1. Introduction
Iodine(v) oxides (i.e. I2O5, HIO3, HI3O8, I4O9) are compounds with unique optical properties and potential for high energy release. The unique optical properties of iodine oxides, specifically iodic acid (HIO3), are a result of the non-centrosymmetric space group P212121 [1] , and have led to the synthesis of different aluminum iodate species [2] [3] . Iodine oxides are also appealing for use as an oxidizer when combined with aluminum fuel particles. In fact, Smith et al. [4] showed detonation velocities could be achieved from aluminum iodate mixtures. The aluminum and iodine oxide reactions also have the potential to disperse high temperatures and aerosolized iodine species that can kill bacterial agents [5] . The optical properties and potential for energy release of iodine oxides are controlled by the crystalline structure of precipitated HIO3 product [2] [3] [4] .
There is only one commonly accepted polymorph for both I2O5 and HI3O8; however, there are four reported polymorphs of HIO3 in literature: α, β, γ, and δ [1] [6] [7] [8] [9] . The most commonly reported structure of iodic acid is α-HIO3. The γ-HIO3 polymorph was first reported by Fischer et al. [1] and structurally determined by single crystal diffraction on crystals produced by mixing iodic acid with chromium. The γ-HIO3 polymorph is thought to be a result of chromium in solution favoring the formation of dimers and trimers of HIO3 that form γ-HIO3 as solutions of HIO3 and chromium precipitate. On the other hand, the β-HIO3 polymorph was first reported in 1960 by Halasz et al. [8] with peak intensities determined from a powder diffraction experiment (Table S1); however, the diffraction pattern was not indexed. The β-HIO3 polymorph is also mentioned briefly in Selte et al. [7] , but is later questioned by Fischer et al. [1] and to our knowledge, these are the only known reports of a β polymorph of HIO3. The δ polymorph was recently reported in Wu et al. [9] , but detailed crystal information was not given. It is suggested that δ-HIO3 is metastable at elevated temperatures and slow heating converts δ-HIO3 directly into I2O5. The dehydration steps for the formation of HIO3 are not well understood, but the formation of dimers, trimers and higher order polymers have been related to concentration of HIO3 [10] [11] [12] . Many studies show polymerization of
from solution [10] [11] [12] , but a link between polymerization and formation of α-HIO3 has not been established.
In this study, we will report on the single crystal structure of β-HIO3 that has not previously been reported and understand the polymorphic physical behavior of β-HIO3. To this end, the objectives were to synthesize β-HIO3 and characterize its structure using single crystal X-ray diffraction (XRD).
2. Experimental
2.1. Sample Preparation
The synthesis method to form β-HIO3 begins with dissolving commercially available I2O5 powder, supplied by Sigma Aldrich (St. Louis, MO), in distilled water. Commercial I2O5 was mixed to a 1:1 wt. % ratio of I2O5 to deionized water and placed in a beaker on a magnetic stirrer and stirred for 10 minutes at 200 RPM to allow complete I2O5 dissolution into solution. The iodate solution was then dried in a vent hood with an average relative humidity (RH) of 35% at 23˚C, until solid crystals precipitated from solution. Multiple samples (i.e., >5) were made to establish repeatability.
2.2. Single Crystal XRD
Single crystal X-ray diffraction data were collected on a Bruker PLATFORM three circle diffractometer equipped with an APEX II CCD detector and operated at 1500 W (50 kV, 30 mA) to generate (graphite monochromated) Mo Kα radiation (λ = 0.71073 Å). Crystals were transferred from the reaction vial and placed on a glass slide in polyisobutylene. A Zeiss Stemi 305 microscope was used to identify a suitable specimen for X-ray diffraction from a representative sample of the material. The crystal and a small amount of the oil were collected on a MῑTiGen cryoloop and transferred to the instrument where it was placed under a cold nitrogen stream (Oxford) maintained at 100 K throughout the duration of the experiment. The sample was optically centered with the aid of a video camera to ensure no translations were observed as the crystal was rotated through all positions.
A unit cell collection was then carried out. After it was determined that the unit cell was not present in either the CCDC or ICSD databases a sphere of data were collected. Omega scans were carried out with a 10 sec/frame exposure time and a rotation of 0.50˚ per frame. After data collection, the crystal was measured for size, morphology, and color. These values are reported in Table S2.
2.3. Single Crystal XRD Refinement Details
After data collection, the unit cell was re-determined using a subset of the full data collection. Intensity data were corrected for Lorentz, polarization, and background effects using the Bruker program APEX 3. A semi-empirical correction for adsorption was applied using the program SADABS [13] . The SHELXL-2014 [14] , series of programs was used for the solution and refinement of the crystal structure. The hydrogen atom bound to O1 was constrained with a DFIX command and a thermal parameter of −1.2. An extinction coefficient of 0.0298 was also applied during the final refinement.
2.4. Powder Diffraction XRD
All powder diffraction data were collected on a Rigaku Ultima III powder diffractometer. X-ray diffraction patterns were obtained by continuously scanning a 2θ range of 15˚ - 60˚, step size = 0.02˚, and scan time ranging from of 1.5 - 3 degrees/minute depending on the scan. The X-ray source was Cu Kα radiation (λ = 1.5418 Å) with an anode voltage of 40 kV and a current of 44 mA. The beam was then discriminated by Rigaku’s Cross Beam parallel beam optics to create a monochromatic parallel beam. Diffraction intensities were recorded on a scintillation detector after being filtered through a Ge monochromator. Samples were prepared as standard powder mounts and diffractograms were processed through the software JADE v9.1.
3. Results
During the initial PXRD studies of what was presumed to be α-HIO3, it was discovered that the resulting patterns did not match any indexed crystal structure. Single crystal studies were then carried out on the large blocky crystals. The β-HIO3 phase reported in Halasz et al. [8] is shown in Supplemental Information. Table S1 and is similar to the lattice parameters shown here. Because of the similarities between our results and Halasz et al. [8] , and to avoid adding any more Greek letters to describe the phases of HIO3, it is assumed crystals shown here are the β-HIO3 polymorph. Since unit cell data is not reported for δ-HIO3 in Wu et al. [9] , only a visual comparison between β-HIO3 and δ-HIO3 could be made and is discussed further below. Unit Cell data, bond lengths and angles, atomic coordinates, anisotropic displacement parameters, hydrogen coordinates, and hydrogen bonds for β-HIO3 determined by single crystal data is shown in Supplemental Information Tables S2-S7. While the crystal structure of β-HIO3 was also solved in the space group α-HIO3 (P212121), it was found that the unit cell axes were all different by about 1 Å. Similar to that of α and γ phases, the unit cell contains only a single HIO3 molecule in the asymmetric unit with I-O bond lengths ranging from 1.786(5) to 1.903(7) Å (Table S3). The I(V) atom is further coordinated by three oxygen atoms of neighboring acid molecules forming a distorted octahedral with a range of I-O distances (2.498(6) - 2.795(7) Å). The one structural difference, shown in Figure 1 (e.g., α-HIO3 shown in Supplemental Information, Figure S1), that separates the β phase from the α and γ phases is that the hydroxyl group is bridging between two I(V) atoms, resulting in a smaller hydrogen bonding distance (O-O distance: 2.559 Å (β), 2.665 Å (γ) and 2.696 Å (α)) and presumably a different crystalline energy.
Figure 1. (a) Coordination environment about the I(V) atom in β-HIO3. The thermal ellipsoids are drawn at 50% probability level with solid and dashed lines representing short or long contacts between I and O atoms, respectively. Red, violet, and light gray ellipsoids are oxygen, iodine, and hydrogen atoms, respectively. (b) Unit cell of β-HIO3 viewed along the axis. Red, violet, and light gray spheres represent oxygen, iodine, and hydrogen atoms, respectively.
4. Discussion
During repeatability testing, β-HIO3 readily converted to α-HIO3, indicating β-HIO3 is metastable, similar to the γ phase. The metastable nature of β-HIO3 is shown in Figure 2. The XRD measurements in Figure 2 are from crystals ground into powder using mortar and pestle before data collection. In Figure 2, the top curve was collected immediately after drying. This sample is referred to as Sample 1 initial. The middle curve in Figure 2 was initially ground after drying and allowed to sit in a vent hood at 20% RH for 8 days before analysis. The middle curve in Figure 2 is pure α-HIO3 and matches PDF 97-006-6643 [15] and the top curve is β-HIO3. The bottom curve in Figure 2 shows XRD measurements that were taken from Sample 1 that was put in a vent hood for 8 days prior to grinding and then ground for XRD measurements. The bottom curve in Figure 2 shows that when Sample 1 is placed in a vent hood for 8 days as a single crystal prior to grinding and then ground for XRD measurements, it remained as β-HIO3.
Figure 2 shows that β-HIO3 is metastable and changes to α-HIO3 over time only if the samples are ground prior to aging. Since all samples are ground before placement in the XRD, the change from β-HIO3 to α-HIO3 is not directly related to the physical force of the grinding and is a time dependent process related to increased surface area from grinding. It is assumed that trapped water results in the formation of β-HIO3. Upon grinding, trapped water is released and drying the solid material readily converts β-HIO3 into α-HIO3. The validity of the assumption that trapped water results in the formation of β-HIO3 is discussed below.
Figure 3 shows Sample 1 with curves from PDF #00-045-0872 labeled as I4O9XH2O from Wikjord et al. [6] . Figure 3 shows that the sample labeled I4O9XH2O is β-HIO3.
Figure 2. Three XRD scans for Sample 1 initially after drying (top scan), and Sample 1 after grinding then drying for 8 days (middle scan), Sample 1 that was dried for 8 days before grinding (bottom scan).
Figure 3. Diffraction pattern from XRD for Sample 1 (Black). Blue curves from PDF #00-045-0872 indicating 2-Theta and relative intensity of I4O9XH2O.
In Wikjord et al. [16] , the sample labeled I4O9XH2O was I4O9 hydrated by atmospheric water. In Smith et al. [17] , it was shown that when I4O9 is exposed to 20% RH for 4 hours, I4O9 hydrates to HIO3. Because β-HIO3 is seen during hydration of I4O9 into α-HIO3 and β-HIO3 is seen here when I2O5 is mixed with water, β-HIO3 is assumed to be a transition step in the formation of α-HIO3. We have shown that β-HIO3 is metastable and converts to α-HIO3 when ground prior to aging. Because β-HIO3 is seen during the formation of α-HIO3 and β-HIO3 is metastable when ground prior to aging, we propose that the formation of β-HIO3 is an intermediate step in the formation of α-HIO3 caused by water trapped inside the precipitated crystals.
Visual comparison of diffraction patterns between δ-HIO3 shown in Wu et al. [9] and diffraction patterns shown here for β-HIO3 are similar. Also, similar comparisons between these phases and I4O9XH2O have been made here and in Wu et al. [9] , suggesting δ-HIO3 and β-HIO3 are the same phase. The 6 reported I-O bond distances for δ-HIO3 are similar (less than 0.03 Å difference) to bond distances reported here for β-HIO3. Similar bond distances and visual comparisons of diffraction patterns suggest δ-HIO3 and β-HIO3 are the same crystal structure. In Wu et al. [9] , it is reported that δ-HIO3 dehydrates directly into I2O5, not HI3O8 during slow heating conditions of DSC analysis. Because of this direct dehydration into I2O5 during slow heating, δ-HIO3 is reported as metastable. This can be explained by the metastable nature of β-HIO3 reported here and results from Smith et al. [18] showing that HIO3 dehydrates directly into I2O5 and HI3O8 is formed during heating. Since comparison between measurements in Halasz et al. [8] and diffraction patterns obtained, using modern XRD are difficult to quantify, it is probable that δ-HIO3 identified by Wu et al. [9] is actually β-HIO3.
5. Conclusion
The β-HIO3 polymorph, previously difficult to detect and whose existence was questioned, has been synthesized and structurally characterized. When crystals are ground before aging, β-HIO3 is metastable and slowly converts to α-HIO3. Experiments were designed to study the metastable nature and based on XRD analysis. β-HIO3 may be a transition step in the formation of α-HIO3 and result from trapped water inside HIO3 crystals.
Acknowledgements
The authors are grateful for support from the Army Research Office under award W911NF-14-1-0250 and encouragement from our program manager, Dr. Ralph Anthenien.
Supplemental Materials
(a) Graphical representation of α-HIO3(b) Graphical representation of γ-HIO3
Figure S1. Graphical Representation of (a) α-HIO3 and (b) γ-HIO3. Red, violet, and pink ellipsoids are oxygen, iodine, and hydrogen atoms, respectively.
Table S1. XRD information from Halasz et al.8
Table S3. Bond lengths [Å] and angles [˚] for β-HIO3.
Table S4. Atomic coordinates (×104) and equivalent isotropic displacement parameters (Å2 × 103) for β-HIO3. U(eq) is defined as one third of the trace of the orthogonalized Uij tensor.
Table S5. Anisotropic displacement parameters (Å2 × 103) for β-HIO3. The anisotropic displacement factor exponent takes the form: −2 π2 [h2 a*2 U11 + ・・・ + 2 h k a* b* U12].
Table S6. Hydrogen coordinates (×104) and isotropic displacement parameters (Å2 × 103) for β-HIO3.
Table S7. Hydrogen bonds for β-HIO3 where hydrogen bonds with H..A < r(A) + 2.000 Å and
110˚ are listed.