A Polymorph of Dipeptide Halide Glycyl-L-Alanine Hydroiodide Monohydrate: Crystal Structure, Optical Second Harmonic Generation, Piezoelectricity and Pyroelectricity

A polymorph of glycyl-L-alanine HI.H2O is synthesized from chiral cyclo-glycyl-L-alanine dipeptide. The dipeptide is known to show molecular flexibility in different environments, which leads to polymorphism. The crystal structure of the glycyl-L-alanine HI.H2O polymorph is determined at room temperature and indicates that the space group is polar (P21), with two molecules per unit cell and unit cell parameters a = 7.747 Å, b = 6.435 Å, c = 10.941 Å, α = 90°, β = 107.53(3)°, γ = 90° and V = 520.1(7) Å3. Crystallization in the polar point group 2, with one polar axis parallel to the b axis, allows pyroelectricity and optical second harmonic generation. Thermal melting of the glycyl-L-alanine HI.H2O polymorph starts at 533 K, close to the melting temperature reported for cyclo-glycyl-L-alanine (531 K) and 32 K lower than that reported for linear glycyl-L-alanine dipeptide (563 K), suggesting that although the dipeptide, when crystallized in the polymorphic form, is not anymore in its cyclic form, it keeps a memory of its initial closed chain and therefore shows a thermal memory effect. Here, we report a pyroelectric coefficient as high as 45 µC/m2K occurring at 345 K, one order of magnitude smaller than that of semi-organic ferroelectric triglycine sulphate (TGS) crystal. Moreover, the glycyl-L-alanine HI.H2O polymorph displays a nonlinear optical effective coefficient of 0.14 pm/V, around 14 times smaller than the value from a phase-matched inorganic barium borate (BBO) single crystal. The new polymorph displays an effective piezoelectric coefficient equal to deff=280 pCN−1, when embedded into electrospun polymer fibers, indicating its suitability as an active system for energy harvesting.


Introduction
Glycyl-L-alanine hydroiodide monohydrate (H 2 N-CH 2 -CO-NH-CH(CH 3 )-COOH·HI·H 2 O is a dipeptide hydrohalide whose crystal structure was determined in 1989 and that is used The piezoelectric behavior displayed by amino acids and dipeptides enables them to be viewed as potential materials to be integrated into energy harvesting devices: it has been reported that the metastable amino acid β-glycine embedded into electrospun polymer fibers displays enhanced piezoelectric and nonlinear optical properties [24]. Chiral diphenylalanine (PhePhe) nanotubes were reported to show a shear component of the piezoelectric tensor of 60 pm/V, and fabricated energy harvesters were able to generate voltage and power up to 2.8 V and 8.2 nW, respectively, with a 42 N force applied periodically [25,26]. Derivatives of PhePhe incorporated into electrospun fibers exhibit strong piezoelectric properties [19,20].

Synthesis
Cyclo-glycyl-L-alanine (1.29 g, 10 mmol) was dissolved in 5 mL HI (57%, stabilized with H 3 PO 3 ) and 10 mL of water. H 3 PO 3 acted as a stabilizer for HI to avoid reduction to elemental iodine. After two weeks of slow evaporation at room temperature, transparent, hexagonal-shaped single crystals of Gly-L-Ala.HI.H 2 O (Poly2) were formed. The crystals were collected and rinsed with acetone, dried and kept in a dissector. An example of the crystals grown is shown in Figure 1. Cyclo-glycyl-L-alanine (cyclo-Gly-L-Ala) was purchased from Bachem AG (Bubendorf, Switzerland). Hydriodic acid (HI) was purchased from Merck (Darmstadt, Germany) and used as received.
be viewed as potential materials to be integrated into energy harvesting devices: it has been reported that the metastable amino acid β-glycine embedded into electrospun polymer fibers displays enhanced piezoelectric and nonlinear optical properties [24]. Chiral diphenylalanine (PhePhe) nanotubes were reported to show a shear component of the piezoelectric tensor of 60 pm/V, and fabricated energy harvesters were able to generate voltage and power up to 2.8 V and 8.2 nW, respectively, with a 42 N force applied periodically [25,26]. Derivatives of PhePhe incorporated into electrospun fibers exhibit strong piezoelectric properties [19,20].

Synthesis
Cyclo-glycyl-L-alanine (1.29 g, 10 mmol) was dissolved in 5 mL HI (57%, stabilized with H3PO3) and 10 mL of water. H3PO3 acted as a stabilizer for HI to avoid reduction to elemental iodine. After two weeks of slow evaporation at room temperature, transparent, hexagonal-shaped single crystals of Gly-L-Ala.HI.H2O (Poly2) were formed. The crystals were collected and rinsed with acetone, dried and kept in a dissector. An example of the crystals grown is shown in Figure 1. Cyclo-glycyl-L-alanine (cyclo-Gly-L-Ala) was purchased from Bachem AG (Bubendorf, Switzerland). Hydriodic acid (HI) was purchased from Merck (Darmstadt, Germany) and used as received.
The synthesis of Gly-L-Ala.HI.H2O (Poly2) started with the cyclic form of the dipeptide (cyclo-glycine-L-alanine) and therefore the crystal growth conditions are different from those reported for Gly-L-Ala.HI.H2O (Poly1). We have also attempted the synthesis of linear glycyl-L-alanine with HI in an aqueous solution as described for Gly-L-Ala.HI.H2O (Poly2) using cyclo-glycyl-L-alanine; however, no crystals were ever formed. After complete evaporation of the solution, an oily residue was obtained. We conclude that Gly-L-Ala.HI.H2O (Poly1) is only obtained following the procedure reported in reference [1].  The synthesis of Gly-L-Ala.HI.H 2 O (Poly2) started with the cyclic form of the dipeptide (cyclo-glycine-L-alanine) and therefore the crystal growth conditions are different from those reported for Gly-L-Ala.HI.H 2 O (Poly1). We have also attempted the synthesis of linear glycyl-L-alanine with HI in an aqueous solution as described for Gly-L-Ala.HI.H 2 O (Poly2) using cyclo-glycyl-L-alanine; however, no crystals were ever formed. After complete evaporation of the solution, an oily residue was obtained. We conclude that Gly-L-Ala.HI.H 2 O (Poly1) is only obtained following the procedure reported in reference [1].

X-ray Crystallography Experimental Conditions Description
A single crystal of Gly-L-Ala.HI.H 2 O (Poly2) was selected, covered with Fomblin (polyfluoro ether oil) and mounted on a nylon loop. Data were collected at 293(2)K on a Bruker D8 Venture diffractometer equipped with a Photon 100 CMOS detector, using graphite monochromated Mo-Kα radiation (λ = 0.71073 Å). The data was processed using the APEX3 suite software package, which includes integration and scaling (SAINT), absorption corrections (SADABS 2016/2) [27] and space group determination (XPREP). The structure solution and refinement were performed using direct methods with the programs SHELXT (version 2014/5) and SHELXL (version 2018/3) [28] contained in the APEX and WinGX-Version 2021.3 [29] software packages. All non-hydrogen atoms were refined anisotropically. Except for NH, OH and water H-atoms, which were located on the difference Fourier map, the remaining hydrogen atoms were inserted in idealized positions and allowed to refine riding on the parent carbon or oxygen atom with C-H distances of 0.96 Å, 0.97 Å and 0.98 Å for methyl, methylene and methine H atoms, respectively. The molecular diagrams were drawn with Mercury [30]. Crystal data for Gly-L-Ala.HI.H 2 O (Poly2) are presented in the Supplementary Information, Table S3. The data for Gly-L-Ala.HI.H 2 O (Poly2) was deposited in the CCDC under deposit number 2247398.

Dielectric Spectroscopy
The dielectric properties of the Gly-L-Ala.HI.H 2 O (Poly2) crystals were characterized by impedance spectroscopy at temperatures of 288-383 K and in the frequency range 20 Hz-3 MHz. The complex permittivity, written as ε = ε − iε , where ε and ε are the real and imaginary parts, respectively, was calculated from the measured capacitance (C) and loss tangent (tan δ), using the equations: Here, A is the electric contact area and d is the crystal thickness. To form the capacitor, the bottom and top electrodes were gold contacts sputtered onto the sample surfaces. A Wayne Kerr 6440A (Wayne Kerr Electronics, London, UK) precision component analyzer was used, together with a dedicated computer and software, to acquire the data. Shielded test leads were employed to avoid parasitic impedances due to connecting cables. Temperature-dependent measurements were performed at a rate of 2 • C/min using a Polymer Labs PL706 PID controller (Polymer Labs, Los Angeles, CA, USA) and furnace.

Pyroelectric Measurements
Pyroelectricity is a property of polar crystalline materials that results from the temperature dependence of their spontaneous polarization. By changing the temperature, an electric field originating from the changes in intrinsic dipoles is compensated for by the surface layer of free charges [31]. The rate of change of the spontaneous polarization ( P s ) with the temperature ( T) is the pyroelectric coefficient, P s = dP s dT . The change in polarization was detected by measuring, at constant stress, the pyroelectric current, I = A dP s dT dT dt , with a Keithley 617 electrometer (Keithley Instruments GmbH, Landsberg, Germany). In the equation, A is the electrode area and dT dt is the rate of temperature change with time ( t). The measurements were performed on a capacitor geometry under short-circuit conditions and the electrode area was 8.11 × 10 −6 m. The temperature interval was between 290 K and 345 K at a heating rate of 2 K/min.

Second Harmonic Generation
The second harmonic measurements were carried out using a mode-locked Ti:sapphire laser (model: Mira, Coherent Inc., Santa Clara, CA, USA) coupled into a Nikon (model: Eclipse Ti2, Nikon Europe B.V., Amstelveen, The Netherlands) inverted microscope, as shown in Figure 2. A calcite Glan-Taylor polarization followed by a zero-order half-wave plate controlled the incident polarization. A Nikon CFI Plan Fluor ×10 objective focused the beam onto the samples, while a Mad City Labs (Madison, WI, USA) xyz piezo-controlled translation stage positioned the samples in the focal plane with sub-micrometer accuracy. Incident powers ranged from 5 mW for the sample to approximately one hundred mi-croWatts for the BBO crystal used for calibration of the system's sensitivity. Although the Fourier limit of the pulse duration is approximately 85 fs, we estimate that the duration stretched to approximately 120 fs when incident on the sample because of the combined effect of the calcite polarizer and the microscope objective. The detection arm along the transillumination direction consisted of a 40 mm focal length best form lens (model:LBF254-040-A, Thorlabs, Newton, NJ, USA) to collimate the second harmonic light followed by a zero-order half-wave plate polarizer combination to analyze the emitted second harmonic light. A long-pass dichroic mirror (Thorlabs DMLP650) filters out most of the incident light while reflecting 98% of the second harmonic light. A short focal length lens then focuses the beam through a narrow band-pass filter (Thorlabs FBH400-40 nm) onto a fiber bundle coupled to an Andor imaging spectrometer (model: Shamrock 300i, Andor Technology Belfast, UKequipped with a cooled CCD array (model: Andor Technology Belfast, Newton, UK).
fs, we estimate that the duration stretched to approximately 120 fs when incident on th sample because of the combined effect of the calcite polarizer and the microscop objective. The detection arm along the transillumination direction consisted of a 40 mm focal length best form lens (model:LBF254-040-A, Thorlabs, Newton, NJ, USA) collimate the second harmonic light followed by a zero-order half-wave plate polariz combination to analyze the emitted second harmonic light. A long-pass dichroic mirro (Thorlabs DMLP650) filters out most of the incident light while reflecting 98% of th second harmonic light. A short focal length lens then focuses the beam through a narro band-pass filter (Thorlabs FBH400-40 nm) onto a fiber bundle coupled to an Ando imaging spectrometer (model: Shamrock 300i, Andor Technology Belfast, UKequippe with a cooled CCD array (model: Andor Technology Belfast, Newton, UK). A crystal plate taken from the crystal sample shown in Figure 1 was mounte perpendicular to the laser beam and the surface was scanned for the most intense SH signal. We acquired the second harmonic signal using the following protocol. At eac position of the half-wave plate controlling the polarization of the incident beam, th analyzer half-wave plate was scanned over 180°. At each analyzer wave plate positio the CCD signal was integrated for 1 s and the subsequent second harmonic light spectr were fitted to a Gaussian profile as shown in Figure S8 in the supplementary informatio The area under the Gaussian fit was taken as the total number of second harmonic sign counts. A crystal plate taken from the crystal sample shown in Figure 1 was mounted perpendicular to the laser beam and the surface was scanned for the most intense SHG signal. We acquired the second harmonic signal using the following protocol. At each position of the half-wave plate controlling the polarization of the incident beam, the analyzer half-wave plate was scanned over 180 • . At each analyzer wave plate position, the CCD signal was integrated for 1 s and the subsequent second harmonic light spectra were fitted to a Gaussian profile as shown in Figure S8 in the supplementary information. The area under the Gaussian fit was taken as the total number of second harmonic signal counts.

Piezoelectric Measurements
The piezoelectric properties of Gly-L-Ala.HI.H 2 O (Poly2) were analyzed by embedding the crystals within fibers fabricated by a conventional electrospinning technique described previously [32]. To produce the fibers, a clear and homogeneous 10 % polymer solution was prepared by dissolving 0.5 g of poly (methyl methacrylate) (PMMA, Mw 996,000, Sigma-Aldrich, Schenlldorf, Germany) in 5 mL of chloroform. To this solution, 0.5 g of Gly-L-Ala.HI.H 2 O (Poly2) powder was added at a 1:1 weight ratio. The resulting mixture was stirred for several hours under ambient conditions before the electrospinning process. This precursor solution was loaded into a syringe and its needle was connected to the anode of a high-voltage power supply (model: CZE2000 Spellmann, Broomers Hill Park, UK). To produce in-plane fibers, the spinning voltage was set at 18 kV, with a distance of 12 cm between the anode and the collector. The flow rate of 0.10 mL/h was controlled by a syringe pump with an attached needle of 0.8 mm diameter. The fiber mat for piezoelectric measurements was collected on high-purity aluminum foil, which served as the electrodes.
The crystallinity and crystallographic orientation of Gly-L-Ala.HI.H 2 O (Poly2) inside the fibers were studied by XRD. The diffraction pattern was recorded between 5 • and 50 • using θ-2θ scans on a Philips (Amsterdam, The Netherlands) PW-1710 X-ray diffractometer with Cu-Kα radiation of wavelength 1.5406 Å. The morphology and fiber thickness were determined using a Nova NanoSEM scanning electron microscope operated at an accelerating voltage of 10 kV. The Gly-L-Ala.HI.H 2 O (Poly2)@PMMA microfibers were deposited on a silica surface previously covered with a thin film (10 nm thick) of Au-Pd (80-20 weight%) using a high-resolution sputter cover, 208 HR Cressington Company, coupled to an MTM-20 Cressington high-resolution thickness controller.
The piezoelectric output voltage and current were measured across a 100 MΩ load resistance connected to a low-pass filter, followed by a low-noise preamplifier (SR560, Stanford Research Systems, Stanford, CA, USA), before being recorded with a digital storage oscilloscope (Agilent Technologies DS0-X-3012A, Waldbronn, Germany). The fiber array sample with a (30 × 40) mm 2 area (200 µm thickness) was subjected to applied periodic mechanical forces imposed by a vibration generator (model: SF2185, Frederiksen Scientific, Olgod, Denmark), with a frequency of 3 Hz determined by a signal generator (model: 33120A, Hewlett Packard, Palo Alto, CA, USA). The applied forces were calibrated using a force-sensing resistor (FSR402, Interlink Electronics Sensor Technology, Graefelfing, Germany). The sample was fixed on a stage, and the forces were applied uniformly and perpendicularly over the surface area.

Crystal Structure
Gly-L-Ala.HI.H 2 O (Poly2) crystallizes as yellowish prisms in the monoclinic system, space group P2 1 , as a glycine-L-alanine hydroiodide salt with one water molecule. Its molecular structure is depicted in Figure 3, and the most relevant bond distances and angles are given in the caption of the corresponding Figure and in Tables S1 and S2 (Supplementary Information). As referred to above, Gly-L-Ala.HI.H2O (Poly2) is a polymorp  Figure S1. If we consider a partial electrical dipole formed by C5-C3-C4, this dipole points in the 2 1 polar axis direction with an inclination of around 30 • , Figure 4a. Moreover, the dipeptide backbone N1-C1-C2-N2-C3 also forms a partial electrical dipole inclined also ca. 45 • to the 2 1 polar axis, Figure 4b.  hand, the carboxylate group shows two distinct C-O bond lengths (C4-O2 C4-O3 1.316(7) Å), clearly indicating that the group is in the carboxylic ac allowing the assignment of the carbonyl and O-H substituents.

023, 16, x FOR PEER REVIEW
The main chain of the molecule presents an almost planar trans conf tween atoms N1 and C3 ( Figure S3), with torsion angles of 167.9(5) ° for N and 177.7(5) ° for C1-C2-N2-C3. In Poly2, the angle for C4-C3-N2-C2 is 11 torsion angle of -77.12º in C4-N1-C2-C1 from [9] indicates a gauche conform The supramolecular arrangement observed in the crystal structur Ala.HI.H2O (Poly2), when viewed along the b axis (ac plane), shows consecu iodide anions and peptide cations parallel to the b axis ( Figure 5), with the pe antiparallel. The supramolecular arrangement observed in the crystal structure of Gl Ala.HI.H2O (Poly2), when viewed along the b axis (ac plane), shows consecutive laye iodide anions and peptide cations parallel to the b axis ( Figure 5), with the peptide cat antiparallel.  Table 1.
The ammonium group (N1) displays a slightly distorted tetrahedral geometry, with the N-H distances within the group varying between 0.885 (3)   The faces of Gly-L-Ala.HI.H2O (Poly2) used for characterization studies and crystal structure determination, with the assigned Miller indices, are indicated in Figure 6a. The complete crystal morphology is depicted in Figure S2.

Pyroelectric Properties
The pyroelectric coefficient is a vector quantity with three components ( p 1 , p 2 , p 3 . For space group P2 1 , with the 2-fold screw axis along the b-axis, the vector has only one component along that crystallographic axis. The pyroelectric coefficient reported in this work was measured on a (010) orientated Gly-L-Ala.HI.H 2 O (Poly2) crystal plate. Its value reached a maximum of p = 45 µC/m 2 K at 345 K, as shown in Figure 7. The pyroelectric coefficient along the polar 2-fold axis reported for the Gly-L-Ala.HI.H 2 O (Poly1) polymorph varied between p = 2 µC/m 2 K at 100 K and p = 15.5 µC/m 2 K at 357 K [2]. Therefore, the pyroelectric coefficient of Poly2 is roughly three times bigger than that reported for Poly1. This is due to the different crystal structure arrangements inside the crystalline unit cell.

Pyroelectric Properties
The pyroelectric coefficient is a vector quantity with three components (p , p , p ). For space group P21, with the 2-fold screw axis along the b-axis, the vector has only one component along that crystallographic axis. The pyroelectric coefficient reported in this work was measured on a (010) orientated Gly-L-Ala.HI.H2O (Poly2) crystal plate. Its value reached a maximum of p = 45 μC m K ⁄ at 345 K, as shown in Figure 7. The pyroelectric coefficient along the polar 2-fold axis reported for the Gly-L-Ala.HI.H2O (Poly1) polymorph varied between p = 2 μC m K ⁄ at 100 K and p = 15.5 μC m K ⁄ at 357 K [2]. Therefore, the pyroelectric coefficient of Poly2 is roughly three times bigger than that reported for Poly1. This is due to the different crystal structure arrangements inside the crystalline unit cell.
There are in the crystal unit cell of Gly-L-Ala.HI.H2O (Poly2) four dipole moments: two from the water molecules and another two formed by NH …I , as shown in Figure  8. The first two dipoles form an angle of approximately 30° with the polar b-axis, whereas the other two dipoles (NH … I ) form an angle of approximately 20°. For Gly-L-Ala.HI.H2O (Poly1), similar dipoles are identified from the reported structure. However, the water dipoles are inclined to the b-axis by 45° [2]. Additionally, for NH …I , the bond length in Poly1 ( 2.64Å ) is shorter than the corresponding bond length in Gly-L-Ala.HI.H2O (Poly 2) (2.97Å). As a consequence, the overall net dipole moment contribution is higher for Poly2 than for Poly1, which explains the higher value for the pyroelectric coefficient reported in this work.
In Table 2, the pyroelectric coefficients of some important inorganic and semi-organic crystals are presented.  ( 2.97Å . As a consequence, the overall net dipole moment contribution is higher for Poly2 than for Poly1, which explains the higher value for the pyroelectric coefficient reported in this work.

Thermal Properties
TGA measurement results indicate that the crystal mass is stable up to initial mass loss of 2% happens at 388 K that results from water molecule e Figure S4), which is also visible in the DSC at 386 K (SI Figure S5). At aro the iodine hydrogen bonds break, visible at the sharp enthalpy peak of th alongside some mass loss. Above 473 K, there is a considerable loss of ma which corresponds to the decomposition of the crystalline compound. Th corresponds to the degradation temperature of the dipeptide glycyl-L-ala esting to note that this temperature, 533 K, is very close to the melting t ported for cyclo-glycyl-L-alanine, which is 531 K, 32 K lower than that for th L-alanine dipeptide (563 K) [39]. This suggests that although the dipeptide in its cyclic form when crystallized as the present hydroiodide salt, it keep In Table 2, the pyroelectric coefficients of some important inorganic and semi-organic crystals are presented.

Thermal Properties
TGA measurement results indicate that the crystal mass is stable up to 380 K. A small initial mass loss of 2% happens at 388 K that results from water molecule evaporation (SI, Figure S4), which is also visible in the DSC at 386 K (SI Figure S5). At around 393-413 K, the iodine hydrogen bonds break, visible at the sharp enthalpy peak of the DSC (393 K), alongside some mass loss. Above 473 K, there is a considerable loss of mass (about 50%), which corresponds to the decomposition of the crystalline compound. The peak at 533 K corresponds to the degradation temperature of the dipeptide glycyl-L-alanine. It is interesting to note that this temperature, 533 K, is very close to the melting temperature reported for cyclo-glycyl-L-alanine, which is 531 K, 32 K lower than that for the linear glycyl-L-alanine dipeptide (563 K) [39]. This suggests that although the dipeptide in Poly2 is not in its cyclic form when crystallized as the present hydroiodide salt, it keeps a memory of its initial cyclic closed chain, therefore showing a thermal memory effect (one should remember that the crystal synthesis started from the cyclic dipeptide form and after reacting with the iodide acid the cyclic chain opened up).

Dielectric Spectroscopy
Figures 9 and 10 show the temperature dependence of the real (ε ) and imaginary (ε ) parts of the dielectric permittivity measured in the range 287 K-370 K for different frequencies from 100 Hz to 100 kHz. Two regimes are identified in both the real and imaginary parts of the dielectric permittivity, with different dependences below and above 350 K. The real part of the electric permittivity is approximately constant and less than 100 from room temperature until 343 K, both for low and high frequencies. However, beyond 350 K it increases very steeply, reaching the value of 1100 for 100 Hz, as shown in Figure 9. Similarly, the imaginary part of the electric permittivity, ε″, is als constant until 350 K and smaller than 25 for all frequencies, as seen in Fig  increases steeply beyond 350 K, reaching 160 000 at T ~ 375 K and a frequ Similarly, the imaginary part of the electric permittivity, ε , is also approximately constant until 350 K and smaller than 25 for all frequencies, as seen in Figure 10. Again, it increases steeply beyond 350 K, reaching 160 000 at T~375 K and a frequency of 100 Hz. to 10 kHz.
Similarly, the imaginary part of the electric permittivity, ε″, is also ap constant until 350 K and smaller than 25 for all frequencies, as seen in Figure  increases Figure 10. The imaginary part of the electric permittivity, ε″, as a function of tempe quencies up to 10 kHz. Figures 11 and 12 show the frequency dependence of the real and imag permittivity at different temperatures. Both ε′ and ε″ present an initial sharp low-frequency region and afterwards attain a slower decrease at high freq Figure 10. The imaginary part of the electric permittivity, ε , as a function of temperature for frequencies up to 10 kHz. Figures 11 and 12 show the frequency dependence of the real and imaginary electric permittivity at different temperatures. Both ε and ε present an initial sharp drop in the low-frequency region and afterwards attain a slower decrease at high frequencies. The initial, low-frequency drop in the imaginary part reveals a contribution from a conductivity term, as the samples are non-ideal capacitors.
ials 2023, 16, x FOR PEER REVIEW initial, low-frequency drop in the imaginary part reveals a contribution from a ity term, as the samples are non-ideal capacitors. Frequency (Hz) Figure 11. The real part of the electric permittivity, ε′, as a function of frequency for d peratures.
40 Figure 11. The real part of the electric permittivity, ε , as a function of frequency for different temperatures. Frequency (Hz) Figure 11. The real part of the electric permittivity, ε′, as a function of frequency for peratures. For purely electronic conductivity, the permittivity is imaginary and = σ DC /(ε 0 ω) [40], where ε 0 is the vacuum dielectric permittivity, ω is the quency and σ DC is the DC conductivity. For ionic charge carriers that caus Maxwell-Wagner polarization effects, this equation can be generalized, so ductivity contribution can be described by the equation ε″ = σ DC /(ε 0 ω s ), wh ponent and s ≤ 1. As such, since the logarithm of the imaginary component tivity as a function of the logarithm of the frequency gives a linear depen fitted with a straight line to determine the conductivity according to: For purely electronic conductivity, the permittivity is imaginary and given by ε = σ DC /(ε 0 ω) [40], where ε 0 is the vacuum dielectric permittivity, ω is the angular frequency and σ DC is the DC conductivity. For ionic charge carriers that cause electrode or Maxwell-Wagner polarization effects, this equation can be generalized, so that the conductivity contribution can be described by the equation ε = σ DC /(ε 0 ω s ), where s is an exponent and s ≤ 1. As such, since the logarithm of the imaginary component of the permittivity as a function of the logarithm of the frequency gives a linear dependence, it was fitted with a straight line to determine the conductivity according to: The inset of Figure 12 shows the linear fit to ln(ε ) as a function of the logarithm of the frequency for different temperatures in the low-frequency region. From the fits, the corresponding calculated DC conductivity (σ DC ) values and their temperature dependence are shown in Figure 13. Again, a small variation in the DC conductivity is observed until 350 K, above which the conductivity rises sharply. The σ DC behavior as a function of temperature in both regions shows characteristic Arrhenius-like processes, with activation energies (E at ) given by the equation [40,41]: where T is the temperature, k B is the Boltzmann constant and σ 0 is a constant. The activation energy can be determined from the slopes of the fittings to the curves of ln(σT) as a function of the inverse of T for the different temperature regions, as shown in the inset of Figure 13. The two temperature regions with different conductivities and the corresponding activation energy (E at ) values are shown in the figure. E at = 0.07 eV corresponds to the low-temperature region and is characteristic of electrical conduction through the polaron transport behavior [42]. At higher temperatures, the activation energy increases to E at = 3.6 eV, which is characteristic of ionic conduction in the samples [40]. As such, the two observed regimes are due to the change from the low-temperature, polaronic transport behavior to the high-temperature ionic conductivity dependence. This region is associated with the onset of temperature-induced changes in the samples (e.g., loss of water as observed from the TGA results, which starts just above 350 K).
T where T is the temperature, kB is the Boltzmann constant and σ0 is a constant. The act tion energy can be determined from the slopes of the fittings to the curves of ln(σT) function of the inverse of T for the different temperature regions, as shown in the inse Figure 13. The two temperature regions with different conductivities and the correspo ing activation energy (E at ) values are shown in the figure. E at = 0.07 eV correspond the low-temperature region and is characteristic of electrical conduction through the laron transport behavior [42]. At higher temperatures, the activation energy increase E at = 3.6 eV, which is characteristic of ionic conduction in the samples [40]. As such, two observed regimes are due to the change from the low-temperature, polaro transport behavior to the high-temperature ionic conductivity dependence. This regio associated with the onset of temperature-induced changes in the samples (e.g., loss of ter as observed from the TGA results, which starts just above 350 K). 1/T (K -1 ) Figure 13. Temperature dependence of the electrical conductivity as determined from the low quency behavior of the imaginary permittivity of Figure 7. The inset shows ln(σT) as a functio the inverse of temperature. The slopes of the linear fits give the activation energies in the two t perature intervals.

Second Harmonic Response
Data for the second harmonic response of an approximately 3 mm-thick Gly Ala.HI.H2O (Poly2) dipeptide crystal (taken from that shown in Figure 1) were acqui using incident fundamental pulses with an average incident power of 5 mW, correspond to roughly 66 pJ of energy per pulse. The beam was incident normal to the as-grown cry surface, and the crystal was scanned over the laboratory x, y and z directions, with z ta Figure 13. Temperature dependence of the electrical conductivity as determined from the lowfrequency behavior of the imaginary permittivity of Figure 7. The inset shows ln(σT) as a function of the inverse of temperature. The slopes of the linear fits give the activation energies in the two temperature intervals.

Second Harmonic Response
Data for the second harmonic response of an approximately 3 mm-thick Gly-L-Ala.HI.H 2 O (Poly2) dipeptide crystal (taken from that shown in Figure 1) were acquired using incident fundamental pulses with an average incident power of 5 mW, corresponding to roughly 66 pJ of energy per pulse. The beam was incident normal to the as-grown crystal surface, and the crystal was scanned over the laboratory x, y and z directions, with z taken to be the direction of the laser beam propagation. The maximum signal values as a function of the polarization of the fundamental beam direction are presented in Figure 14.
Second harmonic signals were observable only for a narrow range of z positions when the fundamental beam waist was within a few tens of µms from the crystal surface. Furthermore, the orientation of the analyzer that resulted in the maximum detected signal was very nearly parallel to the direction of the incident polarization. We believe this to be an indication that the normal of the as-grown crystal is close to the crystallographic b axis. As explained in the supplementary information, for this orientation, the second harmonic light will be generated with nearly the same polarization as the fundamental beam and will suffer from strong phase mismatch, limiting the generation to close to the crystal surface.
We have carried out a study of the second harmonic response as a function of the fundamental beam's waist position by translating the crystal in 10 µm steps along the beam's propagation direction using the MadCity's piezoelectric translation stage. Representative data are shown in Figure 15, along with a theoretical fit as described in the supplementary information ( Figure S8). to be the direction of the laser beam propagation. The maximum signal values as a functio of the polarization of the fundamental beam direction are presented in Figure 14. Second harmonic signals were observable only for a narrow range of z position when the fundamental beam waist was within a few tens of μms from the crystal surface Furthermore, the orientation of the analyzer that resulted in the maximum detected signa was very nearly parallel to the direction of the incident polarization. We believe this to b an indication that the normal of the as-grown crystal is close to the crystallographic b axis As explained in the supplementary information, for this orientation, the second harmoni light will be generated with nearly the same polarization as the fundamental beam an will suffer from strong phase mismatch, limiting the generation to close to the crystal sur face.
We have carried out a study of the second harmonic response as a function of th fundamental beam's waist position by translating the crystal in 10 µm steps along th beam's propagation direction using the MadCity's piezoelectric translation stage. Repre sentative data are shown in Figure 15, along with a theoretical fit as described in the sup plementary information ( Figure S8). assuming Gaussian spatial and temporal profiles for the two beams. The 0 position was chosen to be close to the maximum SHG signal. We estimate that the coherence length due to phase mismatch is approximately 1.7 μm, whereas the Rayleigh range for the fundamental beam is roughly an order of magnitude larger at 16.4 μm.
By calibrating the efficiency of our second harmonic microscope using a 2 mm-thick BBO crystal cut, the phase matching angle for 800 nm incident light, we can estimate a lower bound for the effective nonlinear susceptibility of the dipeptide crystal. Applying the results described in the supplementary information, we estimate that a lower bound Figure 15. The second harmonic response (blue dots) as the crystal was translated in 10 µm steps along the propagation direction of the fundamental beam. Also shown is a theoretical fit (solid line) assuming Gaussian spatial and temporal profiles for the two beams. The 0 position was chosen to be close to the maximum SHG signal. We estimate that the coherence length due to phase mismatch is approximately 1.7 µm, whereas the Rayleigh range for the fundamental beam is roughly an order of magnitude larger at 16.4 µm.
By calibrating the efficiency of our second harmonic microscope using a 2 mm-thick BBO crystal cut, the phase matching angle for 800 nm incident light, we can estimate a lower bound for the effective nonlinear susceptibility of the dipeptide crystal. Applying the results described in the supplementary information, we estimate that a lower bound for the effective second-order nonlinear coefficient of Gly-L-Ala.HI.H 2 O (Poly2) crystals is d eff ≥0.14 pm/V.

Piezoelectric Response
Any pyroelectric material is, by symmetry, also a piezoelectric and nonlinear optical material. Additionally, the tensor describing piezoelectricity and SHG properties of crystalline materials are the same (in this case, that for point group 2), as presented in S6. Therefore, the suitability of Gly-L-Ala.HI.H 2 O (Poly2) crystals to be used as piezoelectric nanogenerators for energy harvestings was investigated by embedding them into electrospun nanofibers as described before.
An interconversion between a mechanical and an electrical stimulus arising due to applied uniform stress, which generates electric polarization inside a dielectric material, is the origin of the piezoelectric effect. For a crystalline solid to display this phenomenon, it must have a crystal structure without inversion symmetry. Gly-L-Ala.HI.H 2 O (Poly2) crystallizes in the polar point group 2, which is acentric. The tensor relationship between the polarization P j and stress σ k , P j = d jk σ k is given by the piezoelectric coefficient d jk [43]. There is no preferential crystallographic orientation of the compound inside the electrospun fibers (see SI7). Therefore, a polarization develops under forces applied repeatedly at regular times perpendicularly to the fiber array, and an effective piezoelectric modulus d eff is measured along the same direction.
The applied stress (force per unit area) ranged between 1.5 × 10 2 Nm −2 and 8.0 × 10 2 Nm −2 . For Gly-L-Ala.HI.H 2 O (Poly2)@PMMA fiber mats, a 1.0 N applied periodical force gave rise to a maximum instantaneous output piezoelectric voltage and current of 28 V and 280 nA, respectively, as shown in Figure 16a. Here, the two opposite peaks correspond to the press and release of the fiber mat. A plot of the output voltage as a function of several applied periodic forces shows an output voltage increasing linearly with the force magnitude as expected (Figure 16b). The PMMA polymer matrix is not piezoelectric and does not contribute to the measured piezoelectric voltage. Taking into account a response time of 1 ms, the magnitude of d eff is obtained from the integration of the induced piezoelectric current over that period of time, Q = Idt, resulting in Q = 280 pC for Gly-L-Ala.HI.H 2 O (Poly2)@PMMA fiber mats. This induced charge, which is related to the applied force by the equation Q = F d eff , allows us to calculate an effective piezoelectric coefficient equal to d eff = 280 pCN −1 . This value is of the same order of magnitude as that obtained for organic-inorganic ferroelectric perovskite trimethylchloromethyl ammonium trichloromanganese ((TMCM)MnCl 3 ), where d 33 = 185 pCN −1 , and barium titanate (BaTiO 3 ), with d 33 = 190 pCN −1 [44,45]. It is also important to compare the present result, d eff = 280 pCN −1 , with that obtained for lead-free organic-inorganic perovskite (N-methyl-N -diazabicyclo [2.2.2]octonium)-ammonium triiodide (MDABCO-NH 4 I 3 ) embedded into PMMA electrospun fibers (MDABCO-NH 4 I 3 @PMMA in a 1:5 ratio), which was reported to be d eff = 64 pCN −1 [46], (Table 3).
It is remarkable that this new organic-inorganic Gly-L-Ala.HI.H 2 O (Poly2) crystalline compound exhibits, when embedded into electrospun fibers, a very highly effective piezoelectric coefficient that is similar in magnitude to an organic-inorganic perovskite also containing the iodide ion. In the present work, we demonstrate that Gly-L-Ala.HI.H 2 O (Poly2)@PMMA fibers may be incorporated into nanogenerators as active piezoelectric materials.

Conclusions
A polymorph (Gly-L-Ala.HI.H 2 O (Poly2)) of a previously reported glycyl-L-alanine HI.H 2 O salt was synthesized from the chiral cyclo-glycyl-L-alanine dipeptide. The dipeptide is known to show molecular flexibility in different environments, which originated the polymorphism. The crystal structure of the glycyl-L-alanine HI.H 2 O polymorph is determined at room temperature in space group P2 1 ; therefore, it is a pyroelectric, piezoelectric and nonlinear optical material.
The pyroelectric coefficient reported in this work on a (010)-orientated Gly-L-Ala.HI.H 2 O (Poly2) crystal plate showed an increase with temperature with no significant abnormalities in the range 300-345 K, reaching a maximum of p = 45 µC/m 2 K at 345 K. Therefore, the pyroelectric coefficient of Poly2 is roughly three times higher than that reported for the Gly-L-Ala.HI.H 2 O (Poly1) polymorph, which was p = 15.5 µC/m 2 K at 357 K. The different orders of magnitude of the measured values for the two polymorphs results from the different atomic coordinates of the dipeptide, water molecules and the iodine ions within the unit cells of both compounds, which creates differences in some bond lengths and the directionality of hydrogen bonds.
Thermal studies showed that Gly-L-Ala.HI.H 2 O (Poly2) begins degradation at 533 K, close to the melting temperature reported for cyclo-glycyl-L-alanine, which is 531 K. That temperature (533 K) is 30 K lower than that reported for linear glycyl-L-alanine dipeptide (563 K), suggesting that although the dipeptide when crystallized in Poly2 is not in its cyclic form anymore, it keeps a memory of its initial closed chain, therefore showing a thermal memory effect.
The DC conductivity behavior as a function of temperature in the regions below and above 350 K indicates characteristics of electrical conduction through the polaron transport behavior in the low-temperature region and at higher temperatures, which is characteristic of ionic conduction in the samples.
The second harmonic generation efficiency of Gly-L-Ala.HI.H 2 O (Poly2) was measured against a state-of-the-art nonlinear optical barium borate (BBO) crystal. A lower bound for the effective second-order nonlinear coefficient of Gly-L-Ala.HI.H 2 O (Poly2) crystals was estimated d eff ≥ 0.14 pm/V.
Finally, an effective piezoelectric coefficient equal to d eff = 280 pCN −1 was measured on an electrospun polymer fiber mat, Gly-L-Ala.HI.H 2 O (Poly2)@PMMA, demonstrating that the fibers are piezoelectrically active systems with great potential to be incorporated into energy harvesting devices.