Role of Kinks in Dyeing Crystals - American Chemical Society

CRYSTAL GROWTH & DESIGN

Role of Kinks in Dyeing Crystals: Confocal Luminescence Microscopy from Single Molecules to Square Centimeters Theresa Bullard, Kristin L. Wustholz, Eric D. Bott, Miranda Robertson, Philip J. Reid,* and Bart Kahr*

2009 VOL. 9, NO. 2 982–990

Department of Chemistry, UniVersity of Washington, Box 351700, Seattle, Washington 98195-1700 ReceiVed July 21, 2008

ABSTRACT: Large (1 cm3) potassium hydrogen phthalate (KAP) crystals grown in the presence of the fluorescent dye 2′,7′dichlorofluorescein (DCF) show patterns of luminescence characteristic of selective inter- and intrasectoral zoning. Despite this selectivity, the polarization dependence of the luminescence of single DCF molecules inside the crystals indicated broad orientational distributions consistent with nonspecific mixed crystal growth mechanisms (Wustholz, K.; Kahr, B.; Reid, P. J. J. Phys. Chem. B, 2005, 109, 16357-16362). In an effort to reconcile this apparent discrepancy, KAP crystals were grown from 10-9 to 10-4 M DCF solutions and examined using confocal luminescence and polarized absorption microscopies, where possible. Single molecules and molecular ensembles were investigated. The fluorescence excitation dichroism was strongly dependent on the concentration of DCF, suggesting that different mixed crystal growth mechanisms were at work at different guest concentrations. Hottenhuis et al. (J. Cryst. Growth, 1986-1989) had earlier established that certain trivalent cations bind preferentially to distinct kink sites of KAP. By blocking particular kink sites with Fe3+ or Ce3+ at high DCF-solution concentrations, the orientation of DCF in the crystals was modulated, thus establishing that the dye not only recognizes some propagating steps as opposed to others but also preferentially chooses between kinks propagating in opposing directions on the same step, evidence that kink selectivity plays a vital role in the dyeing of crystals. Introduction Virtually every civilization has invented or appropriated technologies for dyeing textile fibers.1 The preparation of colorfast cloth has benefited from the experience of protochemical artisans over millennia as well as from physicochemical investigations of adsorption during the past century.2 However, the dyeing of crystals, a general aspect of supramolecular chemistry,3 while investigated sporadically for 150 years, has been only recently framed in terms of specific intermolecular interactions.4 It is now well established that simple molecular and ionic crystals can orient and overgrow a variety of colored or luminescent molecules during growth from solution. The colorant can selectively deposit in only some growth sectors (intersectoral zoning, Figure 1a)4 or inhomogeneously within particular subvolumes of growth sectors (intrasectoral zoning, Figure 1b), a consequence of the chemistry of distinct vicinal facets of spiral hillocks (Figure 1c).5 Symmetry independent kink sites on a given step must also be differentially recognized by dyes, but to date, we have yet to establish this additional selectivity in the dyeing of crystals. The role of kinks in the dyeing of crystals is investigated herein. One of the most remarkable crystals in its ability to orient and overgrow dye molecules6 is potassium hydrogen phthalate (C6H4 · COOH · COO-K+, space group Pca21, hereafter KAP for potassium acid phthalate).7 KAP grows as {010} plates, expressing beautiful spiral hillocks.8 The c direction of KAP is polar;9 the steps that propagate in directions with +c and -c components differ in structure and therefore intrasectoral zoning results when KAP is grown in the presence of a variety of dyes (more than 100).10 Luminescent labels that bind to the widely spaced fast steps in preference to the closely spaced slow steps (Figure 1b) are revealed as bright chevrons (Figure 2). One dye that recognizes the fast steps is 2′,7′-dichlorofluorescein (DCF, * Corresponding author. E-mail: [email protected] (B.K.); preid@ chem.washington.edu (P.J.R.).

Scheme 1. 2′,7′-Dichlorofluorescein

see Scheme 1). (Throughout, an equilibrium mixture of the dye is hereafter referred to as DCF, while the neutral and monoanionic species are distinguished as DCFH and DCF-, respectively.) DCF is overgrown by advancing steps, leaving a fossil-record of hillock evolution in patterns of light that can be dug-out by successive cleavage or with a confocal luminescence microscope. We have previously described the use of confocal luminescence microscopy (CLM)11 to image the spatiotemporal evolution of hillocks in whole, as-grown dyed KAP crystals. By using CLSM to scan from the bottom to the top of heavily dyed crystals, the number and position of the hillocks that form throughout the {010} growth sectors were measured. Figure 2 shows successive optical slices of KAP dyed with DCF (KAP/ DCF) in which the development of {010} is seen in the evolving pattern of luminescence. The optical sections are thin enough such that the z-coordinate can roughly be taken as time of growth. Hillock positions were mapped in each optical section and then compared to the positions in previous and subsequent sections, thereby generating a sequence of growth-active hillocks.10 Fractal analysis of these patterns provided evidence that the birth, death, and evolution of hillocks during crystal growth are correlated. We likewise examined the photophysical properties of DCF in KAP at the single-molecule level using a CLM. The orientations of DCF obtained from ensemble-averaged dichroism measurements in heavily dyed crystals and the average orientations of the corresponding single-molecule distributions12 grown

10.1021/cg800794x CCC: $40.75  2009 American Chemical Society Published on Web 12/29/2008

Role of Kinks in Dyeing Crystals

Crystal Growth & Design, Vol. 9, No. 2, 2009 983

Figure 1. Chemical zoning in KAP. (a) Intersectoral zoning, (b) intrasectoral zoning, and (c) selective kink (k) recognition.

from more dilute dye solutions were at variance. Dichroism measurements and force field calculations were exploited in the analysis of this discrepancy. The dependence of molecular orientation on solution impurity concentration, shown herein, reveals for the first time that the selective recognition of kink sites is a critical step in the mechanism of dyeing crystals and in establishing dye alignment. Results and Discussion Linear Dichroism. Heavily dyed crystals of KAP/DCF (1 dye per 20,000 KAP molecules from a solution concentration of 10-4 M DCF, εDCF,KAP(aq) ∼18,000 M-1 cm-1) exhibit absorption maxima at 473 and 503 nm, corresponding to a mixture of DCFH and DCF-.13 Following 503-nm illumination, the crystals exhibit a fluorescence maximum at 531 nm. Figure 3a shows polarized absorbance spectra for KAP/DCF grown from a 2.5 × 10-5 M DCF solution, exhibiting dichroic ratios of 2.0 and 2.2 at the respective absorbance maxima, corre-

Figure 2. Confocal sections of KAP/DCF. (a) Cartoon of confocal z-sectioning of KAP (010) sector (green) containing screw dislocations (vertical lines). (b) Confocal images read from top to bottom of the crystal, depth in the z-series is indicated. Step size is 15.0 µm. Every fifth optical section is shown. Luminescence develops on the fast slopes of (010) growth hillocks, with vertices that mark the dislocation cores. Crystal size ) 5.5 × 9.2 × 1.0 mm3. Each optical section is 5 frames wide by 10 frames high, where each frame is about 1 mm2. These optical sections were automatically stitched together by the Zeiss LSM 510 confocal control software.

sponding to average orientations (taken as the xanthene ring long axis) of 54 ( 1° and 56 ( 1° relative to [100] for DCFH and DCF-, respectively. The average for many crystals with DCF g 10-6 M was 51 ( 3° and 52 ( 4°, respectively. Comparable orientations were obtained from excitation dichroism of several 10 × 10-µm2 areas using 532-nm illumination (52 ( 4° from [100]). Corresponding transmission and linear dichroism (LD) images made using the rotating polarizer

984 Crystal Growth & Design, Vol. 9, No. 2, 2009

Bullard et al.

Figure 3. Absorption dichroism of DCF (2.5 × 10-5 M) in the fast slopes of KAP (010) hillocks. (a) Polarized absorption anisotropy taken with a microabsorption spectrophotometer. (b) Transmission (%). (c) Absorption anisotropy where tanh(ε) ) (T ′-T ′′)/(T ′+T ′′), with T ′ and T ′′ being the transmittances for orthogonal polarizations. Vertical bars represent azimuthal orientation of the most absorbing eigenmode, showing the transition dipole moments within the fast slope oriented closer to [001].

technique14 are presented in Figures 3b and c, respectively. As expected, KAP/DCF demonstrates the greatest absorption anisotropy in the fast hillock slope that contains most of the dye, and the ensemble-averaged transition dipole moment orientation is closer to [001]. As the dye concentration is decreased, the average orientation approaches [100]. In fact, crystals grown from 10-7 M dye solutions exhibited an average orientation for DCF- of 35 ( 3° from [100], implying that different mixed crystal growth mechanisms are operative at different guest concentrations. Single-Molecule Alignment. The specificity of intrasectoral zoning exhibited by crystals of KAP/DCF speaks to a stereoselective dye-crystal recognition mechanism. We therefore expected a narrow dye orientational distribution associated with a particular supramolecular stereochemical mechanism. The breadth of the orientational distribution measured at the singlemolecule level can reveal information about the extent of environmental heterogeneity in the host and the distribution of inclusion sites. We have previously described the use of singlemolecule microscopy to examine the molecular orientation and photophysics in dyed salt crystals.12,15 Here, we use singlemolecule fluorescence microscopy to measure the average

Figure 4. (a,b) False-color 10 × 10 µm2 images of the fluorescence from a KAP crystal grown from 5 × 10-9 M DCF obtained with 532nm excitation along orthogonal eigenmodes (a) [100] and (b) [001]. Color scale corresponds to counts per 100 ms. (c) Single-molecule orientational histograms for 180 molecules employing 405-nm excitation (DCFH), measured as an angle between 0° and 90° from [100]. The average was 43 ( 18° from [100]. (d) For 151 molecules excited at 532 nm (DCF-), the average orientation was 39 ( 16° from [100].

orientation and distribution of orientations of DCF in KAP at extremely dilute dye concentrations. Figures 4a and b show a typical single-molecule data set, consisting of false-color images of the fluorescence from KAP/DCF grown from a solution containing 5 × 10-9 M dye, with images corresponding to excitation along the [100] and [001] eigenmodes of the crystal. Molecules that are brighter with excitation polarization parallel to [100] and [001] have absorption dipole moments oriented closer to these axes. For 405-nm excitation of DCFH, the


Figure 5. Selective kink adsorption of impurities into KAP hillock fast steps. +c kink blockers such as Ce3+ (left) narrow and elongate hillock fast slopes, mis-orienting steps toward . -c kink blockers such as Fe3+ (right) widen and round out hillock fast steps. The figure was adapted after Hottenhuis et al.16

average orientation of 180 molecules in KAP was 43 ( 18° from [100], ranging from 5.8 to 86.5° (Figure 4c). For 532-nm excitation of DCF-, the average orientation for 151 molecules was 39 ( 16° from [100], ranging from 9.3 to 70.9° (Figure 4d). The averages of both distributions were closer to [100] than [001]. DCFH and DCF- are incorporated within {010} in similar orientations in contrast to the acid-base indicator 2,7diaminoacridine that exhibited differences in the states of dye protonation and orientation in the different growth sectors of KAP.6a,b The average of the single-molecule orientations, measured on crystals containing nanomolar quantities of DCF in the growth solution, is significantly different from ensembleaveraged measurements on heavily dyed crystals. The observation of wide orientational distributions for KAP/DCF suggests that dyes incorporate into multiple conformations and inclusion sites in the {010} hillocks. Hillock Morphology Modification. Hottenhuis et al. observed drastic alterations to KAP step kinetics and hillock morphologies due to trivalent cation additives.16 The impurities were found to slow step movement through pinning and competition with KAP for available binding sites.16 Perhaps hillock morphology changes underlie the concentration-dependent orientation of DCF in KAP. Morphology changes are manifest in the angle spanning the boundary between fast and slow hillock slopes (also seen as the apexes of the hillock chevrons in the fluorescence images, Figures 1b, 2b, and 5). Therefore, to test the hypothesis that DCF modifies KAP morphology, we measured these angles for various DCF concentrations. For the ideal KAP habit with step edges defined by the crystallographic directions [101] and [101j], the angle spanned by the fast-slow boundary is 124°.17 DIC measurements on pure


KAP crystals showed an average fast-slow boundary angle of 124° ( 1°. For KAP/DCF crystals grown from 10-5 M dye solutions, the average fast-slow boundary angle was found to be 127° ( 2°. At the highest dye concentrations (g5 × 10-5 M), the fast-slow boundary angle approached 130°, a relatively modest change compared to the impact of trivalent cations on hillock morphology.16 For example, the fast-slow boundary is as wide as 180° in the presence of Fe3+ and as sweeping as 359° at high concentrations of Ce3+.16 Since the observed morphology changes to KAP hillocks were modest18 even at high dye concentrations, we tested the effects of Ce3+ and Fe3+ cations on DCF orientations. Competition for Kinks. Hottenhuis et al. demonstrated that partially hydrated Ce3+ and Fe3+ function as kink-blocking impurities in KAP, producing dramatic dichotomous changes in hillock morphology.16 On steps, the kinks growing toward ( c are distinct in structure (Figure 5). Upon the addition of Ce3+ to growth solutions, the fast steps of KAP became progressively misoriented until the corner between the [101j] and [1j01j] fast steps was pointed.16 Spirals were elongated along c relative to pure KAP; the angle of the boundary between the fast and slow steps became greater than 180°; the slow steps were unaltered. Therefore, the authors determined that Ce3+ binds preferentially to the +c kink sites, permitting growth only into the -c kinks and along the step edges. With the addition of Fe3+, the fast steps became progressively rounded until the corner between the [101j] and [1j01j] fast steps disappeared.16 Spirals were shortened along c, and the angle of the boundary between the fast and slow steps approached 180°. Hottenhuis et al. concluded that Fe3+ blocks the -c kink sites, resulting in a rounding and shortening of the fast slope. Varying amounts of Fe3+ were added to KAP/DCF growth solutions in order to determine how DCF behaves when the -c kink sites are occupied. Crystals of KAP/DCF/Fe3+ were half as bright as nonferric KAP/DCF crystals obtained under the same growth conditions. Crystals grown from 10-5 M DCF solution with 10 ppm Fe3+ showed an average emission intensity (normalized by crystal thickness) of 7.3 ( 0.6 × 104 counts/ s/mm. In contrast, crystals grown from 10-5 M DCF with no Fe3+ emitted 1.2 ( 0.2 × 105 counts/s/mm. The corresponding solutions of dissolved crystals were likewise less bright, indicating that the reduction in dye incorporation is not simply fluorescence quenching. The average dye orientation in KAP/ DCF/Fe3+ crystals as a function of ion content was 41 ( 3° and 42 ( 2° from [100] for DCFH and DCF-, respectively, as determined by both absorbance and excitation dichroism measurements (Figure 6a). The dye orientation was statistically equivalent for crystals grown from various DCF (10-6 M to 5 × 10-5 M) or Fe3+ (5 to 20 ppm) concentrations. The apparent indifference of DCF orientation to cation concentration suggests that the dye is constrained to a very limited number of docking configurations, unrelated to kink sites, in the presence of Fe3+. This orientation is consistent with the xanthene ring aligned along the fast step edges in KAP hillocks, ideally oriented at 34° from [100]. That DCF orientation was independent of its solution concentration was surprising, given that DCF orientations in KAP crystals grown without trivalent cations are dependent on DCF concentration. We grew crystals of KAP/DCF containing various concentrations of Ce3+, a +c kink blocker. Crystals of KAP/DCF/Ce3+ (g10-6 M DCF) were about 30 times more luminescent (3.6 ( 0.5 × 106 counts/s/mm) than crystals obtained using the same growth conditions but without Ce3+, consistent with the hypothesis that DCF is more effectively overgrown in the -c


Figure 6. Average DCF orientation as a function of (a) Fe3+ and (b) Ce3+ dopant. Error bars correspond to the standard deviation from the mean. (a) Average dye orientation remains the same at 41 ( 3° from [100] for DCFH and 42 ( 2° from [100] for DCF-, regardless of Fe3+ and dye concentrations. (b) Average dye orientation shifts progressively closer to [100] for both DCFH and DCF- with increasing Ce3+ concentration, irrespective of dye concentration. Only DCF- data is shown.

kink site in the fast slopes of KAP hillocks. Figure 6b presents the average orientation of DCF obtained using polarized absorption and excitation measurements in crystals of KAP/ DCF/Ce3+ as a function of Ce3+ concentration. As the Ce3+ concentration is increased up to 5 ppm, the dye orientation moves progressively closer to 25 ( 3° and 21 ( 3° from [100] for DCFH and DCF-, respectively, indicating a switch in orientation from near [001] to closer to [100]. Modeling Dye Incorporation into KAP Steps and Kinks. Force-field calculations were used to predict the transition electric dipole moment orientations of DCF in various docking sites. We used the COMPASS19 force-field because it best matched the crystal structure of KAP to within a volume of 95%. The DCF- and DCFH conformations were obtained ab initio.25 The docking energy was calculated with the following formula:20

Ereaction ) Ef - Ei ) (Estep-phthalate+DCF + Ephthalate)f (Estep + EDCF)i where Estep is the total energy of either the relaxed straight or kinked step for the most stable (010) surface, Estep-phthalate+DCF is the optimized total energy with a DCF docked in place of a KAP molecule, Ephthalate- is the internal energy of the minimized isolated phthalate anion (for DCFH docking this parameter is set to zero; see Experimental Section), and EDCF is the internal energy of the optimized isolated DCF. Steps were considered, both with risers parallel to the b-axis, as well as those with {111j} risers. The reaction energies of many combinations of stepped or kinked surfaces and docking sites were calculated. Exothermic reaction energies were produced from the adsorption of the dye’s carboxylate (or carboxylic acid)

Bullard et al.

group to a K+ ion so as to replace a phthalate carboxylate in the lattice. In the fast steps, K+ ions are readily accessible to the dye at the step edge, whereas the slow steps expose phthalates. This difference in step structure, in addition to growth kinetics, is presumably the reason DCF does not incorporate into the slow steps.16 The structure of the step edges allows for two distinct types of vacancies or docking sites with the aromatic ring of the substituted phthalate perpendicular to the step edge (A in Figure 7), or parallel to the step edge (B in Figure 7). Given that KAPcarboxylate for DCF-carboxylate substitution appears to be the principal replacement motif, docking into the B vacancy forces the DCF xanthene ring system into the step, with consequent endothermicity. Position A docking is always exothermic, and only this site was further considered. Kinks are chemically distinct; the +c kink is acute, exposing COOH, and the -c kink is obtuse, exposing COO-. Kink docking for both DCF- and DCFH enantiomers were considered. Simulations of step and kink-site docking of DCF- and DCFH into the fast steps are summarized in Table 1 and Figure 7. Docking into a straight step vacancy gave an Ereaction ) -20 ( 2 kJ/mol for DCF- and Ereaction ) -366 ( 32 kJ/mol for DCFH. The xanthene long axis of DCF- was oriented at 39 ( 3° from [100], and DCFH was similarly oriented at 38 ( 6°, essentially along the step edge, which is consistent with the average of the single-molecule distribution. The docking of DCF- and DCFH into the kink sites was exothermic in all cases. Overall, the average reaction energies for dye incorporation into the +c kink (-170 ( 11 kJ/mol for DCF- and -196 ( 3 kJ/ mol for DCFH) were more exothermic than for the -c kink (-105 ( 35 kJ/mol for DCF- and -133 ( 21 kJ/mol for DCFH). The average orientation for docking into the +c kink was 64 ( 8° from [100] for DCF- and 54 ( 1° from [100] for DCFH, while that for the -c kink was 25 ( 8° from [100] for DCF- and 20 ( 15° from [100] for DCFH. The modeling was consistent with the wide distribution of orientations among single molecules. The simulations of DCF docked into the +c and -c kink sites as well as the step-edge vacancy support the experimental results for KAP/DCF grown in the presence of Fe3+ and Ce3+. The dye is oriented at roughly 25° from [100] when docked into the -c kink site, consistent with our observations for crystals grown from high Ce3+ concentrations. When docked into the +c kink site, the dye adopts an orientation between 54° and 64° from [100]. However, our experimental observations suggested that Fe3+ inhibits the dye from docking into all but the step-edge holes, limiting its orientation to a narrow 42 ( 2° from [100], consistent with the simulation results for docking DCF into a vacancy in the step edge. While modeling predicts that +c kink adsorption is more exothermic, the force-field calculations do not take into account solvent interactions, growth kinetics, or impurity overgrowth, a critical step that precedes our observables. Force-field calculations of DCF docking into the various KAP steps and kinks also showed surprising indifference in orientation and selectivity with respect to the dye’s charge, in stark contrast to the behavior of the acid-base indicator 2,7-diaminoacridine that exhibited differences in orientation and face selectivity for different protonation states.6a,b The primary difference between the DCF- and DCFH bonding motifs is that the extra proton on the carboxylic acid group of DCFH additionally bonds to an oxygen from a neighboring phthalate anion in the lattice. Therefore, our observations support the hypothesis that the main driving force for DCF incorporation into KAP is carboxylate for carboxylate (or carboxylic acid for carboxylate) substitution.



Figure 7. Docking sites for DCF- in KAP. Dye orientation changes according to the docking site, always preferring to replace a phthalate in position A rather than position B. The -c and +c kinks orient DCF- closer to [100] and [001], respectively, and the dye docked into the step edge is oriented with its transition dipole moment along the fast step, [1j01]. Table 1. Calculated Energies and Orientations for DCF- and DCFH Docking DCFdocking site

Ereaction (kJ/mol)

fast step edge -20 ( 2a site A Fast step +c kink -170 ( 11 Fast step -c kink -105 ( 35 [100] step -25 ( 3b COO termination [100] step -94 ( 5 COOH termination

DCFH

orientation (from [100])

Ereaction (kJ/mol)

orientation (from [100])

39 ( 3°

-366 ( 32

38 ( 6°

64 ( 8° 25 ( 8° 69 ( 1°

-196 ( 3 -133 ( 21 -208 ( 9

54 ( 1° 20 ( 15° 65 ( 1°

77 ( 4°

-220 ( 2

77 ( 2°

a Uncertainties reported for step edge docking are averages for simulations with different step cuts, namely, (1j11j) and (1j01j), and for the enantiomorphous DCF orientations. b Values are averaged between DCF enantiomers, and the uncertainties reported are the corresponding standard deviation of the mean.

The interaction of DCF, Ce3+, and KAP kink sites, leading to a change in overall fluorescence anisotropy of KAP/DCF mixed crystals, was embodied in a simple kinetic model based on a series of four coupled differential equations:

dD+(t) ) k1DK+(t) - k2D+(t) dt

(1)

dD-(t) ) k3DK- - k4D-(t) dt

(2)

dC+(t) ) k5CK+ - k6C+(t) dt

(3)

dK-(t) ) -k7DK-(t) + k8D-(t) + C+(t) dt

(4)

where D+(t) is the population of DCF molecules bound to +c kink sites, K+, and D-(t) is the population of DCF molecules

bound to -c kink sites, K-. Similarly, C+(t) is the Ce3+ bound to the +c kink site. The concentrations of DCF and Ce3+ in solution, D and C, respectively, are held constant for large reservoirs. Equations 1, 2, 3, and 4 are based on simple bimolecular combination terms (k1, k3, k5, and k7) and unimolecular decay terms (k2, k4, k6, and k8). It is assumed in eq 4 that increasing Ce3+ concentration modifies the habit of the hillock creating more K- kink sites in the process. The model was run with the following rate constants and initial conditions: k1 ) 105, k2 ) 101, k3 ) 2 × 105, k4 ) 8, k5 ) 9 × 102, k6 ) 101, k7 ) 30, k8 ) 30, D+(0) ) D-(0) ) C+(0) ) 0 M, K+(0) ) K-(0) ) 10-4, D ) 10-6, and K0+ ) 8 × 10-6. The ratio of the final, equilibrium number densities of D+(t) and D-(t) were then used to calculate the average dye orientation, employing the force-field results that a dye molecule bound in a +c kink will be oriented at 64° and at 25° in a -c kink (K-). The average orientations were calculated as a function of initial Ce3+ concentrations chosen from the experiment. The change in orientation as a function of Ce3+ concentration (Figure 8) is well matched by the modeling, though it is not as linear as the experimental data in Figure 6b. This then semiquantitatively supports the hypothesis that Ce3+ increases the amount of DCF adsorbed into KAP hillocks by generating a surplus of -c kind sites (K-). Role of Steps. Despite underscoring through experiments and computations, the importance of specific kink recognition in crystal dyeing, the motivating question of KAP/ DCF orientational dependence on dye concentration remains to be addressed. Hottenhuis et al. observed the appearance of macro-steps with {100} risers in pure KAP growth experiments.16 In the fluorescence and CLM images of several heavily dyed (g2.5 × 10-5 M) KAP/DCF crystals, we noticed bright strips along the fast-slow boundary of the hillocks, correspond-


Bullard et al.

Figure 8. Computed dependence of excitation dichroism on Ce3+ concentration according the model embodied in equations 1-4. See text. Compare with experimental data and Figure 6b.

ing to step incorporation (Figure 9a). We found one rare but spectacular example where only the steps were dyed exclusively in earlier stages of growth (Figure 9b). LD images indicated an enhanced dichroism toward [001] in this region relative to the fast slopes containing dye (Figure 3c). Since steps propagating in the directions present step edges aligned along , DCF adsorbed to these steps may adopt orientations closer to [001]. Isolating the step inclusion proved difficult. Luminescence from these vicinal faces is always accompanied by a signal from the fast step inclusion. Docking DCF- and DCFH into the steps were carried out to test the hypothesis that DCF orientations will be closer to [001]. Two terminations were modeled for the step: one with the phthalate carboxylate emergent at the step edge and one with the carboxylic acid exposed. Figure 10 gives two examples of the step cuts with DCF- docked into a vacancy. For all simulations, the orientations of DCF- and DCFH ranged from 60° to 85° relative to , indicating that the dyes will orient closer to [001] (Table 1), shifting the ensemble-averaged orientation. Docking into the carboxylic acid-terminated step was more exothermic, and the resulting orientations were closer to [001] than for the carboxylateterminated step cut. Conclusions Crystals of KAP/DCF exhibited dye concentration dependent molecular orientation. Heavily dyed crystals exhibited average orientations closer to [001], while crystals containing nanomolar quantities of dye exhibited varied single-molecule orientations and an average value closer to [100]. To explore the possibility that DCF-induced morphological changes to the crystal hillocks were responsible for changing dye orientation, we measured the angle of hillocks using differential interference contrast (DIC) microscopy. No significant modification to hillock morphology was detected at high dye concentration. Alternative explanations for dye concentration dependent orientation were examined using a combination of experimental and theoretical approaches. By adding trivalent cations to the KAP/DCF growth solutions, we demonstrated that the transition dipole moment orientation of DCF is correlated with specific kink site adsorption within the fast steps of KAP (010) hillocks. In particular, the addition of Ce3+ to growth solutions of KAP/DCF demonstrated that DCF prefers to incorporate into the -c kink sites, progressively pushing the dye orientation closer to [100] as the Ce3+ concentration was increased. The measured dye orientation at

Figure 9. (a) Fluorescence of a typical KAP/DCF (2.5 × 10-5 M) crystal showing the greater amount of dye in the steps relative to the dye included in the fast slopes. (b) Confocal luminescence image of an KAP crystal spectacularly showing DCF primarily in the steps. (Crystal dimensions are 7.4 mm × 8.3 mm.)

high concentrations of Ce3+ agrees with force-field simulations for docking DCF into the -c kink sites, as modeled in the fast steps of the (010) KAP surface. The change in dye orientation as a function of Ce3+ was captured in a simple kinetic model. For crystals of KAP/DCF/Fe3+, the -c kink sites are occupied by Fe3+ cations, and DCF is forced to incorporate into less favorable sites, producing faintly luminescent crystals relative to KAP/DCF and resulting in an average dye orientation independent of Fe3+ concentration consistent with docking into a vacancy along the fast step edge. Overall, this is the first evidence that kink selectivity plays a vital role in the dyeing of crystals. Experimental Section Crystal Growth. Single crystals of KAP/DCF were grown by spontaneous nucleation via evaporation of 110 g/L KAP (Aldrich) in deionized water (Barnsted NANOpure, 18.2 MΩ cm-1) solutions. DCF (Eastman Kodak) was introduced as an ethanol solution. Heavily dyed



Figure 10. DCF- docking into KAP [1j00] step. (a) In the hydroxyl terminated step cut, the carboxylate group of DCF- positions itself in a fashion similar to that of a phthalate, with the transition dipole moment of the dye aligned along the [1j00] step edge. (b) In the carboxylate-terminated step cut, the transition dipole moment is oriented nearly perpendicular to (010), with its projection into the ac-plane close to the c-axis. KAP/DCF crystals were grown from 10-4 M to 10-7 M dye solutions. Solutions containing ∼10-9 M dye produced crystals with DCF densities suitable for single-molecule investigations. Evaporation was performed in temperature-controlled air chambers or water baths held at 30 ( 0.1 °C. After two to four days, crystals were removed from the solution and immediately blown dry using a high-pressure nitrogen jet21 to preserve the as-grown surface. Ce3+ was introduced as cerium chloride (0.1 to 15 ppm CeCl3 · 7H2O (J. T. Baker) with respect to KAP). Fe3+ was introduced as ferric nitrate (1.0 to 20 ppm Fe(NO3)3 · 9H2O (J. T. Baker) with respect to KAP). Bulk Optical Characterization. Heavily dyed crystals were characterized by polarized absorption, fluorescence, and differential interference contrast (DIC) microscopies. The segregation coefficient (σseg), the mole ratio of dye in the crystal to dye in the growth solution, was determined from absorbance measurements on redissolved crystals. Solution and single crystal spectra were obtained with a SI photonics 440 spectrometer coupled to a polarizing microscope (Olympus BX50). The extinction directions [100] and [001] of the birefringent crystals were used to orient the sample relative to the input polarization. Absorption transition dipole moment orientations were computed according to: θ ) tan-1(1.12 · DR)-1/2, where the dichroic ratio (DR ) Ia/Ic) is corrected for sample birefringence.22 Images of the absorption anisotropy were obtained using Metripol, an automated linear birefringence and dichroism imaging system.23 By this method, the LD is displayed as tanh(ε) ) (T ′ - T ′′)/(T ′ + T ′′), with T ′ and T ′′ being the transmittances for orthogonal orientations at 470 nm. Corresponding bulk fluorescence spectra were recorded on a fluorimeter (SPEX FluoroMax-2) employing excitation at 470 nm. Whole-crystal confocal images were obtained using an inverted confocal microscope in tiling mode (Zeiss LSM 510 NLO, Axiovert 200). Dyed crystals were excited at 488-nm with a 30-mW argon ion laser, and the pinhole width was selected to give optical sections of ∼15 µm. Single-Molecule Optical Characterization. Single-molecule microscopy studies were performed using a home-built confocal microscope (Nikon TE2000-U) described in detail elsewhere.12 Briefly, dyed crystals were mounted in an inverted orientation on a scanning stage and excited with low power (∼3 µW) illumination from a 532-nm Nd: YLF (Spectra Physics, Millenia) or 405-nm solid state diode (Photonic Products, PMMF108-4) laser. The excitation field was filtered, reflected toward the sample using an appropriate dichroic mirror, and focused to a diffraction-limited spot on (010) using a 100× oil-immersion objective (Nikon, PlanFluor 1.3 NA). Epi-fluorescence was collected from the sample, passed through the dichroic mirror, spectrally filtered using an emission filter, and spatially filtered with a confocal pinhole. Emission was detected using a single-photon-counting avalanche photodiode (PerkinElmer, SPCM-AQR-16). Total emission intensity following linearly polarized excitation set by a half-waveplate between the [100] and [001] crystal eigenmodes was measured over a 10 × 10 µm2 area by scanning the sample through 100-nm steps. Force-Field Calculations. DCF adsorption to KAP was modeled with the Accelrys Materials Studio suite of programs.19 Crystal habits

and energetically stable (010) faces of KAP were generated from crystallographic data imported from the Cambridge Structural Database.9 Geometry optimizations were performed with the COMPASS force field19 set to a 4 × 10-4 kJ/mol convergence criterion. The molecular coordinates and transition dipole moment orientations of DCF were derived from crystallographic data24 and optimized using density functional theory (B3LYP, 6-31G*).25 For simplicity throughout the discussion, the long axis of DCF’s xanthene ring system was estimated to be the transition electric dipole moment, which is within 5° of the calculated vector in the ground-state for all relevant DCFand DCFH conformations in all local environments. The lowest energy (010) surface26 was terminated by aryl rings. The preference for a C-H (010) termination was confirmed by the observation of smaller total surface (0.70 kJ/mol/Å2) and attachment (-82 kJ/mol) energies relative to the stable (010) surface with partial K+ termination (surface energy ) 3.11 kJ/mol/Å2; attachment energy ) -367 kJ/mol per unit cell). A supercell with 2D periodic boundary conditions was constructed of 3 unit cells along , 5 unit cells along R, and a depth of 3 unit cells in [010]. This supercell was used for all simulations. This cell was split into two regions: region I could relax, while region II was fixed. Steps of unit-cell height were created to model the topography of the hillock surface with cuts corresponding to all possible step orientations of (010) hillocks, namely, the fast and slow step directions on the (010) surface. Step cuts were made based on the lowest energy surface terminations of the {111} and {111j} faces, as well as the {101}, {101j}, and {100}. The stepped and/or kinked surfaces were geometry optimized. For DCF- simulations, the straight or kinked step was relaxed and its Estep calculated. A minimized hydrogen phthalate anion (Ephthalate-) was removed from the upper half of the step to create a docking site. In DCFH simulations, the straight or kinked step was relaxed with a full KAP ion pair already removed from position A to create the docking site. Thus, Estep,“A” already accounts for the removal of a KAP ion pair; Ephthalate- ) 0. In all other respects, the final KAP surfaces for DCFH and DCF- are the same. Next, the lattice was constrained, the dye was inserted ∼5-20 Å from the docking site, and the minimum energy configuration for the docked additive was calculated. For kink-site docking, all neighboring KAP molecules surrounding the kink were also allowed to relax during optimization. After optimization, the resulting lattice was unconstrained and the total energy calculated.

Acknowledgment. We thank A. Rohl for his consultations on force-field calculations, J. Freudenthal for his assistance with AFM measurements, and C. Isborn for ab initio calculations. The authors also thank the National Science Foundation for support of this work through the Center on Materials and Devices for Information Technology Research (DMR-0120967). K.L.W. was a NSF-IGERT and University Initiative Funded fellow, supported by the Center for Nanotechnology at the


University of Washington (DGE-0504573). B.K. acknowledges NSF grant CHE-0349882. We are grateful for the facilities of the University of Washington Nanotechnology User Facility, a member of the NSF National Nanotechnology Infrastructure Network.

Bullard et al.

(14) (15) (16)

References (1) Brunello, F. The Art and Science of Dyeing in the History of Mankind; Neri Pozza: Vicenza, 1973. (2) Rattee, I. D.; Breuer, M. M. Physical Chemistry of Dye Adsorption; Academic Press: New York, 1974. (3) Steed, J. W.; Atwood, J. L. Supramolecular Chemistry, John Wiley: Chichester, U.K., 2000, p. 454. (4) Kahr, B.; Gurney, R. W. Chem. ReV. 2001, 104, 893–951. (5) (a) Zaitseva, N.; Carman, L.; Smolsky, I.; Torres, R.; Yan, M. J. J. Cryst. Growth 1999, 204, 512–524. (b) Gurney, R. W.; Mitchell, C. A.; Ham, S.; Bastin, L. D.; Kahr, B. J. Phys. Chem. B 2000, 104, 878– 892. (c) Gurney, R. W.; Kurimoto, M.; Subramony, J. A.; Bastin, L. D.; Kahr, B. In Anisotropic Organic Materials; Glaser, R.; Kaszynsky, P., Eds.; ACS Symposium Series; American Chemical Society: Washington, DC, 2001. (d) Kurimoto, M.; Subramony, P.; Gurney, R. W.; Lovell, S.; Chmielewski, J.; Kahr, B. J. Am. Chem. Soc. 1999, 121, 6952–6953. (e) Dincer, T. D.; Parkinson, G. M.; Rohl, A. L.; Ogden, M. I. J. Cryst. Growth 1999, 205, 368–374. (f) Kurimoto, M.; Bastin, L. D.; Fredrickson, D.; Gustafson, P. N.; Jang, S.-H.; Kaminsky, W.; Lovell, S.; Mitchell, C. A.; Chmielewski, J.; Kahr, B. Mat. Res. Soc. Symp. 2000, 620, M9.8.1M9.8.10. (g) Intrasectoral zoning in minerals by cathodoluminescence was pioneered by Rakovan, J.; Reeder, R. J. Geochim. Cosmochim. Acta 1996, 60, 4435– 4445, and references therein. (6) (a) Bellinazzi, M.; Barbon, A.; Kahr, B.; Benedict, J. B.; Brustolon, M. Phys. Chem. Chem. Phys. 2006, 8, 379–385. (b) Bellinazzi, M.; Benedict, J. B.; Brustolon, M.; Fleming, S. D.; Jang, S.-H.; Kahr, B.; Rohl, A. L. Angew. Chem., Int. Ed. 2004, 43, 5278–5286. (c) Benedict, J. B.; Wallace, P. M.; Reid, P. J.; Jang, S.-H.; Kahr, B. AdV. Mater. 2003, 15, 1068–1070. (7) (a) Jetten, L. A. M. J.; van der Hoek, B.; van Enckevort, W. J. P. J. Cryst. Growth 1983, 62, 603. (b) Okaya, Y. Acta Crystallogr. 1965, 19, 879–882. (c) van Enckevort, W. J. P.; Jetten, L. A. M. J. J. Cryst. Growth 1982, 60, 275–285. (8) Ester, G. R.; Price, R.; Halfpenny, P. J. J. Phys. D: Appl. Phys. 1999, 32, A128–132. (9) Eremina, T. A.; Furmanova, N. G.; Malakhova, L. F.; Okhrimenko, T. M.; Kuznetsov, V. A. Crystallogr. Rep. 1993, 38, 554–556. (10) (a) Bullard, T.; Kurimoto, M.; Avagyan, S.; Jang, S. H.; Kahr, B. ACA Transactions 2004, 39, 62–72. (b) Bullard, T.; Freudenthal, J.; Avagyan, S.; Kahr, B. Faraday Discuss. 2007, 136, 231–245. (11) (a) Other examples of the use of confocal fluorescence microscopy to study impurity distributions in crystals include Iimura, Y.; Yoshizaki, I.; Nakamura, H.; Yoda, S.; Komatsu, H. Cryst. Growth Des. 2005, 5, 301–305. (b) Iimura, Y.; Yoshizaki, I.; Yoda, S.; Komatsu, H. Cryst. Growth Des. 2005, 5, 295–300. (12) (a) Wustholz, K. L.; Kahr, B.; Reid, P. J. J. Phys. Chem. B. 2005, 109 (34), 16357–16362. (b) Wustholz, K. L.; Bott, E. D.; Isborn, C. M.; Li, X.; Kahr, B.; Reid, P. J. J. Phys. Chem. C 2007, 111, 9146–9156. (c) Wustholz, K. L.; Sluss, D. R. B.; Kahr, B.; Reid, P. J. Int. ReV. Phys. Chem. 2008, 27, 167–200. (13) (a) Mchedlov-Petrossyan, N. O.; Kukhtik, V. I.; Bezugliy, V. D. J. Phys. Org. Chem. 2003, 16, 380–397. (b) Mchedlov-Petrossyan, N. O.;

(17) (18)

(19)

(20)

(21) (22) (23) (24) (25)

(26)

Salamanova, N. V.; Vodolazkaya, N. A.; Gurina, Y. A.; Borodenko, V. I. J. Phys. Org. Chem. 2006, 19, 365–375. Kaminsky, W.; Claborn, K.; Kahr, B. Chem. Soc. ReV. 2004, 33, 514– 525. Wustholz, K. L.; Bott, E. D.; Kahr, B.; Reid, P. J. J. Phys. Chem. C. 2008, 112, 7877–7885. (a) Hottenhuis, M. H. J.; Lucasius, C. B. J. Cryst. Growth 1986, 78, 379388;(b) Hottenhuis, M. H. J.; Gardeniers, J. G. E.; Jeten, L. A. M. J.; Bennema, P. J. Cryst. Growth 1988, 92, 171–188. (c) Hottenhuis, M. H. J.; Oudenampsen, A. J. Cryst. Growth 1988, 92, 513529;(d) Hottenhuis, M. H. J.; Lucasius, C. B. J. Cryst. Growth 1989, 94, 708–720. These values for the ideal hillock morphology are based on KAP habit and lattice parameters, as well as an assumed fast-slow step kinetic ratio of 10:1, as is given by Hottenhuis et al. (See previous reference). Morphological influences of DCF on KAP hillocks were also studied using atomic force microscopy (AFM), confirming negligible differences on hillock morphology in the fast slopes at various concentrations of DCF. COMPASS (Condensed-phase Optimized Molecular Potentials for Atomistic Simulation Studies); Accelrys, Materials Studio, Release 4.0; Accelrys Software, Inc.: San Diego, CA, 2006. (a) Rohl, A. L.; Gay, D. H.; Catlow, C. R. A.; Davey, R. J.; Mackrodt, W. C. J. Chem. Soc., Faraday Discuss. 1993, 95, 273. (b) Rohl, A. L.; Gay, D. H.; Davey, R. J.; Catlow, C. R. A. J. Am. Chem. Soc. 1996, 118, 642–648. (c) Carter, D. J.; Rohl, A. L.; Gale, J. D.; Fogg, A. M.; Gurney, R. W.; Kahr, B. J. Mol. Struct. 2003, 647, 65–73. (d) Carter, D. J.; Ogden, M. I.; Rohl, A. L. J. Phys. Chem. C 2007, 111, 9283– 9289. (e) Also suggested to us by A. L. Rohl during a private consultation. Ester, G.; Price, R.; Halfpenny, P. J. J. Cryst. Growth 1997, 182, 95– 102. Michl, J.; Thulstrup, E. W. Spectroscopy with Polarized Light; VCH: New York, 1986. (a) Glazer, A. M.; Lewis, J. G.; Kaminsky, W. Proc. R. Soc. London, Ser. A 1996, 452, 2751–2765. (b) Geday, M. A.; Kaminsky, W.; Lewis, J. G.; Glazer, M. J. Microscopy 2000, 198, 1–9. Yamaguchi, K.; Tamura, Z.; Maeda, M. Acta Crystallogr., Sect. C 1997, C53, 284–285. Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.;. Montgomery, J. A.; Vreven, J. T.; Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian03, reVision D.02, (Gaussian, Inc.: Wallingford, CT, 2004). Gay, D. H.; Rohl, A. L. J. Chem. Soc. Faraday Trans. 1995, 91, 925–936.

CG800794X

Role of Kinks in Dyeing Crystals - American Chemical Society

Recommend Documents