Drude Electron Materials Having Rutile and Layered Structures


The design of new molecular and non-molecular solids that may be superconducting at high temperatures , in particular at room temperature , continues to challenge me . A new “picture”, or model , of superconductivity based on chemical , rather than physical principles has been gradually emerging . The main features of this hypothesis include :

* In superconductors the electrons that carry the electronic charge and energy through the crystal lattice under an applied potential difference the Cooper pairs are unassociated (free) in nature . They were originally associated with their parent atomic kernels in the valence shell orbitals , but have somehow “escaped” from them and are now completely free and are in the interatomic void spaces . The electronic charge carriers in all superconductors are in effect electrostatic electrons that are freely mobile throughout the lattice . In conventional metals the conduction electrons are associated with valence shell (and sometimes frontier) orbitals comprising the crystal-wide metallic bond (crystal orbital , XO , conduction band) .

* The transition temperature (Tc) of conventional superconductors seems to be directly related to the volume of the void spaces available to the free electrons that have escaped from their parent atomic kernels' valence shell orbitals (among other factors , of course) . The atoms in the elementary metals are very closely packed , usually in a face-centered cubic (fcc) , body-centered cubic (bcc) or hexagonal close-packed (hcp) arrangement . Twenty-nine of the elementary metals are known to superconduct at ambient pressure , and only at transition temperatures close to Absolute Zero . The intermetallic A15 compounds , having the b-tungsten crystal structure , have the highest transition temperatures of the metallic materials . While b-tungsten itself superconducts only at Tc ~ 14 K , the isostructural compound Nb3Ge has the record high Tc of 23.2 K for this class of materials :

Red spheres : Nb ; blue spheres : Ge . Note the relatively large void spaces in this structure .

Similarly the simple compound magnesium diboride , MgB2 , has an unusually high Tc of 39 K , which again may be partially facilitated by the substantial void spaces between the hexagonal boron atom sheets :


Red spheres : Mg ; blue spheres : B (in hexagonal atom sheets , like graphite) .

The famous benchmark cuprate compound YBCO (“1–2–3”, idealized formula Y1Ba2Cu3O7) , the first superconductor discovered having a transition temperature (93 K) above the boiling point of liquid nitrogen (77 K) , has substantial void spaces between its Cu–O layers :

The high Tc of YBCO also undoubtedly benefits from the physical phenomenon of antiferromagnetic (AFM) induction (GIF image , 77 KB) , as discussed in several earlier Chemexplore web pages (eg. the Antiferro one) . The strongly antiferromagnetic copper(II) oxide system induces an antiparallel spin orientation in the free electrons in the void spaces , assisting them to condense into Cooper pairs at an unusually high temperature . These Cooper pairs can “hide” in the relatively large interlayer void spaces , avoiding recapture by their parent Cu2+ kernels until the lattice warms up to ~ 93 K .

In the design of Drude electron materials the crystal structures of many inorganic nonmolecular solids have been surveyed in the search for suitable candidates that might act as the “container” for the free electrons . Generally speaking , the lower the coordinations of the host structure's atoms the better . Such host lattices have considerable open space in their structures which can readily accomodate guest metal atoms and the free electrons they provide to the material .

The rutile crystal structure has a network of void spaces in its lattice that might conceivably host the “popped” electron pairs . The extended rutile structure is portrayed in the following sketch :

A “polyhedral” version of this extended rutile crystal structure is presented in the excellent drawing from the highly recommended website , The Fascination of Crystals and Symmetry . It shows the void spaces more clearly than my own sketch just above :

The rutile crystal structure with 6 : 3 : 3 atomic coordinations has been adopted by a variety of MX2 solid state (non-molecular) compounds , in particular where X = F and M is a small cation (eg. MgF2 , ZnF2 , and Transition metal difluorides) , and by assorted oxides with small M atoms (eg. TiO2 , whose naturally-occurring mineral is “rutile” ; SnO2 ; MnO2 , VO2 , and CrO2 , which was studied in a previous web page) . Many types of rutile oxides were examined by Z. Hiroi , “Structural Instability of the Rutile Compounds and Its Relevance to the Metal–Insulator Transition of VO2”, ArXiv.org , March 8th, 2015 ; see pp. 7-8 [PDF , 5282 KB] .

If M is a large cation the fluorite crystal structure with the higher 8 : 4 : 4 atomic coordinations predominates (eg. CaF2 – the mineral fluorite ; BaF2 ; CdF2 ; HfO2 ; ZrO2 ; CeO2 ; ThO2) . Compounds with the antifluorite structure were surveyed in an earlier web page . MS2 compounds tend to have a layered structure , as the sulfur atoms seem to prefer a trigonal pyramid coordination in them , rather than a trigonal planar coordination assumed by the fluoride and oxide anions in the rutiles .

The large void spaces in the rutile structure portrayed above in the Crystals and Symmetry sketch make it a tempting prospect for hosting pairs of free electrons in novel Drude electron materials . Even more interesting are the channels of void spaces , through which the Cooper pairs can pour downfield under a potential difference applied across the ends of the crystal , which would thus become a high temperature superconductor . Since the free electrons (and electrides in general) are strong chemical reducing agents , the M host atoms must be resistant to reduction and be fairly inert chemically . That consideration rules out any use of the MO2 rutile oxides , since all their M(IV) metal atoms are oxidizing agents to a greater or lesser extent , and they would instantly absorb the free electrons , becoming M(III) and M(II) in the process . That leaves the MF2 rutile fluorides as possible host structures for the free electrons , as various reduction-resistant M2+ cations fortunately are available for use in prospective MF2 rutile hosts .

MX2 solid state compounds having X = chloride , bromide , or iodide often have a layered crystal structure similar to those of the MS2 sulfides , generally of the CdCl2 or CdI2 [GIF image , 24 KB] variety . The interlayer void spaces in these latter MX2 halides would be excellent locations for the free electron pairs . Thus , all the MX2 halides are potentially valuable precursors for chemical modification into new superconductor candidate materials : the fluorides for rutile composites , and the chlorides , bromides , and iodides for CdCl2 or CdI2 layered Drude electron materials .


Electronically Active Tin Compounds


The chemical technique of electride synthesis is probably the best one (and maybe the only one) to obtain Drude electron materials . Generally , zerovalent atoms of a selected metal element are inserted into a suitable host structure , usually in a 1 guest atom to 2 host structure ratio . The first equivalent of host lattice accomodates the guest atom's cationic kernel , while the second equivalent of host provides a suitable large void space (often a cation vacancy or interlayer space) for receiving the valence electron or pairs of electrons that have been separated from the guest atoms . These latter detached electrons , when in the void spaces and surrounded by host anions , are unassociated in nature , having completely severed the orbital connections to their parent atomic kernels . Since they are free , they no longer are obliged to follow Fermi-Dirac statistics , which otherwise would scatter them energy-wise in the Fermi-Dirac distribution , making it very difficult for them to form the Cooper pairs required for superconduction . The pairs of free electrons in the void spaces will be stable over the temperature range at which the Drude electron material containing them is stable . Such a temperature range could in many cases extend from Absolute Zero to well above room temperature , thus potentially making the compound a high temperature superconductor .

As well as using zerovalent metal atoms as the guests in solid state electrides , it might also be possible to use metal cations with inert pairs of valence shell electrons in an electride synthesis of a novel Drude electron material . However , the question of whether or not the inert pairs under consideration can be separated from their underlying electrophilic parent kernels immediately arises . The redox nature of the kernel , usually expressed as its standard reduction potential , E0red , is probably a good measurement of the ability to “pop” the inert pairs into a neighboring void space . Reduction potentials are listed in the following Table for a selection of heavy metal elements with inert pairs :

The example of Hg0 was discussed in the Electrons web page . Even though mercury is a Noble Metal in the redox sense it might still be possible to form its electride in a suitable host lattice . Since sulfide anions form very strong coordinate covalent S>Hg bonds to the underlying Hg2+ kernels , silicon disulfide was selected to host the zerovalent mercury atoms to form the hypothetical compound Hg2+[**]Si2S4 , which could have the wurtzite or zinc blende crystal structure (based on the well known ionic conductor , Cu2HgI4 , ie. Hg2+[ ]Cu2I4 , which is a tetragonal zinc blende at room temperature and a cubic zinc blende above 70 C : GIF image , 52 KB) .

The very low reduction potential of tin(IV) [E0red = 0.151 V to Sn2+] suggests the 5s2 inert pair of Sn(II) might be displaced into suitable void spaces under the appropriate conditions . Several electronically active tin compounds were briefly examined in the Fluorides web page . Such electrically conducting solids are of considerable interest in the present context , as they provide experimental encouragement of a possible use of the non-bonding tin(II) inert pairs in covalent bonds and of them being promoted into higher energy level frontier orbitals .

Tin(III) phosphide , SnP , is a covalentmetallic compound with a high electrical conductivity . Its cubic rocksalt form is superconducting in liquid helium :

Tin(III) phosphide was originally prepared by the direct combination of equimolar quantities of tin metal powder and red phosphorus in a high temperaturehigh pressure (HP–HT , tetrahedral anvil) press , at 800 C and 65 Kbars for one hour : P.C. Donohue , “The Synthesis , Structure , and Superconducting Properties of New High-Pressure Forms of Tin Phosphide”, Inorg. Chem. 9 (2) , pp. 335-337 (1970) [DOI] . The product was pure SnP , but it was a mixture of the cubic rocksalt (minor) and tetragonal (major) phases . “The tetragonal form transformed slowly and irreversibly to the cubic form when heated at ambient pressure between 100 and 200 ”. Tetragonal SnP had the following crystal structure (my sketch is based on Donohue's Figure 1 , p. 336 in his report) :

The phosphorus atoms have a five-fold coordination to the tin atoms , in the form of a tetragonal (square-base) pyramid . The tin atoms similarly have a five-fold coordination to the phosphorus atoms , but a sixth coordination position is occupied by the stereochemical 5s2 inert pairs of electrons , which extend into the interlayer void spaces . From this crystallographic information the following valence electron distribution in the tin and phosphorus atoms in tetragonal SnP can be sketched per the Valence Bond theory , which describes covalent bonds simply and clearly :

Note that the tin atoms must use four hypervalent electrons and their orbitals (4dx2-y2 and 4dz2) in the SnP covalent bonds in order to have enough electrons to complete the twelve-sets of electrons around the octahedrally coordinated atoms . Since the eight tin valence (and hypervalence) and five phosphorus valence electrons add up to a total of 13 system electrons , and only 12 are required for the twelve-sets , one electron is left over and is located to the tin 6s,p frontier orbitals . The continuous overlapping of these latter orbitals throughout the lattice produces the 6s,p metallic bond in it , resulting in SnP being a covalent–metallic solid and an excellent electrical conductor .

Annealing of the tetragonal phase , as noted above , converts it into the cubic rocksalt phase . In the rocksalt crystal structure all the M and X atoms are octahedrally coordinated (and bonded) to each other . Thus , in cubic SnP the Sn and P atoms have six covalent Sn–P bonds each . They require a twelve-set of valence (and hypervalence) electrons to complete these covalent bonds . This is reflected into the valence electron distribution for the tin and phosphorus atoms in cubic SnP :

We see from this simple sketch that tin's 5s2 inert pair in the tetragonal structure is now being used to form the sixth Sn–P covalent bond for the octahedral coordination . As with tetragonal SnP , the “left over thirteenth tin valence electron has been located in the tin atom's 6s,p frontier orbitals , which form the metallic bond in the lattice .

The interesting compound cesium tin(II) tribromide , CsSnBr3 , is another example of the tin(II) 5s2 inert pair losing its inertness and becoming electronically active in the material after it has been surrounded and squeezed by coordinating ligands . In a simple solid state reaction cesium bromide (a white salt , m.p. 636 C) is melted together with tin(II) bromide (a pale yellow salt , m.p. 215 C) to obtain cesium tin(II) tribromide , a black , cubic perovskite having an appreciable electrical conductivity at room temperature :

S.-W. Ng and J.J. Zuckerman , “Where are the Lone-Pair Electrons in Subvalent Fourth Group Compounds ?”, Adv. Inorg. Chem. Radiochem. 29 , pp. 297-325 , H.J. Emelus and H.G. Sharpe (eds.) , Academic Press , Orlando (FL) , 1985 [DOI] ; D.E. Scaife , P.F. Weller , and W.G. Fisher , “Crystal Preparation and Properties of Cesium Tin(II) Trihalides”, J. Solid State Chem. 9 (3) , pp. 308-314 (1974) [DOI] ; J. Barrett et al. , “The Mssbauer Effect in Tin(II) Compounds . Part XI . The Spectra of Cubic Trihalogenostannates(II)”, J. Chem. Soc. A (20) , pp. 3105-3108 (1971) [DOI] .

One explanation of this remarkable phenomenon is that tin(II) has promoted its 5s2 inert pair into higher energy orbitals (probably 6s,pz) , where it can form a bilayer metallic bond with the bromide anions' 4pz (valence shell) orbitals . As pointed out in the ebook [PDF , 6176 KB] , p. 286 ,

“It is interesting to note that if the tolerance factor for CsSnBr3 is calculated using the crystal ionic radius of Sn(II) , 1.22 , it turns out rather low , at t = 0.84 , but if the value for Sn(IV) , 0.83 , is used in the Goldschmidt equation , it will be much higher (t = 0.97) . The former low value would be indicative of a distorted perovskite , while the latter high factor would generally be associated with a perovskite of cubic symmetry . As CsSnBr3 has a cubic symmetry from 19 C and above , the implication is that the SnBr3 supercube framework is indeed based on Sn(IV) , with the tins’ 5s2 inert pair of valence electrons somehow promoted to a higher energy level frontier orbital , where they can interact with the bromide orbitals to form a functional XO metallic bond (conduction band)”.

A valence electron distribution diagram illustrating the promotion of the tin's 5s2 inert pair into the 6s,pz frontier orbitals is sketched as follows :

Unlike the tin atoms in SnP , those in CsSnBr3 have enough normal valence electrons to form the twelve-sets for the six Sn–Br covalent bonds per tin atom . Each Sn contributes two electrons ; the cesium (in covalent structures , formally considered as Cs0) contributes one electron ; and each bromine atom (formally Br0, not bromide anion) contributes three , for a total of nine electrons from the three formula bromines . The total contributions are : 2 + 1 + 9 = 12 electrons for the twelve-set around each tin atom . Each bromine atom has a completed octet , formally making them bromide anions . The large cesium cation nesting in the center of the SnBr3 “cage” has the electronically inert xenon Rare Gas electronic configuration . Meanwhile , the other two remaining tin valence electrons (formally its 5s2 inert pair) have been “pushed upstairs” into the 6s,pz frontier orbitals when the sp5 octahedral hybrid orbital “takes” their 5s orbital for the Sn–Br covalent bonds .

Cesium tin(II) tribromide has a room temperature electrical conductivity of 0.05 ohm-1-cm-1. It shows a “metallic-type behavior between –100 and 350 C with no major change in resistance observed” (Scaife , Weller , and Fisher , cited above , their p. 313) . This low conductivity in the material suggests that the inert pair remains intact in the 6s orbital , but a small amount of electron density “leaks” from it into the adjoining 6pz orbital (the 6px,y orbitals have been “taken” by the sp5 hybrid orbital) . The tin 6pz orbital has the correct size , shape , symmetry , and orientation to overlap throughout the rectilineal SnBr3 framework with the bromide's 4pz valence shell orbitals . The tin 6pz–bromine 6pz bilayer metallic bond is therefore proposed as the conduction band in cubic CsSnBr3 . The relatively low amount of “leaked” electron density in the upper tin layer results in the rather low electrical conductivity of the black , cubic perovskite .

Since CsSnBr3 transitions above 19 C from a distorted perovskite [whose tin(II) atoms have stereochemical 5s2 inert pairs] to a cubic symmetry perovskite with dispersed inert pairs , the implication is that it's fairly easy to “pop” the Sn(II) 5s2 inert pairs off their Sn(IV) kernels in this bromide system (apparently chloride anions aren't as effective as bromides for this purpose , and unfortunately fluorides weren't studied by Scaife , Weller , and Fisher) . In any case the example of CsSnBr3 is very encouraging for the experiments discussed here involving tin(II) .


Rutile Fluorides , [AMF4]rutile (2e)void


The objective here will be to design a series of composite materials consisting of two components . The first is the substrate host lattice , which will be a MF2 fluoride known to have the rutile crystal structure . The second component is the dopant AF2 fluoride whose A atoms have pairs of valence shell electrons that can be “popped” into the void spaces of the resulting rutile composite . The AF2 fluoride is doped , in gradually increasing mole ratios , into the MF2 host fluoride . Hopefully the dopant fluoride will mimic the host and will adopt its rutile structure . When the A atoms are octahedrally coordinated by the fluorides their valence electron pairs will be “popped” into the void spaces , where they will be free electrons , and the [AMF4]rutile (2e)void composite will be a Drude electron material and potentially a high temperature superconductor .

Magnesium fluoride and zinc fluoride both have the rutile crystal structure , and both compounds are commercially available , in high purities , at a modest cost . A number of anhydrous MF2 Transition metal fluorides (those of Cr , Mn , Fe , Co , and Ni) are also interesting as possible host rutile structures , although they are magnetic (antiferromagnetic) . Anhydrous magnesium chloride has the layered CdCl2 crystal structure , while MgBr2 and MgI2 have the related CdI2 layered structure . These latter substrates could also host the AX2 halide dopants (X = Cl , Br , and I) to synthesize layered composites with electron pairs located in the interlayer spaces , as will be discussed in the next section of this essay .

As discussed above , tin(II) is the most promising of the candidate dopants because its 5s2 inert pair can be easily displaced from the 5s orbital location . The tin(II) halides could be doped into the MX2 substrates in increasing mole ratios , for example as with MgF2 and SnF2 :

x MgF2 (m.p. 1263 C) + (1x) SnF2 (m.p. 215 C , b.p. 850 C)

------- [grind together , press pellet , heat , possibly HPHT] ------> MgxSn1-xF2 ,

where x = a mole ratio taken experimentally between 0 (pure SnF2 ) and 1 (pure MgF2 ) .

At x = 0.5 , the composite's empirical formula could be written as [Mg2+SnIVF4]rutile (2e)void . The magnesium component is coordinated as (MgF6)4- octahedrons , while the tin atoms are expected to be coordinated as (SnF6)2- octahedrons . Note that the tin(IV) atoms in SnF4 (a white , crystalline solid , m.p. 442 C) are octahedrally coordinated by their fluoride neighbors in the solid state :

The 5s2 inert pairs could be promoted into the tins' 6s,p frontier orbitals [physically located in the SnF interatomic voids in the (SnF6)2- octahedrons] , where they would probably be localized and unable to form any sort of metallic bondconduction band . Or , they could be located in the vastly larger lattice voids in between the (SnF6)2- and (MgF6)4- octahedrons see the rutile sketch above but as Drude electron pairs . The former case would be facilitated by the relatively small energy gap between the 5s,p and 6s,p orbitals , but the electron pairs would be constrained by the very small SnF interatomic void spaces . The latter scenario would be supported by the very large interpolyhedra voids typically found in the rutile crystal structure , but would be energetically unfavorable because of the coulombic repulsion between the promoted tin(II) electron pairs and the fluoride anions .

Experimentally , the doping of an ionic host substrate (MgF2) with a covalent guest dopant (SnF2) might be problematic , and could require forcing HPHT conditions . Fortunately the crystal ionic radii of the Mg2+ (r = 0.72 , CN = 6) and SnIV (r = 0.69 , CN = 6) are almost similar , resulting in uniformly sized MF6 polyhedra in the lattice . Similarly Zn2+ (r = 0.74 , CN = 6) in the (ZnF6)4- octahedrons would fit acceptably with the (SnF6)2- octahedrons in the rutile composite material :

x ZnF2 (m.p. 872 C) + (1x) SnF2 (m.p. 215 C , b.p. 850 C)

------- [grind together , press pellet , heat , possibly HPHT] ------> ZnxSn1-xF2 ,

where x = a mole ratio taken experimentally between 0 (pure SnF2 ) and 1 (pure ZnF2 ) . At x = 0.5 , the composite's empirical formula could be written as [Zn2+SnIVF4]rutile (2e)void .

A (I + III) = II” combination could replace Sn(II) in the above experiments . Either the I or the III component would have the inert pair for promotion into the lattice voids . The companion III and I component , respectively , could be a chemically inert cation with the Rare Gas electronic configuration . Lithium cation (r = 0.76 , CN = 6) would be such a suitable monovalent , inert cation . All of the lithium halides have the cubic rocksalt structure , with (LiX6)5- octahedrons . Aluminum cation (r = 0.54 , CN = 6) , while somewhat on the small side , is known to form (AlF6)3- octahedrons in AlF3 and in cryolite , Na3AlF6 . The scandium cation Sc3+ (r = 0.75 , CN = 6) would be ideal for these rutile composites as the inert III component , but unfortunately scandium is a rare and very costly commodity , and so is impractical for this study .

Monovalent halides of Family IIIA/13 could be examined as the I component with the inert pair to be displaced into the voids . Thallium(I) halides are stable , well known , and are commercially available at a modest cost :

The underlying Tl(III) kernel is a moderately strong oxidizer , having E0red = 1.252 V , TlIII to Tl1+. Redox calculations in the tabulation suggest there could be some charge transfer between the Tl(III) in the (TlX6)3- octahedrons and coordinating bromide and iodide anions , but not in those octahedrons with X = F and Cl , which would remain purely ionic in nature . Very high pressures would almost certainly be required to displace the 6s2 inert pair in Tl(I) into the void spaces :

x ZnF2 (m.p. 872 C) + (1x) AlF3 (m.p. 1291 C , subl.) + (1x) TlF (m.p. 326 C)

-------- [HPHT] ------> ZnxAl1/2(1-x)Tl1/2(1-x)F2 ,

where x = a mole ratio taken experimentally between 0 and 1 .

At x = 0.5 , the empirical formula of the product could be written as :

[(Zn2+)0.5(Al3+)0.25(TlIII)0.25F2]rutile (e)void , or [(Zn2+Al3+0.5TlIII0.5F4]rutile (e)void , or as {[(Zn2+Al3+0.5TlIII0.5F4]rutile}2 (2e)void .

That is , there would be only one Cooper pair for every two interpolyhedral void spaces in this latter hypothetical rutile composite .

The lighter IIIA/13 elements could also be tried in place of Tl(I) . Indium(I) halides are known , and are even commercially available , but they tend to disproportionate into the more thermodynamically stable In(0) and In(III) : D.G. Tuck , “The Lower Oxidation States of Indium”, Chem. Soc. Rev. 22 (4) , pp. 269-276 (1993) [DOI] . Gallium(I) is known , for example occurring in the mixed-valent compound GaCl2 , which is really GaCl–GaCl3 (GIF image , 11 KB) .

Al(I) compounds are stable only at very high temperatures , and decompose back to Al(0) and Al(III) when cooled down to room temperature : M. Hoch and H.L. Johnston , “Formation , Stability and Crystal Structure of the Solid Aluminum Suboxides : Al2O and AlO”, J. Amer. Chem. Soc. 76 (9) , pp. 2560-2561 (1954) [DOI] ; C.N. Cochran , “Aluminum Suboxide Formed in Reaction of Aluminum with Alumina”, J. Amer. Chem. Soc. 77 (8) , pp. 2190-2191 (1955) [DOI] ; see also T. Forland et al. , “Measurements of Phase Equilibria in the Aluminum – Aluminum Sulfide System”, Acta. Chem. Scand. , Series A28 (2) , pp. 226-228 (1974) [PDF , 375 KB ; DJVU , 109 KB] . The compound AlS has a narrow window of stability between 1010 C and its m.p. of 1060 C .

In(I) , Ga(I) and Al(I) might still conceivably be used in these rutile experiments , but they would have to be generated as transient reaction intermediates by the reproportionation in situ of their M(0) and M(III) precursors :

x ZnF2 (m.p. 872 C) + 1/3(1x) In0 (m.p. 157 C) + 2/3(1x) InF3 (m.p. 1172 C)

------- [HPHT] ------> ZnxIn1-xF2 , where x = a mole ratio taken experimentally between 0 and 1 .

At x = 0.5 , the empirical formula could be written as [(Zn2+)0.5(InIII)0.5F2]rutile (e)void , or as [(Zn2+InIIIF4]rutile (e)void , or as {[(Zn2+InIIIF4]rutile}2 (2e)void . There would be only one Cooper pair for every two interpolyhedral void spaces in this ZnInF rutile composite .

Similar formulations can be written for “gallium(I) fluoride” doped into a ZnF2 host rutile lattice :

x ZnF2 (m.p. 872 C) + 1/3(1x) Ga0 (m.p. 30 C) + 2/3(1x) GaF3 (m.p. 800 C , subl.)

------- [HPHT] ------> ZnxGa1-xF2 ,

where x = a mole ratio taken experimentally between 0 and 1 . At x = 0.5 , its empirical formula could be written as {[(Zn2+Ga3+F4]rutile}2 (2e)void .

The hypothetical “aluminum(I) fluoride” could be analogously formulated with ZnF2 as follows :

x ZnF2 (m.p. 872 C) + 1/3(1x) Al0 (m.p. 660 C) + 2/3(1x) AlF3 (m.p. 1291 C , subl.)

------- [HPHT] ------> ZnxAl1-xF2 ,

where x = a mole ratio taken experimentally between 0 and 1 . At x = 0.5 , its empirical formula could be written as {[(Zn2+Al3+F4]rutile}2 (2e)void .

When an inert monovalent cation is combined with a trivalent cation with a displaceable inert pair , the latter species could be Bi(III) , Sb(III) , and maybe even As(III) and P(III) , as illustrated in the following formulations . Lithium cation , with r = 0.76 (CN = 6) and which prefers an octahedral coordination with halide anions (as in its cubic rocksalt halides) , should be a satisfactory inert monovalent cation in the rutile formation :

x ZnF2 (m.p. 872 C) + (1x) LiF (m.p. 848 C) + (1x) BiF3 (m.p. 230 C)

------- [HPHT] ------> ZnxLi1/2(1-x)Bi1/2(1-x)F2 ,

where x = a mole ratio taken experimentally between 0 and 1 .

At x = 0.5 , its empirical formula could be written as {[(Zn2+Li1+0.5BiV0.5F4]rutile}2 (2e)void . Very high pressures would be reqired to pop the Bi(III) 6s2 inert pairs into the interpolyhedral voids .

A similar formulation could be written for antimony(III) fluoride with LiF and ZnF2 :

x ZnF2 (m.p. 872 C) + (1x) LiF (m.p. 848 C) + (1x) SbF3 (m.p. 287 C)

------- [HPHT] ------> ZnxLi1/2(1-x)Sb1/2(1-x)F2 ,

where x = a mole ratio taken experimentally between 0 and 1 .

At x = 0.5 , its empirical formula could be written as {[(Zn2+Li1+0.5SbV0.5F4]rutile}2 (2e)void .

Arsenic trifluoride (a liquid at room temperature , b.p. 57 C) and phosphorus trifluoride (a gas at STP , b.p. 102 C) are physically unsuitable for use as dopants for MgF2 and ZnF2 . However , it might be possible to reproportionate the M(0) element with one of its M(V) reagents . Lithium hexafluorophosphate , LiPF6 , and lithium hexafluoroarsenate , LiAsF6 , are both commercially available (eg. Alfa-Aesar) at a modest cost , and might be reproportionated with red phosphorus and gray arsenic , respectively , to obtain the transient reaction intermediates LiPF4 and LiAsF4 :

2/5 LiF + 3/5 LiPF6 + 2/5 P0 (red) --------> LiPF4 = LiF + PF3 ; and ,

2/5 LiF + 3/5 LiAsF6 + 2/5 As0 (gray) --------> LiAsF4 = LiF + AsF3 .

For phosphorus ,

x ZnF2 + 1/5(1x) LiF + 3/10(1x) LiPF6 + 1/5(1x) P0 (red)

------- [HPHT] ------> ZnxLi1/2(1-x)P1/2(1-x)F2 ,

where x = a mole ratio taken experimentally between 0 and 1 .

At x = 0.5 , its empirical formula could be written as {[(Zn2+Li1+0.5PV0.5F4]rutile}2 (2e)void .

While phosphorus(V) might be too small (r = 0.38 , CN = 6) for use in a rutile composite , the 3s2 inert pair in phosphorus(III) compounds is quite reactive , as P(III) compounds are very easily oxidized (in a covalent , not ionic manner) to the corresponding P(V) molecules . The underlying P(V) kernel is low energy in redox terms , that is , it isn't oxidizing like Bi(V) is , for example . These considerations make P(III) quite interesting for study in the synthesis of the rutile composites .

Similarly for arsenic ,

x ZnF2 + 1/5(1x) LiF + 3/10(1x) LiAsF6 + 1/5(1x) As0 (gray)

------- [HPHT] ------> ZnxLi1/2(1-x)As1/2(1-x)F2 ,

where x = a mole ratio taken experimentally between 0 and 1 .

At x = 0.5 , its empirical formula could be written as {[(Zn2+Li1+0.5AsV0.5F4]rutile}2 (2e)void .

The crystal ionic radius of arsenic(V) is r = 0.46 , CN = 6 (octahedral) , which is rather small . The As(V) kernel is only mildly oxidizing , with E0red = 0.560 V , for AsV + 2e----> AsIII .


Layered MX2 Compounds with Drude Electron Pairs


Layered MX2 halides could be used as templates for creating new layered structures with Drude electron pairs in the interlayer spaces . Such voluminous void regions would be efficient channels for the flow of the electron pairs through the lattice under an applied potential difference .

As discussed above , the 5s2 inert pair in tin(II) is quite labile , requiring a relatively small amount of energy for promotion into the 6s,p frontier orbitals . Tin(II) halides are all commercially available at a modest cost :

Tin(II) chloride , bromide , and iodide could be doped into the corresponding magnesium halide salts in increasing mole ratios , to prepare [(Mg2+)x(SnIV)1-xX2]layer (2e)void composite materials in which tin's 5s2 inert pairs have been popped into the interpolyhedral void spaces :

x MgCl2 (m.p. 714 C) + (1–x) SnCl2 (m.p. 247 C)

----- [HPHT] -----> [(Mg2+)x(SnIV)1-xCl2]layer (2e)void ,

where x = a mole ratio taken experimentally between 0 (pure SnCl2 ) and 1 (pure MgCl2 ) ;

x MgBr2 (m.p. 711 C) + (1–x) SnBr2 (m.p. 215 C)

----- [HPHT] -----> [(Mg2+)x(SnIV)1-xBr2]layer (2e)void ; and ,

x MgI2 (m.p. 650 C) + (1–x) SnI2 (m.p. 320 C)

----- [HPHT] -----> [(Mg2+)x(SnIV)1-xI2]layer (2e)void .

Hopefully the doped Sn(II) halides will imitate the layered structure (CdCl2 or CdI2 template) of their corresponding host magnesium halides in the doping process . Formation of the (SnIVX6)2- octahedrons will then encourage the promotion of the 5s2 inert pairs into the interpolyhedra void spaces , thereby producing a Drude electron material .

Layered MX2 halides were briefly alluded to in the Electrons web page in connection with the possibility of popping the 6s2 inert pairs of Pb(II) in PbI2 under very forcing HPHT conditions . Lead(II) iodide has the cadmium iodide layered crystal structure :

This sketch was copied from the Wikipedia web page , Lead(II) iodide . Again , my thanks to the author of this sketch , and Wikipedia , for implied permission to reproduce it on this web page .

At this point the energetics of the electride formation should be considered . In order for electrides to be formed by the solution of M0 metal atoms in a host lattice the two individual parts of the M0 guest atoms , the underlying M2+ parent kernels and the pair of electrons that have been popped into the voids , must both be energetically stabilized . Since the electron pairs have been relocated into void spaces whose “walls” are lined with negatively-charge anions , they will tend to be energetically destabilized in this thermodynamically unfavourable environment . The electride synthesis can proceed to completion only if the M2+ parent kernels can form very strong chemical bonds to the anions surrounding them . If this latter chemical bond formation is thermodynamically very favourable a net energetic stabilization of the electride will be achieved and it should be stable at room temperature .

One way the M2+ parent kernels might form those strong M–X bonds is by a charge transfer from the nucleophilic X halide anions to the strongly electrophilic M2+ cations . The extent of such a charge transfer can be roughly estimated from simple High School redox equations , using the well known standard electrode potentials of the interacting species . For example , in PbX2 the underlying “M2+ parent kernel” is Pb(IV) , a strong oxidizing agent (E0red = 1.455 V) . By comparing this standard reduction potential to those of the halide anions , we can calculate the extent to which charge transfer will occur from F, Cl, Br, and I to Pb(IV) . These results are summarized in the following tabulation of the properties of the lead(II) halides :

The substantial positive cell potential of E0T = 0.9195 V for the iodide reduction of Pb(IV) predicts that it would proceed spontaneously at STP , while the equally large negative E0T = – 1.411 V for the fluoride reduction suggests that it would be highly unfavourable thermodynamically .

If lead(II) iodide is compressed and the Pb(II) 6s2 inert pairs are popped into the interlayer void space , the iodide ligands can form Pb–I coordinate covalent bonds with some iodide >Pb(IV) charge transfer . This Pb(IV)–I bonding can energetically stabilize the compressed PbI2 , so that when it is cooled to room temperature and depressurized , the Drude electron pairs might remain in the interlayer channels . Or maybe not : they might flow backward via the iodide anions (partially positive) and settle down around the Pb(IV) kernels as their spherical 6s2 inert pairs again . That should make highly compressed PbI2 a Drude electron material and possibly a high temperature superconductor . No fluoride >Pb(IV) charge transfer is predicted by the redox equations , and therefore no Pb–F coordinate covalent bonds would be present in lead(II) fluoride under HPHT conditions (PbF2 has the fluorite crystal structure and isn't a layered compound) .

To improve the chances of lead iodide popping its 6s2 inert pairs it could be combined with another iodide having the cadmium iodide layered structure ; this could be CdI2 itself or possibly magnesium iodide . The crystal ionic radius of Pb(IV) , CN = 6 , is r = 0.78 ; that of Cd2+ is r = 0.95 ; and that of Mg2+ is r = 0.72 . The underlying Pb(IV) kernels in the compressed PbI2 would fit best with the Mg2+ cations :

PbI2 (m.p. 402 C) + MgI2 (m.p. 650 C) ----- [HPHT] -----> [MgPbIVI4]layered (2e)void .

Commercial MgI2 (eg. from Alfa-Aesar) is quite expensive . Two methods for synthesizing pure , anhydrous MgI2 from its component elements have been described by G. Brauer (ed.) , Handbook of Preparative Inorganic Chemistry , 2nd edition , vol. 1 , Academic Press , New York , 1963 ; pp. 910-911 . This exhaustive inorganic chemistry synthesis compendium (Vols. 1 & 2 combined) can be downloaded for free from the Sciencemadness.org library resources web page [PDF , 19,090 KB] . Note : this PDF file can be opened only with Adobe Acrobat Reader v. 6 or later . If desired , a suitable version of this application can be downloaded for free from Oldversion.com . The simpler method (for me) involves the addition of iodine to a rapidly stirred suspension of magnesium particles in dry ethyl ether . Note that sluggish Grignard reactions can often be started by the addition of an iodine crystal to the reaction vessel containing magnesium chips stirred in ethyl ether or tetrahydrofuran . This produces a stable MgI2–etherate complex which can be heated at 250 C in vacuo to remove the ether ligands . Anhydrous magnesium iodide was described as “hexagonal platelets” with a layer lattice (ie. the CdI2 crystal structure) , m.p. 650 C (in hydrogen or other inert atmosphere ; iodide anions are oxidized to iodine by the oxygen in air : MgI2 + O2 -----> MgO + I2) .

Anhydrous magnesium bromide can be similarly synthesized from the addition of bromine vapour to rapidly stirred magnesium chips suspended in ethyl ether (Brauer comments , “considerable heat of reaction”, pp. 909-910) . MgBr2 prepared by this method was described as a “pure , white salt”, 60-70% yield , 99.5% purity , m.p. 711 C . It's also a layered compound having the CdI2 crystal structure (per Wikipedia) . Magnesium bromide could act as a layered template for lead(II) bromide when the two reagents are combined under HPHT conditions in which the Pb(II) 6s2 inert pairs are popped into the interlayer void spaces :

PbBr2 (m.p. 371 C) + MgBr2 (m.p. 711 C) ----- [HPHT] -----> [MgPbIVBr4]layered (2e)void .

Anhydrous magnesium chloride is used as the feedstock for the electrolytic manufacture of magnesium metal and is quite cheap , even as offered by suppliers of fine chemicals . Fortunately MgCl2 also has a layered structure , that of cadmium chloride , which is similar to the CdI2 crystal form (the CdCl2 layers have a cubic packing arrangement , while those of CdI2 are hexagonally packed in the crystal structure) . Also fortunately , a small charge transfer of electrons (E0T = 0.0967 V) from chloride anions to Pb(IV) is predicted [Table of Lead(II) Halides above] . This opens the door for another interesting experiment : cooking PbCl2 with the inexpensive anhydrous MgCl2 under HPHT conditions :

MgCl2 (m.p. 714 C) + PbCl2 (m.p. 501 C) ----- [HPHT] -----> [MgPbIVCl4]layered (2e)void .

Redox calculations for the thallium(I) halides imply that only TlBr and TlI would be suitable for use in the layered composites [Table of Thallium(I) Halides above] . If the thallium(I) halides are the source of the inert pairs in these experiments there will only be three (and not the required four) halide anions in the resulting composite . The “chemical trick used to resolve this problem is to design a ternary layered composite using half an equivalent of the Tl(I) halide , half an equivalent of a trivalent cation with a poppable 6s2 inert pair , and one equivalent of the corresponding magnesium halide layered host .

Bismuth(III) and possibly antimony(III) might serve as the trivalent cations with the inert pair . Redox calculations suggest that bismuth(III) chloride , bromide , and iodide might be acceptable as the inert pair source in these layered composites :

For example :

TlBr (m.p. 460 C) + BiBr3 (m.p. 219 C) + MgBr2 (m.p. 711 C)

----- [HPHT] -----> [Mg (TlIII)0.5 (BiV)0.5 Br4]layered (2e)void ; and ,

TlI (m.p. 442 C) + BiI3 (m.p. 409 C) + MgI2 (m.p. 650 C)

----- [HPHT] -----> [Mg (TlIII)0.5 (BiV)0.5 I4]layered (2e)void .

The crystal structure and chemical bonding in thallium(I) iodide are quite unusual and deserve some comment . At room temperature orthorhomic TlI has a layered aspect , although its lattice is usually described as “distorted rocksalt” (6 : 6) . Above 175 C yellow TlI has the 8 : 8 CsCl crystal structure , becoming a red color (it's reversibly thermochromic) :

This sketch was copied from the Wikipedia web page , Thallium(I) iodide . Again , my thanks to the author of this sketch , and Wikipedia , for implied permission to reproduce it on this web page .

The interlayer spaces in orthorhombic thallium(I) iodide are highly suggestive as containing inert pairs of electrons in stereochemically active orbitals on both the thallium and iodine atoms . If this conjecture is true , the Tl–I bonds are covalent in nature and not ionic . Such covalent bonds are most simply , easily , and clearly described by the Valence Bond theory . Octahedral hybrid orbitals on the thallium and iodine atoms could contain precisely the right number of valence (and with Tl , hypervalence also) electrons to create five Tl–I covalent bonds between the Tl and I atoms , with an inert pair of electrons on each atom in their sixth orbital lobes . These inert pair lobes extend into the interlayer spaces in the lattice , pushing the Tl–I layers apart somewhat :

The Wikipedia article about thallium(I) iodide states ,

“Under high pressure , 160 Kbar , TlI becomes a metallic conductor”. I was unable to locate the original reference for this statement , but another source comments ,

“...... optical work on thallium iodide is presented and correlated with the unpublished x-ray work of Chen showing that this material undergoes an insulator–metal transition by band-overlap metallization at 15 GPa” (K. Brister , “Static High Pressure Instrumentation , Metallic Xenon , and Band-Overlap Metallization of Thallium Iodide”, Ph.D. thesis , Cornell University , 1989) .

Both elemental iodine and cesium iodide undergo metal–insulator transitions when strongly compressed : H.G. Drickamer , “Pressure and Electronic Structure”, Science 142 (3598) , pp. 1429-1435 (1963) [DOI] ; iodine is discussed on pp. 1432-1434 (it gradually becomes metallic above 160 Kbars) . Cesium iodide metallizes at an applied pressure >100 GPa : Y. Xu et al. , “Superconducting High-Pressure Phase of Cesium Iodide”, Phys. Rev. B 79 (14) , 144110 , 5 pp. (2009) [PDF , 279 KB] . It's therefore possible that the metallic state in compressed thallium(I) iodide may originate from the iodide anions rather than from thalliums' 6s2 inert pairs .

Bismuth(III) iodide also has a remarkable crystal structure in the solid state :

This sketch was copied from the Wikipedia web page , Bismuth(III) iodide . Again , my thanks to the author of this sketch , and Wikipedia , for implied permission to reproduce it on this web page .

Note the octahedral coordination of the iodine atoms around each bismuth atom . Exactly where the Bi(III) 6s2 inert pairs are located is unknown , although given the symmetry of the BiI6 octahedrons they could be spherically distributed around the Bi(V) kernels , which would have outer sp5 hybrid orbitals for the six Bi–I covalent bonds . That is , the inert pairs haven't “gone anywhere” ; they are still in the spherical , unhybridized 6s2 native orbitals around the Bi atoms , surrounded by the six pairs of sp5 valence electrons :


Very high pressure would likely be required to compress the Tl , Bi , Mg , and I atoms together in the reaction mixture for [Mg (TlIII)0.5 (BiV)0.5 I4]layered (2e)void and related layered materials . Hopefully the MgI2 host component would successfully act as a sort of “catalytic template structure” for the formation of the emerging Drude electron material .


Related web pages in this series about Drude electron materials :

A New Picture of Superconductivity : Lightning Bolt Electrons in a Crystal” ;

Perovskites Designed as Drude Metals and Ambient Superconductors” ;

Rocksalts Designed as Super-electrides , Drude Metals , and as Possible High Temperature Superconductors” ;

Chromium as the Guest Atom in Super-electride Drude Metals” ;

Lead , Tin , and Bismuth as the Guest Atoms in Super-electride Drude Metals” ; and ,

* Betaines and Electrides : From Sugar Beets and Baby Shampoo to Superconductors .


[ Index Page ] [ Contact ]