Polar Imperfections in Amino Acid Crystals: Design, Structure, and

3 days ago - Insertion of dopants also induces a polar deformation of neighboring host molecules, resulting in the creation of polar domains within th...
0 downloads 7 Views 7MB Size
Article Cite This: Acc. Chem. Res. XXXX, XXX, XXX−XXX

pubs.acs.org/accounts

Polar Imperfections in Amino Acid Crystals: Design, Structure, and Emerging Functionalities Elena Meirzadeh, Isabelle Weissbuch, David Ehre, Meir Lahav,* and Igor Lubomirsky* Department of Materials and Interfaces, Weizmann Institute of Science, Rehovot 76100, Israel CONSPECTUS: Crystals are physical arrays delineated by polar surfaces and often contain imperfections of a polar nature. Understanding the structure of such defects on the molecular level is of topical importance since they strongly affect the macroscopic properties of materials. Moreover, polar imperfections in crystals can be created intentionally and specifically designed by doping nonpolar crystals with “tailormade” additives as dopants, since their incorporation generally takes place in a polar mode. Insertion of dopants also induces a polar deformation of neighboring host molecules, resulting in the creation of polar domains within the crystals. The contribution of the distorted host molecules to the polarity of such domains should be substantial, particularly in crystals composed of molecules with large dipole moments, such as the zwitterionic amino acids, which possess dipole moments as high as ∼14 D. Polar materials are pyroelectric, i.e., they generate surface charge as a result of temperature change. With the application of recent very sensitive instruments for measuring electric currents, coupled with theoretical computations, it has become possible to determine the structure of polar imperfections, including surfaces, at a molecular level. The detection of pyroelectricity requires attachment of electrodes, which might induce various artifacts and modify the surface of the crystal. Therefore, a new method for contactless pyroelectric measurement using X-ray photoelectron spectroscopy was developed and compared to the traditional periodic temperature change technique. Here we describe the molecular-level determination of the structure of imperfections of different natures in molecular crystals and how they affect the macroscopic properties of the crystals, with the following specific examples: (i) Experimental support for the nonclassical crystal growth mechanism as provided by the detection of pyroelectricity from near-surface solvated polar layers present at different faces of nonpolar amino acid crystals. (ii) Enantiomeric disorder in DL-alanine crystals disclosed by detection of anomalously strong pyroelectricity along their nonpolar directions. The presence of such disorder, which is not revealed by accurate diffraction techniques, explains the riddle of their needlelike morphology. (iii) The design of mixed polar crystals of L-asparagine·H2O/L-aspartic acid with controlled degrees of polarity, as determined by pyroelectricity and X-ray diffraction, and their use in mechanistic studies of electrofreezing of supercooled water. (iv) Pyroelectricity coupled with dispersion-corrected density functional theory calculations and molecular dynamics simulations as an analytical method for the molecular-level determination of the structure of polar domains created by doping of α-glycine crystals with different L-amino acids at concentrations below 0.5%. (v) Selective insertion of minute amounts of alcohols within the bulk of α-glycine crystals, elucidating their role as inducers of the metastable β-glycine polymorph. In conclusion, the various examples demonstrate that although these imperfections are present in minute amounts, they can be detected by the sensitive pyroelectric measurement, and by combining them with theoretical computations one can elucidate their diverse emerging functionalities.



INTRODUCTION Crystals are not perfect; they might contain various defects and imperfections of polar configuration.1 Moreover, their surfaces are intrinsically polar, even when the crystal is composed of ideally nonpolar units, because the environment is necessarily unsymmetrical in the direction normal to the surface.2 Materials with macroscopic polarization are pyroelectric (i.e., they create temporary surface charge upon temperature change, resulting in an external electric current between the hemihedral faces).3 Pyroelectricity was thought to be restricted to the polar directions of the crystals belonging to the 10 polar (out of the all-possible 32) crystal classes. Pyroelectric surface charge develops because temperature variations alter the average © XXXX American Chemical Society

position of the atoms, incurring a change in the spontaneous polarization of the crystal.4 During the past decade, improvement in instrumentation provided means to measure pyroelectric coefficients on the order of 10−13 C·cm−2·K−1, which is 1:100000 with respect to commercially important materials.3 These improvements opened the prospect to use pyroelectricity as an analytical technique, 4,5 in particular for the detection of polar imperfections present at surfaces or polar domains created deliberately within the bulk of nonpolar crystals by doping.6−13 Received: February 1, 2018

A

DOI: 10.1021/acs.accounts.8b00054 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Accounts of Chemical Research

Article



DIFFERENTIATION BETWEEN BULK AND NEAR-SURFACE PYROELECTRICITY One of the most common methods for pyroelectric measurement is the modified periodic temperature change technique (Chynoweth):4,19 the top contact of the sample is irradiated by a modulated heat source (e.g., an IR laser) (Figure 1a), and the pyroelectric coefficient, the derivative of the polarization with respect to temperature (α ≡ ∂P/∂T), is calculated from the external current, I. This technique allows an assessment of the distribution of the pyroelectric coefficient in an inhomogeneous sample.20 When the polarization of the sample is uniform, the pyroelectric current is constant during heating (Figure 1a) and described by αF A I= d Cvd (1)

Studies of the preparation of mixed crystals have demonstrated that the inclusion of a dopant requires two steps: binding of the dopant at specific sites on growing faces of the crystals followed by its inclusion within the bulk of the host.14,15 Such processes imply that the dopant is included in a polar mode. Two factors determine the degree of polarity of the created imperfections: first, the difference between the dipole moment of the dopant and that of the host molecule that it replaces, and second, the polar deformation that the dopant induces in its neighboring environment. Consequently, the insertion of the dopants converts a nonpolar host crystal into a conglomerate composed of different sectors of lower symmetry (Scheme 1, where the additive molecule can be adsorbed only when its protrusion points away from the growing crystal).16,17 Scheme 1. Mode of Insertion of a “Tailor-Made” Additive within a Nonpolar Host To Create Doped Polar Domains within the Different Sectorsa

where Fd is the heat flux at the surface, A is the contact area, Cv is the thermal capacitance per unit volume, and d is the thickness of the sample. If the polarization is not uniform, the current changes with irradiation time as a result of heat propagation through sectors with different polarizations.4 For example, in surface pyroelectricity in nonpolar materials, when the polar region is confined to a thin layer near the surface of the sample, the current decays with time (Figure 1a) and is described as21 I=

a

Adapted from ref 17. Copyright 2003 American Chemical Society.

B t + t0

(2)

B can be extracted from the current−time dependence, and the surface pyroelectric coefficient, αs, then can be calculated from22

In cases where the concentration of the dopant is at least several percent, the structure of the mixed crystal might be determined by neutron and X-ray diffraction studies.9,15 Those methods, however, are not applicable when the dopant concentration is below 1%, as in various functional doped materials. We provide examples where the structures of such polar imperfections can be determined at the molecular level by combining pyroelectric measurements with dispersion-corrected density functional theory (DFT) calculations18 and molecular dynamics (MD) simulations. Here we describe two different methods for pyroelectric measurement applied in the present studies, followed by recent representative examples of the structures of various polar imperfections present within amino acid crystals with relevance to their functional properties.

αs =

B(d + δ)Cv πD FdAδ

(3)

where δ is the thickness of the near-surface polar layer and D is the heat diffusion coefficient. The Chynoweth technique requires attachment of electrodes for measuring the current created during exposure of the crystal to a temperature change. The electrodes, however, might affect the surface. Thus, to exclude the effect of the electrodes, a contactless technique to measure the pyroelectric coefficient in ultrahigh vacuum (UHV) was developed by Ehre and Cohen23 based on X-ray photoelectron spectroscopy (XPS).24 The kinetic energy, Ek, of electrons emitted from a material under monochromatic X-ray irradiation is determined from the condition for energy conservation: E k = hν − E B − ϕ + eVs

(4)

Figure 1. Schematic illustration of the pyroelectric measurement. (a) Chynoweth technique under ambient conditions, emphasizing the different time dependences of the pyroelectric current from inhomogeneous (surface pyro) and homogeneous (bulk pyro) samples. (b) Contactless technique in UHV using XPS. Adapted from ref 11. Copyright 2016 American Chemical Society. B

DOI: 10.1021/acs.accounts.8b00054 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research where hν is the photon energy, EB is the binding energy of the atomic orbital from which the electron was emitted, ϕ is the spectrometer work function, e is the elementary electric charge, and Vs is the surface potential. The pyroelectric coefficient can be measured from changes in Vs by measuring changes in Ek (Figure 1b).

surface pyroelectricity (Figure 2b) with surprisingly large surface charge, reaching ∼1 μC·cm−2, which is comparable to that of strongly polar materials.25 Other possible effects, including trapped charges and photo-, thermo-, or flexoelectric effects, were shown experimentally to be inconsistent with the formation of this current (see the Supporting Information of ref 7.). Therefore, it was concluded that such an anomalous current detected from a centrosymmetric crystal is due to surface pyroelectricity.7,11 Support for this deduction follows from the evidence that in the α-Gly crystals, as opposed to the ordinary polar crystals, pyroelectric charges of the same sign are developed at the two opposite {010} faces.4 Furthermore, the pyroelectric effect vanishes when the crystals are heated above ∼80 °C for ∼2 h (Figure 2c), suggesting that this effect originates from a hydrated polar surface that undergoes reconstruction upon heating. The role played by water was confirmed by the demonstration that freshly cleaved {010} faces do not exhibit surface pyroelectricity but that pyroelectricity appears after the crystals are dipped in water. In order to remove any possible artifacts that might arise from the attachment of the electrodes, we also measured the surface pyroelectricity by the contactless XPS method (Figure 2d,e). Moreover, the comparison between different levels of humidity, ambient and UHV, provides additional insights regarding the interactions between water and Gly molecules at the crystal interface. The Chynoweth technique and the contactless measurements yield similar values of the surface pyroelectric coefficient, demonstrating that removal of the surface water by heating induces reconstruction of the polar layer. On the other hand, the UHV in the XPS experiments is not sufficient to induce such reconstruction. MD simulations of this wetted face performed by the Harries group (Hebrew University of Jerusalem) suggested that water molecules penetrate and deform a few molecular layers near the surface (1−2 nm), creating a near-surface hydrated polar layer.11 However, the effective thickness of the polar layer in the experiment is much larger than that in the simulations (at least ∼100 nm). A possible way to account for the creation of such a thick hydrated layer is the proposition that the crystals grow via a nonclassical crystal growth mechanism. In such circumstances, large amorphous clusters formed within the supersaturated



THE DISCOVERY OF SURFACE PYROELECTRICITY FROM NONPOLAR AMINO ACID CRYSTALS As part of our studies on determining the structures of mixed crystals, we have discovered that the reference system, pure αglycine (α-Gly) crystals (Figure 2a), counterintuitively exhibits

Figure 2. Studies of α-Gly crystals grown in aqueous solution. (a) Different morphologies of the crystals. (b, c) Pyroelectric measurements via the Chynoweth technique: (b) pyroelectric current at room temperature and (c) surface pyroelectric coefficient product αs·δ as a function of surrounding temperature. (d, e) Contactless measurements on (d) cleaved and (e) as-grown crystals. Panel (a) reprinted with permission from ref 7. Copyright 2013 John Wiley and Sons.

Figure 3. (a) Needle-like DL-Ala crystals and (b) packing arrangement showing different functional groups exposed at the {210} faces. (c) Pyroelectric current from naturally developed {210} faces of a DL-Ala crystal, showing the same current direction. (d) Pyroelectric current after the top part of the crystal was scraped. The current direction was reversed after scraping. Panels (a) and (b) reprinted from ref 8. Copyright 2014 American Chemical Society. C

DOI: 10.1021/acs.accounts.8b00054 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research solutions nucleate and grow. Some of those clusters land on the growing {010} faces of α-Gly and are aligned in a polar mode. Because of the large dipole moment of the zwitterionic Gly molecules (∼14.9 D),26 the distorted molecules interacting with water in these clusters create a large macroscopic polarization. The characterization of those polar near-surface layers clarifies the riddle of previously conflicting reports27−29 of anomalous pyroelectricity suspected from the Gly crystals. Similar near-surface hydrated layers were observed in other nonpolar amino acid crystals as well: L- or D-alanine (space group P212121; see below), DL-serine (space group P21/a), and 14,15 DL-glutamic acid monohydrate (space group Pbca). The surface pyroelectricity in all of these systems disappears irreversibly upon heating, similar to that of α-Gly crystals.



Figure 4. (a) Morphology of pure L-Ala crystals (left) vs L-Ala crystals doped with ∼0.3% w/w D-Ala (right). The c direction is elongated upon addition of the D enantiomer. (b) Packing arrangement of an LAla crystal, showing different functional groups exposed at the {210} faces. (c) Temperature dependence of the pyroelectric current generated from the (210) face of a fresh crystal. Inset: subsequent heating−cooling cycles. Adapted from ref 8. Copyright 2014 American Chemical Society.

ENANTIOMERIC DISORDER INDUCED BY SELF-POISONING

DL -Alanine

(Ala) is a polar crystal belonging to the orthorhombic space group Pna21. The crystals display a needlelike morphology along the polar c direction, expressing the {210} faces (Figure 3a,b). An attempt to detect a polar near-surface layer on the nonpolar {210} faces resulted in the discovery of substantial bulk polarity (Figure 3c) with a pyroelectric coefficient an order of magnitude higher than that measured along the polar c axis.16,17 Moreover, the sense of the current is reversed when one of the {210} faces is scraped (Figure 3d). This anomalous polarity was explained by enantiomeric disorder, where a small fraction of L-Ala molecules occupy the sites of the D- enantiomer and vice versa. The decay in the pyroelectric current as a function of irradiation time was very rapid, similar to the surface pyroelectric signal (Figure 1a). However, the reason for the fast decay in this case was that the DL-Ala crystals were very thin (>0.3 mm), and therefore, the time required for the heat to diffuse through the whole crystal was very short.4,8 The energetically favored sites in which the molecules could be interchanged were determined by atom−atom potential energy calculations and are shown in Figure 3b. Such disorder could not be detected by precise low-temperature X-ray or neutron diffraction studies. The large pyroelectric coefficient from the {210} faces compared with the polar c direction can be rationalized by the fact that upon heating DL-Ala expands along the a and b axes but contracts along the c axis.30 Further confirmation of this hypothesis was provided by deliberately contaminating enantiomorphous crystals of L-Ala, which display similar {210} faces as in the DL crystals, with the D enantiomer (Figure 4a,b). After the removal of the nearsurface polar wetted layer by heating, the pure L-Ala crystals did not display pyroelectricity.8 On the other hand, when L-Ala was contaminated with the opposite enantiomer at a concentration of ∼0.3% w/w, the mixed crystals were more elongated and displayed similar bulk pyroelectricity (Figure 4c) as those found in the DL crystal. The lowest energetic cost for docking a molecule of opposite chirality is at site 2, where both (Cα)H and CH3 groups emerge from the (210) plane (Figure 4b). The occupancy of D molecules in L sites and vice versa inhibits the growth of these faces and thus provides a rational explanation for the needlelike morphology of DL-Ala versus the diamond-like morphology of 31 L-Ala, which has a similar structure.



MIXED CRYSTALS WITH VARYING DEGREES OF POLARITY: L-ASPARAGINE·H2O/L-ASPARTIC ACID Pyroelectric crystals were shown to affect the icing temperature of supercooled water.32,33 In order to elucidate the role played by the pyroelectric effect from that of epitaxy, it was indispensable to design an assemblage of crystals displaying different degrees of polarity while exposing the same face on which the icing experiments could be performed. In addition, the pyroelectric measurement enables the influence of electric field to be distinguished from that of the charge. Polar mixed crystals are appropriate systems for such studies since by adjusting the concentration of the dopant one can regulate the degree of polarity.34 L-Asparagine·H2O (L-Asn)/L-aspartic acid (L-Asp) mixed crystals were selected for such studies, as 16% w/w L-Asp can be incorporated within that host.14,15 In addition, they display a well-expressed platelike morphology (Figure 5a) that is wellsuited for performing icing experiments. L-Asn crystallizes from aqueous solutions as a monohydrate in the nonpolar space group P212121. The crystal contains four molecules in the unit cell, forming two ribbons of hydrogen-bonded molecules of opposite polarities, A1/A2 and B3/B4, where molecules A1 and A2 and molecules B3 and B4 are related by 21-fold symmetry (Figure 5b). When mixed crystals are grown in the presence of L-Asp, those molecules are incorporated preferentially at the B3/B4 sites at the {010} faces by a process of surface recognition. The incorporation involves replacement of an N(amide)−H bond that emerges from the (010) face by an O(H) lone-pair electron lobe of a β-carboxylic group of L-Asp (Figure 5b). On the other hand, those L-Asp molecules are rejected from the A1/A2 sites because of lone-pair−lone-pair repulsion between the β-carboxylic O(H) and the CO2− of the L-Asn host. Consequently, such inclusion reduces the symmetry of the mixed crystal from space group P212121 to the pyroelectric space group P21, as independently confirmed by diffraction experiments.15 Pyroelectric measurements revealed the formation of a bipolar structure. However, the top surface, which exposes the (010) face toward the aqueous solution, displays larger pyroelectricity in comparison with the bottom sector, which grew at the glass−water interface (Figure 5c,d). D

DOI: 10.1021/acs.accounts.8b00054 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

Figure 5. (a) Pure L-Asn·H2O crystal (left) and mixed crystal with ∼9% w/w L-Asp (right). The {010} faces become more pronounced as more dopant is occluded into the host. (b) Two ribbons, A1/A2 and B3/B4, of opposite polarities. (c) Gradual shaving reveals a change in the sign of the pyroelectric current. After the first shaving, the pyroelectric current changes sign as a function of time, which implies the presence of a thin layer of one polarity on top of a thick layer of the opposite polarity, as illustrated in (d). Panels (a) and (b) reprinted from ref 9. Copyright 2015 American Chemical Society.

higher by 4 °C than that of the pure nonpolar host. Furthermore, it was found to scale linearly with the pyroelectric current (Figure 6b), demonstrating that electrofreezing is influenced by the electric charge developed upon cooling of the pyroelectric crystals.34

The pyroelectric coefficient of the mixed crystals as a function of the L-Asp concentration is shown in Figure 6a. The plot reveals a linear increase in the pyroelectric coefficient up to 8% w/w L-Asp, suggesting that the dopant molecules do not interact with each other, implying their random distribution within the host. In the region of 8−12%, there is a drastic increase in the pyroelectric coefficient, which can only be a result of enhanced dopant−dopant and dopant−host interactions. However, a further increase in the L-Asp concentration sharply reduces the pyroelectric coefficient of the mixed crystal. This strongly suggests that the concentration of ∼12% is the maximum that can be accommodated in the B ribbons, and further increasing the L-Asp concentration forces some of the dopant molecules to occupy also some of the A sites, with opposite direction of the dipole moment. The concentration of 12% corresponds to every third molecule in the B ribbon being L-Asp, strongly suggesting at least a partial local ordering, which may explain enhancement in the pyroelectric coefficient in the 8−12% region.



STRUCTURE DETERMINATION OF DOPANT-INDUCED DEFORMED CRYSTALLINE DOMAINS AT THE MOLECULAR LEVEL In systems where there are significant structural differences between the dopant and the host, the amount of occlusion is limited and cannot be detected by diffraction techniques. Moreover, in order to gain more detailed knowledge about the structures of polar domains formed as a result of local deformations that different dopants induce in their surroundings, crystals of α-Gly doped with small amounts (