Permanganyl Fluoride: A Brief History of the Molecule MnO3F and of Those Who Cared For It

Abstract Permanganyl fluoride's existence at the stability threshold in the series of oxides and oxide fluorides of the late 3d transition metals is reflected by its experimentally challenging properties and by the difficulties posed in the theoretical description of its bonding characteristics. The history of this molecule is reviewed from early qualitative observations and the growing scattered information on its chemical and physical properties to the accurate determination and interpretation of its molecular structure and spectral features. The still problematic theoretical models for MnO4 − and MnO3F are briefly presented in the broader context of the chemistry of elements in high oxidation states. Short biographies of the scientists engaged in these studies are offered. Related technetium and rhenium compounds are briefly considered for comparison.


Introduction
In ar ecent publication, [1] the oxo chemistry of group VIII element iron in its highest oxidation states has been reviewed focusing in the end on the "non-existence" of iron tetroxide FeO 4 ,w hich could be the first and only example of am olecule containing iron it its highest oxidation state Fe VIII .I ts absence from recordsofs uccessful experiments is particularly intriguing since the tetroxides of the heavierc ongeners of iron are well documented as stable molecules RuO 4 and OsO 4 .T he absence of the cornerstone FeO 4 is evident not only from the series of the transition metal oxides (Scheme 1a), but also from that of the corresponding oxide fluorides (Scheme 1b). As imilar stability barrierl ike that between the iron tetroxide and the acid anhydride and acid halides of permanganic acidH MnO 4 does not exist for the heavier elements of groups VII and VIII (Tc, Ru and Re, Os).
The two manganese(VII) compounds in the two series, Mn 2 O 7 and MnO 3 F, were both first observed in the early 19th century. They could not be missed because both have deep colors, which quickly disappear in fierce explosions developing ad ark browns moke when samples are provoked by shock or heating. While the purple color of Mn 2 O 7 differs only little from that of the permanganates M I [MnO 4 ], the salts of the parent acid HMnO 4 ,t he short-lived green color of MnO 3 Fh as been particularly intriguing to experimentalists, and later also to theoreticians. In contrastt ot he chemistry of Mn 2 O 7, which appears in all textbookso fi ntroductory inorganic chemistry, information on permanganyl fluoride has remained more scattered. The presentr eview therefore tries to summarize both the history and the present state-of-the-art of this "weird molecule". As in the text on the oxo chemistry of iron, [1] the scientists engaged in the few studies carried out overaspan of two centuries are also briefly introduced, because their biographies reflect the genealogical tree of brave chemists who had as pecial predilection for the challenge of workingw ith this kind of compounds. Molecules such as MnO 3 Fo nt he one hand fit well into the framework of simple models of structure and bonding, while on the other hand they pose ac hallenge both for the experimentalist and the theoretical chemist. In af inal chapter therefore the status of the theoreticald escription of the electronic structure is up-dated here.
To complement the story,a nd to place the information in a broader context,t he literature on the corresponding chloride, MnO 3 Cl, ando nt he technetium analogue, Tc O 3 F, have also been included in ac oncise format. The chemistry of rhenium(VII)c ompoundsi sn ot considered because it follows more conventional rules and is devoid of similari rregularities.

AB rief History of Permanganyl Fluoride, MnO 3 F
The acid halides of permanganic acid have remained what is widelyc alled laboratory curiosities for almost two centuries. According to two reports published in 1828, it was Friedrich Wçhler ( Figure 1, right) at the University of Gçttingen who first observed ad eepg reen color which appeared upon treatment of purple potassium permanganate with strong acid containing fluoride,t hat is,b yr eacting am ixture of KMnO 4 and CaF 2 with sulfuric acid [Eq. (1)].T he acid took on ag reen color,a nd ag reen vapor was evolved which exploded to give ad eepbrowns moke. [2,3] After these early observations Wçhlern ever revisited this compound. It had previously long been known that potassium permanganate KMnO 4 when treated with sulfuric acid affords the highly explosive anhydride Mn 2 O 7 of permanganic acid [Eq. (2)]. [4] Therefore the instability of the acid fluoridew as not unexpected. The first more extensive descrip- tion of the properties of Mn 2 O 7 was provided by Oskar Glemser of the University of Gçttingenin1 953 ( Figure 2). [5] Half ac entury after Wçhler's observations, GeorgeG ore FRS (Figure 3), workinga tB irmingham,r eported that "damp crystals of permanganate of potassium hissed and dissolved to a Hubert Schmidbaur received his diploma and PhD in chemistry at the Universityo fM unich (LMU)a nd startedh is academic career at the universities of Marburg and Würzburgb efore he joined the Technical Universityo fM unich (TUM)i n1 973, whereh es erved to his retirement in 2003.A sa ne meritus he since had numerous appointments abroad. Following work in diversef ields of inorganic,o rganometallic and bioinorganicc hemistry he has dedicated much of his research to the chemistry of gold.
W. H. Eugen Schwarz received ad iploma in experimental physical chemistry from Hamburg University and aP hD in natural philosophy from Frankfurt University.F rom 1976 to 2002 he was professor of theoretical chemistry at SiegenU niversity.H ew orked on quantum chemical theory and applications (atomic core potentials, relativistic effects etc.), on the elucidation of basic chemical concepts (bonding, periodic table) and on the history and philosophy of chemistry.A sa n emeritus professor he is still active at the theoretical chemistry center of Tsinghua University Beijing.  1 604-1670) lived, studied,and worked as ap harmacist and chemicalengineer at various places in the German speakingE uropean countries. He first observed the deeply colored permanganates in 1659, whenher eacted the mineral pyrolysite (MnO 2 )w ith sodium carbonate in air.Heo btained ameltcontainingt he green sodium manganate Na 2 MnO 4 ,w hich was converted upon neutralizationi nw ater into sodium permanganate NaMnO 4 with acolor he describeda s" elegans colorp urpureus". Crystallization was optimizedw ith the addition of pottash to give KMnO 4 . [6] Its crystal structure has been determined by Gus Palenik (University of Waterloo) in all its details as late as 1967. In the crystal, the slightly distorted tetrahedral anion has MnÀObondsof1 .607(5) . [7] No further biographic details of Johann F. Glauber and of Friedrich Wçhler (right, 1800-1882, afounder of the modern chemical sciences) are describedh ere becausetheir brief encounterswith the subjecta re only small anecdotesi n their immense contributions to science.  Glemser (1911Glemser ( -2005 studied Chemistry at the Te chnische HochschuleS tuttgart and held positionsa tthe RWTH Aachen before he took the chairofi norganic chemistryatG çttingenw here he dedicated much of his work to fluorine chemistry, attracting inparticular several young talentst othis field (see below). As the president of the GçttingenA cademy of Sciencesand of the German Chemical Society, in his time he was one of the most prominent figures in chemistry. . GeorgeGore (1826-1908) was born as the son of aw orkman in Bristol and as ayoungboy was an apprenticet obecome ac opper-and blacksmith.F ascinated by the wonders of science, he started educating himself and soon was drawn to the city of Birmingham with her rapidly growing chemical industries. Therehef irst joined aphosphorus factory,a nd the invention of the "safety match" has been attributed to him for the work he carried out there. This and other discoveries in typefounding,d yeing, electroplating and other trades madehim afortune. He was laterappointed director of an institute of scientific research in Birmingham where he guided manyp rograms of chemical technology.Exceptfor apension for his daughter,hededicatedh is rich belongings to the Royal Societyand the Royal Institution. green solution" when reacted with hydrofluoric acid [Eq. (3)]. [8] This experiment was just one of the studies he carried out with hydrofluoric acid, ap roduct which he commercialized. His biography shows that this observation was by no means Gore's greatest contribution to science andt echnology.
It then took another half century until Otto Ruff (Figure 4, left), then working at the University of Danzig,i nh is first studies of fluorosulfonic acid included the reaction with permanganate. [9] He againo bserved that the evolvedv apors,w hich sometimes were violet, exploded vehemently upon shock and slight warming.I tl ed him to assume that the volatile product was only manganese heptoxide and would probablyn ot contain any fluorine.
While Otto Ruff thus missedo ut in his search for MnO 3 F, Karl Fredenhagen ( Figure 4, right), workinga tt he Universityo f Greifswald, 20 years later confirmed the formation of green vapors from the reaction of KMnO 4 with hydrofluoric or fluorosulfonica cid that mayb ef ormulatedb yE quations (3) and (4a/ 4b). [10] He assigned the color to am anganese oxyfluoridew ith manganese in the oxidations tate + VII, but gave no formula. It shouldb en oted that as econd publication (of 1939) has Helmut Fredenhagen as the author,w ho also workeda tG reifswald and who thanked Karl Fredenhagen fort he guidance. [11] KMnO 4 þ HSO 3 Relevant work carried out during and after World WarI Ii n the Greifswald laboratories was published in 1950 by Kurt Wiechert (Karl Fredenhagen had died in 1949). [12] As described therein,asecond methodw as employed for the generationo f MnO 3 Fs tartingt his time from manganese metal which was treated with potassium nitrate in hydrogen fluoride. Both green vapors and green solutionsw ere again observed. Fractional condensation of the volatiles at À78 8Ca fforded ad eepgreen, almost black solid which melted upon warming to À45 8 to ag reen liquid and decomposed violently near 0 8C. The elemental analysis of the condensate gave an atomic ratio of Mn:F = 1:1w ithp ercentages approaching those calculated for MnO 3 Fc onfirming finally all previous suggestions. However, reconsidering the properties given for the compound at this stage, the products still were not really pure MnO 3 F.
In 1950, Aynsleigh,P eacocka nd Robinson of the University of Newcastle-upon-Tyne studied the reactionofe lementalfluorine, diluted with nitrogen, with manganese oxides and potassium permanganate at 150 8Ca nd found no evidence for av olatile MnO 3 F. [13] This result led them to question the earlier evidence by Wçhlera nd all the othersm entioned above.T his failure no doubt was due to the appliede levated temperature well beyond the stability range of MnO 3 F.
Only af ew years later (1954), Alfred Engelbrecht and Aristid von Grosse ( Figure 5), workinga tT emple University in Philadelphia, [14] were finally able to isolate MnO 3 Fi napure state and to determine its fundamentalp roperties.T he compound was obtained from potassiump ermanganate and fluorosulfonic acid in am olar ratio of 1:4i nacopperf lask cooled to À78 8C, and separated in av acuum by fractional condensation [Eq. (4b) with an excessofH SO 3 F].Int he reaction, HF is formed in small quantities presumably through the reactionf ormulated in Equation (5).
These traces of HF remaining in the green condensate were removed by the addition of KF which absorbs HF to formK HF 2 but is inert towards MnO 3 F. The green crystals melt at À38 8C and the liquid has av apor pressure of 22.5 mm at À15 8Ca nd of 52 mm at 0 8C. The extrapolated boiling point is + 60 8C. When kept above 0 8C, the compound decomposess lowly,o r occasionally violently,w ith formation of af lame and ab rown smoke. In moist air,v iolet vapors of Mn 2 O 7 appear prior to violent explosions. With excess water more carefully applied, HMnO 4 and HF are formed [Eq. (6)].T his hydrolysisw as used for the elemental analysis of the compound, which gave the  Ruff (left, 1871Ruff (left, -1939 in his time was one of the protagonists of fluorine chemistryinE urope.H eh ad started as ap harmacist, and obtained his PhD with Oskar Piloty in Berlin, but then first joinedE mil Fischer who had newly arrived in Berlin at the turn of the century.After ab rief period with Wilhelm Ostwald in Leipzig, where he received important training in physical chemistry, he returned to Berlinbefore he was awarded his first chair of inorganic chemistryatD anzig (now Poland). His work on fluoroand chlorosulfonic acids was only aminor activity at at ime when he also workedn ot only on the synthesis of diamonds, but also isolated uranium hexafluoride for the first time (1909). After 10 years he took achairatthe University of Breslau (now Poland), where he retired. Ruff's carefully documenteddata on UF 6 founded the basis for much of the laterw ork carried out under the Manhattan Project in the early 1940s in the US, where Aristid von Grosse was one of the prominents cientists (seebelow). Karl Fredenhagen (right, 1877-1949) was born in Loitz (nearG reifswald) in 1877, and had studiedc hemistryatt he University of Gçttingena nd receivedh is PhD with Walther Nernst. He later workeda lso with WilhelmO stwald at the University of Leipzig.H is most widely known work on hydrogen fluoride ledtoanew processf or the production of elemental fluorined ocumented in av aluable patentin1928.I ny et another area of his broad interests, he was one of the pioneers in the development of potassium graphite KC 8 whichu pt ot he presenti sa nextremely important material in the batteryi ndustry and related technologies. expectedv alues for Mn andF .I ti sr emarkable that the authors dared to prepareu pt o4 0gquantities (!)o ft his highly explosive material.
In later work, both von Grosse and Engelbrecht, in collaboration with A. Javan of ColumbiaU niversity in New York, studied the microwaves pectrum of MnO 3 F, and the molecular constants of the symmetric spinning top molecule (point group C 3v )h ave been determined. [15][16][17] In the gas phase the molecule shows three symmetry-related MnÀOd istances of 1.586 and an MnÀFd istance of 1.724 ,w ith O-Mn-F angles of 108.278. The quadrupole hyperfine structure confirmed the nuclear spin of 5/2 for the only stable 55 Mn isotope( 100 %a bundance)w ith aq uadrupole couplingc onstant of eqQ = 16.8 Mc sec À1 and a nuclearq uadrupole moment of Q = 0.55 10 À24 cm 2 .S imilar studies were carriedo ut for perrhenylfluoride ReO 3 F. [18,19] In more recentw ork, the electrical field gradients of the metal atom in MnO 3 Fw erec alculated using density functional theory by Bjornssona nd Bühl. [20] The question arises why this highly decorated protagonist of NuclearC hemistry would care for the synthesis and properties of permanganyl fluoride in 1953. An answer to this question becomes apparent from the social contacts during his days at the Kaiser-Wilhelm (now Max Planck) Institute in Berlin. It was there that he also met Ida Tacke and Walter Noddack who had discovered (together with O. Berg) the element rhenium in 1925, one of the two elements that Mendelejev had predicted half ac entury ago as the yet to be discovered congeners of manganese. Eka-manganese was still missing and was discovered only in 1937 and namedt echnetium. Manganese was the prototypeo ft he intriguing triad Mn-Tc-Re, andw hen finally all three elements had become availablei tw as of natural interest to as enior scientist in the field to study all the fundamentals of their chemistry.I ti sf air to say that all later work on the elusive MnO 3 Fm olecule was based on the experimentalp rogress made by A. v. Grosse and A. Engelbrecht. These two authors also first produced pure chromyl fluorideC rO 2 F 2 ,w hich is even more volatile than the "isoelectronic" and "isosteric" permanganyl fluoride, buta lso much more stable being nonexplosive. [21] The orange crystalso fC rO 2 F 2 melt at 31.6 8Ca nd the vapor pressure reaches 760 mm Hg already at 29.6 8C. The gas phase structure has been determined. [22,23,24] Alfred Engelbrecht was born in 1923 in as mall village in Tyrolia and was educated for his maturai nI nnsbruck, Austria. Being drafted in the early days of World WarI I, he was severely injureda nd allowed to return to Innsbruck in 1944 to study chemistry there for his PhD (1948). Thereafter he went to the US on one of the early UNESCOp rograms "to gain expertise in modern fluorine chemistry" and had post-doctorate positions with George H. Cady at the University of Washington in Seattle and with Aristid von Grossei nP hiladelphia. He returned to Innsbruck in the early 1950s for his habilitation for whichsmall wonder-he chose to work on perchloryl fluoride ClO 3 F on whichh ef irst published in 1952. [25] Colorlessp erchloryl fluoride( m.p. À152.2 8C, b.p. À48.1 8C) has as weetish odor, but is highly toxic. It is thermally much more stable than permanganyl fluoride and decomposition starts as high as 450 8C.
The compound was long considered as ac omponent for powerful rocketp ropellants. In 1965 Engelbrechtw as offered a chair at the University of Innsbruck and established ah ighly productives chool of fluorine chemists there. He had to take early retirement owing to the handicaps inflicted by his war injuries. Unfortunately,w ed id not find ap hotograph of Albert Engelbrecht on public records.
Starting in 1969, the group of Achim Müller( Figure 6) at the University of Dortmund and later at Bielefelds tudied for the first time both the electronic absorption spectrum [26,27] andt he IR spectrumo fM nO 3 Fi nt he gas phase, [28,29] followed by work with the emerging technique of He I photoelectron spectroscopy. [30] Data from all these analytical techniques had still been missing. The results explained the deep green color of the compound arising from O ! Mn charge transfer absorptions similart ot he corresponding transitions of the purple permanganatei on with its idealized T d symmetry,w hich is lowered to point group C 3V for MnO 3 [31] In 1975, this work was followed by another study carried out by J. P. Jasinski, S. L. Holt, J. H. Wood, and J. W. Moskowitz at the University of Wyoming on the calculated and experimentally determined electronic structure of gaseous MnO 3 F. [32] This study was further complemented in 1976 by another microwave investigation of MnO 3 Fp ublished by J. Høg and T. Pedersen of the University of Copenhagen. [33] In 1991, A. L. Brisdon, J. H. Holloway,E .G.H ope, P. J. To wnson,W .Levason,and J. S. Ogdena tthe Universities of Leicester and Southampton presented am ore comprehensive study of the UV/Vis spectra of manganese and rhenium oxide fluorides in low-temperature matrices. [34] The resonance Raman spectrumo fM nO 3 Fh as been registered by E. L. Varetti at liquid air temperature using al aser scanning technique, and the harmonic wavenumbers and anharmonicity constantsh ave been determined. [35] The most recent experimental study of the high resolution vibrational spectra of MnO 3 Finb oth the ground andfirst excited state, followed up by gradient-corrected DFT calculations, has been carriedo ut by the groups of Walter Thiel and Hans Bürger (Figure 7) at the University of Wuppertal and the University of Zürich, respectively.Acomplete set of molecular parameters and vibration-rotation coupling constants hasb een presented. [36] The molecular structure determined by the microwavea nalysis [15] has been well reproduced by single-configuration density-functional corrected BLYP and BP86 calculations, while ab initio configuration-mixing methods still turned out to be less satisfactory.M nO 3 Fw as the first polyatomic molecule where this had been demonstrated not only for the harmonic wave numbers, but also for the anharmonic spectroscopic constants.
In 2006, the group of Konrad Seppelt (Figure 8) at the Free University of Berlin has finally determined the crystal structure of MnO 3 Fa tÀ120 8C. [37] Single crystals were grown from liquid HF at À78 8C( monoclinic, space group C2/c, Z = 4). Owing to the small differences in the MnÀOa nd MnÀFd istances, and because the Oa nd Fa toms have as imilar (negative) charge according to aM ulliken analysis, the molecule is heavily disordered in the crystal and no specific sites can be distinguished for the Oa nd Fa toms. The average MnÀO/F distance was found to be 1.621 ,i ng ood agreement with the gas phase microwaved ata (3 1.586 and 1 1.724 ). [15] In aK -edge EXAFS (extended X-ray absorption fine structure) study carried out by groups at the universitieso fS outhampton and Leices-   ter the MÀOa nd MnÀFb ond lengths were found to be 1.59 and 1.72 . [38] Am olecular orbitald escription of the core excitation spectrumw as provided by ab initio CI calculations published by Decleva, Fronzoni, Lisini, and Stener of the University of Trieste. [39] AG limpse on Related Molecules Permanganyl chlorideMnO 3 Cl Permanganyl chloride was first mentioned in ar eport by B. Franke (University of Leipzig) on oxidesa nd halides of manganese in 1887 [40] indicating ap ossible relationt oW çhler's fluoride, but withoutaclear description.N ol ess than ac entury later,D .M ichel and A. Doiwa reactedp otassium permanganate in sulfuric acid with gaseous HCl and obtained ag reen-violet volatile product that was characterizedb yi ts UV/Vis spectrum. [41] This study was followed by an investigation of T. S. Briggs at CornellU niversity who reacted potassium permanganate with chlorosulfuric acid at À60 8C[ Eq. (7)]. [42] MnO 3 Cl is a volatile liquid,w hich solidifiesn ear À68 8C. The color of solutions in CCl 4 is pinkish orange, but the pure liquid is almost black and deflagrates explosively when allowed to warm above 0 8C, while slow hydrolysis yields HCl and HMnO 4 .
Soon thereafter the group of Achim Müller studied the UV/ Vis absorption spectrum of MnO 3 Cl, [26] followed by the study of its He I photoelectrons pectra. [30] In am ost thorough study carried out by the Seppeltg roup at the Free University of Berlin [37] the compound was prepared by the reaction of HSO 3 Cl andK MnO 4 at À30 8C, separated by vacuum fractionation and finally crystallizeda tÀ196 8C. Its crystal and molecular structure have been determined by single crystal X-ray diffraction (orthorhombic, space group Cmc2 1 , Z = 4). The crystals are green at À150 8C, while the liquid is dark brown at À100 8C. This color change is probably due to the ordered organization of the molecules in the crystal,w here short Cl-Cl contacts are established which affect the charget ransfer excitations.T he average bond lengths are 1.586 for MnÀOa nd 2.099 for MnÀCl. This much larger differencei nduces the ordering of the molecules in the crystal, whereas for MnO 3 Fd isorder had been observed. AM ullikenc harge analysis suggests that the bonds are inverselyp olarizeda s( O 3 Mn) + d ÀF Àd and (O 3 Mn) Àd ÀCl + d ,w ith d % 0.2, reflecting the larger electronegativity of Fa sc ompared to Cl, and the large electronegativity of Mn VII attached to three Oa toms. The positive chargea tt he Cl atoms and their charge deformation makes them electro-and nucleophilic, which is borne out by the intermolecular Cl-Cl donor/acceptor contacts in the crystals. The Ramans pectrum has been observed for the liquid at À120 8C, and the six strong bands at 948 (dominantly:a nti-symmetric MnÀOs tretch), 887 (symmetric MnÀOs tretch), 457 (MnÀCl stretch), 365, 306, and 217 cm À1 have been assigned to the vibrations of the molecule with C 3v symmetry,supported by DFT calculations. [37] Pertechnetyl fluorideT cO 3 F The history of TcO 3 Fi sm uch shorter.T he compound was first synthesized in weighable amountsb yH .S elig and G. Malm in 1963 at ArgonneN ational Laboratory by the reactiono fT cO 2 with elemental fluorine and described asav olatile yellow solid meltinga t1 8.3 8C. The extrapolatedb oiling point is 100 8C. [43] In as ubsequent paper,t he Selig group, workinga tt he HebrewU niversity in Jerusalem, described the synthesis from ammonium pertechnetate and hydrogen fluoride[ Eq. (8)] and reported IR and Raman data of the compound in the gas and liquid phase, respectively.T he bands were assigned to the monomeric molecule with C 3v symmetry.T he 19 Fr esonance has also been recorded at d + 50 ppm (vs. CCl 3 F). No coupling with the 99 Tc isotope (I = 9/2) was observed [44] [Eq. (9)]: The group of Seppelt also studied the Raman spectrum of Tc O 3 Fa nd determined the solid state structure by single crystal X-ray diffraction at À100 8C. The compound has been shown to be weakly associated into adimer with fluoride bridging between the monomers. Further weaki nteractions with oxygen atoms of neighboring dimers lead to an extendeda rray.T he structure is therefore intermediate between MnO 3 F, which is strictly monomeric, andR eO 3 F, which is ap olymer in the crystal (Figure9). The assignment of the vibrational spectra has been reviseda ccordingly.T he samples for this study were obtained by treating KTcO 4 with HF in the presenceo fe xcess BiF 5 added to capturet he water produced in the process as H 3 O + BiF 6 À [Eq. (9)].I nt his work, the crystal structures of CrO 2 F 2 and VOF 3 have also been determined. [45]

Pertechnetyl chlorideT cO 3 Cl
The preparation of pertechnetyl chloride TcO 3 Cl (in several papers also namedp ertechnyl chloride) has been achieved via severalr outes including the reaction of technetium chlorides with oxygen and of technetium metal with chlorinea nd anonspecifiedo rigin for oxygen. The product was described as a volatile yellow liquid. [46,47] In severalr eports the vibrational and the vapor phase UV/Vis absorption spectra have been described anda ssigned in comparison with those of the rhenium analogue. The model of mononuclear molecules with C 3v symmetry allowed ac onsistent analysis of the experimental data. [48][49][50] Later work was also dedicated to the mass spectral behavior of MnO 3 Cl and relatedc ompounds and their adsorption properties on surfaces. [51] The electronic structure of the gas phase monomer has been calculated by relativistic DFT methods, [52] but the crystal structure has not yetb een determined.

The oxidesMn 2 O 7 ,T c 2 O 7 and Re 2 O 7
As mentioned above, [4,5] Mn 2 O 7 hasalong history.H owever,i ts crystal and molecular structure have been determined as late as 1987 by aj oint work of the groups of Arndt Simon and Bernt Krebs ( Figure 10) workinga tt he MPI in Stuttgart and at the University of Münster, [53] respectively,w here the crystal structure of Tc 2 O 7 had first been elucidated earlier by the Krebs group. [54] The interesting structural properties of this compound have recently been revisited. [55,56] The structure of Re 2 O 7 had also been known from early work by Bernt Krebs. [57] In many details, the structuralr esults of the dinuclearu nits (for Mn, Tc )a nd the coordinationp olymer (for Re) of the M 2 O 7 heptoxides resemble those for the corresponding acid fluorides MO 3 F.

Current View of the Electronic Structure of Permanganate and Permanganyl Fluoride
The multifaceted history of permanganylf luoride would be incomplete without as hort account of the current view of its electronic structure. The MnO 3 Fm olecule is an example that throws light both on the helpful role of the molecular orbital and ligand field models to better understand the bondings ituation in amolecule, and on their limitations.
As an introductiont ot he subjecti ti sw orth to quotea caveat from ap aper by To mZ iegler published in 2012 where he writes: "The notion of permanganate being an easy system is completely wrong." [58] If this is true, permanganyl fluoride will also not be an easy task. This situation mayb ecome evident from the following summary of pertinent aspects.
In chemistry's" standard model" of molecular electronic structure, it is assumed that the atomica nd molecular valence shells can be approximated as as et of Ns ingle electrons in N, more or less localized, spin-orbitals (or N/2 spin-coupled pairs in N/2 orbitals). In the Born-Oppenheimer approximation, the dynamical molecular structure is more or less accurately representedb yafixed geometrical arrangement of the nuclei. The quantum-chemical electronic wavefunction is approximated by as ingleN -electronic Slater determinanto fNs pin-orbitals which can often be depicted or cartooned by ac hemical Lewis formula. However,i tm ay happent hat the observeds tructural, spectroscopic and reactivity parameters cannotb ec onsistently combined into such am odel. From the experimental side, a more general model would be preferable in such cases. From the theoretical side,areliable analysis of the model defects is desired.
An important step toward the understanding and interpretation of electric, magnetic and optical properties of transition metal compounds was undertaken around 1930 by Becquerel, [59] Bethe [60] and van Vleck. [61,62] In their crystal field theory,i t were basically the symmetry aspects of the systemst hat were exploited. In the 1950s, this line was resumed by Ilse and Hartmann, [63,64] Griffith and Orgel [65] and merged with Mulliken's [66] concept of "perturbation"b yt he mixing of atomic orbitals to form molecular orbitals, following the early suggestion of VanVleck [67,68] of 1935. This advance was summarized in the famousb ook of Ballhausen [69] in 1962 (Figure 11;c ompare also Schläferund Gliemann,1967). [70] The crystal field and adapted ligand field models worked well for complexes of transition metals (M) in their lower and medium oxidations tates. As an example, the common oneelectron orbital level Schemef or complexes of type ML 4 with a tetrahedral structure is shown on the left-hands ide of  first row of the transition metals)i ntot he partially occupiedM -3d(t 2 )a nd unoccupied M-4s(a 1 )p arts of the valence-shello f the centralm etal atom,e nhanced by weaker L-2p(p)d onation into M-3d(e). In tetrahedral complexes, the d(e) orbitals point toward the ligands, and the d(t 2 )o rbitals are directed between the ligands. This yields the energy order of orbitals ymmetries a 1 < t 2 < e, accordingt oB allhausen's ligand field concept. [69,70,72] The uppers et of six partially occupied or unoccupied valence orbitals forms the mirror image e* < t 2 * < a 1 *, predominantly consisting of M-3d and M-4s atomic orbitals. The intermediate range of energyl evelsh olds six occupiedL -2p ligand orbitals, from where UV-ionizations or excitations may occur (L ! Mc harge transfer).
When this model was applied to interpret and complement the deductions drawn from experimentals tudies of complexes with transition metal atoms in high oxidation states,s uch as permanganate or permanganyl fluoride,i nconsistencies appeared. In 1958, Ballhausen and Liehr [72] therefore performed varioust ypes of approximate and empirically adjusted quantum-chemical calculations which all resulted in ad ifferent order of the molecular orbitale nergies. Ad ecadel ater Ballhausen [71] stressed that it is "very likelyt hat configurationi nteraction between the excited states is important", and deplored: "As to the placement [of spectral lines] on the energy scale we cannote xpect any help from the available calculations". After half ac entury of quantum chemical endeavors, [20, 36-39, 52, 72-74] Ziegler [58] (Figure 12, left) still noted that:" It would be very interestingt oh ave as tudy based on CASPT2", am ore reliable technique which requires very extensive calculation efforts. Substantiating Ballhausen's early discontent, Buijse andB aerends provided indications that any one-electron orbital approach,b ei ta b-initio Hartree-Fock, density-functional Kohn-Sham or semi-empirically corrected NDO self-consistent field approaches, will cause artifacts. [75] The results of Buijse and Baerends suggest that any compoundo ft he later 3d transition metals in ah igh oxidation state will exhibitt his specific electronic structure. Common formalisms feature closed shell cores 1s 2 2s 2 2p 6 3s 2 3p 6 only apparently without (or with very low) de lectron population, and closed shell ligands( halogenide or chalcogenide anions). The (partially oc- Figure 11. Carl Johan Ballhausen (1926Ballhausen ( -2010 was born in København in 1926 and studiedc hemistrythere with Niels J. Bjerrum. After as erious traffic accidentin1 949 and duringt he long hospitalization, he made the best of this situation and taught himselfq uantum mechanics for use in chemistry. After assistantships at Harvard,the Bell Te lephone Laboratories, and the University of Chicago, he eventually becameprofessor (from 1959-1996) and director of the Institute of Physical Chemistry of the University of Copenhagen. Ballhausen contributed deeplyt other enaissance of inorganic chemistry after World WarII, mergingt he fieldsofs pectroscopy,structural chemistry,and molecular quantum mechanics. He combined the symmetry aspects of the electrostatic "crystalfield theory" with the physicalm eaning of the parameters of "ligand field theory", incorporating metal-ligand orbital overlap and introducing molecular orbitalmodels of covalent bonding.
Scheme2.Schematicorbitallevel diagram for transition metalcomplexes of the type ML 4 with at etrahedrals tructure. Left:T raditional order for "M 0 "i n zero or low oxidation state, with the occupied ligand ("Lig") orbital levels energeticallybelow the partially occupiedM -d orbitals. Right:Alternativeo rder for "M q + "inhighoxidation state, with ligand and (nearly)empty M-d orbitals. Baerends, and for applying it to transition metal complexes, catalysis and UV/Vis and NMR spectroscopy.One of his hobbies was the permanganate taken over from his PhDt eacher. [83] Evert Jan Baerends (right, *1945),b orn in aD utch village, studiedc hemistry at the Vrije Universiteit Amsterdam, where he entered ac areer in academia that led him to the chair of theoretical chemistry(1981-2010). Thereafterhew orked for several years at Pohang Universityi nS outh Korea. He developeddensity functional and density matrix functionalt heories and ledthe codingf or application of densityf unctional theory to molecules andperiodic systems in one, two and three dimensions, and to chemical bondinga nalysis,initiatedb yT om Ziegler.T he research tools ADFa nd BANDa re useful in spectroscopy, transition metal and heavy elementsc hemistry, and in the study of bulk crystals, polymers, and surfaces. [75] cupied)3 dv alence shell is largely located inside the Mc oreshell that becomes significantly "polarized" staticallya nd dynamically upon M(d)ÀL(p) bond formation. The anionic ligands are also easily deformed, both meaning significant dispersion interactions that requiret he simulation of extended dynamic many-electron correlations. The repulsions between the closed metal and ligand core-shells lead to stretched M(d)ÀL(p) electron-pair bonds, causings trong static electron correlation. In addition, the high oxidation state of the metal increases its effective electronegativity,inducing ligand to metal chargetransfer in the molecular ground state and some electron-hole character in the ligand shells.T he latter enables ligand-ligand bonding [76] in cases of late 3d transition metal atoms in high oxidation states with small effective radii, [77,78] which requires the simulation of ah uge amount of static many-electron correlations (Scheme 3). This ligand-ligand bondingp henomenon appliesf or example, also to the case of the hypothetical FeO 4 molecule and the [FeO 4 ] 2À dianion, where oxide ligandsu nder the influence of the Fe(VIII/VI) centers may be combined to peroxidel igands leadingt or eady decompositionw ith loss of dioxygen. [1,76] Buijse and Baerends [75] noted that in these cases the simplistic individual-electron or -orbital picture causes somem olecular orbitals to becomep olarized too much towardt he central atom, and other orbitals toward the ligands, so that the charge distribution becomes artifactually biased and the bonding appears too ionic. State-of-the art medium-sized ab-initio CAS-SCF [36] or single-reference CCSD(T) [79] approaches yield insufficiently accurate structures and stabilities. Often the computationally much cheaper density functional approaches such as BP86, B3LYP,M 06L, CAM-B3LYP etc.,g ive satisfactory structures, energies and vibrational frequencies (in particulari f semi-empirical correction factors for the involved bond types are introduced). The correct description of electronic excitations yet remains at heoretical challenge as mentioned half a centurya go.
For interpretative purposes of experimental data, orbital modelsa tv arious levels of the density functionala pproxima-tion may stillb eh elpful, but requireacareful empirically based selection.Atypical alternative orbital level diagrami s shown in Scheme 2o nt he right hand side. Therea re now five low-energy molecular orbitals, the main differenceb eing their nearly homopolar covalent M-3d + L-2p(s,p)c haracter.T he e and t 2 levels are energetically close, whilet he a 1 level is higher and only weakly bonding due to the small admixture of the rather diffuse M-4s. The upper part of the occupied molecular valence band contains six non-bondingl igand-dominated orbitals, with t 1 either above or below t 2 ,l ike on the left hand side. The second main difference is the inverted order t2* < e* < a 1 *o fthe antibonding orbitals, and their strongly mixed M(d)-L(p) character.
This description of permanganate as ac ase with an "inverted ligand field" can be found already in the early work of Jasinski, Holt, Hood, andM oskowitz of 1975 based on SCF-Xa-SW calculations. These authors also considered permanganylf luoride under the symmetry-reduction from T d to C 3v ,a nd based on the revisedd iagram the UV/Vis absorption characteristics have tentatively been assigned for both MnO 4 À and MnO 3 F. [32] Similar results were obtained in 1994 by the group of P. Decleva using ab initio CI methods. [39] Over the past 70 years, more or less crude density-functional or semi-empirical or ab-initio orbital-calculations on Cr VI ,M n VII or Fe VIII complexes yielded orbital energy patterns as on the left or right sideo fS cheme 2, i.e. (i)aminor admixture of M-3d to the lower e,t 2 part of the ligand-dominated valence band, (ii)acorrespondingm inor admixture of ligand character to the spectroscopic e*,t 2 *u pper valence band,( iii)the a 1 M-4s + L-2ps orbitalm ore or less bonding, that is, nearer to the lower or upperp art of the valence band, (iv) with as imilar variation of the antibonding a 1 *p artner in the spectroscopic upper range, and (v) most importantly either the normal e* < t 2 * order, [32,72,80] the inverted ligand fieldo rder t 2 * < e*, [39,74,75,[81][82][83] or the near-degenerate situation e* % t 2 *. [75] In more recent years, extended ab-initio approaches and computer network power have become efficient enough to produce more reliable data in the future. The huge computational effort for such small molecules however may only be acceptable for really important questions that cannot yet be solved experimentally. Concerning the title compound MnO 3 F, its qualitative molecular orbitall evel scheme may then probably be derived from the right side of Scheme 2w ith adaption to the appropriate symmetry (T d ! C 3v ). [32] In summary,t he "weird molecule" MnO 3 Fh as been well taken care of experimentally,w hile it remains atrue challenge for theoretical approaches as to its electronic structure.

AP ersonal Note
More than 60 years ago, one of the authors (HS) prepared trimethylsilyl perrhenate,M e 3 SiOReO 3, the first ester of perrhenic acid, as ah ighlight of his doctorate thesis. [84] The compound was found to be as table crystalline material, andi ts crystal structure has later been determined by the "Sheldrick brothers". [85] However,a ll subsequent attempts to prepare trimethylsilyl permanganate Me 3 SiOMnO 3 failed and ended up in fierce Scheme3.Left:Aclosed-shell complex of al ate 3d transition-metal atom M with formald 0 electronc onfiguration and four ligands L, s and p dativebonding by their formalp 6 closed-shells. Middle:Due to the higheffective electronegativity of the formally naked M(d 0 )c ation, electronic charge is attracted from the ligands toward M. The M-d shell becomes effectively partially occupied (for example d 4 )and the fourligands carry partial holes (for example p 5 ), whiche nablesc ovalentM ÀLand LÀLbond formation in the case of asmall cationic radius, leadingt ol igand-ligand overlap.The formal chargesg iven in red (+ 7and + 3) refer to the case for M = Mn with its corresponding 3d population (d 0 and d 4 ). Right:Many similar "resonating" structures are possible, givingr ise to an extremely complex configuration mixing with ab reak-down of the commonlocalized orbital picture and the corresponding setofL ewis structures. explosions depositing for quite some time pieces of shattered glass in the face and extremities of the experimentalist owing to incomplete safety precautions. When similara ccidents happened with bis(trimethylsilyl) chromate, (Me 3 SiO) 2 CrO 2 ,t his exercise was broughtt oah old by the supervisor.T he experience left the student for the rest of his career with an admiration for those brave scientists who carriedo nt oc ontributet ot his challenging chemistry (above).