Zintl and Intermetallic Phases Grown from Calcium/Lithium Flux Trevor Blankenship

Florida State University Libraries

Electronic Theses, Treatises and Dissertations The Graduate School

2014 Zintl and Intermetallic Phases Grown from Calcium/Lithium Flux Trevor Blankenship

Follow this and additional works at the FSU Digital Library. For more information, please contact [email protected] FLORIDA STATE UNIVERSITY

COLLEGE OF ARTS AND SCIENCES

ZINTL AND INTERMETALLIC PHASES GROWN FROM CALCIUM/LITHIUM FLUX

TREVOR BLANKENSHIP

A Dissertation submitted to the Department of Chemistry and Biochemistry in partial fulfillment of the requirements for the degree of Doctor of Philosophy

Degree Awarded: Fall Semester, 2014

Trevor Blankenship defended this dissertation on September 25, 2014.

The members of the supervisory committee were:

Susan Latturner

Professor Directing Dissertation

Bruce Locke

University Representative

Albert Stiegman

Committee Member

Igor Alabugin

Committee Member

The Graduate School has verified and approved the above-named committee members, and certifies that the dissertation has been approved in accordance with university requirements.

ACKNOWLEDGEMENTS

I would like to foremost give my deep appreciation to Dr. Susan Latturner whose advice and guidance were critical in completion of my degree. She has been committed to her mentorship role and my development as a scientist. The NMR studies were possible due to the work of Dr. Chen who enthusiastically tackled some difficult problems. Dr. Clark has been gracious in giving time and knowledge to assist with solving crystal structures. I would also like to thank the past members of the Latturner group who have assisted me in matters great and small over the years. In particular, Dr. Gina Canfield, Dr. Josiah Matthieu, and Dr. Patricia Tucker have all greatly helped. This work was built on the work of David Lang who gave a great starting point and introduced me to the basics of flux synthesis. Finally, I would like to thank my parents who have given their full support in everything I have wished to do. Unfortunately my father was not able to see me graduate; this work is dedicated to his memory.

iii

TABLE OF CONTENTS List of Tables ...... vi List of Figures ...... vii Abstract ...... x

1. INTRODUCTION ...... 1 1.1 Metal Flux Synthesis...... 1 1.2 Intermetallics...... 2 1.3 Zintl Phases ...... 2

2. EXPERIMENTAL TECHNIQUES ...... 4 2.1 Synthesis in the Ca/Li Flux ...... 4 2.2 SEM-EDS ...... 5 2.3 X-ray Photoelectron Spectroscopy ...... 7 2.4 X-Ray Diffraction ...... 7 2.5 Solid-State Nuclear Magnetic Resonance ...... 8 2.6 Band Structure Calculation ...... 9

3. COMPLEX CARBIDE PHASES CA11E3C8 (E = Sn,Pb) GROWN FROM THE Ca/Li FLUX ...... 10 3.1 Introduction ...... 10 3.2 Experimental ...... 11 3.3 Results and Discussion ...... 15 3.4 Conclusions ...... 23

4. LICA3AS2H AND CA14AS6X7 (X = C, H, N): TWO NEW ARSENIDE HYDRIDE SALTS GROWN FROM Ca/Li METAL FLUX ...... 24 4.1 Introduction ...... 24 4.2 Experimental Section ...... 25 4.3 Results and Discussion ...... 30 4.4 Conclusions ...... 38

5. Ca54In13B4–xH23+x: A COMPLEX METAL SUBHYDRIDE FEATURING IONIC AND METALLIC REGIONS ...... 39 5.1 Introduction ...... 39 5.2 Experimental ...... 40 5.3 Results and Discussion ...... 46 5.4 Conclusions ...... 56

6. ALKALINE EARTH INDIUM ALLENYLIDES SYNTHESIZED IN AE/Li FLUX

(AE = Ca, Ba) ...... 57 6.1 Introduction ...... 57 6.2 Experimental Procedure ...... 58 6.3 Results and Discussion ...... 62 6.4 Conclusions ...... 73

7. Future Work ...... 74 7.1 Introduction ...... 74 7.2 Synthesis ...... 74 7.3 Structure of Ca31H21Al2...... 75 7.4 Structure of Ca4Al2N5 ...... 78 7.5 Structure of Ca24Al9(C1-xHx)N2H16 ...... 79

8. CONCLUSIONS...... 83

REFERENCES ...... 84

BIOGRAPHICAL SKETCH ...... 94

LIST OF TABLES

Table 3.1 Crystallographic data and collection parameters for Ca11E3C8 phases...... 12

Table 3.2 Atomic coordinates and isotropic thermal parameters of Ca11Sn3C8...... 13

Table 3.3 Atomic coordinates and isotropic thermal parameters of Ca11Pb3C8...... 14

Table 3.4 Bond lengths of interest in Ca11Tt3C8 phases, in angstroms...... 20

Table 4.1 Crystallographic data and collection parameters for title phases...... 28

Table 4.2 Atomic positions and site occupancies for Ca14As6C0.46N1.155H5.045...... 29

Table 4.3 Bond lengths of interest in arsenide hydride phases, in angstroms...... 31

Table 5.1 Crystallographic data and collection parameters for two samples of Ca53In13B4H23 ...... 43

Table 5.2 Atom positions and isotropic thermal parameters for Ca53In13B4H23 ...... 44

Table 5.3 Atom positions and isotropic thermal parameters for the second crystal of Ca53In13B4H23 ...... 44

Table 6.1 Crystallographic data collection parameters for the title phases...... 61

Table 6.2 Atomic coordinates and isotropic thermal parameters for the title phases...... 62

Table 6.3 Bond lengths (Å) in title phases...... 66

Table 7.1 Crystal Data and Structure refinement for the presented Al structures ...... 81

LIST OF FIGURES Figure 2.1 Phase diagram of Ca and Li...... 5

Figure 2.2 A crystal imaged from secondary electrons on the left and backscattered electrons on the right...... 6

Figure 3.1 Mass spectrum of products of hydrolysis of Ca11Pb3C8...... 16

Figure 3.2 The structure of Ca11Tt3C8...... 18

Figure 3.3 Carbide anions in Ca11Tt3C8, with carbon−carbon bond lengths indicated...... 19

Figure 3.4 Bicapped square antiprism coordination of an Tt4− anion (red sphere) by calcium (blue spheres) in Ca11E3C8 phases...... 20

Figure 3.5 Raman spectrum of Ca11Sn3C8 showing the modes of the molecular carbon anions...... 21

Figure 3.6 Density of states data for (a) Ca11Sn3C8 and (b) Ca11Pb3C8...... 22

Figure 4.1 The orthorhombic structure of LiCa3As2H, viewed down the b-axis, with different anion coordination polyhedra highlighted...... 32

Figure 4.2 Calculated total and partial density of states data for LiCa3As2H...... 33

Figure 4.3 The tetragonal structure of Ca14As6C0.46N1.155H5.045, viewed down the c-axis, with different anion coordination polyhedra highlighted...... 34

Figure 4.4 Local coordination environments for anions in Ca14As6C0.46N1.155H5.045...... 35

Figure 4.5 The absorbance spectrum of Ca14As6X7 (X = C, H, N) from diffuse reflectance data...... 36

Figure 5.1 X-ray Photoelectron Spectroscopy (XPS) data for Ca54In13B4-xH23+x, highlighting the very strong Ca 2p photoelectron peak, weak B 1s peak, and very weak Li 1s photoelectron peak(likely due to traces of flux residue on the sample surface), in their expected binding energy regions ...... 41

Figure 5.2 Powder X-ray diffraction data for Ca54In13B4-xH23+x, compared to the theoretical pattern calculated based on single crystal structure...... 43

Figure 5.3 Structure of Ca54In13B4–xH23+x...... 47

Figure 5.4 Ordered crystallographic sites of Ca54In13B4–xH23+x...... 48

vii

Figure 5.5 An alternative view of the In(1)-centered cluster inof Ca54In13B4–xH23+x, highlighting its similarity to Bergman clusters found in quasicrystalline approximants such as NaAuSn...... 49

Figure 5.5 Disordered region of the Ca54In13B4–xH23+xstructure...... 50

1 Figure 5.6 H MAS NMR spectra collected on Ca54In13B4–xH23+xusing a 1.3 mm rotor spinning at60 kHz, referenced to TMS at 0 ppm ...... 52

115 Figure 5.7 In NMR spectrum for on Ca54In13B4–xH23+x, collected using the stepped frequency technique (top), compared to calculated total spectrum and individual contributions from In(1) and In(2) sites...... 54

Figure 5.8 Density of states data with the B(3)/H(3) site fully occupied by hydrogen (left), and just the hydrogen density of states.(right)...... 54

Figure 5.9 Density of states data with the B(3)/H(3) site fully occupied by boron in the left graph...... 55

Figure 6.1 Mass spectra of products of hydrolysis reactions of Ca12InC13-x and Ba12InC18H4...... 64

Figure 6.2 The structures of Ca12InC13-x (left) and Ba12InC18H4(right)...... 65

Figure 6.3 The coordination environment of the allenylide anions in Ca12InC13-x (left) and Ba12InC18H4(right)...... 67

Figure 6.4 Raman spectra for Ca12InC13-x and Ba12InC18H4...... 68

Figure 6.5 The calculated electronic density of states for Ca12InC13 (bottom) and Ba12InC18H4 (top)...... 69

Figure 6.6 Partial DOS (left column) for specified atomic orbitals, and COHP data (right column) for interactions between specified atoms in Ca12InC13...... 71

Figure 6.7 Partial DOS (left column) for specified atomic orbitals, and COHP data (right column) for interactions between specified atoms in Ba12InC18H4...... 72

Figure 7.1 The Al@Ca12@H20 cluster in Ca31H21Al2...... 75

Figure 7.2 The clathrate II structure formed from 3 hydride sites...... 76

Figure 7.3 a) The CaH4 unit in the Ca16 cage. b) The Ca16 cage in the H20 hexakaidecahedron...... 77

Figure 7.4 The six-fold split Ca site in Ca31H21Al2...... 77

viii

Figure 7.5 The AlN4 chains shown down the a) a-axis and b) c-axis...... 78

Figure 7.6 The crystal structure of Ca4Al2N5...... 79

Figure 7.7 The Ca24Al4C cluster (left) and the Ca32Al4 cluster (right) in Ca24Al9(C1- xHx)N2H16...... 80

Figure 7.8 A layer of Ca24Al9(C1-xHx)N2H16 looking down the c-axis showing the checkerboard pattern...... 82

Figure 7.9 The structure of Ca24Al9(C1-xHx)N2H16 looking down the [1,1,0] direction...... 82

ABSTRACT

Metal flux synthesis is a useful alternative method to high temperature solid state synthesis; it allows easy diffusion of reactants at lower temperatures, and presents favorable conditions for crystal growth. A mixed flux of calcium and lithium in a 1:1 ratio was explored in this work; this mixture melts at 300°C and is an excellent solvent for main group elements and

CaH2.Reactions of p-block elements in a 1:1 Ca/Li flux have produced several new intermetallic and Zintl phases. Electronegative elements from groups 14 and 15 are reduced to anions in this flux, yielding charge-balanced products. More electropositive metals from group 13 are not fully reduced; the resulting products are complex intermetallics. The reactions of tin or lead and carbon in Ca/Li flux produced the analogous phases

Ca11Tt3C8 (Tt = Sn, Pb) in the monoclinic C21/c space group (a = 13.2117(8) Å, b =10.7029(7) Å, c = 14.2493(9) Å, β = 105.650(1)° for the Sn analog). These compounds are carbide Zintl phases 4- 2- 4- 4- that includes the rare combination of C3 and C2 units as well as Sn or Pb anions. Ca/Li flux

reactions of CaH2 and arsenic have produced the Zintl phases LiCa3As2H in orthorhombic Pnma

(a = 11.4064(7),b = 4.2702(3),c = 11.8762(8) Å), and Ca13As6C0.46N1.155H6.045in tetragonal P4/mbm (a = 15.7493(15), c = 9.1062(9) Å). The complex stoichiometry of the latter phase was caused by incorporation of light element contaminants and was studied by neutron diffraction, showing mixing of anionic sites to achieve charge balance. Ca/Li flux reactions with group 13 metals have resulted in several new intermetallic

phases. Reactions of indium and CaH2 in the Ca/Li flux (with or without boron) formed

Ca53In13B4-xH23+x (2.4 < x < 4.0) in cubic space group Im-3 (a = 16.3608(6) Å) which features metallic indium atoms and ionic hydride sites. The electronic properties of this “subhydride” were confirmed by 1H and 115In NMR spectroscopy. Attempts to replace boron with carbon yielded 4- Ca12InC13-x, (Im-3, a = 9.6055(8)Å) which contains C3 units. A very similar phase, Ba12InC18H4 (Im-3,a = 11.1415(8) Å), was grown from the reaction of indium, carbon, and LiH in Ba/Li flux. 4- This compound also includes C3 units. Preliminary Ca/Li flux reactions of aluminum with other

main group elements have produced several new phases: a hydride clathrate Ca31Al2H25 in cubic

Fd-3m (a=18.0835(15) Å), Ca24Al2(C1-xHx)N2H16 in tetragonal P42/nmc (a=15.9069(12) Å,

c=13.7323(10) Å, and Ca4Al2N5 in orthorhombic Pna21 (a = 11.2331(1) Å, b=9.0768(8) Å, c=6.0093(5) Å.

CHAPTER ONE

INTRODUCTION

1.1 Metal Flux Synthesis

Traditionally, solid-state synthesis involves pressing powders into pellets then heating at high temperatures (over 1000° C) using rf or induction heating which is often followed by a lengthy annealing steps at higher temperature. Because the reaction is limited to the grain boundaries of the powders, high temperatures are necessary to promote sufficient diffusion. These high reaction temperatures typically favor simpler thermodynamically stable products which act as a thermodynamic sink, limiting the synthesis of new phases. Often the reactions need to be reground to expose new surfaces for the reaction to occur, and then further annealed. These conditions are not favorable to crystal growth. Metal flux synthesis is an alternative to traditional high temperature methods. The flux is an excess of a lower melting metal or metal solution which forms an in-situ solvent which dissolves the other reactants much like typical solution chemistry, except the solvent also acts as a reactant. This technique allows much greater diffusion than traditional methods, so the reaction times are generally shorter without need for annealing. Reactions can also be done at lower temperatures which expand the explorable phase-space, and may allow less thermodynamically stable phases to form. Since the reactants are dissolved, reactions involving refractory materials such as carbides and borides can also be done at lower temperatures. Flux synthesis typically 1 uses metallic elements such as Bi, Al, or Pb as the reaction medium. For example, RCo2As2

(R=La, Ce, Pr, Nd) grown from Bi, RE5Mn4Al23-xSix (RE=Ho, Er, Yb) and Er44Mn55(AlSi)237 2,3,4 from Al flux, and URu2Si2 from In flux. However, since flux synthesis requires at least one of the reactants to be in excess to form the solvent, there is not as much stoichiometric control over the reaction. There are several other important factors to consider when using the flux method. The flux should also not be reactive towards the crucible used, and the reactants should be soluble in the flux. After the reaction, the flux needs to be able to be separated from the crystals either through chemical etching or centrifugation.

1.2 Intermetallics Intermetallics are phases composed of two or more metallic or semi-metallic elements with stoichiometry that displays little or no phase width. Unlike alloys, they are typically formed by elements with significantly different sizes and electronegativities which results in atoms with preferred crystallographic sites, limiting site mixing between elements. There is an electronegativity difference which results in some charge transfer; however, the valence electrons are still largely delocalized resulting in metallic or semi-conducting phases, and most do not follow electron-counting rules. Intermetallics then are in a middle ground between localized, charge balanced salts and conducting delocalized metals which can lead to interesting magnetic and electronic properties. There are a very wide variety of intermetallic phases used in many applications. 5 6 7 Superconductors include Nb3Sn , NbTi, and MgB2. Other useful intermetallic materials include 8 9 10 11 strong magnets Nd2Fe14B, Sm2Co17, and SmCo5 and semiconductors ZnSe, and 12 13 HgCr2Se4. LaNi5H6is used in rechargeable batteries, and TiAl and Ni3Al are used as lightweight construction materials as part of alloys.14

1.3 Zintl Phases Zintl phases are a subset of intermetallics which are defined by the Zintl-Klemm concept which was first pioneered by Edouard Zintl in the 1930’s and further developed by Wilhelm Klemm in the 1950’s. They are at the extreme polar end of intermetallics, featuring complete or nearly complete charge transfer between atoms resulting in a charge-balanced structure with localized electrons similar to a salt. These phases typically have a band gap in the semiconducting range (0 – 3 eV) , but may be poor conductors. Since they are closed-shell systems, Zintl phases are almost always diamagnetic or weakly paramagnetic.15 Zintl phases are usually formed by electropositive s-block elements and electronegative p-block metals. Elements from group 13 are much more likely to form intermetallic rather than Zintl phases. This behavior leads to what is commonly called the Zintl border between groups 13 and 14.16 Over time, the traditional definition has broadened; Zintl phases have been discovered containing transition metals and rare earth metals which can lead to interesting magnetic and electronic properties. Furthermore, the lighter p-block elements B, C, and N can also form anions in Zintl phases. Zintl anions are rarely monoatomic, instead forming a wide variety of covalently

2 bonded anionic clusters and networks. The complexity of these anionic p-block clusters range 4- 4- from polyhedral clusters such Sn9 and Sn4 ,which follow Wade’s rules for electron counting, to 17 18 one-dimensional chains such as BaAu2P4 and K2AuSb , layered structures such as K4In4X6 19 20 21 22 (X=As,Sb) and KSi3As3, and 3-D networks such as in Pr4MnSb9 and Na5Sn13.

CHAPTER TWO

EXPERIMENTAL TECHNIQUES

2.1 Synthesis in Ca/Li Flux This work investigates reactions in the Ca/Li binary flux system. Calcium metal has not been heavily explored as a solvent for metal flux reactions due to its high melting point (842 °C) and its volatility above this temperature. However, the addition of lithium to calcium results in mixtures with a dramatically lowered melting point which can be used as fluxes. The Ca/Li phase diagram features one binary intermetallic phase (CaLi2); any Ca/Li mixtures with 50% Li or higher melt below 300 °C.23 A 1:1 Ca:Li ratio melts at about 300°C as shown in Figure 2.1.

The flux dissolves almost all elements and a wide variety of salts including CaH2, Li3N, and

Ca3N2. The products are invariably extremely air-sensitive and must be handled under inert atmosphere. This work focuses on reacting both a heavy p-block element (i.e. Sn, Pb) and a light p-element (i.e. H, C, B) in this flux which will likely form a Zintl or intermetallic phase. In general, Ca from this flux is much more likely to go into a compound than Li, coordinating around a central anion to form a X@Can cluster (where x is an anion surrounded by n Ca2+ cations). Every phase both known and unknown produced from this system contains Ca, while Li always less prevalent.Perhaps the most interesting aspect of this flux is that Zintl phases grown from it have monoatomic anions instead of the much more common polyanions. A polyanion reduces the charge on a particular atom, so the monoatomic anion has a greater charge. Formation of these highly charged monoatomic anions indicates the highly reducing nature of the Ca/Li flux. The reactants are added to a steel crucible (7 cm length, 0.7 cm diameter) in glove box under argon. The reactantsare added first, followed by Ca and Li metal. It is intended that the Li will melt into the Ca, form the flux, then dissolve the other reactants. The steel crucible is sealed with an arc-welder under argon, then sealed into a quartz ampule under vacuum. The ampules are heated to 1050°C then cooled stepwise to 500°C. The flux is removed from the product crystals by centrifugation. It is found that the crystals adhere very well to the sides of the crucible.

Figure 2.1 Phase diagram of Ca and Li.24

There are significant hydrogen impurities in commercial alkaline earth metals that which

can become incorporated in products. The Corbett group has shown that the phases A5Sb3 and

A5Bi3 (A=Ca, Sr, Ba, Sm, Eu, Yb) and Ba5Ga6 are actually the hydrides A5Sb3H, A5Bi3H, and 25,26 Ba5Ga6H2 due tothis issue. In this work, the Ca metal is dehydrogenated in a specialized apparatus by heating at 700°C under a vacuum of 10-4 to 10-5 Torr which will decompose the

CaH2 impurity. CaO however is has fairly stable and has a very low vapor pressure at these temperatures, so this method is not as effective at eliminating oxide impurities. After 4 hours of heating, the vacuum pressure drops to less than 10-5 Torr indicating a significant loss of hydrogen. The Ca metal after heating is softer and has a light golden color.

2.2 SEM-EDS Elemental analysis is performed using SEM-EDS (Scanning electron microscopy – energy dispersive spectroscopy). The instrument (JEOL 5900 scanning electron microscope) focuses a beam of electrons onto the sample which produces various signals by interactions with the surface including secondary electrons, backscattering electrons, Auger electrons, characteristic X-rays, and fluorescent X-rays. Secondary and backscattering electrons are used to produce an image of the sample while the elemental composition is determined by X-rays.

Secondary electrons are ejected from the core shells of the sample atoms from inelastic scattering with the beam electrons. By rastering the beam along over an area, a false image can be produced of the sample using the intensity of the electrons produced. More electrons are detected from parts of the sample that are tilted towards the detector, so these parts appear brighter. This creates a sense of topography in the image. Backscattered electrons are beam electrons which are reflected back from the sample by elastic scattering with an electron cloud. These electrons can create an image similar to secondary electrons; however, heavier atoms will backscatter more electrons, so will be imaged brighter. Since the crystals synthesized in this work are air sensitive, some oxidation of the surface is common. Oxidized areas are oxygen-rich and appear darker on the image. Figure 2.2 compares the image of a sample with secondary and backscattered electrons.

Figure 2.2 A crystal imaged from secondary electrons on the left and backscattered electrons on the right. Residue on the crystal can be clearly seen from the backscattered electrons.

The elemental composition of the sample is determined by emitted X-rays. When a core electron is knocked out by an electron from the beam, an electron from a higher shell will lose energy to fill the hole by emitting an X-ray. The most common emission is the Kα line, an electron moving from the n=2 level or L-shell to the n=1 level or K-shell. Similarly, the second- most common emission is the Kβ from the n=3 or M-shell to the K-shell. Since the energies of the electron levels vary per element, the emitted X-rays will have an energy which is characteristic to that element.

Samples for EDS analysis were affixed to an aluminum SEM stub using carbon tape and analyzed using a 30 kV accelerating voltage. The atomic ratios of only elements heavier than O are determined as the lighter elements are not reliably detected.

2.3 X-ray Photoelectron Spectroscopy X-ray photoelectron spectroscopy was used to determine the identity of light elements in certain products. This technique irradiates the sample with X-ray radiation which causes the emission of core electrons due to the photoelectric effect. The kinetic energy of the emitted electrons is measured, which is related to the fermi level by KE= hν-(EB+φ). Since elements have unique core electron binding energies, they produce characteristic spectral peaks. The intensity of the peaks is directly related to the concentration of the element in the sample. Since this is a surface technique, the sample is sputtered with an ion gun to remove surface contaminants to ensure that the measurement matches the bulk sample. Samples were mounted on carbon tape which has unfortunate side effect of physisorbing water and hydrocarbons, preventing determination of C,H, and O in the sample. A Physical Electronics PHI 5100 series XPS was used with a non-monochromated dual anode (Al and Mg) source having a single channel hemispherical energy analyzer. A Mg Kα source was used.

2.4 X-ray Diffraction The structure of a compound is determined using single crystal X-ray diffraction, and is the main reason why obtaining crystals is very valuable in synthetic chemistry. Crystals are a periodic arrangement of atoms which have a spacing which is similar to an X-ray wavelength. So, X-rays that are passed through the crystal will diffract according to Bragg’s Law described in Equation 2.1: nλ = 2dsinθ (2.1) where λ is the wavelength, d is the spacing between layers of atoms and theta is the angle of incidence. The X-rays will constructively interfere in directions law in which this law is satisfied, so the spacing between atoms can be determined. The strength of the reflection is proportional to Z, allowing the elements to be assigned. However, elements close together in size are difficult to differentiate, and very light elements such as hydrogen can be difficult to detect.

For single crystal samples, an X-ray beam is diffracted in accordance with Bragg’s Law is satisfied which produces spots in a 3-D diffraction pattern. However, powder samples consist of a large number of randomly oriented crystallites causing the diffraction spots to be averaged into a ring. The resulting diffraction pattern can be compared to known phases, but and can determine the space group of unknown phases. It most cases, there is not enough information to fully solve the structure. Sample crystals are brought out of a glovebox under Paratone oil and are mounted in a cryoloop under a microscope. Single-crystal X-ray diffraction data are collected at 200 K in a

stream of nitrogen using a Bruker APEX 2 CCD diffractometer with a Mo Kα radiation source. The data is integrated with the Bruker SAINT software and corrected for absorption effects using the multiscan method (SADABS).27 Refinements of the structures are performed using the SHELXTL package. Neutrons will diffract off a crystal according to Bragg’s Law similarly to X-rays, except neutrons interact with the nucleus rather than the electron cloud. As a result, the scattering strength is independent of Z which allows light elements or closely sized elements to be distinguished. While not used in this work, neutrons have a magnetic moment which will interact with magnetic moments in the sample, providing information about the magnetic structure. Neutron diffraction requires a nuclear reactor or spallation source. For this work, experiments were performed at Oak Ridge National Laboratory.

2.5 Solid-State Nuclear Magnetic Resonance In Nuclear Magnetic Resonance (NMR), the sample is placed within a magnetic field causing the nuclear spins to process around it. A smaller magnetic field is then pulsed perpendicular to the applied magnetic field, tipping the spins to process around this field. The precession speed depends on the strength of the field, characteristics of the nucleus, and its surrounding environment. The spinning magnetic moments that are produced then induce a current in an rf receiver coil. The individual frequencies are then plotted using a Fourier transform. The nuclei are shielded by electrons which create a localized magnetic field which causes the observed magnetization to be different for nuclei in different electronic environments in a phenomenon known as chemical shift. This can be used to determine the chemical

environment around a nucleus. In solid metallic compounds, a similar phenomenon known as the Knight shift occurs from a local magnetic field produced by conduction electrons that shield a particular nucleus. The shielding effect is greatest when the electron is closest to the nucleus, so the Knight shift is almost entirely due to s-electrons. The magnitude of the shift is roughly proportional to the number of conducting s-electrons states at the Fermi level which gives a measure of the metallic nature of the sample.28 Nuclear spins in solid samples have certain anisotropic interactions because the atoms are not free to tumble in space like in a liquid. These interactions are dependent on the term 3cos2 θ - 1 where θ is the angle between the magnetic field and the principal axis about which the chemical shift is defined. In magic angle spinning (MAS), the sample is spun at the “magic angle” of θ = 54.74 at which the above term equals 0, eliminating the chemical shift anisotropy.

2.6 Band Structure Calculation Another way to determine the electronic properties of a material is to look at the band structure. In extended solids, there are effectively infinite discrete energy levels which must fit a finite energy space. The energy levels become so close together that can be treated as continuous bands with varying density of electron states. Electronegative elements in Zintl phases and intermetallics contribute the majority of their electronic states to filled bands while electropositive elements contribute to the majority of empty bands. A greater electronegatvity difference leads to more electrons localized on electronegative elements which will form a band gap leading to semiconducting behavior. A band gap can be confirmed experimentally by optical absorbance or reflectance measurements which can observe an absorption edge in the 0.5 – 4.0 eV range. Band structures were calculated using the Stuttgart TB-LMTO-ASA (tight-bonding linear muffin-tin orbital atomic sphere approximation) software package29 Interstitial space was covered by empty spheres. Integration over the Brillouin zone was performed using the tetrahedron method30 Details on the calculations are given in their respective chapters. The LMTO method approximates the energy potential for each orbital as a spherical space around an atom, known as a muffin-tin sphere.31 The interstitial space between muffin-tin spheres is approximated as a constant energy potential, but in the atomic sphere approximation, the spheres around atoms are blown up to cover the full interstitial volume or empty spheres are added.32

CHAPTER THREE

COMPLEX CARBIDE PHASES Ca11E3C8 (E = Sn, Pb) GROWN FROM THE Ca/Li FLUX

3.1 Introduction Metal carbides, comprised of metal or metalloid elements in combination with carbon, possess very useful properties, including extreme hardness (WC, B4C, Fe3C), superconductivity 33,34 (LaNi2B2C, Y2C3), and chemical inertness and hig h melting point (SiC). These compounds can be grouped into several classes, depending on the nature of the metallic element in the compound. Strongly electropositive metals form salt-like carbides, ionic compounds containing 2– 4– discrete carbide anions such as C2 in calcium acetylide and C3 in Mg2C3. Transition metal or rare earth metal carbides (TiC, LaC2) are usually metallic and do not follow charge-balancing rules. Carbides of semimetals such as silicon or boron are essentially covalent; the resulting network of strong localized bonds yields refractory compounds such as SiC.30 Due to the refractory nature of elemental carbon (often used as a crucible material for this reason), reactions to form metal carbides usually involve high temperature methods such as arc melting. However, metal flux reactions have proven to be an excellent alternative for the synthesis of new carbide phases. Reactions of carbon with other metals in eutectic melts of La/Ni

(88 wt % La, mp 532 °C) have produced complex metallic carbides such as La21Fe8Sn7C12 and 35,36 La11(MnC6)3. If a flux comprised of more electropositive metals is used, salt-like carbides are formed. Ca/Li melts have been used as intercalation media for graphite, forming superconducting 37 CaC6 when graphite is soaked in Ca/Li at low temperatures (below 400 °C). Our explorations of higher temperature reactions in Ca/Li mixtures have indicated that these fluxes are excellent solvents for carbon, breaking carbon sources into small reduced anions. Reactions of carbon and

CaH2 in Ca/Li flux recently yielded a complex carbide hydride phase, LiCa2C3H, containing the 4– 38 relatively rare C3 anion. We have continued our explorations of carbon reactions in Ca/Li by adding heavy tetrelide reactants (tin or lead) to the flux; these reactions have produced new Zintl phase carbides Ca11Tt3C8 (Tt = Sn, Pb). These compounds contain three different anionic species:

4– 2– 4– monatomic Tt ions and carbide anions C2 and C3 . The presence of two different carbide anions is highly unusual among the known metal carbides, with most examples being 4– 4– combinations of monatomic C and another carbide anion (for instance, Ca4Ni3C5 contains C 2– 4– 4– 39,40 and C2 anions; Lu4C7 and Sc5Re2C7 contain C and C3 anions). Besides Sc3C4 and its 4– 2– 4– 41 analogs (which contain C , C2 , and C3 ), the title compounds Ca11Tt3C8 are the only known 2– 4– phases to feature both acetylide (C2 ) and allenylide (C3 ) anions.

3.2 Experimental 3.2.1 Synthesis Calcium slugs (99.5%, Alfa Aesar) were purified by heating in a steel tube under a 10– 5Torr vacuum at 600 °C for 3 h. Heating was continued under 10–3 Torr for 12 h. This process decomposes any calcium hydride and calcium oxide present and removes the resulting gaseous hydrogen and water. Chunks of Li (99.8% Strem), acetylene carbon black powder (99.5% Alfa Aesar), and either Sn (99.9% Alfa Aesar) or Pb (99.7% Fisher) granules were used as received. Reactants and flux metals were added to stainless steel crucibles (7.0 cm length/0.7 cm diameter) in a 7:7:0.7:1.9 mmol Ca/Li/Tt/C ratio in an argon-filled glovebox. The crucibles were sealed by arc-welding under argon and were placed in silica tubes which were flame-sealed under a vacuum. The ampoules were heated from room temperature to 1050 °C in 3 h and held there for 4 h. The reactions were cooled stepwise to 800 °C over 48 h, 600 °C over 144 h, and 500 °C over 72 h. The reactions were held at 500 °C and then were removed from the furnace, inverted, and centrifuged for 2 min to separate the crystalline products from the Ca/Li melt. The solid product adheres to the side of the crucible. The steel crucibles were cut open in an argon-filled glovebox. These compounds can also be synthesized in niobium crucibles to avoid the possibility of incorporation of traces of paramagnetic impurities from steel crucibles.

3.2.2 Elemental Analysis Elemental analyses were performed using a JEOL 5900 scanning electron microscope with energy dispersive spectroscopy (SEM-EDS) capabilities. Samples were affixed to an aluminum SEM stub using carbon tape and analyzed using a 30 kV accelerating voltage. The atomic ratios of the tetrelide and Ca were determined with this method, but C was too light to be

detected. The Ca/tetrelide ratio determined from EDS was within 5% of 70%/30%. No incorporation of elements from the steel crucible was observed in any of the samples.

Table 3.1. Crystallographic data and collection parameters for Ca11Tt3C8 phases.

Ca11Sn3C8 Ca11Pb3C8 Formula weight 893.07 1158.54 Crystal System Monoclinic Monoclinic

Space group P21/c (No.14) P21/c (No. 14) a (Å) 13.1877(9) 13.2117(8) b (Å) 10.6915(7) 10.7029(7) c (Å) 14.2148(9) 14.2493(9) β(°) 105.649(1) 105.650(1) Z 4 4 Volume 1929(2) 1940.2 Index ranges -16 < h < 16, -13 < k < 13, -17 < h < 17, -14 < k < 14, -18 < l < -18 < l < 18 17 Reflections collected 21736 22034 Unique data/parameters 4463/200 4536/199 μ (mm-1) 8.067 46.76 R1/wR2a 0.0190 / 0.0429 0.0289 / 0.069 R1/wR2 (all data) 0.0202 / 0.0429 0.0343 / 0.069 Residual peak/hole (e- A-3) 0.78 and -0.64 1.49 and -3.14

a 2 2 2 2 2 1/2 R1 = Σ||Fo|-|Fc||/Σ|Fo|; wR2 = [Σ[w(Fo - Fc ) ]/Σ[w(Fo ) ]]

3.2.3 Structural Characterization Sample crystals were brought out of the glovebox under Paratone oil and were mounted in a cryoloop. Single-crystal X-ray diffraction data were collected at 170 K in a stream of nitrogen using a Bruker APEX 2 CCD diffractometer with a Mo Kα radiation source. The data were integrated with the Bruker SAINT software and corrected for absorption effects using the multiscan method (SADABS).23 Refinement of the structure was performed using the SHELXTL

42 package. The structure was solved in monoclinic space group P21/c (No. 14); data collection and refinement parameters are shown in Table 3.1. Powder X-ray diffraction studies were carried out on reaction products to identify byproducts using a Rigaku Ultima III X-ray powder diffractometer. In a glovebox, samples of solid products from each reaction were ground with a small amount of Si as an internal standard and placed in an airtight sample holder to prevent oxidation. The MDI JADE software suite was used for analysis of the powder patterns.

Table 3.2 Atomic coordinates and isotropic thermal parameters of Ca11Sn3C8.

a x y z Ueq Sn(1) 0.134488(13) 0.518226(16) 0.416366(12) 0.00759(5) Sn(2) 0.144280(13) 0.019099(16) 0.400100(12) 0.00727(5) Sn(3) 0.502100(12) 0.248789(16) 0.000673(11) 0.00685(5) Ca(1) 0.01834(4) 0.75387(5) 0.27375(4) 0.00983(10) Ca(2) 0.02543(4) 0.27005(5) 0.48563(4) 0.00957(10) Ca(3) 0.10591(4) 0.50387(4) 0.14584(4) 0.00957(10) Ca(4) 0.14247(4) 0.00360(4) 0.16427(4) 0.00806(10) Ca(5) 0.26117(4) 0.70101(5) 0.03885(4) 0.1150(11) Ca(6) 0.26604(4) 0.30335(5) 0.34284(4) 0.01233(11) Ca(7) 0.30576(4) 0.34280(5) 0.07150(4) 0.00956(10) Ca(8) 0.39042(4) 0.00662(4) 0.09261(4) 0.00869(10) Ca(9) 0.50430(4) 0.2457(5) 0.26940(4) 0.00995(10) Ca(10) 0.60945(4) 0.49289(4) 0.14662(4) 0.00823(10) Ca(11) 0.70252(4) 0.16284(5) 0.17078(4) 0.01023(11) C(1) 0.67900(18) 0.3456(2) 0.29661(17) 0.0077(5) C(2) 0.22681(19) 0.7953(2) 0.18327(17) 0.0074(5) C(3) 0.13352(19) 0.7391(2) 0.15870(17) 0.0078(5) C(4) 0.14035(19) 0.2699(2) 0.15817(17) 0.0078(5) C(5) 0.2322(2) 0.2105(2) 0.18021(17) 0.0076(5) C(6) 0.32602(18) 0.1578(2) 0.20861(17) 0.0077(5) C(7) 0.4285(2) 0.4921(2) 0.2349(2) 0.0155(6) C(8) 0.3405(3) 0.5159(4) 0.2029(3) 0.0390(9)

a Ueq is defined as one-third of the trace of the orthogonalizedUij tensor.

Table 3.3 Atomic coordinates and isotropic thermal parameters of Ca11Pb3C8.

a x y z Ueq

Sn(1) 0.365683(19) 0.51768(2) 0.082976(18) 0.00687(7) Sn(2) 0.643889(19) 0.51844(2) 0.399597(18) 0.00652(7) Sn(3) 1.001940(17) 0.74875(2) 0.000529(16) 0.00565(7) Ca(1) 1.00401(10) 0.74768(13) 0.26987(10) 0.0093(3) Ca(2) 0.64319(10) 0.50338(11) 0.16450(10) 0.0075(3) Ca(3) 0.47523(10) 0.26923(12) 0.01429(9) 0.0078(3) Ca(4) 0.48139(10) 0.75315(12) 0.22616(10) 0.0083(3) Ca(5) 0.23400(11) 0.30360(13) 0.15774(10) 0.0116(3) Ca(6) 1.23850(10) 0.70057(13) 0.46090(10) 0.0108(3) Ca(7) 0.80538(10) 0.84273(12) 0.07154(10) 0.0094(3) Ca(8) 1.39359(11) 0.50365(12) 0.35371(10) 0.0083(3) Ca(9) 1.10925(11) 0.99245(12) 0.14738(9) 0.0071(3) Ca(10) 1.20209(10) 0.66221(13) 0.17116(10) 0.0099(3) Ca(11) 0.89023(11) 0.50710(12) 0.09337(10) 0.0078(3) C(1) 0.3660(5) 0.7397(6) 0.3410(4) 0.0065(13) C(2) 0.3601(5) 0.2691(6) 0.3418(4) 0.0040(12) C(3) 0.8269(5) 0.6576(6) 0.2092(4) 0.0068(12) C(4) 1.1786(5) 0.8446(6) 0.2971(4) 0.0071(13) C(5) 0.7271(6) 0.2972(6) 0.1834(5) 0.0094(13) C(6) 0.7338(6) 0.7080(6) 0.1806(5) 0.0112(14) C(7) 0.9328(6) 0.9929(6) 0.2358(6) 0.0161(16) C(8) 0.8399(8) 1.0124(9) 0.2028(7) 0.039(3) a Ueq is defined as one-third of the trace of the orthogonalizedUij tensor.

3.2.3 Raman Spectroscopy

A crystal of Ca11Sn3C8 was sandwiched between quartz slides which were sealed together with TorrSeal epoxy under argon. The Raman measurements were carried out using a

commercial JY Horiba LabRam HR800 system excited by a HeNe laser emitting at 633 nm. The power at the sample was 28 mW. The spectrograph uses a holographic notch filter to couple the laser beam into the microscope (Olympus BX30) by total reflection. The beam is focused on the sample through a 50× IR (Leica N.A. 0.80) microscope objective. Scattered radiation is collected by the objective, and the laser radiation is filtered out by the notch filter with Raman scattering coupled into the spectrograph CCD through a confocal hole. Spectra were collected under ambient conditions over the spectral range of 50 to 3500 cm–1. Attempts to collect a spectrum for the Pb analog were not successful, possibly due to its more metallic nature.

3.2.4 Electronic Structure Calculations Density of states calculations were carried out using the Stuttgart TB-LMTO-ASA 25,26 software package. The structural models for the Ca11Tt3C8 phases were based on the unit cell dimensions and atomic coordinates derived from single crystal diffraction data. Empty spheres were added by the program where appropriate to fill the unit cell volume. An 8 × 8 × 8 κ-point mesh was used and integrated using the tetrahedron method. The basis sets consisted of 5s/5p/5d/4f for Sn, 4s/4p/3d for Ca, and 2s/2p/3d for C. The Sn 5d/4f, Ca 4p/3d, and C 3d orbitals were downfolded.

3.2.5 Reactivity Studies

Ca11Tt3C8 samples were reacted with protolytic compounds (water or NH4Cl) to explore their interaction with the carbide anions in the structure. The reaction with water is instant and vigorous at room temperature, forming gaseous products upon the addition of 5 μL of water to 20 mg of sample in a 100 mL Schlenk flask sealed with a rubber septum under argon. Aliquots of the product gases were taken by syringe and analyzed by injecting them into a HP 6890 series GC system coupled to a HP 5973 mass selective detector.

3.3 Results and Discussion 3.3.1 Synthesis and Reactivity

Ca11Tt3C8 phases form as non-faceted elongated black chunks of about 1–2 mm in size, although some rod-like crystals are also observed. These phases were initially found as minor

products from reactions of Ca/Li/Tt/C at a 10:10:1:1 mmol ratio, along with unreacted Sn or Pb and CaSn3 or CaPb3. The optimized synthesis ratio of carbon to heavy tetrelide is stoichiometric (7:7:0.7:1.9 mmol Ca/Li/Tt/C ratio), and based on these elements the yield is about 70%. The use of purified calcium metal is necessary to avoid the incorporation of adventitious hydride. If commercial (hydride-contaminated) calcium metal is used in the reaction, LiCa2C3H is observed as a byproduct.

The Ca11Tt3C8 phases are extremely reactive. They are air-sensitive and must be handled under an inert atmosphere. Exposure to water results in a rapid reaction to form acetylene and 3– allene. Similar behavior is reported for other carbide phases containing C4 units; for instance,

Ln4C7 and LiCa2C3H both produce C3H4 upon reacting with water, and hydrolysis of Sc3C4 yields a mixture of C1, C2, and C3 hydrocarbons in accordance with the mixture of carbide anions 35,44a,48 in its structure. The mass spectra of hydrolysis products of both Ca11Tt3C8 analogues show evidence of acetylene and propadiene (shown in Figure 3.1), referenced to the corresponding spectra in the NIST database.43 Neither stannane nor plumbane was observed. Additional peaks at 42 and 43 m/z may indicate a small amount of propene formation.

Figure 3.1 Mass spectrum of products of hydrolysis of Ca11Pb3C8. Allene (C3H4) and acetylene (C2H2) peaks are present at 40 and 28 m/z respectively. Additional peaks are due to O2 and trace H2O.

3.3.2 Structure Description

Ca11Tt3C8 (Tt = Sn, Pb) are, to the best of our knowledge, the only known ternary phases comprised of these elements. It is notable that no binary tin or lead carbides are known. Ternary

carbides containing these elements are also rare, with M2SnC (M = Ti, Zr, Hf, V) and the

perovskite carbides R3SnC (R = rare earth) the only reported examples; in these structures, the 44,45 tin and carbon are separated by the electropositive metal. Similarly, the new Ca11Tt3C8 phases reported here do not feature any Tt–C interactions; the heavy tetrelide and carbide anions are separated from each other and coordinated only to calcium cations. This is also what is seen

in Ba3Ge4C2, the only other reported ternary carbide containing an alkaline earth metal and a 2– 4– heavier tetrelide element; its structure features C2 anions and Ge4 tetrahedral units, each surrounded by barium cations.46

The Ca11Tt3C8 structure is shown in Figure 3.2; it forms in monoclinic space group P21/c and features Tt4– anions, Ca2+ cations, and two different types of carbide anions. The carbide 2– 4– anions are present as triple-bonded units of C2 and double-bonded units of C3 , which are deprotonated forms of acetylene and allene (propadiene), respectively. The resulting 2+ 4– 2– 4– stoichiometry is charge-balanced ((Ca )11(Tt )3(C2 )(C3 )2), with the metalloids all achieving an octet electron configuration, so these compounds can be viewed as carbide Zintl phases. While there are many known inorganic phases which incorporate organic carbon anion units, 2– 4– such as the metal acetylide salts with C2 units (CaC2, Na2C2, LaC2, etc.) and Mg2C3 with C3 29,47 39 units, the only other reported phase that contains both C2 and C3 units is Sc3C4. As shown in

Figure 3.2b, the carbide anions in Ca11Tt3C8 are grouped together in the same plane of near- monatomic thickness and are each surrounded by a cage of Ca cations. 4– The bond lengths within both of the crystallographically unique C3 allenylide units in

the Ca11Tt3C8 structures are in the range 1.310–1.349 Å (see Figure 3.3 and Table 3.2). These values are similar to bond lengths reported for allene (1.31 Å) and other allenylide phases,29,37,38,46 indicating that the interactions between the carbons in these anions are double bonds. It is notable that the bond lengths from the central carbon to the two terminal carbon atoms are slightly different. Most other reported ionic allenylides have identical bond lengths, 4– with the central carbon atom often lying on an inversion center or mirror plane. All C3 anions

in the Ca11Tt3C8 phases are slightly bent with angles ranging from 174.4 to 176.1°. The

17 allenylide ions are in a cage of nine Ca2+ cations; terminal carbon atoms are bonded to three Ca2+ ions in a pyramidal geometry, and the central carbon atom is bonded to three Ca2+ ions in a trigonal planar geometry. These cations are in a staggered arrangement with the ones on the end.

The Ca–C bonds range from 2.437(2) to 2.657(2)Å for Ca11Sn3C8 and 2.446(6) to 2.663(6)Å for 2+ 4– Ca11Pb3C8. A similar arrangement of Ca cations around a C3 unit is found in Ca3C3Cl2, which has three Ca2+ ions coordinating the end carbon atoms but only two Ca2+ ions coordinated to the central carbon atom.48,49

a) b)

Figure 3.2 The structure of Ca11Tt3C8. Carbon is represented by black spheres, Tt = Sn or Pb by large red spheres in coordination polyhedra, and calcium cations by blue spheres. (a) Overall monoclinic structure, viewed down the b axis. (b) Layer of carbide anions, viewed down thec- axis.

2– The bond distance for the C2 unit in Ca11Sn3C8 (1.155(4) Å) is smaller than the usual bond length of 1.19 Å seen in ionic acetylides.29 This can be explained by disorder in the atom positions of this unit, indicated by the large displacement parameters of the carbon atoms (see Figure 3.3c). The bond distance is calculated from the centroid of the displacement ellipsoids around the carbon atoms, but depending on the actual location of the carbon atom the effective 2– bond length will be longer. This effect has also been reported for C2 units in Ca5Cl3(C2)(CBC) and Ca15(CBN)6(C2)2O, which have disorder-averaged bond lengths of 1.08 and 1.095 Å,

respectively.50,51 The acetylide anion in the Pb analogue does not exhibit as much disorder. Accordingly, its bond distance (1.21(1) Å) is within the range expected for a triple bond. The heavy tetrelide elements form isolated Tt4– anions bonded solely to calcium cations in 10- coordinate arrangements. The high negative charge on these anions is evidence of the strong reducing power and large excess of the Ca/Li flux reaction medium. This is in contrast to the 4– 4– larger anionic clusters often found in tetrelide Zintl phases, such as Pb4 (seen in KPb) and Sn9 52,53 (seen in Rb12Sn17), which have a smaller negative charge per atom. The calcium coordination of all three crystallographically unique tetrelide sites can be described roughly as bi-capped square antiprismatic; an example is shown in Figure 3.4. The Sn–Ca bond distances range from 3.1388(5) to 3.8122(6) Å, and the Pb–Ca bond distances range from 3.153(1) to 3.828(1) Å. Bond distances from literature range between 3.1 and 3.72 Å for Ca–Sn bonds and 3.1 and 3.76 for Ca–Pb bonds but are usually below 3.5 Å.54,55

Figure 3.3 Carbide anions in Ca11Sn3C8, with carbon−carbon bond lengths indicated. Calcium 4− 2− cations depicted as blue spheres. (a) Allenylide (C3 ) anions. (b) Acetylide (C2 ) anion. (c) Acetylide anions for the tin and lead analogs, displaying displacement ellipsoids.

Table 3.4 Bond lengths of interest in Ca11Tt3C8 phases, in angstroms.

Bond Ca11Sn3C8 Ca11Pb3C8 Tt1 - Ca 3.1594(5) – 3.7644(6) 3.1795(13) – 3.7780(14) Tt2 - Ca 3.1388(5) – 3.7612(6) 3.1532(13) - 3.7636(14) Tt3 - Ca 3.1983(5) – 3.8122(6) 3.1933(13) - 3.8307(15) Ca - C 2.437(2)−2.896(2) 2.446(6) −2.899(6) C1 – C2 1.313(3) 1.303(9) C2 – C3 1.329(3) 1.363(3) C4 – C5 1.328(3) 1.336(9) C5 – C6 1.320(3) 1.304(9) C7 – C8 1.155(5) 1.207(12)

Figure 3.4 Bicapped square antiprism coordination of an Tt4− anion (red sphere) by calcium (blue spheres) in Ca11Tt3C8 phases.

3.3.3 Raman Spectrum

The Raman spectrum of Ca11Sn3C8 (Figure 3.5) confirms the presence of double and triple bonds in the carbide anions. The acetylide anion stretch (1879 cm–1) is within the 1800– 1900 cm–1 range observed for binary alkali metal acetylides, which is significantly downshifted from the range typical for neutral organic alkynes (2100–2300 cm–1) due to the higher negative

29 4– charge for anionic species. Similarly, the C3 allenylide stretching modes in Ca11Sn3C8 (νsym –1 –1 968 cm , νasym 1482, 1542 cm ) occur at lower energy than the typical alkene range of 1600– –1 4– 1700 cm . The only previous report on vibrational spectra of a phase containing the C3 anion –1 –1 38,56 is for Ca3C3Cl2, which exhibits νsym 1159 cm and νasym 1660 cm . While the observed allenylide modes for Ca11Sn3C8 are significantly different from those stated for Ca3C3Cl2, the

Raman spectrum in Figure 3.5 is very similar to those seen for Ca15(CBN)6(C2)2X2 (X = F or H). 4– 4– These complex salts feature linear, double bonded CBN anions (isoelectronic with C3 ), 57 2– acetylide anions, and either fluoride or hydride anions. The hydride analog exhibits a C2 –1 4– stretching mode at 1879 cm , identical to that for Ca11Sn3C8. The CBN anion νasym modes (at –1 –1 –1 1482 and 1543 cm ), νsym stretch (970 cm ), and bending mode (605 cm ) all correspond very 56 well to the Raman modes for the allenylide anion in Ca11Sn3C8. The Raman spectrum in Figure 3.5 shows an additional peak at 80 cm–1. This may be due to a rattling mode of the stannide anions. No vibrational spectra are available for Ca2Sn, but –1 Mg2Sn exhibits a Raman mode at 222 cm , and the corresponding mode for the heavier calcium analog would be expected to be shifted to lower energy.58 This mode might be analogous to that – of heavy anions in inverse clathrate cages (such as I anions in the cages of I8Sb8Ge38), which exhibit rattling vibrations below 100 cm–1.59

Figure 3.5 Raman spectrum of Ca11Sn3C8 showing the modes of the molecular carbon anions.

3.3.4 Electronic Structure

Density of states data for the Ca11Tt3C8 phases is shown in Figure 3.6. DOS diagrams of both analogs feature small band gaps or pseudo band gaps at the Fermi level (Eg = 0.1 eV for the tin phase and a pseudogap for the lead analog), in agreement with their charge-balanced stoichiometries and black color. Bands associated with calcium and the heavy tetrelide anions are

prevalent just below Ef. States derived from carbide anion orbitals are found at lower energies

(between −2 and −4 eV below Ef for the tin analog). This is in agreement with the higher electronegativity of carbon compared to the heavier tetrelide elements; the carbide associated bands are also considerably narrower (electrons more localized) than the Tt4– bands. The data

compare well to calculations carried out on binary phases Ca2Sn and CaC2. Ca2Sn features tin anions surrounded by nine calcium cations in a tricapped trigonal prism coordination. It is a very

narrow band gap semiconductor (Eg = 0.1 eV), with states derived from Sn p orbitals 60 predominant just below Ef and empty Ca bands above the gap, as is also seen for Ca11Sn3C8. 2– CaC2 has a tetragonal distorted NaCl structure type, with C2 anions aligned along the c axis in octahedral coordination by Ca2+ ions. The energy difference between states derived from filled acetylide anion orbitals and empty calcium bands is considerable; this colorless compound has a bandgap over 3 eV.61 This supports the observed relative energies of carbon- and tin-derived valence bands with respect to the calcium-based states in the conduction band of Ca11Sn3C8.

Figure 3.6 Density of states data for (a) Ca11Sn3C8and (b) Ca11Pb3C8.

3.4 Conclusions Calcium/lithium melts have proven to be excellent reaction media for the synthesis of new complex carbide phases. Reactions at high temperatures have produced several new 4– compounds featuring the relatively rare C3 anion, including the Ca11Tt3C8 title compounds and 35 LiCa2C3H. Variation of the reaction temperature and carbon source may allow for the formation of phases with larger carbide anions. Ca11Tt3C8 phases are highly reactive toward electrophiles. The presence of two different carbide anions and their packing in the unit cell may make these compounds useful precursors for the synthesis of complex organic molecules.

CHAPTER FOUR

LiCa3As2H AND Ca14As6X7 (X = C, H, N): TWO NEW ARSENIDE HYDRIDE SALTS GROWN FROM CA/LI METAL FLUX

4.1 Introduction Metal fluxes have proven to be excellent solvents for the synthesis and crystal growth of new intermetallic compounds.1 The flux method eliminates the diffusion barrier present in traditional solid state methods, and enables crystals to be easily formed without lengthy annealing steps. We are investigating the use of mixed metal fluxes which have a lowered melting point and will dissolve a larger variety of elements. Reactions in mixed fluxes have yielded new intermetallic phases including R3-δFeAl4-xMgxSi2 (R = Yb, Dy) and 62 R5Mg5Fe4Al12Si8 (R = Gd, Dy) grown from the Mg/Al eutectic, La21Fe8M7C12 (M = Ge/Al, Sn, 35 Sb, Te, Bi) and LaNi2Ru2Alfrom the La/Ni eutectic, and Co7+xZn3−xSn8 from the Zn/Sn eutectic.63 The flux method is not a stoichiometric synthesis technique. The presence of excess flux metal can introduce a number of complications, including the possibility of contamination. This is often observed as unwanted incorporation of the flux element into products, whether as inclusions of flux droplets within bulk products or as substitutions of the flux element for others in the structure. An example of both of these phenomena is observed in tin flux growth of KxBa1- 64,65 xFe2As2. The crystals contain tin occlusions (evidenced by observation of the superconducting transition of tin) as well as some tin substitution on barium sites. Another contamination problem occurs when the flux metal(s) themselves are impure; since the flux is used in such a large excess, these contaminants can be present in significant amounts. This is particularly true for reactive metals such as alkaline earth metals, which are known to be susceptible to hydride contamination, and often have significant oxide coating. Inadvertent hydride contamination from alkaline earth metal reactants has been responsible for the discovery 38,66,26 of phases such as LiCa2C3H, Ba21Ge2O5H24, and Ba5Sb3H. We are exploring the use of calcium / lithium melts as synthesis media. Ca/Li mixtures with 50 mol % lithium or greater have melting points below 300°C and will dissolve most main

group elements and many binary hydrides including CaH2. Reactions such as Ca/Li/CaH2/M (M

= main group metalloid) lead to crystal growth of complex hydride phases such as LiCa2C3H and 38,24 2+ LiCa7Ge3H3. These compounds feature main group and hydride anions surrounded by Ca and/or Li+cations, and are usually charge-balanced; they fall under the classification of Zintl phase hydrides. Zintl phases are typically comprised of electropositive metals (from groups 1 or 2) combined with main group p-block metals or metalloids, yielding charge-balanced narrow band gap semiconductors. Zintl phase structures feature metal cations surrounding p-block 4- element anions which can range from monatomic (Sn in Ca11Sn3C8detailed in chapter 3) to 4- clusters (Sn4 in Na4Sn4) to networks (K8Sn44); the p-block element accepts electrons and/or forms bonds to satisfy the octet rule.67 Zintl phase hydrides contain both p-block anions and

hydride anions, both surrounded by cations, and are exemplified by compounds such as Ca3SnH2 2+ 4- - 2+ 3- - 68,69 ((Ca )3(Sn )(H )2) and Ca5Sb3H ((Ca )5(Sb )3(H )).

The Zintl phase hydride phases we have isolated from Ca/Li/CaH2/M (M = C, Si, Ge, Sn) reactions have all been small band gap semiconductors. Extending this exploration to more electronegative M = arsenic has led to formation of two new larger band gap phases, LiCa3As2H and Ca14As6X7 (X = C, H, N), with the latter incorporating contaminant atoms from the flux. Single crystal neutron diffraction data were crucial in determination of these light atom impurities.

4.2 Experimental Section 4.2.1 Synthesis Chunks of lithium (99.8% Strem), acetylene carbon black powder (99.5% Alfa Aesar), arsenic powder (99.9%, Alfa Aesar, stored under Ar), and calcium hydride powder (98%, Alfa Aesar) were used as received. Calcium shot (99.5% Alfa Aesar) was purified by heating in a steel tube under 10-5 Torr vacuum at 600 °C for 3 hours. Heating was continued under 10-3 Torr for 12 hours; the purified calcium was then cooled and stored under argon. This process decomposes any calcium hydride and calcium hydroxide present and removes the resulting gaseous hydrogen and water. Reactants and flux metals were added to stainless steel crucibles

(7.0 cm length/0.7 cm diameter) in a 7:7:0.7:1.4:0.7mmol Ca/Li/As/C/CaH2 ratio in an argon- filled glovebox. The crucibles were sealed by arc-welding under argon and were placed in silica tubes which were flame-sealed under vacuum. The ampoules were heated from room

temperature to 1050 °C in 3 hours, and held there for 2 hours. The reactions were cooled stepwise to 850 °C in 36 hours, to 600 °C in 36 hours, and then to 500 °C in 24 hours. The reactions were held at 500 °C, then were removed from the furnace, inverted, and centrifuged for 2 minutes to separate the crystalline products from the Ca/Li melt. The solid product adheres to the side of the crucible. The steel crucibles were cut open in an argon-filled glovebox.

4.2.2 Elemental Analysis Elemental analyses were performed using a JEOL 5900 scanning electron microscope with energy dispersive spectroscopy (SEM-EDS) capabilities. Samples were affixed to an aluminum SEM stub using carbon tape and analyzed using a 30 kV accelerating voltage. The atomic ratios of calcium and arsenic could be determined with this method, but it is not sensitive to the lighter elements. The average calcium/arsenic ratio determined from EDS was found to be

70%/30%in agreement with the stoichiometry of the Ca14As6X7phase. (The minority phase

LiCa3As2H has a similar Ca:As ratio, and could not be distinguished from the predominant phase based on EDS measurements.)To analyze the lighter elements and confirm the presence of

nitrogen indicated by neutron diffraction data (vide infra), a sample of Ca14As6X7 was sent to Atlantic Microlabs for CHN analysis. This confirmed the presence of all three of these elements (0.37, 0.61, and 0.83 % by weight respectively).

4.2.3 Structural Characterization Sample crystals were brought out of theglovebox under Paratone oil and were mounted in a cryoloop. Single-crystal X-ray diffraction data were collected at 170 K in a stream of nitrogen using a Bruker APEX 2 CCD diffractometer with a Mo Kα radiation source. The data were integrated with the Bruker SAINT software and corrected for absorption effects using the multiscan method (SADABS).24 Refinements of the structureswere performed using the 39 SHELXTL package. The structure of LiCa3As2H was solved in monoclinic space group Pnma;

that of Ca14As6X7 was solved in tetragonal space group P4/mbm.The heavy atom positions (Ca, As) in both structures were located using direct methods, and the light element sites were refined by least squares analysis of the electron density map and assigned using consideration of bond

lengths, charge balancing, and neutron diffraction data for Ca14As6X7. . Powder X-ray diffraction studies were carried out on reaction products to identify byproducts using a PANalytical X’Pert

Pro X-ray powder diffractometer. In a glovebox, samples of solid products from each reaction were ground into powder and placed in an airtight sample holde rto prevent oxidation. Due to the presence of six light element sites and the possibility of mixed occupancy, single crystal neutron diffraction data were collected on a crystal of Ca14As6X7 using the TOPAZ instrument at Oak Ridge National Laboratory70 In a glove box, the crystal was coated in Krytox grease and adhered to a kapton tube on a sample pin with a magnetic base. This was attached to the goniometer head inside the TOPAZ sample chamber. The crystal was cooled to 100 K under a stream of nitrogen. Data were collected using 15 crystal orientations optimized with CrystalPlan software for an estimated 98 % coverage of symmetry-equivalent reflections of the tetragonal cell.71 Each orientation was measured for approximately 4.5 hrs. The integrated raw Bragg intensities were obtained using the 3-D ellipsoidal Q-space integration method in Mantid.72 Data reduction including neutron TOF spectrum, detector efficiency, and absorption corrections was carried out with the ANVRED2 program.73 The reduced data were exported to GSAS format and the structure was refined using the GSAS program suite74. Refinement parameters were the structure parameters, wavelength dependent detector scaling, and extinction parameters. The extinction corrected data were converted to SHELX HKLF4 format in WinGX and refined to convergence.75,76 The final R-factors reduced to R1 = 0.0547, wR2 = 0.0930 for 7364 reflections after removing 179 outlier reflections with a maximum | |/ ( )= 5σ 2 2 2 cut-off (see Table 4.1). �� − ��

4.2.4 Electronic Structure Calculations

Density of states calculations were carried out on LiCa3As2H using the Stuttgart TB- LMTO-ASA software package.26,27,3 The structural model for this phase was based on the unit cell dimensions and atomic coordinates derived from single crystal diffraction data. Empty spheres were added by the program where appropriate the fill the unit cell volume. An 18×6×6 K-point mesh was used and integrated using the tetrahedron method. The basis sets consisted of 4s/4p/4d for As, 4s/4p/3d for Ca, 2s for Li, and 1s/2p/3d for H. The As4d, Ca 4p, Li 2p/2d and the H 2p/3d orbitals were downfolded.Similar calculations were not carried out on Ca14As6X7 because of the disorder in the structure; the presence of several mixed occupancy sites would require an extremely large supercell as a model.

Table 4.1 Crystallographic data and collection parameters for title phases.

LiCa3As2H Ca14As6C0.46N1.155H5.045 Ca14As6C0.46N1.155H5.045 X-ray data Neutron data Formula weight (g/mol) 278.03 1037.35 1037.35 Crystal System Orthorhombic Tetragonal Tetragonal Space group Pnma P4/mbm P4/mbm a (Å) 11.4064(7) 15.7493(15) 15.7027(2) b (Å) 4.2702(3) c (Å) 11.8762(8) 9.1062(9) 9.1043(2) Z 4 4 4 Volume (Å3) 578.46(7) 2258.7(4) 2244.89(6) Density (g/cm3, calc) 3.078 3.2171 3.068 Index ranges -15 ≤ h ≤ 14, -20 ≤ h ≤ 19, -24 ≤ h ≤ 25, -5 ≤ k ≤ 5, -19 ≤ k ≤ 20, -25 ≤ k ≤ 26, -15 ≤ l ≤ 15 -12 ≤ l ≤ 11 -14 ≤ l ≤ 14 Collection Temp (K) 200 200 100 Reflections collected 6356 25411 7364 Unique data/parameters 783/41 1530 / 92 1330/88 μ (mm-1) 9.868 12.230 0.0333 + 0.0367×λ R1/wR2a 0.0201/0.0398 0.0177/0.0436 0.0541/0.0927 R1/wR2 (all data) 0.0173/0.0428 0.0182/0.0436 0.0547/0.0930 Residual peak / hole 0.57/-0.82 0.89/-1.07 1.02/-1.08 (barn A-3) (e- A-3) The neutron linear absorption coefficient is wavelength dependent and it is calculated as: µ = 0.0333 + 0.0367λ mm-1 a 2 2 2 2 2 1/2 R1 = Σ||Fo|-|Fc||/Σ|Fo|; wR2 = [Σ[w(Fo - Fc ) ]/Σ[w(Fo ) ]]

4.2.5 UV/VIS Spectroscopy

A sample of Ca14As6X7 was ground under argon and loaded into an air-tight holder which was placed into a PerkinElmer Lambda 900 spectrometer with a Labsphere PELA-1000 integration sphere. The spectrum was taken from 300 nm to 1200 nm with a slit width of 2.0 nm.

Reflectance data were converted to absorption using the Kubelka-Munk equation. The absorbance edge is observed at 780 nm from the linear extrapolation. Similar measurements could not be carried out on the minority phase LiCa3As2H due to the small quantity available.

Table 4.2 Atomic positions and site occupancies for Ca14As6C0.46N1.155H5.045 a b Site x y z Occup Ueq As(1) 4g 0.2845(1) 0.2155(1) 0 0.00865(6) As(2) 4h 0.2365(1) 0.2635(1) ½ 0.00730(9) As(3) 16l 0.1979(1) -0.0081(1) 0.2442(1) 0.00720(9) Ca(1) 4f ½ 0 0.3198(1) 0.01153(12) Ca(2) 8k 0.3424(1) 0.1576(1) 0.3059(1) 0.01100(12) Ca(3) 8j 0.0310(1) 0.3130(1) ½ 0.01262(12) Ca(4) 8i 0.3215(1) 0.0195(1) 0 0.00953(11) Ca(5) 8k 0.1834(1) 0.3166(1) -0.2104(1) 0.00818(11) Ca(6) 4e 0 0 0.2512(1) 0.01155(12) Ca(7) 8i 0.1213(1) 0.1139(1) 0 0.00674(15) Ca(8) 8j 0.1189(1) 0.1090(1) ½ 0.00901(16) X(1) 4h 0.4162(2) 0.0838(2) ½ 100% N 0.0148(9) X(2) 2b 0 0 ½ 89(2)% C 0.018(2) 27(2)% N, 43(2)% X(3) 2a 0 0 0 0.0166(13) H H(1) 4g 0.3805(10) -0.1195(10) 0 70(3)% H 0.073(11) H(2) 8k 0.0985(19) 0.4015(19) -0.3490(60) 0.064(11) H(3) 8k 0.4224(13) 0.0776(13) 0.1470(40) 0.038(6) a All occupancies 100% except where noted. b Ueq is defined as one-third of the trace of the orthogonalizedUij tensor. a 2 2 2 2 2 1/2 R1 = Σ||Fo|-|Fc||/Σ|Fo|; wR2 = [Σ[w(Fo - Fc ) ]/Σ[w(Fo ) ]]

4.3 Results and Discussion 4.3.1 Synthesis and Reactivity Reactions of arsenic with light element (X = H, C, N) sources in a Ca/Li flux leads to the formation of two new hydride phases, Ca14As6X7 and LiCa3As2H, as well as traces of LiCaAs which is an analog of the recently reported LiSrAs (with the TiNiSi structure type).177,4 Both of the hydrides form as dark red facetted crystals which are air-sensitive and must be handled under inert atmosphere. Preparing these phases in isolation is challenging because the carbide- containing phase, Ca14As6X7, can form without adding C to the reaction by leaching carbon from the steel crucible. Deliberate addition of higher amounts of carbon reactant favors the carbide.

This. TheLiCa3As2H phase is present as a small minority phase that is usually less than 10% of the total product. A powder X-ray diffraction pattern of the solid product of a reaction of

(Ca/Li)/As/C/CaH2 in a 10:10:1:2:1 mmol ratio indicates the presence of both phases and a small amount of LiCaAs byproduct.

4.3.2 Structure and Properties of LiCa3As2H

LiCa3As2H forms with a new structure type in the monoclinic space group Pnma. Figure 4.1 displays the structure, highlighting the connectivity of the arsenide-centered clusters (Figure 4.1). The strongly reducing flux converts the arsenic reactant to monatomic arsenide anions. The As1 site is coordinated to 7 Ca cations and 2 Li cationsforming a capped square antiprismaticAs@Ca7Li2cluster which is distorted by the short As-Li bond lengths (2.511(4)Å) compared to the much longer As-Ca bonds (2.9374(8) – 3.2014(7)Å4.).The As2 site is coordinated to 7 Ca atoms and one Li atom in a bi-capped trigonal prismatic As@Ca7Liunit with the Li atom on one of the caps (As-Li bond 2.538(7)Å, and As-Ca bonds 3.0006(6) –

3.0883(7)Å). These As@Ca7Li2and As@Ca7Li clusters share edges and faces to form a network that defines channels containing the hydride anions. The bond distances compare well to those 78,79,15,16 reported for Li3As (Li-As distance 2.57Å)and Ca5As3 (Ca-As distances 2.90 – 3.13Å). The hydride site is coordinated by 4 Ca cations and 1 Li cation, forming trigonal bipyramidal H@Ca4Li clusters. These units share equatorial corners to form a chain running along the a-axis. The H-Li bonds (2.01(7)Å) are axial and all point in the same direction in the chain. While each individual chain is polar, the symmetry equivalents are aligned in opposing directions, so the net dipole is cancelled. The Ca-H bond lengths range from 2.17(2) – 2.396(10)

Å. These bond lengths are consistent with those seen in phases such as BaLiH3 (Li – H distance 7, 17 2.01Å) and CaH2 (Ca – H distances 2.27 – 2.65 Å). An interesting point of comparison is

LiCa2C3H which is also grown from a Ca/Li flux. It features chains of corner-sharing H@Li2Ca4 units, with the hydride coordinated by both Ca (Ca-H = 2.502 Å) and Li (Li-H = 1.876 Å) ions in an octahedral arrangement.7

Table 4.3 Bondlengths of interest in arsenide hydride phases, in angstroms.

Bond LiCa3As2H As1 – Ca 2.9691(13) – 3.2318(4) As2 - Ca 2.9700(13) – 3.2362(4) As3 - Ca 3.0325(8) – 3.1223(3) As4 – Ca 3.033(8) – 3.1216(3) H1 - Ca 2.2863(10) – 2.4391(12) H2 - Ca 2.3233(10) – 2.4704(12) H1 – Li1 2.0250(9) H2 – Li2 2.0347(10)

Bond Ca14As6X7 As1 – Ca 2.9556(2) – 3.1421(3) As2 – Ca 2.8893(2) – 3.3283(3) As3 – Ca 2.9733(2) – 3.5076(2) H1 – Ca 2.3674(2) – 2.4051(2) H2 – Ca 2.2017(1) – 2.2834(1) H3 – Ca 2.2496(1) – 2.3472(1) X1 – Ca 2.4145(1) – 2.4852(1) X2 – Ca 2.2654(2) – 2.5403(2) X3 - Ca 2.2877(2) – 2.6201(2)

LiCa3As2H is a Zintl phase, a complex salt which can be charge-balanced as follows: + 2+ 3- - (Li )(Ca )3(As )2(H ). Density of states data are shown in Figure 4.2 which indicate a band gap

of 1.4 eV, in agreement with the red color of the crystals. The binary salt Ca3As2 is also reported 18 to have a reddish-brown coloration. The arsenic states of LiCa3As2H make the dominant

contributions to broad bands at and just below EF (from 0 to -3 eV); calcium states are dominant

above the band gap. The hydride states are localized well below EF, between -4 and -5 eV. This is comparable to the calculated DOS for other Zintl phase hydrides; the hydride states in 38,24.86 Ca3SnH2, LiCa2C3H, and LiCa7Ge3H3 are also found 4 eV below the Fermi level. These Zintl hydrides have significantly smaller band gaps (all below 0.5 eV) than the arsenides reported here. In these phases, the band gap is determined by the relative energies of the metalloid anion states which comprise the valence band, and the calcium states which make up the conduction band. The greater electronegativity of arsenic leads to more stabilized valence bands and a larger band gap.

Figure 4.1 The orthorhombic structure of LiCa3As2H, viewed down the b-axis,with different anion coordination polyhedra highlighted. Blue and magenta spheres are calcium and lithium cations, respectively. Left: connectivity of arsenic-centered polyhedra (green). Right: connectivity of corner-sharing hydride-centered trigonal bipyramids (yellow).

60 Total 50 As ) )

-1 Ca 40 Li cell

-1 H 30 20

DOS, (eV 10 0 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 E, eV

Figure 4.2 Calculated total and partial density of states data for LiCa3As2H.

4.3.3 Structure of Ca14As6X7 (X = C, H, N)

Ca14As6X7 also exhibits a new structure type, in the tetragonal space group P4/mbm, shown in Figure 4.3.It is composed of a network of edge and face-sharing As@Ca8 and As@Ca9 clusters (shown in Figure 4.3a) interspersed with X@Ca6octahedra and X@Ca4tetrahedra which are linked into chains running along the c-axis (Figure 4.3b). Of the three unique arsenide sites, two can be described as being surrounded by a snub disphenoid coordination of 8 calcium cations, forming As@Ca8 units. The other arsenide site has 9 neighboring calcium sites, creating a monocapped square antiprismatic As@Ca9 unit (Figure 4.4). The Ca-As bond distances in

Ca14As6X7 range from 2.9556(5) – 3.5076(4) Å. The longest bond distance is As1 – Ca2 bond, and is considerably longer than the next longest Ca-As bond of 3.3491(3) Å. This range of Ca-As bonds is larger than that observed forLiCa3As2H4.3, but not unusual compared to other phases with arsenic sites coordinated by 8 or more calcium atoms. The binary phase Ca2As has arsenic coordinated by 8 Ca atoms in a square antiprism with bond distances ranging from 3.00 – 3.32 Å, and the superconducting phase Ca10(Pt3As8)(FeAs2)5 features Ca-As bonds ranging from 2.957 – 80,81,19,20 3.387 Å. Ca14MnAs11 features 10-coordinate As with Ca-As bond lengths ranging from 2.895 – 3.562 Å.82,1

The Ca14As6X7 structure has six unique light atom positions. Three of these sites are tetrahedrally coordinated by calcium cations; occupancy refinements and bond lengths indicate these are most likely hydride sites (Ca – H distances in the range 2.20(2) –2.405(8)Å). These

H@Ca4 tetrahedra share corners and edges in order to form tubes running in the c-axis direction, surrounded by the calcium arsenide network. Another light atom site, X(1), coordinated by six

calcium cations, is also contained within these calcium hydride tubes. The occupancy and bond lengths (in a small range of 2.415(1) – 2.485(1)Å) suggest that this site is occupied by either carbon or a similar light element, X(1) = C or N. Each of these X(1)@Ca6 units share edges with

a symmetry equivalent, and they also share edges with H@Ca4 units. The two remaining light atom sites, X(2) (on special position 2b at (0,0,½)) and X(3) (on special position 2a, at the origin of unit cell), are also octahedrally coordinated by 6 calcium cations; these X@Ca6 units share opposing corners to form a chain running along the c-axis of the unit cell. Both of these octahedra are compressed along this axis; the X-Ca bonds to the axial (shared) corners are 2.2654(6) Å or 2.2877(6) Å, while the four equatorial X-Ca bonds are longer

in each case (2.5403(5) or 2.6200(5) Å). This chain is surrounded by As@Ca9 capped square antiprismclusters.

H As

Figure 4.3 The tetragonal structure of Ca14As6C0.46N1.155H5.045, viewed down the c-axis, with different anion coordination polyhedra highlighted. Calcium cations represented by blue spheres. Left: connectivity of arsenic-centered polyhedra (green). Right: connectivity of corner- and edge-sharing polyhedra centered by light atoms; H@Ca4tetrahedra are yellow and X@Ca6octahedra (X = C, H, N) are black.

If the tetrahedrally coordinated light atoms are assigned as hydrides, and the octahedrally coordinated light atom positions as carbides, full occupancy of all anionic sites would lead to a

2+ 3- 4- - stoichiometry of Ca14As6C2H5(Z = 4). This is not charge-balanced; (Ca )14(As )6(C )2(H )5 has a net negative charge. However, semiconducting charge-balanced behavior is indicated by the red coloration of the compound, and the charge-balanced nature (and similar red color) of the co- product LiCa3As2H. Partial occupancy of the arsenic sites was not indicated in the crystallographic data refinement. Partial or mixed occupancy on the carbide and hydride sites is therefore mandated, but occupancies of light elements are difficult to analyze from single crystal X-ray diffraction data due to their low X-ray scattering factors. However, the neutron scattering lengths of carbon, hydrogen, and nitrogen are significantly different enough (6.6460 fm, -3.7390 fm, and 9.36fm, respectively) to enable refinement of mixed or partial occupancies of these sites.22 The single crystal neutron diffraction data were analyzed using the constraints that all calcium and arsenic sites were fully occupied (as indicated by SCXRD) and that the charges should balance (as indicated by the red color of the crystals).

Figure 4.4 Local coordination environments for anions in Ca14As6C0.46N1.155H5.045.

Of the 6 light element sites in the structure, the three that aretetrahedrally coordinated by Ca2+ ions (a 4g site and two 8k sites) had the smallest electron density peaks in the single crystal XRD refinement, indicative of hydride occupancy as mentioned above. This was supported by the neutron diffraction data; the two 8k sites could be refined as fully occupied by hydride with stable thermal parameters. The 4g site on the other hand refinedasas slightly less than fully occupied (82% hydride occupancy).

Intensity (au) Intensity 1

0 300 500 700 900 1100 Wavelength (nm)

Figure 4.5 The absorbance spectrum of Ca14As6X7 (X = C, H, N) from diffuse reflectance data.

Refinement of the three octahedrally coordinated light element sites (X(1), X(2), and X(3) on 4h, 2b, and 2a sites respectively)proved more complicated. These were originally assigned as carbon in the refinement of the SCXRD data; when the occupancy was allowed to refine, X(1) refined as over 100% occupied, X(2) as close to 100% occupied, and X(3) as below 100% occupied. However, it was immediately evident from the neutron diffraction data that the 4h site could not be occupied by carbon, as it had a much larger neutron scattering factor. This is also clearly distinct from the negative scattering factor of hydrogen; occupancy by oxygen is also precluded, since its neutron scattering factor of 5.803 fm is lower than that of carbon. The data clearly indicate occupancy by nitrogen, with the site refining as 100% occupied. These

N(1)@Ca6octahedra are linked to one symmetry equivalent and six surrounding

H@Ca4tetrahedra by edge sharing, forming the chain depicted in Figure 5c. The Ca-N bond lengths are in a small range of 2.4145(1) – 2.4852(1)Å. Similar distances are seen in Ca3AsN and

Ca3N2both of which feature N@Ca6octahedra with bond lengths of 2.39 – 2.42 Å and 2.41 – 2.48 Å respectively.83,843,24

The other two light element sites center X(2)@Ca6 and X(3)@Ca6octahedra which share opposing corners to form a chain running along the c-axis (Figure 5d). Refinement of the neutron diffraction data indicates that X(2) is 92% occupied by carbon, in agreement with the X-ray refinement which also supports carbon at this site. The X(3) site had a low neutron scattering factor; charge-balancing considerations were used to aid in assigning the site occupancy. With three hydride sites (and 82% occupancy of the 4g site), one nitride site, and one carbon site with

92% occupancy, an overall stoichiometry of Ca14As6N1C0.46H4.82(X(3))0.5-d is indicated, allowing for the fact that the X(3) site may be partially occupied. The charge and occupancy of the anions on the X(3) site must contribute -0.34 to the overall formula to allow for charge balancing. Unfortunately there are a number of variables involved, the anion mixture and site occupancy, but it is likely that this site is possibly partially occupied by a mixture of N and H. This site was refined as containing a mixture of31% N and 46% H; this yielded stable thermal parameters, This produces a low overall neutron scattering factor with stable thermal parameters for this site, and would also yield an apparent occupancy in the X-ray refinement lower than carbide. The resulting overall stoichiometry of Ca14As6C0.46N1.155H5.045 is close to being charge balanced.

The presence of nitrogen in Ca14As6X7was confirmed by CHN elemental analysis,. indicating a nitrogen mass percentage of 0.83%, slightly lower than the 1.3% expected from the refined stoichiometry. The acetylene carbon black used as a reactant was also analyzed by CHN analysis and contained no nitrogen, sothe source of the nitrogen appears to be the flux metals. Both lithium and calcium are known to form surface nitrides, and the large quantity of these metals used could lead to significant amount of nitride present in the melt.

In the flux synthesis of Ca14As6C0.46N1.155H5.045, the growing crystals may scavenge from the molten metal solution whatever small anions are needed to maintain charge neutrality. The 4- 3- - Ca6octahedra are large enough to incorporate C , N , and H . Similar behavior likely occurs during the formation of other mixed anion phases such LiCa11Ge3OHx and 24,8525 (Ca,Y)2Si4(N,C)7. Another mixed anion compound, Ba21Ge2O5H24, provides an interesting case study.61 While this phase does not feature O2- and H- mixing on one site, a structural analog

- 4- (Sr21Si2O5C6) can be synthesized by substituting the H anions for C anions, with partial occupancy of the carbide anions to balance the charge.2,86,6 In these alkaline-earth rich phases, the AE6octahedra are of suitable size to incorporate different small anions (and mixtures of anions), as is also seen in Ca14As6C0.46N1.155H5.045.

4.4 Conclusions

The ability of calcium/lithium melts to dissolve ionic hydrides and main group metalloids make them rich growth media for complex metal hydrides. The high reactivity of these flux metals does lead to the possibility of impurities being incorporated into products. The use of purified (distilled or dentritic) alkali metals and alkaline earth metals is recommended to avoid accidental incorporation of oxide, nitride, and hydride anions. However, incorporation of several different monatomic anions leads to new compounds often characterized by high levels of structural complexity.

CHAPTER FIVE

Ca54In13B4–XH23+X: A COMPLEX METAL SUBHYDRIDE FEATURING IONIC AND METALLIC REGIONS

5.1 Introduction Reactions with heavy elements past the Zintl border (Group 14 and 15) in previous chapters have resulted in Zintl phases. Group 13 elements are less likely to form charge balanced phases which will be seen to form subhydride phases in this and later chapters. Metal hydrides, of interest for potential use as battery and hydrogen storage materials, exhibit a broad range of bonding types and electronic characteristics.87 Ionic hydrides form with electropositive metals; these charge-balanced semiconductors or insulators include simple binaries like CaH2 as well as complex Zintl phase hydrides such as Ba21Ge2O5H24 and 66,88 Yb3PbH2. Covalent metal hydrides contain hydrogen bonded to main group metals or metalloids, exemplified by LiBH4 and B2H6 as well as BaAlSiH and other polyanionic hydride 89,90 Zintl phases. Transition metal hydrides such as PtHx and LaNi5Hx are metallic and can contain a variable amount of interstitial hydrogen. We report here the metal flux synthesis of a complex metal hydride that falls in between these classifications. Flux synthesis makes use of a low-melting element or compound present in large excess which acts as a solvent for the other reactants. This technique has been used to grow large crystals of known phases and to discover new compounds.91 Calcium metal (mp 845 °C) is too high melting to be useful as a flux, but a 1:1 mol ratio of Ca and Li melts at 300 °C. This Ca/Li flux mixture dissolves CaH2. It is also strongly reducing, and it will convert most main group metals and metalloids to their anionic state. As a result, reactions of CaH2 with metalloids in

Ca/Li flux yield salt-like Zintl phase hydrides such as LiCa2C3H and LiCa7Ge3H3; these phases + 2+ 4– contain Li and Ca cations surrounding the hydride anions and the tetrelide anions (C3 and Ge4–, respectively).20,35

The exploration of Ca/Li/CaH2/M reactions with M = group 13 metals has led to the discovery of Ca54In13B4–xH23+x. While group 14 reactants are sufficiently electronegative to be reduced by Ca/Li flux to form anions, metals such as indium are not likely to be reduced to the −5 state. Instead, the title phase has ionic calcium hydride regions separated by a metallic

calcium/indium network; conduction electrons are confined to the latter regions of the structure, 92,93,94,95 similar to the behavior seen for suboxides and subnitrides such as Cs11O3 and Na16Ba6N.

We have found solid-state NMR to be extremely useful in characterizing the Ca54In13B4–xH23+x “subhydride”, since both chemical shift and relaxation time are affected by the interaction of a nucleus with conduction electrons.96 The 1H and 115In NMR spectra, as well as electronic structure calculations, confirm the presence of conducting and insulating regions in this compound.

5.2 Experimental 5.2.1 Synthesis Boron powder (95–97%, Strem), calcium hydride powder (98%, Alfa Aesar), and indium powder (99.9%, Alfa Aesar) were used as received. Chunks of lithium (99.8%, Strem) were soaked in hexanes to remove their mineral oil coating and then stored and handled under argon. Calcium chunks (99%, Alfa Aesar) were purified by heating at 700 °C under a dynamic high –5 vacuum of 10 Torr for 10 h to decompose any CaH2 and Ca(OH)2 (common contaminants in commercial calcium metal) and then stored and handled under argon. The procedure will not eliminate trace CaO contamination in the calcium metal, but XPS studies did not indicate any incorporation of oxygen into the products. Reactants and flux metals were added to stainless steel crucibles (7.0 cm length/0.7 cm diameter) in a 10:10:1:1:1 mmol Ca/Li/In/B/CaH2 ratio in an argon-filled glovebox. The crucibles were sealed by arc-welding under argon and were placed in silica tubes that were flame-sealed under vacuum. The ampules were heated from room temperature to 1050 °C in 4 h and held there for 2 h. The reactions were then cooled stepwise to 800 °C over 36 h, 600 °C over 36 h, and 500 °C over 24 h. The reactions were held at 500 °C and then were removed from the furnace, inverted, and centrifuged for 2 min to separate the crystalline products from the Ca/Li melt. The steel crucibles were cut open in an argon-filled glovebox to isolate the highly air- sensitive product, which adheres to the sides of the crucible. It was subsequently determined that the best yield was found for reactions with a starting Ca/Li/In/B/CaH2 millimole ratio of 10:10:3:1:2. The reaction was repeated in a sealed niobium crucible to eliminate the possibility of contaminants leaching from the steel crucible.

5.2.2 Elemental Analysis Elemental analyses were performed using a JEOL 5900 scanning electron microscope with energy dispersive spectroscopy (SEM-EDS) capabilities. Samples of product crystals were affixed to an aluminum SEM stub using carbon tape and analyzed using a 30 kV accelerating voltage. The observed Ca:In atomic ratios ranged from 80:20 to 90:10. No incorporation of metals from the steel crucible was observed in any of the samples.

17000 120000 B 1s Ca 2p 16500 110000 16000 100000 15500 90000 Counts 15000 Counts 80000 14500 70000 14000 60000 195 200 205 210 340 345 350 355 360 Binding Energy (ev) Binding Energy (ev)

11000 Li 1s 10500 10000 9500

Counts 9000 8500 8000 45 50 55 60 65 Binding Energy (ev)

Figure 5.1 X-ray Photoelectron Spectroscopy (XPS) data for Ca54In13B4-xH23+x, highlighting the very strong Ca 2p photoelectron peak, weak B 1s peak, and very weak Li 1s photoelectron peak (likely due to traces of flux residue on the sample surface), in their expected binding energy regions.

This technique is not sensitive to the presence of light elements, so X-ray photoelectron spectroscopy measurements were carried out, using a Physical Electronics PHI 5100 series XPS with a non-monochromated dual anode (Al and Mg) source having a single channel

hemispherical energy analyzer. A Mg Kα source was used. Sputtering of the sample was carried out to remove any surface oxidation. XPS spectra were taken after every 5 min of sputtering until the oxide peak disappeared, and no further changes in the spectra were observed. These measurements showed the clear presence of Ca and B in the sample, but the lithium 1s peak was very small (see Figure 5.1) While the presence of a small amount of lithium was indicated by XPS spectra, 7Li MAS NMR studies did not show any lithium peaks. Lithium is likely present only in trace amounts on the surface of the crystals; the powder XRD pattern (Figure 4.2) shows

CaLi2 as a contaminant, formed from the freezing of Ca/Li flux residue on the product. Since the XPS samples were mounted on graphite tape, which physisorbs water and hydrocarbons, the presence of H was not confirmed with this method.

5.2.3 Crystallographic Studies

Crystals of Ca54In13B4–xH23+x were brought out of the glovebox under Paratone oil and were mounted in a cryoloop. Single-crystal X-ray diffraction data were collected at 173 K in a stream of nitrogen using a Bruker APEX 2 CCD diffractometer with a Mo Kα radiation source. An absorption correction was applied to the data using the SADABS program.97 Refinement of the structure was performed using the SHELXTL package.98 The structure was solved in cubic space group Im3 (No. 204). Fully occupied calcium and indium sites were determined using direct methods. Partially̅ occupied calcium sites and light element positions were located using difference Fourier calculations. Hydride ions were modeled as helium atoms to account for the extra electron on the atom; this yielded more stable thermal parameters on these sites and improved the fit.67b,99 Crystallographic data and collection parameters are shown in Table 5.1, and atomic positions and thermal parameters can be found in Table 5.2 and 5.3 for two example crystals. Powder X-ray diffraction studies were carried out on reaction products to identify byproducts using a PANalytical X’Pert Pro X-ray powder diffractometer equipped with a Cu Kα Xray source. In a glovebox, samples of solid products from the optimized reaction

(Ca/Li/In/B/CaH2 millimole ratio of 10:10:3:1:2) were ground and placed in an airtight sample holder with a Kapton cover. The pattern (shown in Figure 5.2) indicates that the product is

predominantly Ca54In13B4–xH23+x; a few small extra peaks indicate the presence of CaLi2 (from

solidified flux residue) and a small amount of Ca3In byproduct.

Figure 5.2 Powder X-ray diffraction data for Ca54In13B4-xH23+x, compared to the theoretical pattern calculated based on single crystal structure. Analysis of the pattern indicates a unit cell parameter a = 16.35(3) Å. Trace impurities include CaLi2 (from solidified residual flux) and Ca3In byproduct; these peaks are indicated by red symbols.

Table 5.1 Crystallographic data and collection parameters for two samples of Ca53In13B4H23.

Ca53In13B4-xH23+x Ca52.9In13B1.3H25.7 Formula weight (g/mol) 3683 3648.3 Crystal System Cubic Cubic Space group Im3¯(No.204) Im3¯(No.204) a (Å) 16.3608(6) 16.358(1) Z 2 2 Volume (Å3) 4379.4(3) 4376.7(6) Density (g/cm3, calc) 2.767 2.768 Index ranges -21< h < 21 21< h < 20 Temperature (K) 200 200 Wavelength (Å) 0.71073 0.71073 Reflections collected 24720 22925 Unique data/parameters 988 / 62 977/62 μ (mm-1) 6.47 6.47 R1/wR2a 0.0165/ 0.0394 0.169/0.0368 R1/wR2 (all data) 0.0171 / 0.0396 0.0173/0.0369 Residual peak / hole (e- A-3) 0.98 / -0.55 0.93/-0.53

a 2 2 2 2 2 1/2 R1 = Σ||Fo|-|Fc||/Σ|Fo|; wR2 = [Σ[w(Fo - Fc ) ]/Σ[w(Fo ) ]]

Table 5.2 Atom positions and isotropic thermal parameters for Ca53In13B4H23

a Wyckoff x y z occ Ueq Site In(1) 2a 0 0 0 0.0113(1) In(2) 24g 0 0.20929(1) 0.33803(1) 0.01563(9) Ca(1) 24g 0 0.10896(4) 0.17264(4) 0.0161(1) Ca(2) 12d 0 0 0.35830(6) 0.0167(2) Ca(3) 48h 0.18401(3) 0.11348(3) 0.29807(3) 0.0232(1) Ca(4) 24g 0 0.3981(1) 0.3819(2) 0.568(6) 0.0388(8) Ca(5) 12e 0 0.3222(2) ½ 0.282(4) 0.028(1) Ca(6) 24g 0 0.4078(3) 0.3214(7) 0.181(6) 0.041(3) H(1) 16f 0.1471(6) 0.1471(6) 0.1471(6) 0.048(4) H(2) 24h 0.089(1) 0 0.235(1) 0.082(6) H(3)/B(3) 8c ¼ ¼ ¼ 0.59(6)/0.41(6) 0.024(4) H(4) 6b 0 0 ½ 0.014(4) a Ueq is defined as one-third of the trace of the orthogonalized Uij tensor.

Table 5.3 Atom positions and isotropic thermal parameters for the second crystal of Ca53In13B4H23 a Wyckoff x y z occ Ueq Site In(1) 2a 0 0 0 0.01131(15) In(2) 24g 0 0.209290(14) 0.338035(14) 0.01563(9) Ca(1) 24g 0 0.10895(4) 0.17262(4) 0.01610(14) Ca(2) 12d 0 0 0.35828(6) 0.01667(19) Ca(3) 48h 0.18400(3) 0.11349(3) 0.29806(3) 0.02318(13) Ca(4) 24g 0 0.39808(12) 0.3819(2) 0.595(5) 0.0385(9) Ca(5) 12e 0 0.3224(3) ½ 0.242(4) 0.0278(12) Ca(6) 24g 0 0.4078(3) 0.3213(8) 0.191(5) 0.042(3) H(1) 16f 0.1469(6) 0.1469(6) 0.1469(6) -0.072(4) H(2) 24h 0.0900(12) 0 0.2352(12) -0.032(4) H(3)/B(3) 8c ¼ ¼ ¼ 0.69(6)/0.31(6) 0.023(4) H(4) 6b 0 0 ¼ 0.004(3) a Ueq is defined as one-third of the trace of the orthogonalized Uij tensor.

5.2.4 NMR Spectroscopy

For solid-state NMR studies, crystals from several batches of Ca/Li/In/B/CaH2 reactions in a ratio of 10:10:3:1:2 mmol were ground together with NaCl (a 50:50 by volume mixture of

Ca54In13B4–xH23+x and NaCl) to facilitate the spinning of the conducting sample in the magnetic field. For magic angle spinning (MAS) 1H NMR data collections, the samples were loaded into a 1.3 mm zirconia rotor in a glovebox, for use in an Ultrafast spinning MAS probe with spinning speeds of 60 kHz on a Bruker AVIII HD 500 MHz WB spectrometer (B0 = 11.7 T). Data were collected after 32 scans with a recycle delay of 30s (T1 relaxation time was measured to be 5 s). An empty rotor was measured under the same conditions to subtract the background. Spectra were referenced to TMS at 0 ppm, with adamantane (1.6 ppm) used as a second reference. For 115In NMR experiments, the sample was packed in a 4 mm diameter zirconia rotor in a

glovebox. For quadruple nuclei with large values of CQ, the highest practical applied magnetic field strength is usually advantageous. Therefore, a Bruker Avance II 800 MHz spectrometer with an 115In frequency of 175.4 MHz was used for the measurements. A 90° pulse width of 0.75

μs and relaxation delay of 1 s were used. Aqueous 0.1 M In(NO3)3 in 0.5 M HNO3 was used as a reference at 0 ppm. Due to the extreme broadness of the 115In spectrum, MAS was not effective. Instead, the QCPMG pulse sequence and stepped-frequency technique was used. The QCPMG (quadrupolar Carr–Purcell–Meiboom–Gill) experiment is a modern NMR technique that adapts the CPMG sequence to enhance the NMR signal of quadrupolar nuclei. This method is particularly useful for the acquisition of data for dilute or unreceptive quadrupolar nuclei. In the QCPMG pulse sequence, the magnetization is refocused repetitively, giving rise to a train of echoes upon Fourier transformation, which leads to a set of spikelets with manifold resembling the conventional powder pattern. Since all the intensity is allocated into sharp spikelets, a large gain in signal-to-noise ratio is obtained.100,101 Experimental conditions for QCPMG were set following literature procedures, with 4096 scans for each step.102 The individual stepped- frequency spectra were coded using the skyline projection method. 115In NMR parameters were determined by visual comparison of experimental NMR spectra with those simulated using the DMFit software package.103

5.2.5 Electronic Structure Calculations Density of states calculations were carried out using the TB-LMTO-ASA technique with the Stuttgart TB-LMTO 4.7 software package.104 Exchange and correlation effects were calculated within the local density approximation (LDA) using the von Barth–Hedin local exchange correlation potential.25,26 ,105 A self-consistent field calculation is performed by the

program. Two structural models for the Ca54In13B4–xH23+x phase were used, based on the unit cell dimensions and atomic coordinates derived from single crystal diffraction data. For both models, instead of the three partially occupied calcium split sites [Ca(4), Ca(5), and Ca(6)], it was assumed that the Ca(5) and Ca(6) sites were empty and the Ca(4) site was fully occupied. In one model, the B/H mixed occupancy 8c site was occupied only by boron, and in the other model, it

was occupied by hydrogen. The stoichiometries of the resulting models are Ca54In13B4H23 and

Ca54In13H27. The Wigner–Seitz radii were determined by an automatic procedure as follows: Ca = 1.703–1.883 Å, In = 1.766–2.058 Å, H = 0.977–1.00 Å, B = 1.361 Å. Empty spheres were added by the program where appropriate to fill the unit cell volume. A 12 × 12 × 12 k-point mesh was used and integrated using the tetrahedron method. The basis sets consisted of 5s/5p/5d/4f for In, 4s/4p/3d for Ca, 2s/2p/3d for B, and 1s/2p/3d for H. The downfolded orbitals consist of 5d/4f for In, 4p for Ca, 3d for B, and 2p/3d for H.

5.3 Results and Discussion 5.3.1 Structure Description

Ca54In13B4–xH23+x (2.4 silver spheroid crystals up to 1 mm in diameter. It exhibits a new structure type in the cubic space group Im3 with a unit cell length of a = 16.289(1)–16.393(1) Å, as shown in Figure 5.3.

The variation in ̅unit cell parameter stems from the presence of several calcium split sites and B/H mixing on another site. The structure is built upon a body-centered cubic array of indium- centered calcium icosahedra; the In(1) atoms in the corners and center of the unit cell (2a Wyckoff sites) are coordinated by 12 Ca(1) ions at a distance of 3.3400(7) Å (Figure 5.4a). This 106,107 is within the 3.2–3.7 Å range reported for Ca–In distances in Ca18Li5In25 and Ca2In. Each triangular face of the In@Ca12 icosahedron is capped by a hydride ion [H(1) and H(2) ions, in 24g and 16f Wyckoff sites, respectively], as shown in Figure 4.4b. The resulting pentagonal dodecahedron of hydride ions is in turn encapsulated by a sphere of 30 calcium ions (Ca(2) and

Ca(3), in 12d and 48h Wyckoff sites, respectively). Each hydride site thereby obtains an octahedral coordination of calcium cations, with Ca–H distances ranging from 2.47(2) to 2.63(2) 67 Å. Similar bond lengths are seen in CaH2, LiCa2C3H, and LiCa7Ge3H3 (2.23–2.67 Å). The resulting In@Ca12@H20@Ca30 cluster (Figure 5.4c) can be viewed as an indium atom coated with a calcium hydride shell.

Figure 5.3 Structure of Ca54In13B4–xH23+x. Indium atoms are shown as red spheres. Ordered and disordered calcium sites are blue and gray spheres, respectively. Mixed B(3)/H(3) sites are shown as black polyhedra and hydride sites as yellow polyhedra.

The calcium surface of the In@Ca12@H20@Ca30 cluster is capped by In(2) atoms (24g sites) and an additional light element in an 8c site. Twelve In(2) atoms form an icosahedron around the cluster; the resulting In@Ca12@H20@Ca30@In12 unit contains three of the four concentric shells of Bergman clusters seen in quasicrystals and approximants as shown in Figure 4.5.108 The electron density at the light element 8c site is consistently higher than the H(1) and H(2) sites, but lower than expected for a boron atom. This position is assigned as a B/H mixture, denoted as B(3)/H(3). Refinements of single crystal XRD data for several crystals from different syntheses show boron ratios on this site ranging from 0% to 41% (yielding stoichiometries from

Ca54In13H27 to Ca54In13B1.6H25.4). While it is inherently problematic to refine light element site occupancies in the presence of surrounding heavy elements, it was observed that the unit cell

parameter increases with boron content. This 8c site links neighboring clusters together, as shown in Figure 5.4d, achieving an octahedral coordination from three calcium atoms on one cluster and three on a neighboring cluster, with B–Ca distances of 2.6026(5) Å. This is somewhat short compared to other Ca–B bond lengths; the observed range in CaB4 and CaB6 is 2.738– 3.141 Å.109 There are a few phases with shorter Ca–B bonds: A Ca–B bond of 2.554 Å is found

in Ca[B(OH)4]2(H2O)2, 2.579 Å in Ca5Cl3C2(CBC), 2.459 Å in Ca9Cl8(BC2)2, and 2.458 Å in 110 Ca2(BN2)F. The relatively short 2.6026 Å Ca–B/H bond length in Ca54In13B4–xH23+x may be stabilized by the high ratio of hydride ions mixed on this site.

a b

d c

Figure 5.4 Ordered crystallographic sites of Ca54In13B4–xH23+x. (a) The In(1) site is coordinated by an icosahedron of calcium ions. (b) Each face of the icosahedron is capped by a hydride anion. (c) The hydride anions are capped by calcium sites, to form the In(1)-centered cluster, In@Ca12@H20@Ca30. Octahedrally coordinated hydride sites are depicted as yellow polyhedra. (d) The In(2) atoms (red) and B(3)/H(3) sites (black polyhedra) cap the clusters, with the B(3)/H(3) sites bridging two clusters.

Figure 5.5 An alternative view of the In(1)-centered cluster in Ca54In13B4-xH23+x, highlighting its similarity to Bergman clusters found in quasicrystalline approximants such as NaAuSn. Dashed lines are not bonds but guidelines to highlight polyhedra. Variations arise in the 2nd and 4th shells. The 2nd coordination shell in Ca54In13B4-xH23+x is not a simple pentagonal dodecahedron; it is a pentagonaldodecahedron of Ca-coordinated hydride sites. The extra Ca30 shell (blue spheres) is viewed as part of the 20 H@Ca6 octahedra which link to form a pentagonal th dodecahedron. The expected buckyball-shaped 4 shell is not seen in Ca54In13B4-xH23+x due to the presence of three disordered calcium sites (grey spheres).

The In@Ca12@H20@Ca30 clusters and the boride/hydride sites that link them together form a 3-D network; the voids defined by this network (see Figure 5.4d) contain disordered calcium sites. This is similar to the encapsulation of disordered solvent molecules in the cages of framework compounds such as MOFs and zeolites grown solvothermally. In the synthesis of

Ca54In13B4–xH23+x, the calcium-rich metal flux provides the disordered “solvent” atoms. The partially occupied Ca(4), Ca(5), and Ca(6) sites were consistently observed in structural refinements of several crystals. These sites are too close to each other to be fully occupied; their occupancies were therefore constrained to sum to 1.

a b

Figure 5.6 Disordered region of the Ca54In13B4–xH23+x structure. (a) The In(2) site is coordinated by an icosahedron of calcium ions (blue and gray spheres represent ordered and disordered Ca sites, respectively). (b) H(4) hydride site. (c) Network of disordered Ca sites, viewed along the [111] direction.

An additional hydride ion H(4), on the 6b Wyckoff site, is found in the midst of these disordered calcium cations, as seen in Figure 5.5b. Depending on the calcium site occupancy, this hydride site can be coordinated by four Ca2+ ions in a square planar configuration or by an octahedron of Ca2+ ions. These three partially occupied calcium sites also complete the coordination sphere of the In(2) atoms, producing a distorted icosahedral configuration around this site, with Ca–In bonds in the 3.089–3.820 Å range. The 3-D connectivity of this disordered calcium ion network, and the hydride and indium atoms coordinated by these ions, is highlighted in Figure 5.5c. The voids in this figure are along the body diagonal of the cubic unit cell and are filled by chains of ordered In@Ca12@H20@Ca30 clusters linked by the 8c boride/hydride site.

5.3.2 Solid-State NMR Spectroscopy

The view of the structure of Ca54In13B4–xH23+x as interpenetrating lattices of linked

In@Ca12@H20@Ca30 clusters and a disordered calcium indide/calcium hydride network may also extend to its electrical properties. 1H and 115In NMR spectra were collected to observe the effects of conduction electrons on the chemical shifts and relaxation times of the nuclei. There are four hydride sites in this compound. H(1) and H(2) form the ordered hydride shell around the In(1) site (Figure 5.4b). These two hydrides have very similar sites and local coordination environments and are expected to have very similar chemical shifts. H(3) and H(4) have much lower multiplicity, with H(3)/B(3) mixing further reducing the hydride content on this 8c site. The 1H MAS NMR spectrum is shown in Figure 5.6. The dominant resonance at 7.7 ppm has a spin–lattice relaxation time T1 of 5 s and is likely due to the combination of H(1) and H(2) sites. This peak was broad at low spinning speeds, with a wide envelope of spinning side bands, but ultrafast MAS at 60 kHz narrowed the peak and collapsed the side bands. The observed chemical shift of 7.7 ppm is in the 3–9 ppm range observed for other ionic metal hydrides. For instance, 1 CaH2, SrH2, and BaH2 have reported H NMR resonances at 4.5, 6.7, and 8.7 ppm, respectively.111,112 Hydride anions surrounded by Ca2+ cations are also found in H–-doped 4+ – 1 1131 mayenite [Ca24Al28O64] ·4H , yielding a H chemical shift of 5.1 ppm. H resonances for more 67,114 covalent hydrides such as LiH, MgH2, and LiCa2C3H are found around 3 ppm. The 7.7 ppm chemical shift for the hydride sites in Ca54In13B4–xH23+x is also similar to the 8.1 ppm shift reported for BaInGeH, although that compound is highly disordered and contains amorphous

inclusions and the exact position of the (likely interstitial) hydride sites was not clear.115 No

Knight shift is observed for the hydride resonance of Ca54In13B4–xH23+x, indicating that these hydrides are not interacting with conduction electrons. The presence of conduction electrons also

typically facilitates fast relaxation of the nucleus; metallic hydrides such as LaNi5H6.8 and PdHx 116 have T1 values below 500 ms. The 5 s T1 for this Ca54In13B4–xH23+x hydride resonance is 117 similar to those reported for insulating ionic hydrides such as NaH and NaMgH3 (T1 = 5–50 s). Other features in the 1H spectrum include a small narrow peak at 4.4 ppm, and a very small broad peak at −6.2 ppm. The resonance at 4.4 ppm was also very narrow under slower spinning conditions (Figure 5.7), and it is likely due to H2 gas. Metal hydrides may be in equilibrium with small amounts of hydrogen gas, even at room temperature; sharp peaks at 4.3 – 118 4.5 ppm have been reported in studies of other phases including AlH3 and MgH2. The small −6.2 ppm peak may be due to an impurity; it was not observed in 1H spectra collected of other samples of this compound.

1 Figure 5.6 H MAS NMR spectra collected on Ca54In13B4–xH23+x using a 1.3 mm rotor spinning at 60 kHz, referenced to TMS at 0 ppm.

Indium solid-state NMR is complicated by the very large quadrupolar moment of the 9 indium nucleus (I = /2); the interaction of this moment with the electric field gradient of the surrounding environment leads to broadening and shifting of the central transition resonance.

This is quantified by the quadrupolar coupling constant CQ and the asymmetry parameter η. Of the two indium sites in Ca54In13B4–xH23+x, the In(1) atom in the low multiplicity 2a site has a

highly symmetric coordination environment, in the center of the In@Ca12@H20@Ca30 spherical cluster. The high symmetry should minimize the electric field gradient at the In(1) nucleus. The In(2) atoms have a lower overall symmetry, but will dominate the 115In spectrum due to their much higher multiplicity. The 115In spectrum, shown in Figure 5.7, can be fitted to two sites: the

In(1) resonance at 1071 ppm, with a CQ of 1.2 ± 0.2 MHz and a very small asymmetry parameter 79 η = 0.1, and the In(2) resonance at 1207 ppm, with a CQ of 6.5 ± 0.2 MHz and η = 0.88. It is notable that both indium resonances exhibit a significant paramagnetic shift to values far higher than the negative shifts or small positive shifts seen for indium nuclei in ionic species (such as – 115 InCl4 ions in solution) or coordination compounds such as In(acac)3, which typically have In 119 chemical shifts of less than 200 ppm with respect to the In(NO3)3(aq). Semiconducting indium 120 phases such as Ba2In5P5 and InAs have resonances in the 300–800 ppm range. Heavy doping into the metallic regime can shift the resonances of indium III–V phases above 900 ppm. The conduction electrons in metallic indium-containing alloys are polarized by an applied magnetic field, causing a large Knight shift of the 115In resonance to the 3000–9000 ppm range, as observed for intermetallics such as Ni2In3, as well as indium metal itself, which has a resonance 72 at 8000 ppm. The indium sites in Ca54In13B4–xH23+x have shifts in the metallic regime, indicating that these nuclei are interacting with conduction electrons at the Fermi level.

5.3.3 Electronic Structure Calculations To further investigate this, density of states (DOS) calculations were carried out on ordered model compounds of the title phase — one with the B/H mixed site fully occupied by boron, and one with it fully occupied by hydrogen. Since this site is richer in hydrogen than boron, the calculation on the Ca54In13H27 model compound is more reflective of the real phase; the resulting DOS data is shown in Figure 5.8. A pseudogap is found just below the Fermi level, separating the states above EF (dominated by empty calcium bands) from the filled states below

(dominated by indium and hydrogen bands). The small but nonzero DOS at EF indicates that this compound is metallic, which is supported by its silver coloration and luster. The presence of a 121 pseudogap at EF is typical of polar intermetallic compounds. Incorporation of a small amount of boron on the 8c site may stabilize the phase by modifying the valence electron count (VEC) to position the Fermi level in the pseudogap. The DOS diagram of the completely boron substituted

Ca54In13B4H23 is shown in Figure 5.9. The boron states appear at EF, eliminating the pseudogap.

The Ca54In13B4–xH23+x system may incorporate enough boron on the 8c site to position the Fermi level in the pseudogap, but not eliminate it.

115 Figure 5.7 In NMR spectrum for Ca54In13B4–xH23+x, collected using the stepped frequency technique (top), compared to calculated total spectrum and individual contributions from In(1) and In(2) sites.

350 200 Total 1) - 1) - 300 In

cell 150 cell

250 Ca -1 -1 200 H 100 150 100 50

50 eV (states DOS DOS (states eV (states DOS 0 0 -8 -6 -4 -2 0 2 4 6 8 -8 -6 -4 -2 0 2 4 6 8 Energy (ev) Energy (eV) Figure 5.8 Density of states data with the B(3)/H(3) site fully occupied by hydrogen (left) and just the hydrogen density of states.(right)..

250 Total 70 In B 60 1)

Ca - H 1) 200 - H 50 cell cell

B -1 -1 150 40 30 100 20 50 10

DOS (states eV (states DOS 0 DOS (states eV (states DOS 0 -8 -6 -4 -2 0 2 4 6 8 -8 -6 -4 -2 0 2 4 6 8 Energy (eV) Energy (ev) Figure 5.9 Density of states data with the B(3)/H(3) site fully occupied by boron (left), and the hydrogen and boron density of states (right)..

The Ca54In13H27 bands at the Fermi level are derived from indium and calcium orbitals, with no contribution from hydride states. The hydride states are found in narrow bands localized well below EF in the −4.5 to −7.5 eV range, in agreement with their anionic nature. This is similar to hydrides states in the charge balanced LiCa3As2H density of states (Fig 4.3). This supports the hypothesis that the conduction electrons are interacting with the indium sites and causing the Knight shift observed in the 115In NMR data, while the hydride sites in the structure are largely ionic, with their 1H nuclei exhibiting chemical shifts and relaxation times typical of ionic insulators.

The NMR spectra and DOS calculations indicate that Ca54In13B4–xH23+x is composed of ionic calcium hydride regions and metallic Ca/In regions. This is similar to the structural and electronic behavior observed by Simon et al. in many suboxides and subnitrides. Suboxide phases such as Cs7O can be described as ionic Cs11O3 clusters embedded in a metallic matrix of 71 excess Cs atoms (Cs11O3 + Cs10 = Cs7O). Likewise, the subnitrides Na16Ba6N and Na5Ba3N can 70 be viewed as ionic clusters of Ba6N in a metallic sodium matrix. Photoelectron spectroscopy studies and band structure calculations on these materials show anionic nitride states localized 122 well below EF. These subvalent phases behave as “void metals”; the conduction electrons are repelled by the ionic clusters and are confined to the spaces between them. This confinement raises the energy of the conduction electrons and is demonstrated by the lower than expected work function observed for suboxides.68,69

5.4 Conclusions

Analogous to the suboxides and subnitrides, Ca53In13B4–xH23+x can be viewed as a main group metal “subhydride”, a metallic compound in which the conduction electrons avoid the hydridic regions in the structure. It is notable that no main group subhydride has been reported until now, although recent computational work predicts the stability of several lithium 123 subhydrides (LimH, 4

Further AE/Li (AE = Ca, Sr, Ba) flux reactions of AEH2 with other group 13 metals are being explored as a potential source of new subhydrides; reactions of aluminum have already shown promising results. In addition to having complex structures, unique electronic properties, and potential use as hydrogen storage materials, the subhydrides are particularly amenable to solid-state NMR studies. While the known alkali metal suboxides do not contain suitable nuclei, 23 15 it would be of interest to collect Na and N MAS spectra on subnitrides such as Na5Ba3N; the sodium resonances would be expected to show Knight shifts, while the 15N resonances should not.

CHAPTER SIX

ALKALINE EARTH INDIUM ALLENYLIDES SYNTHESIZED IN AE/Li FLUX (AE = Ca, Ba)

6.1 Introduction Reactions of electropositive elements from groups I and II with main group p-block metals and metalloids often result in electron transfer and the formation of charge-balanced Zintl phases. Depending on the reactant ratio, the p-block elements accept electrons and/or form M-M bonds to fill their valence shells. For instance, tin can exist in Zintl phases as isolated Sn4- anions 4- 4- (in Ca2Sn), or clusters such as Sn4 (in Na4Sn4) and Sn9 (K4Sn9). However, a “Zintl border” exists between groups 13 and 14; group 13 metals are typically not electronegative enough to form Zintl anions. Elements such as indium therefore exhibit a rich and unpredictable chemistry. The Corbett group explored this extensively, discovering binary and ternary polar-but-not- charge-balanced intermetallic phases of indium with alkali and alkaline earth metals which 7- feature unusual clusters such as the In11 cluster in K8In11; the linked In12, In15, and In16 species 125,126,127 in K39In80; and the In28 species in K34In92.3Li12.7, K14Na20In91.82Li13.8, and K14Na20In96. In the process of exploring reactions of alkaline earth metals with group 13 and 14 elements, the Corbett group also investigated inadvertent hydride incorporation. Closer examination of reported phases such as Ba5Ga6and Ba21Ge2O5indicated that they were actually hydrides (Ba5Ga6H2 and Ba21Ge2O5H24, respectively); their formation was facilitated by the use of hydrogen-contaminated barium.128,129 Alkaline earth metals react with trace water vapor to form hydrogen; this diffuses readily into the bulk metal to form a solid solution. As a result, commercially available heavier alkaline earths (Ca, Sr, Ba) may contain 5 – 20 atomic % hydrogen, with barium being the most susceptible to a high level of contamination. We are deliberately using the high solubility of hydrogen in alkaline earth metal-rich fluxes to synthesize new complex hydrides of group 13 and group 14 metals. Metal flux synthesis involves the use of an excess of low melting metal as a reaction medium. Upon melting, this metal acts as a solvent for other reactants present, bringing them into solution to react with each other (and potentially with the flux metal itself) to form products.1 Alkaline

earth metals are higher melting than the more commonly used metal fluxes (including Sn, In, Ga, Al); however, the melting point is lowered upon addition of lithium. A 50:50% mixture of Ca/Li melts at around 300°C, as does a similar Ba/Li mixture.130 Main group elements are highly

soluble in alkaline earth/lithium mixtures. (Ca/Li)/M/CaH2 reactions with M = group 14 or group

15 elements have produced charge-balanced Zintl phase hydrides such as LiCa2C3H, 38,24,131 LiCa7Ge3H3, and LiCa3As2H. Our initial investigations into similar reactions with group

13 metals led to a very unusual “subhydride” compound, Ca54In13B4H23, which contains ionic hydride regions and metallic Ca/In regions.132 Further reactions of indium with other elements in

AE/Li fluxes has led to two new phases, Ca12InC13-xandBa12InC18H4. Both of these phases contain indium coordinated by 12 alkaline earth cations; these clusters are packed in a bcc array - 4- 4- and linked by either H or C anions. The remaining void spaces are filled by C3 allenylide anions, which were observable by Raman spectroscopy and by mass spectroscopic studies of hydrolysis products. Density of states and COHP calculations confirm the metallic nature of these phases.

6.2 Experimental Procedure 6.2.1 Synthesis

Ba12InC18H4 was synthesized from reactions of In, LiH, and C in Ba/Li flux. Chunks of lithium (99.8% Strem), barium (99%, Alfa Aesar), acetylene carbon black powder (99.5% Alfa Aesar), indium powder (99.9%, Alfa Aesar), and lithium hydride powder (97%, Alfa Aesar) were used as received. Reactants and flux metals were added to stainless steel crucibles (7.0 cm length/0.7 cm diameter) in a 8 : 8 : 0.8 : 0.8 : 0.8 : 0.8 mmol Ba/Li/In/C/LiH ratio in an argon- filled glovebox. The crucibles were sealed by arc-welding under argon and were placed in silica tubes which were flame-sealed under vacuum. The ampoules were heated from room temperature to 1050 °C in 3 hours, and held there for 2 hours. The reactions were cooled stepwise to 800 °C in 72 hours, to 500 °C in 36 hours, and then held at 500 °C for 24 hours. The reactions were then were removed from the furnace, inverted, and centrifuged for 2 minutes to separate the crystalline products from the Ca/Li melt. The solid product adheres to the side of the crucible. The steel crucibles were cut open in an argon-filled glovebox.

Ca12InC13-xwas initially found as a product of reactions of indium and carbon in Ca/Li flux, which used the calcium metal as received (99.5%, Alfa Aesar), without subsequent

purification to remove hydride contamination. The reactants Ca/Li/In/C were combined in a 7:7:1:2 mmol ratio and placed in a stainless steel crucible. This was arc-welded shut under argon and then placed into a silica tube which was flame-sealed under vacuum. The ampoule was heatedusing the same heating profile as the Ba12InC18H4 synthesis. After crystallographic studies (see below), further syntheses were carried out to determine the likely occupancy of a possible

mixed C/H site. Syntheses using added CaH2 were explored, with reactions such as

Ca/Li/In/C/CaH2 producing Ca54In13H27 (a boron-free variant of Ca54In13B4-xH23+x in Im-3, a = 16.28Å)14 instead of the desired carbide phase. Subsequent syntheses were carried out without any hydride added, and using calcium that was purified by heating to 700 °C under high vacuum to remove any hydride contaminants. These Ca/Li/In/C reactions yielded the desired phase, supporting its identification as a carbide (Ca12InC13-x) and not a carbide hydride (Ca12InC9+xH4-x). The optimized Ca/Li/In/C reactant ratio of 7:7:1:3 was combined and heated as described above.

6.2.2 Elemental Analysis Elemental analyses were performed using a JEOL 5900 scanning electron microscope with energy dispersive spectroscopy (SEM-EDS) capabilities. Samples of product crystals were affixed to an aluminum SEM stub using carbon tape and analyzed using a 30 kV accelerating voltage. The EDS detector is not sensitive to the presence of light elements such as carbon and hydrogen, so only the relative ratios of alkaline earth and indium were observed. The observed

Ba:In atomic ratios averaged 90:10 for Ba12InC9H4. No incorporation of metals from the steel crucible (Fe, Ni, Mo) was detected in any of the samples.

6.2.3 Crystallographic Studies

Samples of Ca12InC9H4and Ba12InC18H4 were brought out of the glovebox under Paratone oil and examined under a microscope to select crystals for diffraction studies. Spheroid crystals of suitable size were mounted in cryoloops. Single-crystal X-ray diffraction data were collected at 200 K under a stream of nitrogen using a Bruker APEX 2 CCD diffractometer with a Mo Kα radiation source. Absorption corrections were applied to the datasets using the SADABS program.27 Refinements of the structures were performed using the SHELXTL package.42The structure solutions for both compounds were complicated by possible lower symmetry cells (with the software recommending a rhombohedral setting for the calcium phase and a monoclinic

setting for the barium phase), but the apparent cubic metrics of the unit cell parameters enabled

the selection of the proper cell symmetry. The structure of Ba12InC18H4 was initially solved in cubic space group I23, but use of the AddSym program in the PLATON software suite indicated the presence of additional symmetry elements and converted the structure to space group Im-3

(No. 204). It was subsequently found that the Ca12InC9H4 structure could also be solved in this space group. For both structures,alkaline earth and indium sites were determined using direct methods. Light element positions were located using difference Fourier calculations. Refinement of the allenylidecarbon sites was straightforward, with occupancies refining at/near 100% and bond lengths confirming their identity as carbon atoms. The 8c sites presented the most difficulty in both structures, being occupied by a light atom octahedrally coordinated by surrounding heavy alkaline earth cations. These sites were initially assigned as carbon, with partial occupancy indicated in both phases; if assigned as hydrogen, both refinements indicated greater than 100% occupancy. This could be due either to the fact that this site is occupied by a hydride anion (with 2 electrons instead of 1), or by a mixture of H- and C4-, or by partially occupied C4-. The barium phase refinement proved stable with this site assigned as a fully occupied hydride, leading to the stoichiometry Ba12InC18H4. For the calcium compound, this site was refined as a partially occupied carbon atom after hydride-free syntheses were carried out which successfully yielded the target product. Several data sets were collected, with the carbon occupancy on this site averaging 70% (for a stoichiometry of Ca12InC11.8).Crystallographic data and collection parameters are shown in Table 6.1, and atomic positions are listed in Table 6.2.

6.2.4 Band Structure Calculations Density of states (DOS)and crystal orbital Hamilton population (COHP) calculations for both title compounds were carried out using the Stuttgart TB-LMTO-ASA software package, based on the unit cell dimensions and atomic coordinates derived from single crystal diffraction 18 data. For Ca12InC13-x, an ordered model compound with the carbide 8c site fully occupied was used (Ca12InC13). Empty spheres were added by the program where appropriate to fill the unit cell volume. A 12×12×12 k-point mesh was used and integrated using the tetrahedron method.

The basis sets for Ba12InC18H4 consisted of In 5s/5p, Ba 6s/5d/4f, C 2s/2p, and H 1s orbitals. The

In 5d/4f, Ba 6p, C 3d, and H 2s/2p orbitals were downfolded. For Ca12InC13, the basis sets consisted of In 5s/5p, Ca 4s/3d, and C 2s/2p orbitals.

Table 6.1 Crystallographic data collection parameters for the title phases.

Ca12InC13-x Ba12InC18H4 Formula weight 707.90 1983.11 Crystal System Cubic Cubic Space group Im-3 (#204) Im-3(#204) a (Å) 9.6055(8) 11.1447(7) Z 2 2 Volume (Å3) 886.3(1) 1384.2(1) Density, calc (g/cm3) 2.653 4.758 Index ranges -12 ≤ h ≤ 12, -12 ≤ k ≤ 12, -14 ≤ h ≤ 14, -14 ≤ k ≤ 14, -12 ≤ l ≤ 12 -14 ≤ l ≤ 14 Reflections collected 4634 7912 Unique data/parameters 223 / 19 331/21 μ (mm-1) 4.79 17.57 R1/wR2a 0.0198 / 00381 0.0178 / 0.0447 R1/wR2 (all data) 0.0259 / 0.0396 0.0178 / 0.0447 Residual peak/hole (e- A- 0.37 / -0.67 0.62 / -1.86 3)

a 2 2 2 2 2 1/2 R1 = Σ||Fo|-|Fc||/Σ|Fo|; wR2 = [Σ[w(Fo - Fc ) ]/Σ[w(Fo ) ]]

6.2.5 Raman Spectroscopy

Crystals of Ca12InC13-xand Ba12InC18H4 were sandwiched between quartz slides which were sealed together with TorrSeal epoxy under argon. The Raman measurements were carried out using a JY Horiba LabRam HR800 system excited by a HeNe laser emitting at 633 nm. The spectrograph uses an edge filter to couple the laser beam into the microscope (Olympus BX30) by total reflection. The beam is focused on the sample through a microscope objective 50x IR (Leica N.A. 0.80). Scattered radiation is collected by the objective and the laser radiation is filtered out by the edge filter with Raman scattering coupled into the spectrograph CCD through a confocal hole. Spectra were collected under ambient condition over the spectral range of 100 to 3200 cm-1.

Table 6.2 Atomic coordinates and isotropic thermal parameters for the title phases. Atom Wyckoff Site x y z Ueqa

Ba12InC18H4 In(1) 2a 0 0 0 0.0209(3) Ba(1) 24g 0.30273(3) 0.18009(3) 0 0.0168(2) C(1) 12e 0.1517(10) ½ 0 0.040(2) C(2) 24g 0.1606(6) 0.3815(7) 0 0.030(1) H(1) 8c ¼ ¼ ¼ 0.03(3)

Ca12InC13-x In(1) 2a 0 0 0 0.0090(2) Ca(1) 24g 0.30553(6) 0.18892(6) 0 0.0159(2) C(1) 6b ½ 0 0 0.012(1) C(2) 12e ½ 0 0.1357(4) 0.026(1) C(3)* 8c ¼ ¼ ¼ 0.015(2) a Ueq is defined as one-third of the trace of the orthogonalized Uij tensor.

6.2.6 Hydrolysis Studies

Ca12InC13-xand Ba12InC18H4 samples were reacted with water to explore the hydrolysis of the carbide anions in their structures. Crystals were placed in a 200 mL Schlenk flask sealed with a rubber septum under argon. The reaction with water occurs instantly at room temperature, forming gaseous products upon addition of 5μL of water. Aliquots of the product gases were taken by syringe and analyzed by injecting them into a HP 6890 series GC system coupled to a HP 5973 mass selective detector.

6.3 Results and Discussion 6.3.1 Synthesis

Ca12InC13-x crystallizes as small silver cubes up to 1 mm on a side. The yield is 70% based on carbon; byproducts include CaC2 and an unidentified reddish phase in small quantities.

TheBa12InC18H4 phase forms as bronze non-faceted chunks of about 0.5 – 1 mm in size.

Reactions averaged about a 30% yield based on carbon. BaC2 and Ba3FeN3 are commonly seen as byproducts. The latter phase is probably due to nitrogen impurities in the lithium metal and iron that is leached from the crucible. Carbon is leached from the crucible as well to form the title phase; it was initially isolated in small amounts from a Ba/Li/In/LiH reaction without added carbon. The formation of BaC2 makes it difficult to improve yields as adding more carbon pushes

the reaction to favor BaC2. Both Ca12InC13-x and Ba12InC18H4 are highly air-sensitive, reacting readily with air or water. Mass spectrometry on the products of hydrolysis indicates the

formation of allene (C3H4) for both compounds (see Figure 6.1). Traces of acetylene were also

seen in the mass spectra of both phases due to hydrolysis of the AEC2 byproducts.

6.3.2 Structure Description Both title phases exhibit new crystal structures in space group Im-3 based on a body

centered cubic packing of indium-centered alkaline earth icosahedra, In@AE12, shown as red polyhedra in Figure 6.2. The indium atoms of both structures occupy a 2a site (in the corners and center of the unit cell) and are surrounded by twelve alkaline earth cations which occupy 24g sites. A similar BCC packing of In@AE12 clusters is found in Ca54In13B4-xH23+x, also grown from

Ca/Li flux; the In@Ca12 clusters in that structure have a somewhat shorter In-Ca distance

(3.3400(7)Å) compared to that in Ca12InC13-x(3.4505(7)Å).The binary phase Ca8In3 also features indium sites surrounded by calcium cations, exhibiting Ca-In distances in the 3.25 – 3.81Å range.133 Other calcium indide phases have more In-In bonding, but feature similar Ca-In bond 134,135 distances (ranging from 3.2 – 3.7 Å in Ca18Li5In25 and Ca2In). The In@Ba12 clusters in

Ba12InC18H4 can be compared to similar indium-centered clusters in Ba9In4H and 136,66 Ba21In2O5Hx. While bond length data for the distorted icosahedral In@Ba12 units in the latter compound are not available, the In@Ba10 units in Ba9In4H exhibit In-Ba distances ranging

from3.52-4.25Å. The observed In –Ba distance in the close-to-ideal In@Ba12 icosahedra of

Ba12InC18H4 is in the middle of this range at 3.9259(4)Å. It is also within the 3.53 – 4.19 Å range 137 seen for the 11-coordinate In sites found in Ba5Al4.1In1.9.

50000 Ca InC 40000 12 13-x

30000

20000 counts

10000

0 10 20 30 40 50 40000 Ba12InC18H4 30000

20000 counts

10000

0 10 20 30 40 50 250000

200000 blank

150000

100000 counts

50000

0 10 20 30 40 50 m/z

Figure 6.1 Mass spectra of products of hydrolysis reactions of Ca12InC13-x and Ba12InC18H4. Expected allene peaks are denoted by green shading; acetylene peaks indicated by yellow shading. The large peak at m/z = 43 in the Ba12InC18H4 spectrum indicates possible production of propene (or possibly traces of acetone).

Figure 6.2 The structures of Ca12InC13-x (left) and Ba12InC18H4(right).Alkaline earth atoms are blue, C atoms are black, C/H atoms are yellow, and In atoms are red; the local coordination environments of C4-/H- and indium atoms are highlighted as yellow and red polyhedra, respectively.

Eight of the triangular faces of the In@AE12 icosahedra are capped by light atom sites on

8c Wyckoff sites. This position is occupied by hydride ions in Ba12InC18H4 and partially occupied by carbide ions in Ca12InC13-x. These sites bridge the icosahedral clusters, with three alkaline earth cations from one cluster and three from a neighboring cluster forming an overall octahedral coordination around this site (shown as yellow polyhedra in Figure 6.2). In

Ba12InC18H4, the resulting H@Ba6 units have a Ba-H distance of 2.952Å, very similar to the

2.955 and 2.970Å distances observed for the same building blocks in Ba9In4H. It is also comparable to BaH2 which has Ba-H bond lengths from 2.60 – 3.00 Å. The presence of carbide anions on this site in Ca12InC13-x was confirmed in synthetic experiments (this compound forms in the absence of hydride sources) and in the crystallographic data refinement. Full occupancy of this site would lead to a stoichiometry of Ca12InC13, but partial occupancy was indicated, with 4- several data sets supporting a stoichiometry of Ca12InC13-x (x = 1.2). The octahedral C @Ca6 units have a Ca-C distance of 2.5289(3)Å. It is notable that no calcium methanide exists with 4- comparable monatomic C anions. While the methanide Be2C is the stable binary carbide for beryllium and a high pressure magnesium methanide Mg2C was recently reported, no methanides are known for the heavier alkaline earths; those MC2 metal carbides (M = Ca, Sr, Ba) contain

138 33 acetylide anions instead. , Ca12InC13-x is also one of a select few metal carbides featuring two 4- 4- different types of carbide anion; monatomic C and the triatomic allenylide anion C3 (vide infra) are also found in Ln4C7 (Ln = Ho, Er, Tm, Lu), Sc3C4, and R5Re2C7 (R = Er, Tm, Lu, Sc).139,21,140

Table 6.3 Bond lengths (Å) in title phases

Bond Ca12InC13-x Ba12InC18H4 In – AE 3.4505(7) 3.9257(4) C1 – C2 1.304(4) 1.325(7) C1 – AE 2.6043(6) 2.889(7), 3.423(5) C2 – AE 2.515(3), 2.9124(2) 2.747(7), 2.829(5), 3.120(5) 8c site - AE 2.5289(3) 2.9521(2)

The remainder of the space in the structures of Ca12InC13-x and Ba12InC18H4 is filled by 4- allenylide anions, C3 . Their identity as allenylides is confirmed by the fact that hydrolysis 4- reactions of both phases produce allene gas. In the solids, these C3 species are nearly linear, with C-C bond lengths which are clearly double bonds (1.304(4) Å and linear in the calcium compound, 1.325(7) Å with a 171° angle in the barium phase). Very similar bond lengths (in the

1.30 – 1.33 Å range) are seen in charge-balanced allenylide salts such as Mg2C3, Ca3C3Cl2,

LiCa2C3H, and Ca11Sn3C8, as well as metallic carbides including Ln4C7,Sc3C4 and 141,142,24,142,143,143 4- R5Re2C7. The packing of these C3 anions is different in Ca12InC13-x and

Ba12InC18H4, as shown in Figure 6.3. In both compounds, these linear anions are aligned parallel to the axes of the unit cell and positioned between the In@AE12 icosahedral clusters. The barium phase has larger voids between the In@Ba12 units, allowing for incorporation of twice as many 4- carbide anions. In the calcium compound, the C3 anions are surrounded by two face-sharing square antiprisms of calcium cations; one encapsulating each terminal carbon atom. The same 4- arrangement of barium atoms (two face-sharing square antiprisms) can encapsulate two C3 units, one enveloped in each square antiprism.

Figure 6.3 The coordination environment of the allenylide anions in Ca12InC13-x (left) and Ba12InC18H4(right). Alkaline earth atoms are blue, C atoms are black; atoms in the 8c site (C or H) are yellow.

The Raman spectra of both phases, shown in Figure 6.4, are both dominated by the stretching modes of the allenylide anions. The spectrum of Ca12InC13-x shows the symmetric and -1 -1 asymmetric stretching modes at νsym = 1100 cm and νasym = 1480 cm ; weaker overtone and

-1 combination bands were observed at higher wavenumbers (not shown; 2νsym≈ 2180 cm , νsym + -1 -1 νasym≈ 2550 cm ; 2νasym≈ 2940 cm ). This data is very similar to that reported for Ca3C3Cl2, 4- which also contains the C3 anion surrounded by calcium cations and had Raman peaks at νsym =

-1 -1 31 1159 cm and νasym = 1660 cm . The allenylide stretching modes are shifted to lower

-1 -1 wavenumbers for Ba12InC18H4 (νsym = 1020 cm and νasym = 1197, 1234 cm ); a bending mode

-1 4- is also seen (δ = 593 cm ). This is comparable to the C3 bending mode observed for Ca11Sn3C8

-1 2 (δ = 589 cm ). No corresponding peak is seen for Ca12InC13-x; the higher site symmetry of the 4- C3 anions in this structure may make this mode Raman inactive. The energies of the stretching modes of the allenylide anion are likely correlated to the electronegativity of the surrounding -1 cations. Raman spectra reported for Mg2C3 show a band at νsym = 1213 cm and a calculated but

-1 4- not observed νasym = 1700 cm . These modes shift to lower energy as the C3 anion is

surrounded by calcium (Ca12InC13-x) and further shift for the barium phase Ba12InC18H4 (νsym

67 changing from 1213 cm-1 to 1100 cm-1 to 1020 cm-1 respectively). This is similar to what is reported for acetylide anion stretches in the alkaline earth acetylides, which show theνsymmode -1 -1 -1 33 shifting from 1860 cm (CaC2) to 1852 cm (SrC2) to 1831 cm (BaC2).

Ba12InC18H4

Ca12InC13-x

400 600 800 1000 1200 1400 1600 -1 wavenumber (cm )

Figure 6.4 Raman spectra for Ca12InC13-x and Ba12InC18H4.

6.3.3 Electronic Structure

While both Ca12InC13-x and Ba12InC18H4 contain recognizable cations and anions which have commonly accepted formal charges, attempts to derive a charge-balanced state results in a 2+ 4+ 4- 4- 2+ 4+ 4- - required charge of +4 for indium ((Ca )12(In )(C3 )3(C )4 and (Ba )12(In )(C3 )6(H )4). Thus, the structures cannot possibly be stabilized by charge-balancing. They appear to be aiming instead for similar stabilizing valence electron counts (VEC); while the barium phase has twice 4- - as many allenylide anions, the substitution of C for H in Ca12InC13-x allows for both compounds to have identical total formal charges on the anions. The calculated density of states data for the two title phases are shown in Figure 6.5. Both phases have a non-zero DOS at the

Fermi level and are therefore metallic. The states near EFin Ba12InC18H4are derived from contributions from the carbide anions and their associated barium coordination sphere (which produce bands at and just below the Fermi level) and the indium orbitals (which produce bands

at and just above EF). The hydrogen-derived states are for the most part localized in a narrow band at around -1.5 eV below EF, indicating they are largely anionic.

100 Total In ) 80 -1 Ba

cell 60 C -1 H 40

DOS (eV 20

0 -6 -4 -2 0 2 4 6 Energy (eV)

40 35 )

-1 30 25 cell -1 20 15 10 DOS (eV 5 0 -6 -4 -2 0 2 4 6 Energy (eV)

Figure 6.5 The calculated electronic density of states for Ca12InC13 (bottom) and Ba12InC18H4 (top). The Fermi level is set at 0 eV.

The density of states for Ca12InC13was calculated using a model structure with full occupancy of all sites; the data show some similarities to that of the barium compound (including a non-zero DOS at EF, and a bandgap at around 1 eV above EF) but some significant differences.

There are fewer states at EF for the calcium phase (11 states vs 21 states for Ba12InC18H4), indicating that while the calcium compound is metallic, it is a poorer metal than the barium

phase. The states at the Fermi level are largely derived from carbon orbitals, with smaller contributions from calcium and indium. The Fermi level cuts through a sharp peak in the DOS which is right above a pseudogap. This might be the driving force for the vacancies on the

8ccarbide site; fewer electrons may position EF in the pseudogap, stabilizing the structure.

Similar stabilizing effects of vacancies have been postulated for LaZn0.67As2 and La21- 144,145 dMn8Te7C12. More specific contributions to the DOS near the Fermi level and corresponding COHP

data for Ca12InC13 are shown in Figure 6.6. The dominant states at EF are derived from indium p- orbitals and surrounding calcium states, and from the p-orbitals of the C4- anion (C3 carbon on 8c site) and associated calcium states. The COHP calculations indicate that the Ca-In and Ca-C3

interactions are strongly bonding at EF. States from the allenylide anion are predominantly located well below the Fermi level (showing well-stabilized C-C bonding interactions at -4 to -7 eV, and antibonding interactions above 2 eV), although the p-orbitals of the terminal carbon

atoms of this group do make a small (C-C non-bonding) contribution to states at EF.

Similar DOS and COHP data for Ba12InC18H4 is shown in Figure 6.7. As was the case with the calcium compound, indium p-orbitals contribute to a very narrow band at the Fermi level; these orbitals have a bonding interaction with surrounding barium ions. The other

dominant contribution at EF are from the p-orbitals of the terminal carbon atoms (C2) of the allenylide unit. These states are strongly bonding to surrounding barium ions, but are non- 3- bonding to the center carbon of the C4 anion. The COHP for the C-C interactions of these

anions in Ba12InC18H4 looks very similar to that of the allenylides in Ca12InC13: strong C-C

bonding interactions below -4 eV, and antibonding interactions well above EF. While there are a small number of hydride states at the Fermi level, they are largely non-bonding; most of the

hydride states are instead located 1.5 eV below EF. This hydride band is higher in energy than

typically seen for calcium-based hydrides such as Ca54In13B4-xH23+x, LiCa2C3H, and LiCa3As2H, all of which have their hydride states 4 eV below the Fermi level.38,24,132 This may reflect the - 2+ longer distance between the H anion and surrounding Ba cations in Ba12InC18H4, which would lead to a weaker interaction.

5 3 Ca s In s 4 Ca p In p 2 3 In d DOS DOS DOS DOS 2 1

1 0 0 -10 0 10 -10 0 10

8 C-C 0.8 6 Ca-C Ca-C1 2.604 Ca-C2 4 0.4 2.515 Ca-C2 2.912 COHP 2 COHP 0 0 -10 -5 0 5 10 -10 -5 0 5 10 -2 -0.4 Energy (ev) 6 Ca-In 4

COHP -10 -5 0 5 10 -2

-4

-6 Energy (eV) Figure 6.6 Partial DOS (left column) for specified atomic orbitals, and COHP data (right column) for interactions between specified atoms in Ca12InC13.Bond lengths are indicated to distinguish certain bonds.

20 C1 p 6 C-C 15 C2 p C1 s 4 10 C2 s DOS 2 COHP 5 0 -6 -4 -2 0 2 4 6 8 10 0 -6 -4 -2 0 2 4 6 8 10 -2 10 20 Ba-In In-s 15 In-p 5 In-d 10 DOS

COHP 0 5 -6 -4 -2 0 2 4 6 8 10 0 -5 -6 -4 -2 0 2 4 6 8 10 25 Ba-s Ba-C Ba-C1 2.78 20 Ba-C1 2.84 Ba-p 0.5 Ba-C2 2.89 15 Ba-d Ba-C 3.11 10

0 DOS

COHP -6 -4 -2 0 2 4 6 8 10 5 -0.5 0 -1 -6 -4 -2 0 2 4 6 8 10 Energy (ev) Energy (eV) Figure 6.7 Partial DOS (left column) for specified atomic orbitals, and COHP data (right column) for interactions between specified atoms in Ba12InC18H4. Bond lengths are indicated to distinguish certain bonds.

6.4 Conclusions Alkaline-earth rich melts are excellent solvents for carbon, allowing for the formation and crystal growth of new complex carbides. Flux reactions of carbon in Ca/Li and Ba/Li have 4- enabled the formation of many new phases containing the allenylide anion (C3 ), which has been rarely seen otherwise. Raman spectroscopy and DOS calculations indicate that these species are largely anionic although they are building blocks of metallic compounds. Solid state NMR spectroscopy will be useful in further exploring the nature of these carbide anions.

CHAPTER SEVEN

FUTURE WORK

7.1 Introduction Given the unusual subhydrides formed from the reaction of indium and CaH2 in the Ca/Li flux, similar reactions were carried out with the other group 13 metals. Reactions with aluminum and CaH2 in the Ca/Li flux have resulted in a number of novel phases which are introduced in this chapter. However, carbon and nitrogen impurities are incorporated in some of the products which has complicated the reaction. The yields for these reactions are very low, so for these phases have only been characterized by XRD. Further investigation will hopefully yield provide better and more consistent yields of these phases allowing additional characterization techniques. In particular, neutron diffraction should clearly identify the light elements (H,C,N). In this chapter, these phases will be described, but the identity of the light elements is based solely on X-ray diffraction data and bond length considerations, so must be treated as preliminary.

7.2 Synthesis Chunks of lithium (99.8% Strem), acetylene carbon black powder (99.5% Alfa Aesar), aluminum powder (??), and calcium hydride powder (98%, Alfa Aesar) were used as received. Calcium shot (99.5% Alfa Aesar) was purified by heating in a steel tube under 10-5 Torr vacuum at 600 °C for 3 hours. Heating was continued under 10-3 Torr for 12 hours. This process decomposes any calcium hydride and calcium hydroxide present and removes the resulting gaseous hydrogen and water. Reactants and flux metals were added to stainless steel crucibles

(7.0 cm length / 0.7 cm diameter) in a 7 : 7 : 0.7 : 0.7 mmol Ca/Li/Al/CaH2 ratio in an argon- filled glovebox. The crucibles were sealed by arc-welding under argon and were placed in silica tubes which were flame-sealed under vacuum. The ampoules were heated from room temperature to 1050 °C in 3 hours, and held there for 2 hours. The reactions were cooled stepwise to 800 °C over 72 hours, 500 °C over 36 hours. The reactions were held at 500 °C, then were removed from the furnace, inverted, and centrifuged for 2 minutes to separate the

crystalline products from the Ca/Li melt. The steel crucibles were cut open in an argon-filled glovebox. The solid product adheres to the side of the crucible.

7.3 Structure of Ca31H21Al2

The Ca31H21Al2 structure crystallizes in the cubic Fd-3m space group (a = 18.0835(15) Å). The Al site is in the center of a cage of 12 calcium atoms in a icosahedral arrangement (Fig 6.1) with bond lengths of 3.2001(3) and 3.4345(10) Å. This is similar to bond lengths seen in 146 147 CaAl2 (3.325 Å)146F and Ca8Al3 (3.185 – 3.554 Å).147F Each trigonal face of the icosahedron is capped by hydride sites which form a H20 pentagonal dodecahedron. Since the Al- H distance (4.04 – 4.1 Å) is similar to the Ca-Al bond length, the calcium atoms are found in the faces of the of H20 cage.

Figure 7.1 The Al@Ca12@H20 cluster in Ca31H21Al2. The Al atom is red, Ca atoms are blue, and H atoms are yellow.

The structure can be viewed as a framework of the H20 pentagonal dodecahedra and H28 hexakaidodecahedra which form a clathrate II structure shown in Figure 7.2. Clathrates are composed of a polyhedral framework incorporate guest molecules. The type II structure is most commonly a framework of hydrogen-bonded water molecules in encapsulating gas molecules 148 such as methane or H2. Analogous inorganic compounds for the type II structure are rare and

149 150 include guest-free structures such as Si1361 and Ge136 , and encapsulated single atoms such 151 152 153 154 as Rb8Na16Si136 and Ba16Ga32Sn104. Ca21Ni2Zn36 and Mg35Cu24Ga53 have encapsulated icosahedra with icosahedra of Ni@Zn12 and Cu6Ge6 respectively centered in the 20 atom polyhedra. A similar cluster is also seen in In@Ca12@H20 cluster observed in Ca53In13B4-xH23+x in chapter 4. However, for these compounds, the bond lengths of the icosahedra are much smaller, and the vertex atoms do not lie in the plane of the faces of the pentagonal dodecahedra.

Figure 7.2 The clathrate II structure formed from 3 hydride sites.

The H28 cage is filled with a 16 member calcium cluster that can be seen as a hexagonal antiprism with three capping atoms on one hexagonal face and one capping atom on the other.

As in the smaller polyhedral, the calcium atoms lie in the faces of the encapsulated H28 cage. This cluster encapsulates a calcium ion on the 8b site is surrounded by 4 hydride sites in a tetrahedral configuration. These hydride positions tend to move around during refinement steps and have a large thermal parameter. The best refinement gives a Ca-H bond length of 2.25(12) Å 155 which is comparable to the bond length in CaH2 of 2.26 – 2.65 Å. The overall result can be

viewed as concentric clusters of average stoichiometry CaH4@Ca16@H28.

b a

Figure 7.3 a) The CaH4 unit in the Ca16 cage b) The Ca16 cage in the H20 hexakaidecahedron.

This calcium site originally positioned on the 16d special position in the hexagonal face of the H28 hexakaidecadehedron was found to have unusual thermal parameters and a large peak in the residual electron density is consistently seen in several refinements of this structure. The bond length from the original site would be 2.816 Å which is longer than typical literature values. It appears that this calcium position is a 6-fold split site off of the special position which 1 th is /6 occupied as shown in Figure 7.4. This allows the calcium to position itself with a more suitable bond length of 2.44(2) Å to adjacent hydride sites.

Figure 7.4 The six-fold split Ca site in Ca31H21Al2. The thermal parameters show a ring of electron density which total to the equivalent of a Ca atom.

7.4 Structure of Ca4Al2N5

The Ca4Al2N5 phase forms in the Pna21space group (a=11.2331(10), b = 9.0768(8), c= 6.0093(5)). The structure features two aluminum sites, each tetrahedrally coordinated by four nitrogen atoms, with bond lengths of 1.7767(1) – 1.8923(1). The AlN4 tetrahedra alternate along a chain formed from a corner-shared nitrogen atom as shown in figure 6.5. The AlN4 tetrahedra around the Al2 site has all of its corners shared with the other AlN4 (Al1) group. The Al1 tetrahedra shares two corners and the other two are terminal.

a) b)

Figure 7.5. The AlN4 chains shown down the a) a-axis and b) c-axis. The Al1 site is red, the Al2 site is orange, and N atoms are green

Ternary nitrides are common for main group nitrides.156 For known main group nitrides with alkali earth metals, the nitride tetrahedra all form edge sharing chains. Phases with edge- sharing chains of alternating tetrahedra as seen in Ca4Al2N5 can be found in various tungsten 157 nitride phases: Na3MN3 (M=Mo,W), Na2K(WN3), Na11Rb[(WN3)4],and Na5A[(WN3)2] (A=Rb,Cs),158 but the chains are not interconnected.

The space between AlN4 chains is filled with calcium atoms that are coordinated to 5 or 6 nitrogen atoms in distorted octahedral or square pyramidal clusters respectively. The bond lengths around each calcium atom range from 2.3545(1) – 2.6398(2) Å which is reasonable compared to other calcium nitride phases. Ca11N8 have bond lengths of 2.308 – 2.503 Å and longer Ca-N bonds are seen in Be4Ca2N4 (2.689 Å) and BaCa2N12P6 (2.631 Å).

The assignment of nitrogen is based on the q-peak value since carbon, nitrogen, and hydrogen have similar bond distances to calcium. Although, nitrogen was not deliberately added to the reaction, it has been seen as a contaminant, most likely from the lithium, metal. Since this phase has only been seen once in various aluminum reactions in the flux, it is possible that it is due to contamination. However, it was not seen in reactions with both aluminum and nitrogen. Further studies, likely neutron XRD, will likely have to be done to determine the identity of the light element. The crystals have a red color which suggests a charge-balanced phase Zintl phase, but the given formula has an overall -1 charge. The substitution of O2- on the N3- sites can balance the charge.

Figure 7.6 The crystal structure of Ca4Al2N5. Ca atoms are light blue, Al atoms are red, and N atoms are green.

7.5 Structure of Ca24Al9(C1-xHx)N2H16

Ca24Al9(C1-xHx)N2H16 crystallizes in the tetragonal P2/nmc space group (a = 15.9069(13), c=13.7323(10)). The structure is composed of ABA layers of Ca/Al clusters arranged in a

checkerboard pattern. There are two different atoms on 16h sites which form Ca24Al4C and

Ca32Al4 clusters. These clusters are arranged diagonal to each other through shared edges in a

layer as shown in Figure 7.7. The second layer is a 180° rotation of the first so that the Ca24Al4C

and Ca32Al4 clusters alternate down the c-axis.

In the Ca24Al4C cluster is formed from four Al@Ca10C clusters arranged together in a square so that they are sharing an edge. This configuration is shown in Figure 7.7. The calcium atoms are in the Al@Ca10C arranged in a bicapped pentagonal prism around the central

aluminum atom with two of the vertices another Al@Ca10C cluster which each shares 4 calcium

atoms. The carbon atom is in an octahedral site in the center of the Ca24Al4C cluster on the 2a Wycoff site.

The Ca32Al4 cluster shown in figure 7.8 is composed of 4 Al@Ca10 clusters also in a

square arrangement where each shares two edges with another Al@Ca10 cluster. The calcium atoms are in an icosahedral arrangement around the aluminum atom with 2 calcium atoms missing. The missing vertex is a void space caused by steric hindrance between the calcium

atoms of a neighboring Al@Ca10 cluster.

Figure 7.7 The Ca24Al4C cluster (left) and the Ca32Al4 cluster (right) in Ca24Al9(C1-xHx)N2H16. The Al atom is light red, the Ca atom is blue, and the C atom is black. The Ca32Al4 cluster.

There is a channel along the c-axis in the space between the Ca24Al4C and Ca32Al4

clusters which can be described in two parts. The first is an Al2Ca16 cluster from two Al@Ca9 clusters in monocapped square antiprisms which share a trigonal face. The second is H/Ca

clusters which form two distinct hydrogen sites. The hydrogen on the 8g site is in a H2Ca7 cluster

from two HCa4 tetrahedra which bridges opposite Ca32Al4 clusters. The other H on the 16h site is

80 in a square pyramid of Ca atoms and shares a face with a Ca24Al4C cluster. There are four of these clusters on either side of the H2Ca7cluster. The Ca-H bond lengths range from 2.1698 – 2.6774 Å. The nitrogen site is in a N@Ca6 octahedron in an interstitial site between the Al/Ca clusters. The bond lengths range from 2.317 – 2.4137 Å. This is comparable to other calcium 159,83 nitrides such as Ca3N2 (2.457 – 2.479 Å) and Ca3NP (2.365 Å).

Table 7.1 Crystal Data and Structure refinement for the presented Al structures

Ca31Al2H25 Ca24Al9(C1-xHx)N2H16 Ca4Al2N5 Formula weight (g/mol) 1324.58 1260.86 507.73 Crystal System Cubic Tetragonal Orthorhombic

Space group Fd-3m P42/nmc Pna21 a (Å) 18.0835(15) 15.9069(12) 11.2331(1) b (Å) 9.0768(8) c (Å) 13.7323(10) 6.0093(5) Z 8 4 4 Volume (Å3) 5913.5(8) 3474.7(4) 612.71(9) Density (g/cm3, calc) 2.1334 2.2243 3.0823 Index ranges -24 ≤ h ≤ 23, -20 ≤ h ≤ 20, -14 ≤ h ≤ 14, -23 ≤ k ≤ 23, -20 ≤ k ≤ 21, -11 ≤ k ≤ 11, -23 ≤ l ≤ 23 -18 ≤ l ≤ 18 -8 ≤ l ≤ 7 Collection Temp (K) 200 200 200 Reflections collected 17055 38031 5634 Unique data/parameters 394/36 2287 / 108 1438/100 μ (mm-1) 1.764 1.743 3.744 R1/wR2a 0.0333/0.0841 0.0336/0.1100 0.0143/0.0403 R1/wR2 (all data) 0.0366/0.0856 0.0352/0.1111 0.0144/0.0403 Residual peak / hole (e- A-3) 0.54/-0.42 0.53/-0.42 0.32/-0.24 a 2 2 2 2 2 1/2 R1 = Σ||Fo|-|Fc||/Σ|Fo|; wR2 = [Σ[w(Fo - Fc ) ]/Σ[w(Fo ) ]]

Figrue 7.8 A layer of Ca24Al9(C1-xHx)N2H16 looking down the c-axis showing the checkerboard pattern. The Al2Ca16 cluster is red, the Ca24Al4C cluster is dark blue, the Ca32Al4 cluster is green, the H atoms are yellow and the Ca atoms are light blue.

Figure 7.9 The structure of Ca24Al9(C1-xHx)N2H16 looking down the [1,1,0] direction. Both layers are seen with the second a 180· rotation of the first layer. The Al2Ca16 cluster is red, the Ca24Al4C cluster is dark blue, the Ca32Al4 cluster is green, the H atoms are yellow and the Ca atoms are light blue.

CHAPTER EIGHT

CONCLUSIONS

Several complex intermetallic phases have been crystallized from the Ca/Li flux system. The presented research has focused on reactions with a combination of heavy and light p-block elements in the flux. As the research presented has focused mainly on the reaction of hydrogen and carbon, there is plenty of potential in using other light elements such as boron and nitrogen as seen in Ca54In13B4-xH23+x. However, contamination has been a problem which has lead to difficulties in determining the composition of the observed phases. While a bit of mixed blessing, as it has also led to the discovery of novel phases, in future work this should be addressed. Other possibilities will be to use transition metals or Sr/Li and Ba/Li fluxes on which there have been little exploration. There is also potential in using different reaction ratios, heating profiles, etc. in previously explored reactions. This work has focused on a shallow approach of using a 1 mmol ratio of reactions added to 10mmol of flux on a wide range of elements. The heating profile used has been adjusted to spend considerably more time at higher temperatures which was not done for earlier work with carbides. Also, the reactions were loaded so that flux was on top of the added reactants with the intention of the low melting lithium flowing onto the calcium to form the flux. A sandwich-like arrangement with flux below and above the added reactants may result in better mixing. 4- The formation of the rare allenylide C3 unit Ca11Sn3C8 and Ca2LiC3H is and warrants further work with carbides in this flux. It also possible that the allenylide may be used as a

reactant much like in CaC2. Attempts to react the phase with gaseous iodine and sulfur did not result in any observed products, but experiments in solution with reactants such as benzyl bromide have shown evidence of reactivity.

REFERENCES

1. Kanatzidis, M.G; Pottgen, R; Jeitschko, W. Angew. Chem. Int. Ed. 2005, 44, 6996-6702

2. Thompson, C.M. et al. Chem. Mater. 2014, 26, 3825-3837

3. Calta, N.P; Kanatzidis, M.G. Inorg. Chem.2013, 52, 9931 - 9940

4. Baumbach, R.E; Fisk,Z; Ronning,F; Movshovich, R; Thompson, J.D; Bauer, E.D. Phil. Mag.

5.Matthias, B.T; Geballe, T.H; Gellar, S; Corenwit, E. Phys. Rev. 1954, 95, 1435

6. Berlincourt, T.G; Hake, R.R. Bull. Am. Phys. Soc. II. 1962, 7, 408

7. Vinod, K; Kumar,R.G.A; Syamapradad,U. Supercond. Scit. Technol.2007, 20, R1-R13

8. Shoemaker, C.B; Schomaker, D.P. Acta. Cryst. 1984, C40, 1665-1668

9. Strnat, K; Hoffer, G; Ostertag,W; Olson, J.C. J Appl. Phys. 1966, 37, 1252-1253

10. Nassau, K; Cherry, L.V; Wallace, W.E. Physics and Chemistry of Solids, 1960, 21, 44 - 49

11. Masumoto, K; McGahan, W.A. MRS Bulletin. 1996, 21(5), 44 -49

12. Lehmann, H.W; Emmenegger, F.P. Solid State Comm. 1969, 40, 1017 - 1019

13. Geng, M; Han, J; Feng, F; Northwood, D.O. Int. J. Hydrog. Energy, 1998, 23, 1055

14. Stoloff, N.S; Liu, C.T; Deevi, S.C. Intermetallics. 2000, 8, 1313-1320

15. Sevov, S.C. Intermetallic Compounds, Vol. 3. Principles and Proactice. Ed. Westley, J.H, 2002

16. Nikolaus, K. Z. Anorg. Chem.2012, 638, 1057-1060

17. Fuller, J; Kaseman, D.C; Dolyniuk, J-A; Lee, K; Sabyasachi, S; Kovnir, K. Inorg. Chem.2013, 52, 7061 - 7067

18. Kim, S-J; Miller, G. J; Corbett, J.D. Z. Anorg. Allg. Chem. 2010, 636, 67-73

19. Birdwhistell, T.L.T; Klein, C.L; Jeffries, T; Stevens, E.D; O’Conner, C.J. J. Mater Chem. 1991, 1, 555-558

20. Hurng, W-M; Corbett,J.D; Wang, S-L; Jacobsen, R.A. Inorg. Chem. 1987, 26, 2392-2395

21. Chen, X; Shen, J-N; Wu,L-M; Chen, L. Inorg. Chem. 2013, 52, 7441-7447

22. Vaughey, J. T; Corbett, J. D. Inorg. Chem. 1997, 36, 4316-4320.

23. Massalski, T. B.; Okamoto, H. Binary Alloy Phase Diagrams, 2nd ed.; ASM International: Materials Park, OH, 1990.

24. Lang, D. A.; Latturner, S. E. Eur. J. Inorg. Chem. 2011, 26, 4006– 4011

25. Henning, R.W; Leon-Escamilla, E.A; Zhao, J-T; Corbett, J.D. Inorg. Chem. 1997, 36, 1282- 1285

26. Leon-Escamilla, E.A; Corbett, J.D. J. Alloys and Cmpds. 1998, 265, 104-114

27. SAINT, version 6.02a, Bruker AXS Inc., Madison WI, 2000.

28. Bennet, L.H; Watson, R.E; Carter, G.C; Nat. Bur. Stand. (U. S.), Spec. Publ.1971, 323, 601- 643

29. Jepsen, O.; Burkhardt, A; Andersen, O.K. The program TB-LMTO-ASA, version 4.7 Max- Planck-Institut fur Festkorperforschung: Stuttgart, Germany, 2000.

30. Blöch, P.E.; Jepsen, O.; Andersen, O.K. Phys. Rev. B, 1994, 49, 16223-16233.

31. O.K. Andersen, Phys. Rev. B 12, 3060 (1975); O.K. Andersen, Europhysics News 12, 4

32. Ferreira, S; Duarte, J; Frote-Pessoa, S. Phys. Rev B. 1990, 41, 5627-5635

33. Ruschewitz, U. Coord. Chem. Rev. 2003, 244, 115– 136

34. Simon, A. Angew. Chem., Int. Ed. 1997, 36, 1788– 1806

35. Benbow, E. M.; Dalal, N.; Latturner, S. E. J. Am. Chem. Soc. 2009, 131, 3349– 3354

36. Zaikina, J. V.; Zhou, H.; Latturner, S. E. J. Solid State Chem. 2010, 183, 2987– 2994

37. Emery, N.; Herold, C.; Lagrange, P. Prog. Solid State Chem. 2008, 36, 213– 222

38. Lang, D. A.; Zaikina, J. V.; Lovingood, D. D.; Gedris, T. E.; Latturner, S. E. J. Am. Chem. Soc. 2010, 132, 17532– 17530

39. Musanke, U. E.; Jeitschko, W. Z. Naturforsch. B 1991, 46, 1177

40. (a) Czekalla, R.; Jeitschko, W.; Hoffmann, R. D.; Rabeneck, H. Z. Naturforsch. B 1996, 51, 646– 654 (b) Pöttgen, R.; Jeitschko, W. Z. Naturforsch. B 1992, 47, 358– 364

41. Pöttgen, R.; Jeitschko, W. Inorg Chem. 1991, 30, 427– 431

42. Sheldrick, G. M. SHELXTL NT/2000, version 6.1; Bruker AXS, Inc.: Madison, WI, 2000.

43. Stein, S. E.Mass Spectra. In NIST Chemistry WebBook, NIST Standard Reference Database Number 69; Linstrom, P. J., Mallard, W. G., Eds.; National Institute of Standards and Technology: Gaithersburg, MD. http://webbook.nist.gov (accessed Nov. 2012).

44. Jeitschko, W.; Nowotny, H. N.; Benesovsky, F. Monatsh. Chem. 1963, 94, 672– 676

45. Haschke, H.; Nowotny, H. N.; Benesovsky, F. Monatsh. Chem. 1966, 97, 1045

46. Curda, J.; Carrillo-Cabrera, W.; Schmeding, A.; Peters, K.; Somer, M.; von Schnering, H. G. Z. Anorg. Allg. Chem. 1997, 623, 929– 936

47. Fjellvaag, H.; Karen, P. Inorg. Chem. 1992, 31, 3260– 3263

48. Hoffmann, R.; Meyer, H. J. Z. Anorg. Allg. Chem. 1992, 607, 57– 71

49. Meyer, H. J. Z. Anorg. Allg. Chem. 1991, 593, 185– 192

50. Reckeweg, O.; Meyer, H. J. Angew. Chem., Int. Ed. 1998, 37, 3407– 3410

51. Wörle, M.; Muhr, H. J.; Altenschildesche, H. M. Z.; Nesper, R. J. Alloys Compd. 1997, 260, 80– 87

52. Röhr, C. Z. Naturforsch. B 1995, 50, 802– 808

53. Hoch, C.; Wendorff, M.; Röhr, C. J. Alloys Compd. 2003, 361, 206– 221

54. Eckerlin, P.; Leicht, E.; Wolfel, E. Z. Anorg. Allg. Chem. 1961, 307, 145– 156

55. Bruzzone, G.; Merlo, F. J. J. Less-Common Met. 1976, 48, 103– 109

56. Reckeweg, O.; Schulz, A.; DiSalvo, F. J. Z. Naturforsch. 2010, 65b, 1409– 1415

57. Reckeweg, O.; Schulz, A.; DiSalvo, F. J. Z. Naturforsch. 2011, 66b, 1092– 1096

58. Buchenauer, C. J.; Cardona, M. Phys. Rev. B 1971, 3, 2504– 2507

59. Shimizu, H.; Oe, R.; Ohno, S.; Kume, T.; Sasaki, S.; Kishimoto, K.; Koyanagi, T.; Ohishi, Y. J. Appl. Phys. 2009, 105, 043522

60. (a) Yang, Z.; Shi, D.; Wen, B.; Melnik, R.; Yao, S.; Li, T. J. Solid State Chem. 2010, 183, 136– 143

(b) Migas, D. B.; Miglio, L.; Shaposhnikov, V. L.; Borisenko, V. E. Phys. Rev. B.2003, 67, 205203

61. Ruiz, E.; Alemany, P. J. Phys. Chem. 1995, 99, 3114– 3119

62. a) Ma, X.; Whalen, J.B.; Cao, H.; Latturner, S.E. Chem. Mater.2013, 25, 3363-3372. b) Ma, X.; Chen, B.; Latturner, S.E. Inorg. Chem., 2012, 51, 6089-6095.

63. Reynolds, P.C.; Stojanovic, M.; Latturner, S.E. J. Solid State Chem., 2011, 184, 1875-1881.

64. Ni, N.; Bud’ko, S. L.; Kreyssig, A.; Nandi, S.; Rustan, G. E.; Goldman, A. I.; Gupta, S.; Corbett, J. D.; Kracher, A.; Canfield, P. C. Phys. Rev. B,2008, 78, 014507.

65. Mathieu, J.; Latturner, S.E.Chem. Commun., 2009, 33, 4965-4967.

66. Huang, B. Q.; Corbett, J. D. Inorg. Chem. 1998, 37, 1892–1899.

67. Kauzlarich, S.M. Chemistry, structure, and bonding of Zintl phases and ions. VCH Publishers, Inc., New York, 1996.

68. a) Huang, B.; Corbett, J.D. Inorg. Chem.1997, 36, 3730-3734. B) Orgaz, E.; Aburto, A. Int. J. Quantum Chem. 2005, 101, 783-792

69. Häussermann, U. Z. Kristallogr.2008, 223, 628-635.

70. Schultz, A. J.; Jorgensen, M. R. V.; Wang, X.; Mikkelson, R. L.; Mikkelson, D. J.; Lynch, V. E.; Peterson, P. F.; Green, M. L.; Hoffmann, C. M. J. Appl. Cryst. 2014, 47, 915.

71. Zikovsky, J.; Peterson, P. F.; Wang, X. P.; Frost, M.; Hoffmann, C. J. Appl. Cryst. 2011, 44, 418.

72. Schultz, A. J.; Jorgensen, M. R. V.; Wang, X.; Mikkelson, R. L.; Mikkelson, D. J.; Lynch, V. E.; Peterson, P. F.; Green, M. L.; Hoffmann, C. M. J. Appl. Cryst. 2014, 47, 915.

73. Schultz, A. J.; Srinivasan, K.; Teller, R. G.; Williams, J. M.; Lukehart, C. M. J. Am. Chem. Soc.1984, 106, 999.

74. Larson, A. C.; Von Dreele, R. B. General Structure Analysis System (GSAS), Los Alamos National Laboratory Report LAUR 86-748, 2004.

75. Sheldrick, G.M. Acta Cryst.A. 2008, 64, 112.

76. Farrugia, L. J. Appl. Cryst. 2012, 45, 849.

77. Feng, X.J.; Prots, Y.; Schmidt, M.P.; Hoffmann, S.; Veremchuk, I.; Schnelle, W.; Burkhardt, U.; Zhao, J.T.; Grin, Y. Inorg. Chem. 2013, 52, 8971-8978.

78. Von Benda, K.; Juza, R. Z. Anorg. Allg. Chem. 1969, 371, 172-192. Li3As

79. Bruzzone, G.; Franceschi, E.; Merlo, F. J. Less-Common Metals. 1978, 60(1), 59-63.

80. Huetz, A.; Nagorsen, G. Z. Metallkunde. 1974, 65, 618-618.

81. Ni, N.; Allred, J.M.; Chan, B. C.; Cava, R.J. PNAS. 2011, 108, E1019 - E1026.

82. Rehr, A.; Kuromoto, T.Y.; Kauzlarichm S.M.; del Castillo, J.; Webb, D.J. Chem. Mater.1994, 6, 93-99.

83. Chern, M.Y.; DiSalvo, F.J.; Parise, J.B.; Goldstone, J.A. J. Sol. State Chem. 1992, 96, 426 – 435

84. Reckeweg, O.; DiSalvo, F.J. Z. Anorg. Allg. Chem. 2001, 627, 371-377.

85. Liddell, K; Thompson, D.P; Teat, S.J. J. Europ. Ceramic Soc. 2005, 25, 49-54.

86. Knoth, M.; Rossler, U.; Eisenmann, B. Z. Anorg. Allg. Chem.2005, 631, 1237-1240

87. Sakintuna, B.; Lamari-Darkrim, F.; Hirscher, M. Int. J. Hydrogen Ener. 2007, 32, 1121– 1140

88. Huang, B.; Corbett, J. D. Inorg. Chem. 1997, 36, 3730– 3734

89. Zuttel, A.; Rentsch, S.; Fischer, P.; Wenger, P.; Sudan, P.; Mauron, P.; Emmenegger, C. J. Alloys Compds 2003, 356, 515– 520

90. Evans, M. J.; Holland, G. P.; Garcia-Garcia, F. J.; Häussermann, U.J. Am. Chem. Soc. 2008, 130, 12139

91. (a) Kanatzidis, M. G.; Pöttgen, R.; Jeitschko, W. Angew. Chem., Int. Ed. 2005, 44, 6996– 7023 (b) Bugaris, D. E.; zur Loye, H. C. Angew. Chem., Int. Ed. 2012, 51, 3780– 3811

92. Simon, A. Coord. Chem. Rev. 1997, 163, 253– 270

93. Simon, A. Philos. Trans. R. Soc. A.2010, 368, 1285– 1299

94. Snyder, G. J.; Simon, A. Angew. Chem., Int. Ed. 1994, 33, 689– 691

95. Simon, A.; Westerbeck, E. Z. Anorg. Allg. Chem. 1977, 428, 187– 198

96. Carter, G. C.; Bennett, L. H.; Kahan, D. J. Metallic Shifts in NMR; Progress in Materials Science; Pergamon-Elsevier Science: Oxford, U.K., 1977; Vol. 20.

97. SADABS; Bruker AXS Inc.: Madison, WI, 2008.

98. Sheldrick, G. M. Acta Crystallogr., Sect. A. 2008, 64, 112

99. Bailey, M. S.; Obrovac, M. N.; Baillet, E.; Reynolds, T. K.; Zax, D. B.; DiSalvo, F. J. Inorg. Chem. 2003, 42, 5572– 5578

100. (a) Carr, H. Y.; Purcell, E. M. Phys. Rev. 1954, 94, 630– 638 (b) Meiboom, S.; Gill, D. Rev. Sci. Instrum. 1958, 29, 688– 691

101. Huang, I.; Gan, Z. J. Magn. Reson. 2010, 204, 256– 265

102. (a) Kennedy, M. A.; Vold, R. L.; Vold, R. R. J. Magn. Reson. 1991, 92, 320– 331 (b) Larsen, F. H.; Jakobsen, H. J.; Ellis, P. D.; Nielsen, N. C. J. Phys. Chem. A.1997, 101, 8597– 8606 (c) Nagayama, K.; Bachmann, P.; Wüthrich, K.; Ernst, R. R. J. Magn. Reson. 1978, 31, 133– 148 (d) Blümich, B.; Ziessow, D. J. Magn. Reson. 1982, 49, 151– 154

103. Massiot, D.; Fayon, F.; Capron, M.; King, I.; Le Calvé, S.; Alonso, B.; Durand, J. O.; Bujoli, B.; Gan, Z.; Hoatson, G. Magn. Reson. Chem. 2002, 40, 70– 76

104. (a) Jepsen, O.; Burkhardt, A.; Andersen, O. K. The Program TB-LMTO-ASA, version 4.7; Max-Planck-Institut fur Festkorperforschung: Stuttgart, Germany, 2000. (b) Blochl, P. E.; Jepsen, O.; Andersen, O. K. Phys. Rev. B.1994, 49, 16223– 16233

105. Anderson, O. K.; Jepsen, O. Phys. Rev. Lett. 1984, 53, 2571– 2574

106. Mao, J. G.; Goodey, J.; Guloy, A. M. Inorg. Chem. 2004, 43, 282– 289

107. Wendorff, M.; Röhr, C. Z. Naturforsch. B.2004, 59b, 619– 628

108. Lin, Q.; Smetana, V.; Miller, G. J.; Corbett, J. D. Inorg. Chem. 2012, 51, 8882– 8889

109. Villars, P.; Calvert, L. D. Pearson’s Handbook—Crystallographic Data for Intermetallic Phases; ASM International: Materials Park, OH, 1998.

110. (a) Rohrer, F. E.; Nesper, R. J. Solid State Chem. 1998, 135, 194– 200 (b) Reckeweg, O.; Meyer, H. J. Angew. Chem., Int. Ed. 1998, 37, 3407– 3410 (c) Reckeweg, O.; DiSalvo, F. J.; Meyer, H. J. Z. Anorg. Allg. Chem. 1999, 625, 1408– 1410 (d) Majzoub, E. H.; Ronnebro, E. J. Phys. Chem. C 2009, 113, 3352– 3358

111. Nicol, A. T.; Vaughan, R. W. J. Chem. Phys. 1978, 69, 5211– 5213

112. Franco, F.; Baricco, M.; Chierotti, M. R.; Gobetto, R.; Nervi, C. J. Phys. Chem. C. 2013, 117, 9991– 9998

113. Hayashi, K.; Sushko, P. V.; Hashimoto, Y.; Shluger, A. L.; Hosono, H. Nat. Commun. 2014, 10.1038/ncomms4515

114. (a) Bowman, R. C.; Hwang, S. J.; Ahn, C. C.; Vajo, J. J. Mater. Res. Soc. Symp. Proc. 2005, 837, 59– 64 (b) Magusin, P. C.; Kalisvaart, W.; Notten, P.; Van Santen, R. Chem. Phys. Lett. 2008, 456, 55– 58

115. Evans, M. J.; Kranak, V. F.; Garcia-Garcia, F. J.; Holland, G. P.; Daemon, L. L.; Proffen, T.; Lee, M. H.; Sankey, O. F.; Haussermann, U. Inorg. Chem. 2009, 48, 5602– 5604

116. (a) Rayhel, L. H.; Corey, R. L.; Shane, D. T.; Cowgill, D. F.; Conradi, M. S. J. Phys. Chem. C. 2011, 115, 4966– 4970 (b) Conradi, M. S.; Mendenhall, M. P.; Ivancic, T. M.; Carl, E. A.; Browning, C. D.; Notten, P. H.; Kalisvaart, W. P.; Magusin, P. C.; Bowman, R. C.; Hwang, S. J.; Adolphi, N. L. J. Alloys Compd. 2007, 446–447, 499– 503

117. (a) Shane, D. T.; Corey, R. L.; Bowman, R. C.; Zidan, R.; Stowe, A. C.; Hwang, S. J.; Kim, C.; Conradi, M. S. J. Phys. Chem. C. 2009, 113, 18414– 18419 (b) Sorte, E. G.; Corey, R. L.; Bowman, R. C.; Birkmire, D.; Zidan, R.; Conradi, M. S. J. Phys. Chem. C. 2012, 116, 18649– 18654

118. (a) Senadheera, L.; Carl, E. M.; Ivancic, T. M.; Conradi, M. S.; Bowman, R. C.; Hwang, S. J.; Udovic, T. J. J. Alloys Compd. 2008, 463, 1– 5 (b) Corey, R. L.; Ivancic, T. M.; Shane, D. T.; Carl, E. A.; Bowman, R. C.; Bellosta von Colbe, J. M.; Dornheim, M.; Bormann, R.; Huot, J.; Zidan, R.; Stowe, A. C.; Conradi, M. S. J. Phys. Chem. C.2008, 112, 19784– 19790

119. (a) Harris, R. K.; Mann, B. E. NMR and the Periodic Table; Academic Press: London, 1979. (b) Chen, F.; Ma, G.; Cavell, R. G.; Terskikh, V. V.; Wasylishen, R. E. Chem. Commun. 2008, 5933– 5935

120. Mathieu, J.; Achey, R.; Park, J. H.; Purcell, K. M.; Tozer, S. W.; Latturner, S. E. Chem. Mater. 2008, 20, 5675– 5681

121. (a) Ravindran, P.; Asokamani, R. Bull. Mater. Sci. 1997, 20, 613– 622 (b) Dshemuchadse, J.; Jung, D. Y.; Steurer, W. Acta Crystallogr, Sect. B. 2011, B67, 269– 292

122. (a) Steinbrenner, U.; Adler, P.; Holle, W.; Simon, A. J. Phys. Chem. Solids. 1998, 59, 1527– 1536 (b) Ebbinghaus, G.; Simon, A. Chem. Phys. 1979, 43, 117

123. Hooper, J.; Zurek, E. ChemPlusChem.2012, 77, 969– 972

124. Wendorff, M.; Scherer, H.; Röhr, C. Z. Anorg. Allg. Chem. 2010, 636, 1038– 1044

125. Sevov, S. C.; Corbett, J. D. Inorg. Chem. 1991, 30, 4875.

126. Li, B.; Corbett, J. D. Inorg. Chem. 2003, 42, 8768-8772.

127. Li, B; Corbett, J.D. J. Am. Chem. Soc. 2005, 127, 926-932.

128. Henning, R.W.; Leon-Escamilla, E.A.; Zhao, J.T.; Corbett, J.D. Inorg. Chem. 1997, 36, 1282-1285.

129. Huang, B., Corbett, J.D. Inorg. Chem. 1998, 37, 1892-1899.

130. Massalski, T.B.; Okamoto, H. Binary Alloy Phase Diagrams, 2nd ed., ASM International, Materials Park, OH, 1990

131. Blankenship, T.V.; Wang, X.; Hoffmann, C.; Latturner, S. E. Inorg. Chem.,2014, in press.

132. Blankenship, T.V.; Chen, B.; Latturner, S.E. Chem. Mater.2014, 26, 3202-3208. Ca54In13B4H23

133. a) Fornasini, M.L. ActaCryst. C1987, 43, 613-616. b) Marsh, R.E.; Slagle, K.M. ActaCryst. C1988, 44, 395-396. Ca8In3

134. a) Fornasini, M.L. ActaCryst. C1987, 43, 613-616. b) Marsh, R.E.; Slagle, K.M. ActaCryst. C1988, 44, 395-396.

135. Wendorff, M.; Röhr, C. Z. Naturforsch. B. 2004, 59, 619-628.

136. Wendorff, M.; Scherer, H.; Röhr, C.Z. Anorg. Allg. Chem. 2010, 636, 1038-1044.

137. Wendorff, M.; Röhr, C. Z. Naturforsch. B. 2008, 63, 804 – 818.

138. Kurakevych, O.O., Strobel, T.A.; Kim, D.Y.; Cody, G.D. Angew. Chem. Int. Ed.2013, 52, 8930-8933.

139. Czekalla, R.; Jeitschko, W.; Hoffmann, R.D.; Rabeneck, H. Z. Naturforsch. B, 1996, 51, 646-654.

140. Pöttgen, R; Wachtmann, K.H.; Jeitschko, W.; Lang, A.; Ebel, T.Z. Naturforsch. B, 1997, 52, 231-236.

141. (a) Fjellvag, H.; Karen, P. Inorg. Chem. 1992, 31, 3260–3263. (b) Karen, P.; Fjellvag, H. J. Alloys Compd. 1992, 178, 285–295. (c) Disch,S.; Cheetham, A. K.; Ruschewitz, U. Inorg. Chem. 2008, 47, 969–973.

142. (a) Hoffmann, R.; Meyer, H. J. Z. Anorg. Allg. Chem. 1992, 607, 57–71. (b) Meyer, H. J. Z. Anorg. Allg. Chem. 1991, 593, 185–192

143. Blankenship, T.V.; Lita, A.; Latturner, S.E. Inorg. Chem.2012, 51, 13345-13350.

144. Stoyko, S.S.; Mar, A. J. Solid State Chem. 2011, 184, 2360-2367

145. Zaikina, J.V.; Schellenberg, I.; Benbow, E.M.; Pöttgen, R.; Latturner, S.E. Chem. Mater. 2011, 23, 1768-1778.

146. Soederberg, K.; Kubota, Y.; Muroyama, N.; Gruener, D.; Yoshimura, A.;Terasaki, O. J. Solid. State. Chem.2008, 181, 1998 - 2005

147. Huang, B; Corbett, J.D. Inorg. Chem. 1998, 37, 5827 - 5833

148. Schafer, M.C; Bobev, S. J. Am. Chem. Soc. 2013, 135, 1696 - 1699

149. Gryko, J; McMillan, P.F; Marzke, R.F; Ramachandran, G.K; Patton, D;Deb,, S.K; Sankey, O.F. Phys. Rev. B. 2000, 62, R7707 – R7710

150. Guloy, A. M.; Ramlau, R.; Tang, Z.; Schnelle, W.; Baitinger, M.;Grin, Yu. Nature. 2006, 443, 320.

151. Nolas, G.S; Vanderveer, D.G; Wilkinson, A.P; Cohn, J.L. J. Appl. Phys. 2002, 11, 8970 - 8973

152. Kroener, R; Peters, K; Von Snering, H.G; Nesper, R. Z. Kristallogr – New Cryst. Struct. 1998, 213, 664

153. Milorad, S; Latturner, S.E. J. Sol. State Chem. 2007, 180, 907 - 914

154. Qishing, L; Corbett, J.D. Inorg Chem. 2008, 44, 512 -518

155. Wu, H; Zhou, W; Udovic, T.J; Rush, J.J; Yildirim, T. J. Alloys. Compound. 2007, 436, 51 - 55

156. Niewa, R; DiSalvo, F.J; Chem. Mater. 1998, 10, 2733 - 2752

157. Ostermann, D; Zacheieja, U; Jacobs, H. J of Alloys and Compounds. 1992, 190, 137 - 140

158. Niewa, R; Jacobs, H.J of Alloys and Compounds. 1996, 232, 61 – 68

159. Baker, C; Barker, M; Blake, A. Acta Crystallographica Section E-Structure Reports Online. . 2001, 57, 6 - 7

BIOGRAPHICAL SKETCH

Education

Florida State University Tallahassee, Fl

Ph.D. in Inorganic Chemistry, expected September 2014

University of California, Davis

B.S. in Chemistry, December 2008

Experience

Florida State University, Dept of Chemistry: 2008 – present

Graduate Student with Dr. Susan Latturner

• Synthesized MOFs and zeolites using hydrothermal methods and performed BET and UV-VIS measurements.

• Grew Zintl and intermetallic phases from Alkaline earth/Li flux. Characterized phases using PXRD, Single-crystal XRD, FTIR, and SEM-EDS, and TGA-DSC.

University of California, Davis, Dept of Chemistry 2007 – 2008

Research Assistant with Dr. Phillip Power

• Used schlenk line techniques to synthesize bulky ligands to create a boron bisamine complex.

Technical Experience

Synthesis: Schlenk line, hydrothermal and flux crystal growth. Experience working in glove box and air- sensitive reactions.

Characterization: Powder XRD, Single-Crystal X-Ray and Neutron Diffraction, Liquid and Solid-State NMR, FTIR, Raman, and UV/VIS spectroscopy, SEM-EDS, TGA-DSC, GC-MS, BET

Analysis: Crystal structure refinement using SHELX-T, Band structure calculations using TB-LMTO- ASA

Teaching Experience

3 years experience - Lectured in recitation and lab courses; Wrote quizzes; Graded exams and lab reports.

______

Publications

Blankenship, T.V; Lita, A; Latturner, S.E. “Ca11E3C8 (E=Sn,Pb): New Complex Carbide Zintl Phases Grown from Ca/Li Flux” Inorg. Chem. 2012, 51, 13345–13350.

Blankenship, T.V; Banghao, C; Latturner, S.E.“Ca54In13B4-xH23+x: A Complex Metal Subhydride Featuring Ionic and Metallic Region.” Chem. Mater. 2014, 26, 3202 - 3208

Blankenship, T.V; Laturner, S.E. “New Zintl Phases Ca14As6(C/H/N)7 and LiCa3As2H Grown From a Ca/Li Flux” Inorg. Chem. 2014, 53, 10620 - 10626

Blankenship, T.V; Dickman, M.J; van de Burgt, L.J. Latturner, S.E. “Alkaline earth indium allenylides synthesized in AE/Li flux (AE=Ca, Ba)”, in press