Crystal structure determination as part of an undergraduate laboratory experiment: 1′,3′,3′-trimethylspiro[chromene-2,2′-indoline] and 1′,3′,3′-trimethyl-4-[(E)-(1,3,3-trimethylindolin-2-ylidene)methyl]spiro[chroman-2,2′-indoline]

The crystal structures of the title compounds were determined as part of an experiment in an undergraduate teaching laboratory that demonstrates the relationship between molecular structure and function. 1′,3′,3′-Trimethylspiro[chromene-2,2′-indoline] is both a photoswitch and thermochromic molecule. Students synthesized it and a bis-indoline adduct and compared the crystallographically determined structures to computed gas-phase models.


Chemical context
In an ever evolving pursuit to improve the educational experience in undergraduate organic chemistry laboratory courses, we introduced an experiment in which students prepare a 'functional molecule,' in this case spiropyran 1.
Compounds such as 1 are broadly characterized as 'responsive,' due to their ability to be actuated by a range of stimuli, including light, heat, metal ions, pH, mechanical force, and changes in solvent polarity (Klajn, 2014). An advantage of the spiropyran system over other photochromic/thermochromic materials is the strongly differentiated electronic forms between which equilibrium is shifted. The closed-ring isomer of 1 comprises an indoline and a chromene ring bound together at a spiro junction, while the open-ring form is a zwitterionic merocyanine 1a (Scheme 1).
Although a variety of substituted spiropyran derivatives are known in the literature, for simplicity, we elected to focus on the unsubstituted parent compound, which is colorless in its closed form and red in its open form. The molecule was synthesized in a single step by condensation of 1,3,3-trimethyl-2-methyleneindoline with salicylaldehyde (Koelsch & Workman, 1952). The methyleneindoline nucleophile can also react a second time with 1 to give the bis adduct 2 as a side product (Scheme 2).
Since this experiment was oriented around the functional attributes of 1, it presented an ideal opportunity to introduce structural characterization methods into the laboratory course, since the function of 1 is directly linked to its structure. Students first model the two forms of 1 using both molecular mechanics and semi-empirical quantum mechanical methods. These calculations indicate that the spiropyran form of 1 is more stable than the open form 1a. They then grow crystals of 1 by slow evaporation from acetone, resulting in most cases in large (up to 10 mm Â 10 mm), thin pink plates. Although the students do not themselves determine the X-ray crystal structure, crystallographic characterization of 1 has allowed students to compare gas-phase models with condensed-state empirical data.
Differences between molecular mechanics force field MM2 calculations and the semi-empirical quantum mechanical methods PM6 and PDDG versus experimental X-ray values for selected bond lengths and angles can be seen in Table 1. A clear trend in the data is reflected in the fact that thermal motion in low-temperature X-ray diffraction experiments tends to lead to an apparent bond shortening. Considering only those distances not involving phenyl carbon atoms, the data indicate that MM2 shows the poorest mean agreement with X-ray in bond lengths (AE0.043 Å ), while PDDG (AE0.021 Å ) and PM6 (AE0.017 Å ) perform better. The most serious modeling failure was in the MM2 N1-C2 bond which, at 1.270 Å , was interpreted by molecular mechanics to be a double bond, but which was clearly a single bond in the X-ray structure at 1.405 (2)  The molecular structure of 1. Displacement parameters are shown at the 50% probability level.

Figure 2
The molecular structure of 2. Displacement parameters are shown at the 50% probability level. deviations ranged from 0 to 5 and averaged ca 2 for all three methods. Interestingly, if the two angles in poor agreement around C1 are discarded, MM2 actually performs somewhat better than the semi-empirical models for angle data. If all data in Table 1 are taken into account, PM6 is seen to outperform both PDDG and MM2.

Supramolecular features
The KPI of 1 is 68.7% and that of 2 is 69.6% (van der Sluis & Spek, 1990). Neither structure has significant directional intermolecular interactions.

Database survey
There are 67 structures in the CSD (Groom et al., 2016) with the basic skeleton of compound 1. All of these are substituted in one way or another. There are no unusual differences among these structures. Since the C1-O1 bond is broken in the transformation to the merocyanine form, it is of interest to examine this bond length. Of the 82 hits with similar geometry, the mean C-O distance in the CSD is 1.479 (15) . For 1, this distance is 1.4708 (19) Å . For 2, the same distance is 1.4648 (12) Å . There are five structures in the CSD that involve further methyleneindoline substitution, similar to 2. In all cases, the structures are racemic and the chirality is either RR or SS. Two of the deposits (NESZOC and NESZOC01; Ashraf et al., 2012) describe the results from two different crystals, two different radiations (Cu K and Mo K), and two different temperatures (153 and 113 K), respectively. Structurally, there is no significant difference between them, but the higher temperature crystal is described as a red prism while the lower temperature crystal is a pink plate. This feature was not discussed, but it raises the possibility of a merocyanine impurity arising due to the thermochromic effect.

Refinement
Crystal data, data collection and structure refinement details are summarized in Table 2. The hydrogen atoms bonded to carbon were located by geometry and refined using a riding model. Distances were fixed at 0.95 Å for C-H bonds in phenyl rings and 0.98 Å in methyl groups. In structure 2, primary C-H bonds were assigned C-H distances of 1.00 Å while secondary C-H distances were given values of 0.99 Å . The U iso (H) parameters were set equal to 1.5U eq for the methyl groups and to 1.2U eq of the parent carbon for all others.  Table 1 Comparison of modeled (MM2, PDDG, PM6) bond lengths, angles, and dihedral angles (Å , ) with X-ray crystallographic data.  Computer programs: SMART (Bruker, 2002), SAINT (Bruker, 2013(Bruker, , 2014, APEX2 (Bruker, 2014), SHELXTL (Sheldrick, 2008), SHELXT (Sheldrick, 2015a) and SHELXL2014 (Sheldrick, 2015b).

Fractional atomic coordinates and isotropic or equivalent isotropic displacement parameters (Å 2 )
x y z U iso */U eq O1 0.80339 (11) 0.34663 (10) (17)  Special details Geometry. All esds (except the esd in the dihedral angle between two l.s. planes) are estimated using the full covariance matrix. The cell esds are taken into account individually in the estimation of esds in distances, angles and torsion angles; correlations between esds in cell parameters are only used when they are defined by crystal symmetry. An approximate (isotropic) treatment of cell esds is used for estimating esds involving l.s. planes.

Fractional atomic coordinates and isotropic or equivalent isotropic displacement parameters (Å 2 )
x y z U iso */U eq O1 0.34762 (