively.Unfortunately, application of this reaction to the actual synthesis ofcompletely syndiotactic polymethylphenylsiloxane has not beensuccessful yet.30

Scheme 7. Stereochemistry of siloxane bond formation from opticallyactive methoxysilane and silane.

CONCLUSIONSOptically active silicon compounds have been used to synthesizeoptically active and/or stereoregular silicon-containing polymersby polyaddition, ring-opening polymerization and polyconden-sation reactions. Optically active silyl anions can be synthesizedfrom optically active chlorosilanes, and can be further used tosynthesize optical active oligosilanes.

REFERENCES1 Szwarc M, Nature 178:1168–1169 (1956).2 Szwarc M, Levy M and Milkovich R, J Am Chem Soc 78:2656 (1956).3 Szwarc M and Beylen MV, Ionic Polymerization and Living Polymers.

Chapman & Hall, New York (1993).4 Corriu RJP, Guerin C and Moreau JJE, Dynamic stereochemistry at

silicon, in The Chemistry of Organic Silicon Compounds, vol. 1, ed. byPatai S and Rappoport Z. Wiley, Chichester, p. 305 (1989).

5 Brook MA, Silicon in Organic, Organometallic, and Polymer Chemistry,part 1. Wiley, New York, p. 115 (2000).

6 Sommer LH, Stereochemistry, Mechanism and Silicon. McGraw-Hill, NewYork (1965).

7 Shinke S and Kawakami Y, Chromatographia 53:140 (2001).8 Kawakami Y, Takeyama K, Komuro K and Ooi O, Macromolecules

31:551 (1998).9 Kawakami Y, Takahashi T, Yada Y and Imae I, Polym J 30:1001 (1998).

10 Kakihana Y, Uenishi K, Imae I and Kawakami Y, Macromolecules38:6321 (2005).

11 Uenishi K, Imae I, Shirakawa E and Kawakami Y, Chem Lett 30:986(2001).

12 Uenishi K, Imae I, Shirakawa E and Kawakami Y, Macromolecules35:2455 (2002).

13 Gilman H and Lichtenwalter GD, J Am Chem Soc 80:607 (1958).14 George MV, Peterson DJ and Gilman H, J Am Chem Soc 82:403 (1960).

Polym Int 2009; 58: 279–284 c© 2009 Society of Chemical Industry www.interscience.wiley.com/journal/pi

28

4

www.soci.org Y Kawakami et al.

15 Chalk AJ and Harrod JF, J Am Chem Soc 89:1640 (1967).16 Hengger E and Holtschmidt N, J Organomet Chem 12:5–7 (1968).17 Kloninger C and Rehahn M, Macromolecules 37:1720–1727 (2004).18 Wang H, Winnik MA and Manners I, Macromolecules 40:3784 (2007).19 Kloninger C, Knecht D and Rehahn M, Polymer 45:8323 (2004).20 Datta U and Rehahn M, Macromol Rapid Commun 25:1615 (2004).21 Bellas V and Rehahn M, Angew Chem Int Ed 46:5082 (2007).22 Kloninger C and Rehahn M, Macromol Chem Phys 208:833 (2007).23 Buchi G and Wuest H, J Org Chem 44:546 (1979).24 Choi J-K, Hart DJ and Tsai Y-M, Tetrahedron Lett 23:4765 (1982).25 Tamao K and Kawachi A, Angew Chem Int Ed Engl 34:818 (1995).26 Kawachi A and Tamao K, Bull Chem Soc Jpn 70:945 (1997) and

references therein.

27 Omote M, Tokita T, Shimizu Y, Imae I, Shirakawa E and Kawakami Y,J Organomet Chem 611:20 (2000).

28 Suzuki K and Kawakami Y, Organometallics 22:2367 (2003).29 Suzuki K, Kawakami Y, Velmurugan D and Yamane T, J Org Chem

69:5383–5389 (2004).30 Shinke S, Tsuchimoto T and Kawakami Y, Silicon Chem 3:243 (2007).31 Zhou DQ and Kawakami Y, Macromolecules 38:6902 (2005).32 Chojnowski J, Rubinsztajn S, Cella JA, Fortuniak W, Cypryk M and

Kurjata J, et al., Organometallics 24:6077–6084 (2005).33 Oishi M and Kawakami Y, Org Lett 1:549–551 (1999).34 Oishi M, Moon JY, Janvikul W and Kawakami Y, Polym Int 50:135 (2001).

www.interscience.wiley.com/journal/pi c© 2009 Society of Chemical Industry Polym Int 2009; 58: 279–284