SciELO - Scientific Electronic Library Online

 
vol.45 número3SEPARATION BY CAPILLARY ELECTROPHORESIS OF C-GLYCOSYLFLAVONOIDS IN Passiflora sp. EXTRACTSTEMPLATE SYNTHESIS, STRUCTURAL AND BIOLOGICAL STUDIES OF NEW TETRAAZAMACROCYCLIC COMPLEXES OF LEAD(II) índice de autoresíndice de assuntospesquisa de artigos
Home Pagelista alfabética de periódicos  

Serviços Personalizados

Journal

Artigo

Indicadores

Links relacionados

Compartilhar


Boletín de la Sociedad Chilena de Química

versão impressa ISSN 0366-1644

Bol. Soc. Chil. Quím. v.45 n.3 Concepción set. 2000

http://dx.doi.org/10.4067/S0366-16442000000300018 

PSEUDO-DIEQUATORIAL CONFORMATION PREFERENCE
IN 3-SUBSTITUTED THIETANE – 1-OXIDE :
A THEORETICAL STUDY

J. Guillermo Contreras*(1) and Sandra T.Madariaga(2)

1) Facultad de Ciencias Químicas, Universidad de Concepción
Casilla 160-C , Concepción, Chile

2) Facultad de Oceanografia y Pesquería, Universidad Austral
(Puerto Montt), Chile
(Received: March 15, 2000 - Accepted: June 28, 2000)

In memorian of Dr. Guido S. Canessa C.

ABSTRACT

The interconversion between the two isomeric forms (cis and trans) of a series of 3-substituted thietane –1-oxide derivatives have been theoretically studied by means of ab initio molecular orbital theory. The optimized geometries were obtained at the HF/6-31G** level, whereas energies were calculated using several basis sets all including electron correlation at the second order Moller Plesset theory. To derive the thermodynamics of the cis Û trans interconversion, the energies calculated at MP2/6-311++G** were used. To determine the effect of the substitution on C3 , the chloro, methyl, ethyl and t-buthyl derivatives have been studied. Ab initio calculations show that in all cases the cis isomer is more stable than the trans regardless the size of the substituent. The axial preference (trans) of the oxygen atom of the S=O group, found in six-membered heterocycles is not found in the four-membered thietane-1-oxide (TOX). The 1,3 repulsive interaction and the Hb - electron pair on sulfur repulsion are unimportant in the determination of the preferred geometry of these species. The Hb - O9 non-bonded interaction does determine the cis geometry to be the most stable one.

A small solvent effect on the cis Û trans interconversion is predicted for the 3-chloro and 3-methyl TOX, regardless the polarity of the medium. Temperature exerts an important role inthe 3-chloro-TOX system but just a small effect on the 3-alkyl derivatives. The use of the IR spectra in the S=O stretching modes regions is proposed to distinguish between cis and trans isomeric forms in 3-substituted TOX.

Keywords: Ab Initio Calculations, Conformational Analisis, 3-Substituted Thietane

RESUMEN

Se ha estudiado teoricamente en el marco de la Teoria de Orbitales Moleculares, la interconversion entre las dos formas isomericas (cis y trans) de una serie de derivados del tietano-1-oxido substituidos en 3. En este estudio las geometrías optimizadas se obtuvieron al nivel HF/6-31G**,mientras que las energías se calcularon usando varios set de bases con inclusion de correlación electronica Moller-Plesset de segundo orden. Para calcular la termodinámica de la interconversión cis Û trans ,se usaron las energías determinadas al nivel MP2/6-311++G**. Para la determinación del efecto de la substitucion en C3 , se han estudiado los derivados clorados,metilados, etilados y t-butilados. Los calculos ab initio muestran que el isomero cis es mas estable que el trans al margen del tamaño del substituyente. La preferencia axial (trans) del átomo de oxígeno del grupo S=O que se observa en los heterociclos de seis miembros no se encuentra en los anillos de cuatro miembros como el tietano-1-óxido (TOX). La interacción repulsiva 1,3 y la repulsión entre el Hb---- par electrones sobre el átomo de azufre no es importante en la determinación de la geometría preferencial de estas especies. La interaccion no-enlazante Hb----O9 determina que la geometria cis sea la mas estable.

Se predice un pequeño efecto del solvente sobre la interconversión cis Û trans, para los derivados 3-cloro y 3-metil TOX ,sin importar la polaridad del medio. La temperatura ejerce un efecto importante sobre el sistema 3-cloro TOX, pero sólo un efecto modesto sobre los derivados alquílicos. Se propone el uso del espectro IR, en la región de los modos de tensión S=O, para distinguir entre las formas isoméricas cis y trans en derivados del TOX substituidos en 3.

PALABRAS CLAVES: Calculos ab initio, análisis conformacional, tietano-1-oxido-3-substituído.

INTRODUCTION

It is well known that the oxygen atom of a sulfoxide group is axially located in sulfur-containing six-membered rings (1-7). In fact all 4-substituted sulfur, phosphites and dioxanes the "cis" isomer is more stable than the "trans". The cis form is a chair conformer with the oxygen ( or a methoxy group) axial and the substituent in C4 is equatorial, whereas the less stable trans is also a rigid chair with both oxygen and the substituent axially located. The relative greater stability of the sulfite with respect to the phosphite analog is due to the larger DGo value for the cis Û trans equilibrium in the former system (8). It seems reasonable to think that the structural differences between phosphites, sulfites and dioxanes come from the changes in bond distances and angles around the phosphorus, sulfur and carbon atoms. The DGo (cis –trans) is ca. 1.8 kcal/mol at 40 oC for 4-t-buthyl-trimethylene cyclic sulfite which compare well with the calculated value of 1.4 – 1,9 kcal/mol for 1,4 –dioxane (9) and the 1,4-phosphite systems (6).

The axial preference shown for the S=O group in these six-memberered rings is not found in four membered heterocyclies where a pseudo-equatorial preference has been observed (10). In fact, chemical or thermal equilibration studies (10) of 3-t-buthyl thietane-1-oxide in dioxane/HCl at 25ºC produce a cis/trans ratio of 85%/15% no matter the ratio of the starting material, whereas in 3-p-chlorophenyl thietane-1-oxide the starting trans isomer renders predominantly the cis isomer, under the same conditions. The different isomers were separated by vapor phase chromatography and assigned by either dipole moment measurements (10) or 1H-NMR spectroscopy (11). The dipole moment method has its obvious limitation and in particular in the later example where the isomerization at room temperature yields predominantly the cis isomer. Since the dipole moments were measured in benzene and no isomerization studies in this solvent were carried out the question that arises is: is the trans form isomerizating to the cis one in this solvent as well?. The calculated m values for the cis and trans isomeric forms are in agreement with the experimental ones within 7 – 12%, assuming a puckering angle of 37º. Our calculations produce different m values and the puckering angles are ca. 25º.

On the other hand, a series of 3-substituted thietane-1-oxide were assigned either to the cis or trans structures based on the 1H-NMR chemical shifts of the ring protons (11) . The assignments were based on the different chemical shifts of the Ha and Hb protons. In fact, in the trans conformation (Hb cis to S=O) , a significant deshielding effect of the S=O group on the Hb would produce the corresponding signal to show up 47 –68 Hz downfield with respect to the Hb , cis the lone electron pair on sulfur atom. However, in most cases the Hb is burried under the Ha multiplet and an a-deuterated derivative needs to be examined to determine its position. The similarity of the spectra argues against a variation in the conformational equilibrium with the change in the substituent (11).

In order to stablish which isomeric form is the most stable species both in the gas phase and in solution, we have carried out ab initio calculations on the 3-chloro, 3-methyl , 3-ethyl and 3-t-buthyl thietane-1-oxide. The results indicate that the cis geometry is the most stable species either in the gas phase and in solution of various solvents and that in the 3-chloro derivative, the trans form is present in ca. 12%.

COMPUTATIONAL METHODS

Ab initio geometry optimization at HF/6-31G** level were carried out using GAUSSIAN 98 codes (12). The initial geometry’s of the cis and trans isomers were those calculated previously for thietane-1-oxide (13) and /or from AM1 calculations. Frequency calculations and IR intensities were predicted at the equilibrium geometries yielding all real frequencies; hence the calculated structures are real minima. Energy calculations were performed using various basis sets to determine the inclusion of both polarization and diffuse functions effect. In all cases, electronic correlation at the second order Moller Plesset theory was taken into account. The frozen core approximation was used. The single point energies calculated at MP2/6-311++G** were corrected for zero point vibrational energies (unscaled). Scaling by 0.8829 , as usual, to account for the overestimation of the Hartree-Fock vibrational frequencies produced an almost constant value that makes no differences on the conclusions of this work. Isomerization enthalpies were calculated by adding ZPE and thermal corrections to the relative energies calculated at MP2/6-311++G** level. The DGº values were obtained, as usual, from DGº = DHo - TDSº .

The solute-solvent effect on the cis – trans interconversion reaction was taken into account by using Tomasi´s polarizable continuum model (PCM) (14) modified by Wiberg et al (15,16 ). The isodensity polarizable continuum method (IPCM) calculates the electric field analytically and the cavity is defined upon an isosurface of the total electron density calculated at the level being used. Therefore, the cavity is uniquely derived from the electronic environment and just the isosurface level, i.e., charge density (0.0004-0.001 e/B3) need to be specified. In this model, the solvent effect is derived from the surface potentials and the dielectric continuum interactions. This, is equivalent to going to infinite order in the electric moments expansion. The IPCM method was applied to the gas phase molecular geometries since structural parameters change very little in going from the gas phase to solution and accordingly no large effect on solvation energies can be expected (17). The free energies of solvation (DGos) were calculated from Ds = DE( soln. – gas) and the free energies in solution ( Gosoln.) from Gosoln. = DGgas + DGos (18).

RESULTS AND DISCUSSION

3.substituted thietane-1-oxide can exist in its cis and trans isomeric forms. In the cis, the oxygen of the sulfoxide group is in a quasi-equatorial position and the lone pair of electrons on the sulfur atom is cis to the b-hydrogen (Hb). In the trans form,the sulfoxide oxygen atom is axially located, is cis with Hb and the lone pair of electron on sulfur is trans to this hydrogen atom. For the sake of clearness, we have call cis to the isomer where the oxygen atom is cis to the substituent on C3 and hence the trans one is that having the oxygen trans to the substituent . Figure 1. Shows the cis and trans structures of 3-chloro thietane-1-oxide with its atom numbering ( methyl,ethyl and t-buthyl derivatives of TOX drawings can be obtain from the authors upon request). From table 1, it can be inferred that in general the ring geometry remains fairly constant for all the cis and trans isomers of methyl ,ethyl and t-buthyl derivatives. Very small variations are observed for the 3-chloro species (ca. 0.014 Å for the C – C bond distances). In all cis isomers the S-O distances are ca. 1.478 Å (av), whereas in the trans compounds the average distance is ca. 1.484 Å. These S-O distances are typical of a normal sulfoxide group(18,19). The calculated non-bonded S1---C3 distances are ca. 2.40 Å. These short distances are a remarkable feature ,also found in dithietane (13), come from the a puckering S1C4C3C2 angle of ca 24º. The S1---Hb non-bonded distance would approximately be a measurement of the distance between the electron pair on sulfur and the hydrogen atom.. In the 3-chloro TOX, this distance is ca. 2.84 Å both in the cis and trans forms,whereas for the alkyl derivatives small differences (ca. 0.06 Å) are observed. This would imply that roughly the electron lone pair is far away in the trans as compared with the cis isomers.


Fig. 1 Optimized geometries for-3-Chloro Thietane-1-Oxide and Atom Numbering for all derivatives.


Table I. Selected Bond Distances (Å) and Angles (º) for the 3-Substituted Thietane 1-Oxide (TOX) Calculated at HF/6-31G**

It is known that at least for the 3-chloro and 3-t-buthyl TOX, the cis amd trans isomeric forms are separable by vapor phase chromatography (10). Samples prepared with different cis/trans ratios dissolved in HCl/dioxane produce the same proportion of the cis and trans forms (5:1) after equilibration at room temperature. The Cis Û trans isomerization would catalyzed by the HCl. Johnson and Siegel (11) have used the 1H-NMR spectra to infer the cis or trans configuration of several 3-substituted TOX.. The method to distinguish between cis and trans geometries for these species is based on the different chemical shifts of the Ha and Hb protons. However the Hb signal always shows up burried into the broad Ha multiplet. In an attempt to determine the stability of these isomeric species in the gas phase and in solution and see whether the theoretical results give or not support to the experimental ones, we have carried out detailed ab initio calculations using various basis sets including electron correlation. From table 2, it can be inferred that the inclusion of electron correlation at the second order Moller-Plesset theory at the frozen core approximation, produces a stabilization of both isomeric forms and that the use of a flexible basis set like 6-311++G** is necessary for a good description of the energetics of these systems. In the alkyl TOX derivatives, the free energy changes, taking the cis isomer as reference, reveal that the trans form is less stable than de cis by ca. 2.5 kcal/mol. This implies that the trans isomer is present in just ca. 2%, being the cis the predominant species. In the 3.chloto TOX, the cis Û trans equilibrium is also displaced to the cis formation, though the isomerization DGº value indicates thet the trans isomer is present in ca. 16%. These results support the experimental fact that the cis is the most stable species regardless the size of the substituent. In fact, 2.5 kcal/mol is not a large difference enough to prevent the independent existence of both isomeric forms. The smallest chlorine substituent seems to alter the above trend since the DGoisomerization is ca. 0.9 kcal/mol in the gas phase. Johnson and Siegel (10) have suggested that thietane behaves likewise 1.3-disubstituted non-planar cyclobutane where the preferred geometry is also cis diequatorial. The cis conformation preference has been attributed to the tendency to minimize the 1,3 non-bonded repulsion (20,21). This would imply that in TOX the repulsive interaction between S1 and C3 and or its H attached to it produce the trans isomer to become less stable than the diequatorial cis form. This is clearly a difference of the non-planar four-membered rings with respect to the six-membered heterocycles where the 1,3 repulsive interaction must be weaker as a consequence that the heteroatom and C3 are far away one from the other. This fact also produces a larger puckering angle in the later as compared with the former. In fact, we have found puckered angles of ca. 62º in isomeric trimethylene cyclic sulfites(18) , whereas in 1,3-substituted dithietane this angle is ca 20º. In 3-substituted TOX the puckered angles vary from 21 – 25º. It is worth noting that in all alkyl TOX derivatives the differences in the cis and trans puckered angles are ca. 4º, whereas in the 3-choloro TOX, this difference is just 1º. This implies that in this case S1 and C3 are closer one to another and that the 1,3 repulsive interactions should be greater. Although, this is just the only case studied here where the cis Û trans isomerization takes place in a reasonable extension and therefore we think this is not the effect that produces the stabilization of the cis isomer over the trans one. As we stated before, if we take the S1----Hb as the approximate distance between Hb and the lone pair of electron on sulfur, from table 1 it can be seen that in the cis-trans pair for the 3-chloro TOX this distance is just above the same in both isomers, whereas in the alkyl derivatives the electron pair is closer to Hb in the cis than in the trans. Accordingly, the Hb---.electron pair repulsion is not either determining the preferred conformation in these species. It seems clear that the Hb---O9 non-bonded interaction determines the diequatorial (cis) geometry in these compounds.


Table II. Energeticsa,b and Dipole Moments for the 3-Substituted Thietane 1-Oxide (TOX) in the Gas Phase

TEMPERATURE EFFECT

Since equilibration studies in decalin at 170 – 175 ºC of cis/trans mixtures in different ratios always produce ca. 85% of cis and 15% trans, we have also carried out calculations at different temperatures to see whether the cis Û trans equilibrium can be reversed. Table 3 shows that at least for the 3-chloro TOX the % of the trans isomer increases from a 16.4% at 25 ºC to ca. 29% at 225 ºC. For the alkyl derivatives the temperature effect is not as strong as in the 3-chloro derivative. In fact, for the 3-t-buthyl TOX just an 8% conversion takes place at 225 ºC, whereas for the 3-chloro TOX ca. 29 % of the trans is predicted at 225 oC. In other words, for the alkyl derivatives an increase in the temperature produces a modest reverse of the equilibrium ,while in the chloro derivative a strong effect is predicted.


Table III. Temperature Effect on the Cis Û Trans Equilibrium of 3-chloro TOX

SOLVENT EFFECT

To study the effect of the solvent on the cis to trans interconversion, we have calculated the Dsoln. values for all species studied here. We have used carbon tetrachloride (e = 2.228) , acetone (e = 20.70) , acetonitrile (e =36.64) and dimethylsulfoxide (e = 46.70) as solvents. Table 5 gives the free energies of solvation (Ds) and the free energies in solution (Gºsoln). The relative Dsoln. for the cis ¤ trans equilibrium is also listed. From table 4, it can be inferred that for the 3-chloro TOX the % of trans isomer remains constant at ca. 11% regardless the change in the polarity of the solvent, whereas for the methyl the concentration of this species is just 2% , i.e., similar to the gas phase results . In the ethyl and t-buthyl TOX, the solvent exerts no influence at all. The above trends can be rationalized in terms of the similar dipole moments of the cis and trans isomeric forms calculated at MP2/6-311++G**for these species. The cis and trans 3-chloro TOX show a slightly different dipole moments and that is what makes the difference with respect to the alkyl derivatives.


Table V.
v (S = O) Infrared Bands (cm-1)a

INFRARED SPECTRA

Trimethylene cyclic sulfite (1,18) shows bands at ca. 1190 and 1230 cm-1 that have been assigned to the axial and equatorial S=O groups, respectively. In fact, in cyclic sulfites the S=O stretching vibrations show up in the range 1195 – 1215 cm-1 in six membered rings (22) , whereas in five-membered rings a band at 1213 – 1215 cm-1 has been attributed to this vibration. In TOX derivatives, the S=O group behaves as a typical sulfoxide and hence a strong band at ca 1050 cm-1 is to be observed. This band that is remarkably constant for non-cyclic sulfoxides shifts up to 50 cm-1 in cyclic ones depending on the geometry of the S=O group. In fact, the range 1000 – 1090 cm-1 defines roughly the extreme frequencies at which the S=O is likely to absorb (22, 23). The good agreement between the calculated and the above range of absorption indicates that in TOX this group behaves as a typical cyclic sulfoxide . Table 5 gives the calculated S=O IR bands for the conpounds studied here. Careful analysis of the atomic Cartesian displacements and the predicted IR intensities show that the n (S=O) is strongly coupled to some kind of H-C-H bending modes in both isomers. From table 5 it can be inferred that for the cis forms the n (S=O) should be observed in the range 1055 – 1080 cm-1, whereas for the trans the range is 1040 – 1050 cm-1. Accordingly, we feel that IR spectra can also be used to distinguish between cis and trans geometries. Experimental IR work is called for.

CONCLUSIONS

1.The axial preference for the oxygen atom of the sulfoxide group found in six-membered rings is not found in 3-substituted thietane- 1-oxide.

2. In the four non-planar sulfur-containing heterocycles like thietane. The cis isomeric form (oxygen pseudo- equatorial) is more stable than the trans form.

3. The size of the substituent on C3 has no effect on the the stability of isomeric forms.

4. 1,3- repulsive interaction is also not important to the stabilization of the cis form over the trans one.

5. Hb-----electron lone pair seems also not to be important effect to be taken into account.

6. Hb-----O9 non-bonded interaction would be the most important factor in determining the diequatorial (cis) geometry in these species.

7. Temperature plays an important role for the 3-chloro TOX, but exerts just a small effect on the alkyl derivatives.

8. Solvents of different polarities exert a small influence in the cis ¤ trans equilibrium. An exception is the 3-chloro TOX whose trans isomer is in ca. 11 %.

9. IR spectra in the S=O stretching modes zone can also be used to distinguish between cis and trans geometries.

ACKNOWLEDGEMENT

This work was supported by an operating grant (No 97.21.006-1.2) from the Universidad de concepcion Chile)

REFERENCES

1. D.G.Hellier, J.G.Tillet , H.F.Van Woerden and R.F.M. White , Chem & Indust. 1956 (1963)        [ Links ]

2. P.C.Lauterbur , J.G.Pritchard and R.L.Vollmer , J. Chem. Soc. 5307 (1963)        [ Links ]

3. C.G.Overberger , T.Kurtz and S.Yaroslavsky , J. Org. Chem. 30, 4363 (1965)        [ Links ]

4. J.G.Pritchard , and R.L.Vollmer , J. Org. Chem. 28, 1545 (1963)        [ Links ]

5. J.G.Martin and J.J.Uebel , J. Am. Chem. Soc. 86, 2935 (1964)        [ Links ]

6. W.G.Bertrude and J.H.Hengis , J. Am. Chem. Soc. 92, 7136 (1970)        [ Links ]

7. J.G.Contreras and S.T.Madariaga , J. Phys. Org. Chem. 11, 392 (1998)        [ Links ]

8. H.F.Van Woerden , H.Cerfontain , C.H.Gree and R.J.Riejerkerk, Tetrahedron Letters 22, 6107 (1968)        [ Links ]

9. E.L.Eliel and M.C.Knoeber , J. Am. Chem. Soc. 90, 3444 (1968)        [ Links ]

10. C.R.Johson and W.O.Siegl, J. Am. Chem. Soc. 91, 2796 (1969)         [ Links ]

11. C.R.Johson and W.O.Siegl, Tetrahedron Letters 23, 1879 (1969)        [ Links ]

12. M.J.Frisch, G.W., Trusch, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheesman, V.G. Zakrzewski, J.A. Montgomery, Jr., R.E. Stratmann, J.C. Burant, S. Dapprich, J.M. Millam, A.D. Daniels, K.N. Kudin, M.C. Strain, O. Farkas, J. Tomasi, V. Barone, M.Cossi, R.Cammi, B. Mennucci, C. Pomelli, C. Adamo, S. Clifford, J.Ochterski, G.A. Petersson, P.Y. Ayala, Q. Cui, K.Morokuma, D.K. Malick, A.D.Rabuck, K. Raghavachari, J. B. Foresman, J. Cioslowski, J.V. Ortiz, B.B. Stefanov, G.Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. Gomperts, R.L. Martin, D.J. Fox, T. Keith, M.A. Al-Laham, C. Y. Peng, A. Nanayakkara, C. Gonzalez, M. Challacombe, P.M.W. Gill, B. Johnson, W. Chen, M.W. Wong, J.L. Andres, C. Gonzalez, M. Head-Gordon, E.S. Replogle, and J.A. Pople, GAUSSIAN 98 revision A.7. Gaussian, Inc., Pittsburgh PA, 1998.        [ Links ]

13. J.G.Contreras and S.T.Madariaga, J. Mol.Struct. (THEOCHEM) 466, 111 (1999)        [ Links ]

14. a) S.Mertius , B.Scrocco and J. Tomasi, Chem. Phys. , 55, 117 (1981)        [ Links ]

b) J.Tomasi , R.Bonaccorsi R.Cammi and F.J.O del Valle, J. Mol. Struct. 234, 401(1991)        [ Links ]

c) J.Tomasi and M.Persico, Chem. Rev. 94, 2027 (1994)        [ Links ]

15. J.B.Foresman , T.A.Keith, K.B.Wiberg, J.Snoolan and M.J.Frisch, J. Chem. Phys. 100, 16090 (1996)        [ Links ]

16. a) K.B.Wiberg, P.R.Ralden and T.A.Keith, J. Am. Chem. Soc. 117, 4261 (1995)        [ Links ]

b)K.B.Wiberg, T.A.Keith, M.J.Frisch and M.Murko, J. Phys. Chem. 99, 9072 (1995)        [ Links ]

c)K.B.Wiberg, H.Castejon and T.A.Keith, J. Comput. Chem. 17, 185 (1996)         [ Links ]

17. J.G.Contreras and S.T.Madariaga, Biorg. Chem. 26, 345 (1998)        [ Links ]

18. J.G.Contreras and S.T.Madariaga, J. Phys.Org. Chem. (in press)        [ Links ]

19. Molecular Structures and Dimensions, vol A1,Crystallographic Data Centre Cambridge, 1972.        [ Links ]

20. J.B.Lambert and J.D.Roberts, J. Am. Chem. Soc. 87, 3884 (1965)        [ Links ]

21. K.B.Wiberg and G.M.Lampman, J. Am. Chem. Soc. 88, 449 (1966)        [ Links ]

22. L.J.Bellamy, "The Infra-Red Spectra of Complex Molecules" , Methuen & Co LTD., 1962.        [ Links ]

23. N.B.Colthup, L.H.Daly and S.E.Wiberley , "Introduction to Infrared and Raman Spectroscopy", Academic Press, New York, 1964.        [ Links ]

Creative Commons License Todo o conteúdo deste periódico, exceto onde está identificado, está licenciado sob uma Licença Creative Commons