Metallic Perovskite Fluorides


This web page borrows an idea from an earlier Chemexplore essay , CuNiO , Ni2OAs , and Related Compounds”, and develops it into a new concept involving superconductivity [underlined blue hyperlinks can be clicked when online to retrieve the article cited . The requested document will open in a new window] . The hypothetical compound CuAlF3 , formed by the insertion of zerovalent copper atoms into the chemically inert host lattice of AlF3 , should be metallic and possibly superconducting , but only at a very low temperature close to Absolute Zero (if at all) . The new idea related to CuAlF3 is very simple : synthesize and study analogous metallic compounds with other MX3 host lattices ; however , these latter hosts would have unpaired , singlet valence electrons in an antiferromagnetic (AFM) spin ordering . The AFM ordering would be transmitted into the Pauli paramagnetic free electrons of the Cu0 atoms in the CuMF3 (copper 4s1 orbitals) . The random electron orientations of the Pauli paramagnetism would be reorganized into the alternating antiparallel spin orientations required for the formation of Cooper pairs and the appearance of superconductivity at a higher temperature than without the AFM induction ; this is the AFM model of high temperature superconductivity (HTS) (GIF image , 77 KB) .

Possible AFM MX3 host lattices include : CrF3 (TN = 80 K) ; MoF3 (TN = 185 K) ; MnF3 (TN = 43 K) ; FeF3 (TN = 394 K) ; and CoF3 (TN = 460 K) [data are from the CRC Handbook of Chemistry and Physics , 87th edition , CRC Press / Taylor & Francis , Boca Raton (FL) , 2006 ; “Selected Antiferromagnetic Solids”, pp. 12-107 to 12-108] . Iron(III) fluoride looks quite interesting in this regard . It has a high Nel temperature , TN ; its crystal structure is similar to AlF3 (Hargittai and co-workers , 1990) ; and iron is , of course , a cheap and abundant element , making FeF3 a practical sort of chemical reagent to use (the references are presented at the end of the text , below) .

Unfortunately the redox chemistry of FeF3 disqualifies it from consideration . Fe(III) is a natural oxidizer , and would instantly oxidize the Cu0 to Cu1+ in the attempted preparation of CuFeF3 . Cu1+ is redox-unstable in a fluoride environment , and would disproportionate to Cu0 and Cu2+ :

Cu0 – e -------> Cu1+ ; E0ox = – 0.521 V ;

Fe3+ + e -------> Fe2+ ; E0red = 0.771 V ;

Net reaction : Cu0 + Fe3+ -------> Cu1+ + Fe2+ ; E0T = 0.250 V ; then ,

Cu1+ ---------> Cu0 + Cu2+ ; E0T = 0.368 V .

Overall reaction : Cu0 + Fe3+ -------> Cu0 + Cu2+ + Fe2+ ; E0T = 0.618 V .

The standard redox potentials cited are versus the SHE at STP , and were from : D.R. Lide (ed.) , CRC Handbook of Chemistry and Physics , 87th edition , CRC Press / Taylor & Francis , Boca Raton (FL) , 2006 ; P. Vansek (ed.) , “Electrochemical Series”, pp. 8-20 to 8-25 . Another useful reference with redox potentials is : A.J. Bard (ed.) , Encyclopedia of Electrochemistry of the Elements , various volumes , Marcel Dekker , New York , ca. 1973-1986 .The Wikipedia web page “Table of Standard Electrode Potentials” is a helpful online reference in this regard . For a convenient tabulation of oxidizing metal oxides and their E0red values , see this GIF image (45 KB) .

The hypothetical compound Cu0Fe3+F3 is predicted to be a cubic symmetry perovskite (t = 0.99) ; however , the mixture of Cu0 + Cu2+ + Fe2+ metal atoms in the lattice would undoubtedly ruin the perovskite structure . It's hard to predict exactly what sort of crystal structure and electronic behaviour would result from the Cu0 / Fe3+ combination in CuFeF3 .

Mn3+ (E0red = 1.5415 V to Mn2+) and Co3+ (E0red = 1.92 V to Co2+) are also strong natural oxidizers and are therefore unsuitable for combination with Cu0 . Chromium(III) and molybdenum (III) are low energy redox species (neither oxidizing nor reducing in nature) , so it should be possible to insert Cu0 into the lattices of CrF3 and MoF3 to form the cubic perovskites CuCrF3 (t = 0.95) and CuMoF3 (t = 0.91) , respectively . Like CuAlF3 these latter compounds should be metallic solids . How would the AFM ordering in the surrounding MX3 cages” affect the spin ordering of the copper 4s1 valence electrons , those responsible for the material's electrical conductivity ? If there is at least some AFM induction into these Pauli paramagnetic electrons , CuCrF3 and CuMoF3 could become superconducting substantially above Absolute Zero , although probably not in the liquid nitrogen range . CuMoF3 is predicted to have a higher transition temperature Tc than CuCrF3 , since its MoF3 host has a higher Nel temperature (TN = 185 K) than CrF3 (TN = 80 K) .

However , CrF3 is chemically more practical” than MoF3 as a precursor reagent for the synthesis of the perovskites . It is commercially available in a pure , anhydrous form (eg. from Alfa-Aesar) but rather expensive or as the much cheaper hydrated CrF3 . Such hydrated inorganic salts can often be efficiently dehydrated by a reagent that chemically reacts with and removes their water content (such as acetyl chloride , acetic anhydride , or thionyl chloride) . Azeotropic distillation of their water of crystallization (possibly with isopropanol over 3A molecular sieves in a Soxhlet apparatus) might be another way of chemically dehydrating certain sensitive halide salts that hydrolyze when heated .

Anhydrous chromium(III) fluoride is a green , crystalline solid , thermally stable (m.p. 1425 C) , and insoluble in water .

Molybdenum(III) fluoride is a somewhat obscure , little-known compound . It has been described as yellow-brown hexagonal crystals with a melting point > 600 C , and insoluble in water (Handbook of Chemistry & Physics) . A careful characterization (crystal structure , by X-ray diffraction) of MoF3 was described by LaValle and co-workers in 1960 , who prepared it by the reproportionation of molybdenum metal powder and the very volatile MoF6 (b.p. 34 C) . MoF3 is offered commercially by the company American Elements , but I suspect it would be very expensive !

Both CrF3 and MoF3 have the rhombohedral VF3 crystal structure :

My thanks to WebElements for the reproductions of their sketches of chromium(III) fluoride and molybdenum(III) fluoride above (the latter apparently is incorrect) .

It's interesting to note in this regard that although CrF3 has a non-cubic symmetry , its component Cr3+ cations have a regular , undistorted octahedral coordination by the fluoride anions . This suggests that crystalline CrF3 is composed of CrF63- octahedrons , packed into the lattice in the most energetically stabilizing manner possible , which happens to be the rhombohedral VF3 crystal structure . That in turn implies that the hypothetical cubic symmetry perovskite CuCrF3 would be a stable structure comprised of those regular CrF63- octahedrons packed into a cubic lattice .

Another related MX3 host lattice that might be examined in this context of metallic perovskite fluorides is TiF3 , described as a violet-colored (or purplish-brown) crystalline solid (m.p. 1200 C) . It has a crystal structure similar to AlF3 and FeF3 with a rhombohedral symmetry . Ti(III) is a mild natural reducing agent [E0ox ~ 0.055 V to Ti(IV)] , so it would be compatible with Cu0 in the perovskite . TiF3 is paramagnetic , as its Ti3+ cations have the 3d1 electronic configuration . The crystal ionic radius of Ti3+ (6-coordinate octahedral , per Shannon-Prewitt) is 0.67 ; the tolerance factor of the hypothetical CuTiF3 perovskite derived from it is calculated to be t = 0.94 , suggesting it would have a cubic symmetry .

Crystal ionic radii were from the CRC Handbook of Chemistry and Physics , “Ionic Radii in Crystals”, pp. 12-11 to 12-12 . The tolerance factor for CuAlF3 was previously calculated as t = 0.99 , but the radius of the fluoride anion for this calculation was taken as 1.33 , which is for 6-coordinate octahedral fluoride (as in the rocksalt crystal structure , for example) . However , in AMF3 perovskites the fluorides have a 2-fold linear coordination to the M cations . The 2-fold linear value of the fluoride radius isn't provided by the Handbook . In an effort to obtain a more accurate radius value for fluoride anions in perovskites I took a ratio with the known radius values for the oxide anion , making the optimistic assumption that the radii values of fluoride and oxide anions are approximately proportional . The new , improved (I hope) fluoride 2-fold linear radius is thus = 1.33 [6] x 1.21 [2 , oxide] / 1.40 [6 , oxide] = 1.15 [2 , fluoride] .

Recalculating the tolerance factors for the metallic perovskite fluorides discussed above , we now have :

CuAlF3 : t = 1.01 ; there could be some distortion in the cubic structure of this compound ; diamagnetic lattice ; dubious superconductor (Tc close to Absolute Zero , if at all) .

CuCrF3 : t = 0.97 , cubic ; AFM lattice (TN = 80 K) ; possible superconductor , guesstimated Tc ~ 1020 K) ;

CuMoF3 ; t = 0.93 , cubic ; AFM lattice (TN = 185 K) ; possible superconductor , guesstimatedTc ~ 3050 K) ;

CuTiF3 ; t = 0.94 , cubic ; paramagnetic lattice ; no superconductivity at any temperature .

These four perovskites , which could be considered as copper synthetic metals (their metallic bonds are comprised exclusively of the copper atoms' 4s1 orbitals) , would be excellent subjects for a comparative study of a possible modulating matrix effect on the free electrons in a simple metallic bond system .


Tin(III) Metallic Perovskite Fluorides


As there are only a limited number of MF3 reagents suitable for noble metal insertion to form metallic perovskite fluorides , another strategy might be to reproportionate equimolar quantities of MF2 and MF4 to the intermediate M(III) valence state MF3 . Tin is probably the only practical” element that might be suitable for such a reproportionation . The two tin reagents SnF2 and SnF4 are stable (they don't decompose or disproportionate when heated) and are reasonably accessible (ie. are commercially available at a moderate cost) .

Tin(II) fluoride is known to the general public under the tradename Fluoristan™ , used as an anticaries preventative in several brands of toothpaste . SnF2 is a white , crystalline , hygroscopic solid with a low melting point (215 C) and boiling point (850 C) . The crystal structure of SnF2 shows that it is a tinfluorine inorganic polymer with covalent SnF bonds , and comprised of Sn4F8 tetramer units . Each of these tetramer units is in turn formed from two SnF3 and two SnF5 structures , whose Valence Bond description is fairly straightforward (sp3 tetrahedral and sp5 octahedral , respectively) :

I found it quite difficult to draw the Sn4F8 tetramer unit with my chemistry software , so I abstracted a portion of the sketch of the SnF2 crystal structure from the Wikipedia web page , Tin(II) fluoride , for my own drawing above . I thank the author of the original SnF2 sketch , and Wikipedia , for implied permission to reproduce it here on this web page .

Tin(IV) fluoride is a white , crystalline solid [m.p. 442 C , b.p. 705 C (sublimes)] that vigorously reacts with water . It also has an infinite atomic lattice with covalent SnF bonds . The tin atoms have a uniformly octahedral coordination by the fluorines :

The SnF4 lattice consists of layers of fused SnF6 octahedrons ; each layer is offset from its upper and lower neighbour by one fluorine each in the x and y planes . The Valence Bond description of tin(IV) fluoride again is fairly simple :

The tin(IV) fluoride lattice can be thought of as an assembly of conventional SnF4 molecules (like the more familiar SnCl4 molecule) which have been flattened out in their x-y planes ; they are joined together to form the SnF6 octahedrons in a three-dimensional structure by fluorine-to-tin coordinate covalent bonds .

Suppose SnF2 and SnF4 were combined together in equimolar quantities ; would they form a homogeneous , discrete new tin(III) fluoride , SnF3 ?

SnF2 (m.p. 215 C) + SnF4 (m.p. 442 C) ---------> SnF3 .

As shown in the sketches above , the tin atoms in both SnF2 and SnF4 can have an octahedral coordination by neighbouring fluorines . If the tin(III)” atoms in SnF3 are really bonded as tin(IV) , with an exclusive octahedral coordination by fluorines as in SnF4 , the hypothetical SnF3 could adopt the rhenium trioxide crystal structure :

Tin(III) fluoride would be interesting to study from the point of view of its electronic nature . If its Sn(II) and Sn(IV) components were genuinely reproportionated to Sn(III) , it should be a metallic solid having a Sn(IV)F skeletal framework with free , itinerant electrons resonating in its lattice . Its chemical formula could then be represented as Sn(IV)F3 (e) . A suggested Valence Bond electronic structure for this metallic SnF3 is presented in the following sketch :

Tin must use two hypervalent orbitals , together with its four “normal valence shell orbitals (5s2 5p2) when it forms octahedral covalent bonds with various electronegative elements , eg. as in the SnX62- anions [X = halogen] , and in the perovskite BaSnO3 . Since tin is a p-block element , it will most likely use two empty 6p frontier orbitals to combine with its 5 s-p orbitals . The sp5 composite orbital combines the 5s , 5p , and two of the 6p orbitals for the construction of the strong Sn–F covalent bond skeleton of the SnF3 lattice .

Each tin atom has six Sn–F covalent bonds , requiring a total of 12 electrons for completion . Each of the three formula fluorine atoms can provide three of its valence electrons these bonds , or nine in total . Tin provides three of its four valence electrons , to complete the octahedral 12-set . The fourth tin valence electron is promoted up above the skeleton into the 6s and remaining 6pz native frontier orbitals . The voluminous 6s1 orbitals can overlap throughout the lattice to form a tin-only sigma XO (crystal orbital = polymerized MO = metallic bond = conduction band) . The smaller 6pz1 orbitals can overlap continuously throughout the lattice by piggybacking” over the fluorine 2pz2 orbitals , which have a similar shape , symmetry , and orientation as them . This latter Sn 6pzF 2pz composite pi XO could conceivably form a bilayer metallic bond in the SnF3 lattice .

The metallic tin(III) phosphide , SnP , first synthesized by Donohue in 1970 by the high pressurehigh temperature [HPHT] combination of equimolar quantities of tin metal powder and red phosphorus , is an interesting example of a genuine tin(III) compound . Two forms of SnP , both having high electrical conductivities , were produced in Donohue's experiments . The tetragonal form displayed the crystallographic bulging” suggestive of the voluminous tin(II) 5s2 inert pairs . That is , this material contained partially disproportionated tin(II) and (IV) . The cubic rocksalt form of SnP apparently had fully reproportionated Sn(III) , i.e. Sn(IV) + e . Unlike the tetragonal form , it was able to become superconducting somewhere in the range of 2.8 and 4.0 K .

As indicated in the above sketch of the electronic structure of SnF3 , tin(III) fluoride might similarly be a metallic solid if its component Sn(II) and Sn(IV) are fully reproportionated to Sn(IV) + e , as in the cubic rocksalt form of SnP , and if it had a cubic ReO3 crystal structure .

A second possibility is that tin(III) fluoride might form a stable ReO3 lattice , but having disproportionated Sn(II) and Sn(IV) . It would then be a semiconductor or even an insulator whose chemical formula would be better written as Sn(II)0.5Sn(IV)0.5F3 . It could be either a Robin-Day Class II mixed-valent compound (semiconductor) if it had the ReO3 crystal structure as shown , or a Class I compound (insulator) , if its Sn(II) and Sn(IV) atoms were coordinated differently by the fluorines , resulting in a distorted variation of the ReO3 structure . The distortion would be caused by the stereochemically prominent 5s2 pair of electrons in the tin(II) atoms in the lattice .

As one of the heavier p-block elements tin exhibits a noticeable inert pair effect in its divalent state . That is , the 5s2 pair of electrons in tin(II) compounds is often quite stereochemically prominent , causing a characteristic bulging” around the tin atoms and unusually long Sn–X interatomic distances in the lattice , as measured by X-ray diffraction [eg. tin(II) oxide , SnO , which has the litharge crystal structure : GIF image , 39 KB] . A number of tin(II) solid state compounds are known in which the inert pair effect is mysteriously absent ; several of them have a cubic symmetry with no observable bond length distortions . Ng and Zuckerman (1985) have reviewed the chemistry of a variety of compounds of heavy metal p-block elements in which their ns2 inert pairs have seemingly been dispersed into empty , higher energy level frontier orbitals .

The series of cesium tin(II) trihalides , CsSnX3 , is particularly pertinent to a discussion of the hypothetical SnF3 . The following tabulation lists several properties of the CsSnX3 compounds :

While CsSnF3 wasn't investigated , the experience with CsSnCl3 , a white , monoclinic crystalline solid prepared by a chemie douce procedure around room temperature , isn't encouraging for the proposed tin(III) fluoride . When heated it transformed into the yellow cubic perovskite form at 117 C . Its bromide analogue , CsSnBr3 , is a black solid with the cubic perovskite structure at room temperature , and was found to be a semiconductor , with a “metallic-type behavior between –100 and 350 C with no major change in resistance observed” (Scaife , Weller , and Fisher , p. 313) . The substitution of some of the bromides by chlorides attenuated the metallic properties of CsSnBr3 , changing it from black to red and inhibiting its electrical conductivity .

The tolerance factor for CsSnF3 as a perovskite can be calculated from the following data : the crystal ionic radius of Cs1+ , 12-coordinate (as in perovskites) is r = 1.88 , per Shannon and Bierstedt ; that of Sn(II) , 6-coordinate (octahedral) , was estimated as r = 1.16 ; and that of the fluoride anion , 2-coordinate (linear) , was estimated as r = 1.15 . Applying these values in the Goldschmidt equation , the tolerance factor for CsSnF3 is calculated as t = 0.93 , which suggests that it would be a cubic symmetry perovskite . The tin(II) 5s2 inert pairs would not be dispersed in this system , but would spherically surround the Sn(IV) kernel . CsSnF3 would undoubtedly be a white , electrically insulating solid .

A striking example of the dispersion of the ns2 inert pairs of a heavy metal p-block element is provided by the mineral galena , which chemically is lead(II) sulfide , PbS :

The above picture was copied from the Wikipedia web page , Galena . I thank the author of this photo , and Wikipedia , for implied permission to reproduce it here .

Chemically pure lead(II) sulfide has the cubic rocksalt crystal structure (a = 5.936 ) . Its cubic lattice at the atomic level is reflected in its mineral crystal form in which the macroscopic cubes are plainly visible , as in the above picture . Galena is a semiconductor , and was used as the radio frequency detetecting “crystal” in the “crystal sets” in the early days of radio transmission before the introduction of transistors .

One way of looking at the chemical bonding in the CsSnX3 compounds and in galena is to consider their Sn–X and Pb–S bonds , respectively , as being covalent . The puzzle is then to try to determine into which empty frontier orbitals the tin 5s2 and lead 6s2 inert pairs have been dispersed . Another simpler and more straightforward approach is to consider CsSnX3 and PbS to be ionic solids with no covalent bonding . Then the answer is , quite simply , that the inert pairs don't go (or are dispersed) anywhere ! The tin(II) and lead(II) in these compounds would be the Sn(IV) and Pb(IV) kernels , surrounded by their spherical 5s2 and 6s2 inert pairs . Another model (other than Valence Bond) must then be devised to describe the extraordinary physical properties , such as the remarkable colors and semiconductivity , of the CsSnX3 compounds and galena .

Undoubtedly charge transfer occurs quite extensively in PbS , as shown by the following redox equations . Note that sulfide anion is a natural reducing agent :

S2- – 2e -------> S0 ; E0ox = 0.476 V ;

Pb2+ + 2e -------> Pb0 ; E0red = – 0.1262 V ;

Net reaction : Pb2+ + S2- -------> Pb0–S0 ; E0T = 0.3498 V .

The positive cell potential E0T for the Pb2+ / S2- interaction indicates that the redox reaction as written would be thermodynamically spontaneous at STP , and suggests that PbS (the pretty galena crystals shown in the Wikipedia picture above) is actually composed of a mass of zerovalent lead and sulfur atoms bonded together electrostatically ! In fact , this may not be too far from the truth :

“The direct integration of charge density , the observed atomic scattering factors , and the population analysis of the valence electrons all indicate that the lead atom is negatively charged , i.e. electrons are transferred from sulfur to lead [in galena] ” – Ng and Zuckerman (p. 313) .

Very likely charge transfer from the sulfide to the Pb2+ cations is only partial , and not quantitative as implied in the above redox equations . The transfer may also involve a certain amount of valence electron resonance , and such a resonance (Mott-Wannier excitons) may cause the bluish-gray color and metallic luster of macroscopic crystals of galena mineral (powdered reagent PbS is black) .

In the CsSnX3 compounds the component Cs1+ and Sn2+ cations and halide anions are all low energy redox species (neither oxidizing nor reducing in nature) , so a redox-mediated charge transfer from halide to tin – or vice-versa – can't be the basis of the unusual physical properties of these remarkable tin(II) perovskites .

In any case , we don't have the luxury of the simple ionic model for chemical bonding in the hypothetical SnF3 . Its tin atoms really are tin(IV) , which are very small [crystal ionic radius = 0.69 , 6-coordinate octahedral , per Shannon and Prewitt] and highly electrophilic . As such , their Sn–X bonds are always covalent in nature . I don't know of any ionic Sn(IV) salts , do you ? The Valence Bond approach , as discussed above , is probably the best way to analyze the covalent bonding in the Sn–F skeletal framework , and thereby derive a reasonable approximation of what and where the metallic bond in it might be .

The intriguing properties of the CsSnX3 perovskites suggests another interesting project related to them , and to the proposed reproportionation of SnF2 and SnF4 to SnF3 . Would the entire series of SnX3 + SnX3 -----> SnX3 provide analogous results to the the CsSnX3 compounds , that is , with optimum metallic properties displayed in SnBr3 ? The following tabulation lists several physical properties of the tin(II) and (IV) halides :

Equimolar quantities of each (II)(IV) halide pair would be combined together , perhaps by simple melting in a round-bottom flask under dry nitrogen (the tetrahalides are highly sensitive to water , with very rapid hydrolysis ; and the dihalides are quite hygroscopic as well) . Only a white , electrically-insulating product would probably be obtained in the fluoride case . However , if the supposed SnF3 was subjected to HP–HP conditions , its Sn(II)–(IV) might be forcibly reproportionated to metallic Sn(III) in an ReO3 lattice (cf. SnP , Donohue) . All eight of the tabulated tin reagents are commercially available , eg. from Alfa-Aesar ; the tetrahalides are moderately expensive , except for the commonest of them , tin tetrachloride , a versatile electrophilic Lewis acid catalyst in organic synthesis .

A new family of metallic halide perovskites based on these SnX3 compounds could be investigated , in which zerovalent metal atoms would be inserted into them , as proposed above for the Transition metal MF3 host lattices . The crystal ionic radius of tin(III) is unknown in any coordination environment . That of tin(II) for CN = 8 is 1.36 . “Normalizing” it for CN = 6 (octahedral , as in the MSnX3 perovskites) : 1.36 x 0.69 [Sn(IV) , CN = 6] / 0.81 [Sn(IV) , CN = 8] = 1.16 . The crystal ionic radius for Sn(III) is assumed to be the average of that for Sn(II) and Sn(IV) , with CN = 6 : (0.69 + 1.16) / 2 = 0.925 .

Using the Goldschmidt equation for calculating the tolerance factors (t) for perovskites , we see that because of the relatively large crystal ionic radius of Sn(III) , only the p-block heavy metal elements might form cubic symmetry perovskites with SnF3 :

Hg0, 6s2 , m.p. –39 C : metallic radius = 1.62 ; for Hg0SnF3 , t = 0.94 ;

Tl0, 6s2 6p1 , m.p. 304 C : metallic radius = 1.70 ; for Tl0SnF3 , t = 0.97 ;

Pb0, 6s2 6p2 , m.p. 327 C : metallic radius = 1.75 ; for Pb0SnF3 , t = 0.99 ;

Bi0, 6s2 6p3 , m.p. 271 C : metallic radius = 1.82 ; for Bi0SnF3 , t = 1.01 .

The synthesis of HgSnF3 would best be carried out in a high pressure and temperature (HPHT) apparatus to safely contain the toxic mercury vapor :

Hg0 (b.p. 357 C) + SnF2 (m.p. 215 C) + SnF4 (m.p. 442 C) ----- [HPHT] ----> Hg0SnF3 .

The interactions between the heavy metal valence shell electrons and those of the metallic tin(III) in these novel tin halide perovskites would be quite interesting to study . The large zerovalent atoms might actually stabilize the formation of the perovskite structures around them by a sort of “template effect”.


Zinc and Cadmium(0–II) Fluorides


Zinc is an interesting – and challenging ! – element for incorporating into metallic solids , and possibly superconductors . It is an inexpensive , abundant , and readily available base metal , and so is appealing from an economic point of view . Zinc is a fairly strong reducing agent (E0ox = 0.7618 V , Zn0 to Zn2+) , so it's difficult to formulate with many container crystal structures . However , mixed-valent , low-valent zinc fluorides might be available from solid state syntheses .

For example , the combination of (Zn0)0.5 + (Zn2+)0.5 = Zn1+ could be incorporated into a fluoride perovskite as the large A cation in the centers , with a small B cation such as Ca2+ :

Zn0 (m.p. 420 C , b.p. 907 C) + ZnF2 (m.p. 872 C) + CaF2 (m.p. 1418 C)

----- [HPHT] ----> Zn1+Ca2+F3 .

Assuming the crystal ionic radius of Zn1+ is about the same as that of K1+ (both are 4s1 electronically) , ie. r = 1.64 (CN = 12) , and r = 1.00 for Ca2+ (CN = 6) , the tolerance factor for ZnCaF3 is calculated as t = 0.92 , indicating a cubic symmetry for it .

The mixed-valent Zn0Zn2+Zn0, having low-valent Zn0.67+, might be incorporated into a fluoride antiperovskite , having a large central anion (iodide) , and smaller corner atoms (fluorides) :

2 Zn0 (m.p. 420 C , b.p. 907 C) + ZnI2 (m.p. 450 C , b.p. 625 C) + ZnF2 (m.p. 872 C)

----- [HPHT] ----> I1F1(Zn0.67+)3 .

Fluoride spinels having Zn1+ as the larger octahedral cation might be investigated :

MgF2 (m.p. 1263 C) + Zn0 (m.p. 420 C , b.p. 907 C) + ZnF2 (m.p. 872 C)

----- [HPHT] ----> (Mg2+)tet(Zn1+Zn1+)octF4 .

The mixed-valent Zn0Zn0Zn0Zn2+, having low-valent Zn0.5+, might be incorporated into a fluoride anti-spinel :

3 Zn0 + ZnF2 + ZnI2 ----- [HPHT] ----> FtetIoct(Zn0.5+)4 .

The smaller fluoride anion would be tetrahedrally coordinated by the zinc(II) kernels , while the much larger iodide anion would be octahedrally coordinated by them . The small zinc cations would be tetrahedrally coordinated by the halide anions .

Cadmium is a heavier analogue to zinc in the IIB/12 family of elements . It is a much rarer , more expensive commodity than zinc , and is considered to be a toxic heavy element , like mercury and lead . Cd2+ is considerably larger than Zn2+ (its crystal ionic radius per Shannon and Prewitt , six-coordinate , is 0.95 versus 0.74 for Zn2+) , and it generally prefers an octahedral coordination by anions and other ligands . Twelve-coordinate Cd2+ is quite large (r = 1.31 ) , and can be used as the central A cation in perovskites . Thus , Cd1+ [from (Cd0)0.5 + (Cd2+)0.5] should also be satisfactory for this purpose , combined with a smaller B cation such as Ca2+ :

Cd0 (m.p. 321 C , b.p. 767 C) + CdF2 (m.p. 1075 C) + CaF2 (m.p. 1418 C)

----- [HPHT] ----> (Cd1+)A(Ca2+)BF3 .

Assuming the crystal ionic radius of Cd1+ is about the same as that of Rb1+ (both are 5s1 electronically) , ie. r = 1.72 (CN = 12) , and r = 1.00 for Ca2+ (CN = 6) , the tolerance factor for CdCaF3 is calculated as t = 0.94 , indicating a cubic symmetry for it .

Fluoride spinels with octahedral Cd1+ are conceivable , such as (Zn2+)tet(Cd1+Cd1+)octF4 :

ZnF2 (m.p. 872 C) + Cd0 (m.p. 321 C , b.p. 767 C) + CdF2 (m.p. 1075 C)

----- [HPHT] ----> (Zn2+)tet(Cd1+Cd1+)octF4 .

Zinc and cadmium thus look promising for the design and synthesis of a variety of fascinating new materials as metallic solids and possible superconductors (the latter only with low transition temperatures , though) .


Mercury(0) Fluoride Perovskites


The tolerance factor for Hg0SnF3 [above] was calculated as t = 0.94 , suggesting it would have a cubic symmetry if it was a perovskite . In this compound the 6s2 inert pairs would not be promoted into the 7s orbitals ; they would remain spherically surrounding the Hg2+ kernel .

Suppose Hg0 was inserted into the lattices of FeF3 and AlF3 in an equimolar proportion . The metallic radius of Hg0 , 1.62 , is quite large . The crystal ionic radii (per Shannon and Prewitt) of Fe3+ and Al3+ (6-coordinate) are 0.55 and 0.54 , respectively . The crystal ionic radius of 2-coordinate fluoride anion (ie. linear coordinating) was estimated above as 1.15 . Applying these values in the Goldschmidt equation , the tolerance factors for the hypothetical perovskite fluorides Hg0FeF3 and Hg0AlF3 are respectively calculated as t = 1.15 and 1.16 . These tolerance factors are well above the limits for a cubic symmetry crystal structure . It would be so distorted as to no longer resemble a perovskite .

However , suppose the FeF3 and AlF3 host lattices accepted the mercury atoms , not as Hg0 but as Hg2+ , with the 6s2 inert pairs promoted into the 7s orbitals . The crystal ionic radius of 8-coordinate Hg2+ is 1.14 . The 12-coordinate radius of Hg2+ – as in a perovskite – can be estimated by comparing it to that of K1+ : 1.14 x 1.64 (12-coordinate K1+) / 1.51 (8-coordinate K1+) = 1.238 (12-coordinate Hg2+) . The tolerance factors of Hg2+FeF3 and Hg2+AlF3 are now calculated as t = 0.99 and 1.00 respectively , which are comfortably in the cubic symmetry range for perovskites .

At this point two different models or “pictures” of mercury's promoted 6s0 inert pairs must be considered . In the first model they will be relocated into the empty 7s frontier orbitals above the mercury atoms . The wave functions of the inert pair electrons will be physically voluminous as they are coupled to the wave functions of the kernels . Where in the lattice can they be located , since all the 12-coordinate voids are occupied by the mercury atoms ? A possible solution to this problem might be to insert only half of an equivalent of Hg0 into one equivalent of the FeF3 and AlF3 . That will leave the second half of an equivalent of 12-coordinate void spaces for the promoted inert pairs [denoted as (e22)] :

Hg0 (b.p. 357 C) + FeF3 (sublimes >1000 C) ------ [HPHT] ------> Hg2+0.5(e22)0.5FeF3 ;

Hg0 + AlF3 (sublimes ~12601291 C) ------ [HPHT] ------> Hg2+0.5(e22)0.5AlF3 .

In this picture two factors should assist in the promotion of the inert pairs into the Hg 7s orbitals in these compounds : first , the thermodynamically stabilizing cubic symmetry of the Hg2+FeF3 and Hg2+AlF3 perovskite structures ; and second , provision of spacious , comfortable “homes” for the promoted inert pairs in the host lattices' 12-coordinate voids .

In the second picture , when the inert pairs are promoted into the 7s frontier orbitals they rise above the Fermi level , EF , in the sigma XO metallic bond (this concept is discussed in the inert-pairs Chemexplore web page) . They are thus quantitatively converted into Cooper pairs , and Hg0FeF3 and Hg0AlF3 become superconducting at a low temperature . In each Cooper pair the electrons have identical wave functions , but as they are in a precise antiparallel orientation with respect to each other , their wave functions exactly cancel each other . That is , the Cooper pair electrons have lost their wave properties and now are simply charged particles .

As pointed out in several earlier Chemexplore web pages , these electrons resemble tiny spherical magnets with a north and south pole (GIF image , 12 KB) . Their magnetic attractive force is vastly stronger (possibly 1200–2400 times , GIF image , 34 KB) than their repulsive electrostatic force when they are in a perfectly antiparallel orientation , as they are known to be in Cooper pairs . This attractive magnetic force is the universal prerequisite – “cause” – for superconductivity in all physical and chemical conditions . The inert pair electrons are thus magnetically bonded , as charged particles without wave properties , as the Cooper pairs . Of course , electrons are fantastically small physical entities , immensely smaller than atoms , so they require virtually no physical space in the lattice , certainly not the 12-coordinate voids occupied by the Hg2+ kernels . They will resemble the “electron gas” first hypothesized by the German physicist Paul Drude (1863-1906) in his groundbreaking electron theory of metals in 1900 .

Because such an electron gas , now of Cooper pairs rather than the singlet electrons in conventional electrical conduction , has essentially no spatial requirements , it should be possible to synthesize Hg0FeF3 and Hg0AlF3 as cubic symmetry perovskites , using equimolar quantities of the Hg0 and FeF3 and AlF3 precursor reagents . The Fe3+ (E0red = 0.771 V to Fe2+) is incapable of oxidizing the Hg0 (E0ox = 0.851 V to Hg2+) , and of course Al3+ is neither an oxidizer nor reducer .

The two hosts FeF3 and AlF3 are suggested for a comparative study of the mercury insertion . As mentioned above , iron(III) fluoride is antiferromagnetic (AFN) , with a fairly high Nel temperature of TN = 394 K , while AlF3 is non-magnetic , with diamagnetic Al3+ cations . Would the AFN FeF3 framework assist the inert pairs in remaining coupled together magnetically , even at very high temperatures – possibly even at ambient temperature ! – as predicted by the AFN induction enabling mechanism of HTS ? The AlF3 host , as a framework that wouldn't provide any sort of AFN induction into the inert pair electrons , is proposed as a sort of “baseline comparison” to FeF3 .

In the discussion above of the CsSnX3 compounds the tolerance factor of CsSnF3 was calculated as t = 0.93 , indicating that it would likely be a cubic symmetry perovskite . The Sn(II) cation in the compound is “comfortable” and unstressed , so its 5s2 inert pair would remain spherically surrounding the Sn(IV) kernel . Suppose the analogous hypothetical compound NaSnF3 was synthesized under HPHT conditions , intended to “pop” the 5s2 inert pairs up into the 6s frontier orbitals :

NaF (m.p. 996 C) + SnF2 (m.p. 215 C) ------ [HPHT] ------> NaSn(IV)(e22)F3 .

The tolerance factor for NaSn(IV)(e22)F3 is calculated from the following data : the crystal ionic radius of Na1+ , 12-coordinate (as in perovskites) is r = 1.39 , per Shannon and Bierstedt ; that of Sn(IV) , 6-coordinate (octahedral) , is r = 0.69 , again per Shannon and Bierstedt ; and that of the fluoride anion , 2-coordinate (linear) , was estimated as r = 1.15 . Applying these values in the Goldschmidt equation , the tolerance factor for NaSn(IV)(e22)F3 is t = 0.98 . This suggests that if the “normal” compound NaSn(II)F3 was strongly heated and compressed during its formation in an anvil press the cubic symmetry perovskite NaSn(IV)(e22)F3 might form , with the promotion of the 5s2 inert pairs up into the 6s frontier orbitals . That would make the compound a metallic solid , and possibly superconducting , maybe even at a very high temperature if the inert pairs were able to retain a substantial degree of their “two-ness” nature .

Only NaSn(IV)(e22)F3 seems to have a cubic symmetry ; the tolerance factors of KSn(IV)(e22)F3 (t = 1.07) and LiSn(IV)(e22)F3 (t = 0.86) are outside the cubic symmetry range .

The chloride analogue NaSnCl3 would also be quite interesting to prepare and study :

NaCl (m.p. 801 C) + SnCl2 (m.p. 247 C , b.p. 623 C) ---- [HPHT] ----> NaSnCl3 .

The two precursor reagents NaCl and SnCl2 are quite common and inexpensive .

The crystal ionic radius of six-coordinate Cl is 1.81 ; that of two-coordinate , linear Cl is estimated as 1.564 by taking a proportionality ratio with six- and two-coordinate oxides . The tolerance ratio for the perovskite NaSn(II)Cl3 is then calculated as t = 0.77 , well below the cubic symmetry range . That of NaSn(IV)(e22)Cl3 is calculated as t = 0.93 , in the middle of the cubic symmetry range . It should be possible to synthesize this latter cubic form under HPHT conditions . What makes this (and the other metallic fluorides discussed in this web page) so interesting is what happens to them when they are cooled to room temperature and the pressure is released . Will they be stable under STP conditions ?

NaSn(IV)(e22)Cl3 would undoubtedly be a metallic solid and possibly even a superconductor . However , at room temperature the electron gas (the tins' 5s2 inert pairs) might infiltrate back through the lattice to bond again with the electrophilic kernels as stereochemically prominent inert pairs . The chloride anions are redox-inert , neither oxidizing nor reducing in nature . Unlike chemically reducing chalcogenides like sulfide they can't neutralize any oxidizing nature of the tin kernels . Fortunately , as pointed out above tin(IV) is only a very mild oxidizer , but even so NaSnCl3 might have to be cooled substantially in order to observe the appearance of superconductivity in it . It may be an “all-or-nothing” proposition : the compound will be either metallic NaSn(IV)(e22)Cl3 , superconducting at some temperature , or nonmetallic and insulating NaSn(II)Cl3 when cooled to ambient conditions from synthesis in the anvil press .

This redox-related problem becomes more acute with the mercury compounds , in which the Hg2+ kernels are even more strongly electrophilic and oxidizing than the rather mild and innocuous tin(II) kernels . It may be that only chemically-reducing anion ligands such as sulfide can successfully be used with mercury(0) systems to “neutralize” the oxidizing nature of the Hg2+ kernels and so permit a promotion of the 6s2 inert pairs up into the 7s orbitals , thereby producing the electron gas for the metallic bond and for the appearance of superconductivity in the material .

The exploration of these novel fluoride (and generally , halide) solid state systems has the potential to provide researchers with many new fascinating metallic materials for future study and development . An investigation of the two systems outlined above , Hg2+0.5(e22)0.5FeF3 and Hg2+0.5(e22)0.5AlF3 versus Hg2+(e22)FeF3 and Hg2+(e22)AlF3 , might in particular provide some revealing insight into the quantum nature of the Cooper pair electrons . If they can be prepared and prove to be stable at STP , the compounds NaSn(IV)(e22)F3 and NaSn(IV)(e22)Cl3 might also be superconducting .


References and Notes


Hargittai and co-workers : M. Hargittai , M. Kolonits , J. Tremmel , J.-L. Fourquet , and G. Ferey , “The Molecular Geometry of Iron Trifluoride from Electron Diffraction and a Reinvestigation of Aluminum Trifluoride”, Struct. Chem. 1 (1) , pp. 75-78 (1990) .

dehydrated by a reagent : G.W. Watt , P.S. Gentile , and E.P. Helvenston , “Synthesis of Anhydrous Metal Halides”, J. Amer. Chem. Soc. 77 (10) , pp. 2752-2753 (1955) . The preparation of anhydrous chromium(III) chloride from CrCl3.6H2O by refluxing with thionyl chloride is described by G.G. Schlessinger , Inorganic Laboratory Preparations , Chemical Publishing Co. , New York , 1962 ; p. 20 :

This valuable inorganic chemistry synthesis compendium can be downloaded for free from the library resources web page [DJVU , 4990 KB ; a suitable DjVu reader for your computer can be downloaded for free from . The WinDjView reader v. 1.0.3 for older FAT 32 Windows OS can be downloaded for free from FileHorse] .

LaValle and co-workers : D.E. LaValle , R.M. Steele , M.K. Wilkinson , and H.L. Yakel Jr. , The Preparation and Crystal Structure of Molybdenum (III) Fluoride”, J. Amer. Chem. Soc. 82 (10) , pp. 2433-2434 (1960) .

Donohue : 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) .

inert pair effect : A.R. West , Basic Solid State Chemistry , John Wiley , New York (1988) ; pp. 106-107 ; idem. , Solid State Chemistry and Its Applications , John Wiley , Chichester (UK) , 1984 ; pp. 314-315 . Inert pairs of electrons in inorganic compounds are similar to the lone pairs of electrons in both inorganic and organic compounds (for example , the two lone pairs on the oxygen atom of the water molecule) . The presence of inert pairs in a crystal structure is a reliable diagnostic of covalent bonding in it .

Ng and Zuckerman : 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 .

Scaife , Weller , and Fisher : 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) ; 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) .

Goldschmidt equation : U. Mller , Inorganic Structural Chemistry , John Wiley , Chichester , UK , 1993 ; p. 200 ; A.F. Wells , Structural Inorganic Chemistry , third ed. , Clarendon Press , Oxford , UK , 1962 ; p. 497 . It's also mentioned in the WolfWikis web page , “Perovskite”.

Mott-Wannier excitons : anon. , Excitons Types , Energy Transfer, Organic Optoelectronics , Lecture 7 , 24 pp. , MIT , Harvard (MA) ; February 27th , 2003 [PDF , 424 KB] .



[ Index Page ] [ Contact ]