Experimental and Theoretical Analysis of the Luminescence Spectroscopy of atomic Mercury and atomic Manganese isolated in Rare Gas Solids

Experimental and Theoretical Analysis of the Luminescence Spectroscopy of atomic Mercury and atomic Manganese isolated in Rare Gas Solids A Thesis submitted by Martin A. Collier, B.Sc. (Hons.) to the National University of Ireland in fulfilment of the requirements for the Degree of Doctor of Philosophy Based on research carried out in the Low Temperature Laboratory, Department of Chemistry, National University of Ireland, Maynooth. Research Supervisor: Dr. John G. McCaffrey Head of Department: Prof. Charles M. Quinn Maynooth, August, 2004 Co. Kildare, Eire. Contents I Table of contents Page Abstract XI Chapter I The optical spectroscopy of metal atoms isolated in rare gas solids Introduction I.1 Overview 1 I.2 Matrix-isolation; History and development 1 I.3 Rare Gas solids, (RG) 4 I.4 Matrix Effects 7 I.4.I Multiple metal atom trapping sites 8 I.4.II Gas Phase – Matrix transition frequency shifts 9 I.4.III Dynamic Jahn-Teller effect 11 I.5 Luminescence spectroscopy of Hg/RG and Mn/RG solids 13 I.5.I Hg/RG 13 I.5.I Mn/RG 14 References 17 Chapter II Experimental II.1 Introduction 20 II.2 Matrix – isolation apparatus 20 II.3 Gas handling system, (GHS) 23 II.4 M/RG sample preparation 24 II.4.I Metal vapour generation, Mercury 24 II.4.II Metal vapour generation, Manganese 25 II.5 Luminescence measurements 28 II.5.I Steady-state spectroscopy (continuous lamp excitation) 28 II.5.II Time-resolved spectroscopy 31 II.5.III Excited state lifetime measurements 37 References 39 Contents II Chapter III Luminescence spectroscopy of 3P1 and 3P0 state atomic mercury isolated in solid Ar, Kr and Xe III.1 Introduction 41 III.2 Results 42 III.2.I Hg 3P1 ← 1S0 absorption spectra 42 III.2.II Hg 3P1 ↔ 1S0 excitation and emission spectra 43 III.2.III Hg 3P0 → 1S0 emission spectra 57 III.3 Discussion 63 III.3.I Hg 3P1 → 1S0 emission 63 III.3.II Hg 3P0 → 1S0 emission 64 III.4 Conclusion 67 References 68 Chapter IV A pair-potentials analysis of the optical spectroscopy of 3P1 state atomic mercury isolated in solid Ar, Kr and Xe IV.1 Introduction 70 IV.2 Methods and Results 72 IV.2.I Ground 1S0 state 74 IV.2.II Excited 3P1 state 75 IV.2.II.I Tetragonal (4-atom) symmetry modes 79 IV.2.II.II Trigonal (6-atom) symmetry modes 85 IV.2.II.III Two-fold (2-atom) symmetry modes 91 IV.3 Discussion 94 IV.3.I Absorption Energies 96 IV.3.II Emission Energies 97 IV.4 Conclusions 101 References 102 Contents III Chapter V A pair-potentials analysis of I) the Hg(3P1 ↔ 1S0)/Ne luminescence and II) Hg(3P0 ↔ 1S0)/RG (RG = Ar, Kr and Xe) emission spectroscopy V.I A pair potentials analysis of the Hg(3P1 ↔ 1S0)/Ne luminescence 104 V.I.1 Introduction 104 V.I.2 Methods and Results 107 V.I.2.I Ground State Site occupancy 107 V.I.2.II Excited 3P1 state 111 V.I.3 Absorption and Emission Energies 118 V.I.4 Discussion 119 V.I.5 Conclusion 122 V.II Hg(3P0 → 1S0)/RG emission spectroscopy (Ar, Kr and Xe) 123 V.II.1 Introduction 123 V.II.2 Method and Results 125 V.II.2.I Ground 1S0 and Excited 3P0 states 125 V.II.2.II Hg(3P0 ↔ 1S0)/RG18 Absorption and Emission Energies 128 V.II.3 Discussion 129 V.II.4 Conclusion 130 References 131 Chapter VI The absorption spectroscopy of atomic manganese isolated in solid Ar, Kr and Xe. VI.1 Introduction 132 VI.2 Results – Mn/RG UV/Vis absorption spectroscopy 134 VI.2.I Mn/Ar 134 VI.2.II Discussion Mn/Ar absorption spectroscopy 138 VI.2.III Mn/Kr 141 VI.2.IV Discussion Mn/Kr absorption spectroscopy 145 VI.2.V Mn/Xe 146 VI.2.VI Discussion Mn/Xe absorption spectroscopy 149 Contents IV VI.3 Discussion Mn/RG UV/Vis absorption spectroscopy 150 VI.4 Conclusion 154 References 155 Chapter VII Luminescence spectroscopy of the z6P state of atomic manganese isolated in rare gas solids, (RG = Ar, Kr and Xe) VII.1 Introduction 156 VII.2 Results Mn(z6P)/RG luminescence 158 VII.2.I Mn(z6P)/Xe 159 VII.2.I.I Discussion Mn(z6P)/Xe 167 VII.2.II Mn(z6P)/Ar 168 VII.2.II.I Mn(z6P)/Ar Site Specific Emission spectroscopy 174 VII.2.II.II Mn(z6P)/Ar – Red (1°) site luminescence 175 VII.2.II.III Mn(z6P)/Ar – Blue (2°) site luminescence 190 VII.2.II.IV Discussion Mn(z6P)/Ar 199 VII.2.III Mn(z6P)/Kr 202 VII.2.III.I Mn(z6P)/Kr Site Specific Emission spectroscopy 207 VII.2.III.II Mn(z6P)/Kr – Blue (1°) site luminescence 208 VII.2.III.III Mn(z6P)/Kr – Red (2°) site luminescence 218 VII.2.III.IV Discussion Mn(z6P)/Kr 226 VII.3 Discussion Mn(z6P)/RG luminescence 229 VII.4 Conclusion 232 References 234 Chapter VIII Direct laser excitation of the ‘forbidden’ z8P ↔ a6S and a6D ↔ a6S transitions of atomic Mn/RG solids, (RG = Ar, Kr and Xe) VIII.1 Introduction 235 VIII.2 Results Mn(a6D)/RG luminescence 238 VIII.2.I Mn(a6D)/Xe 239 VIII.2.II Mn(a6D)/Kr 246 VIII.2.III Mn(a6D)/Ar 257 VIII.3 Discussion Mn(a6D)/RG luminescence 264 VIII.3.I Mn(a6D)/RG Excitation spectroscopy 264 VIII.3.II Mn(a6D)/RG Emission spectroscopy 266 Contents V VIII.4 Conclusion Mn(a6D)/RG luminescence 267 VIII.5 Results Mn(z8P)/RG luminescence 268 VIII.5.I Mn(z8P)/Xe 269 VIII.5.II Mn(z8P)/Kr 273 VIII.5.II.I Mn(z8P)/Kr – Blue (1°) site luminescence 276 VIII.5.III Mn(z8P)/Ar 280 VIII.5.III.I Mn(z8P)/Ar – Red (1°) site luminescence 282 VIII.5.III.II Mn(z8P)/Ar – Blue (2°) site luminescence 284 VIII.6 Discussion Mn(z8P)/RG luminescence 287 VIII.6.I Mn(z8P)/RG Excitation spectroscopy 287 VIII.6.II Mn(z8P)/RG Emission spectroscopy 288 VIII.7 Conclusion Mn(z8P)/RG luminescence 290 VIII.8 Mn/RG Discussion – Blue site-specific luminescence 290 VIII.9 Conclusion Mn(a6D and z8P ↔ a6S)/RG 294 References 296 Chapter IX Sites of manganese atom isolation in RG solids, RG = Ar, Kr and Xe IX.1 Introduction 297 IX.2 Site Analysis Mn(z6P ← a6S)/RG 299 IX.3 Site Analysis Mn(y6P and z8P ← a6S)/RG 303 IX.3.I Mn y6P ← a6S Excitation spectroscopy 303 IX.3.II Mn z8P ← a6S Excitation spectroscopy 305 IX.4 Discussion 307 IX.5 Conclusion 311 References 312 Contents VI Chapter X Conclusion IX.1 Hg/RG 313 IX.2 Mn/RG 316 IX.3 Summary 320 Summary Experimental and Theoretical analysis of the Luminescence Spectroscopy of atomic Mercury and atomic Manganese isolated in solid Rare Gases The work presented in this thesis is primarily experimental but also contains an important theoretical extension to gain further insight into the optical spectroscopy of atomic ns2 metal atoms, mercury and manganese isolated in cryogenic thin films of rare gases argon, krypton and xenon. The luminescence spectroscopy of solid-state M/RG (M = Hg and Mn; RG = Ar, Kr and Xe) samples has been recorded employing both time-integrated (steady-state) and time-resolved methods. The impetus for the selection of the Hg/RG systems being the availability of solid-state spectroscopic and gas phase pair-potentials data that allowed the development of a theoretical model. The theoretical analysis conducted on the Hg/RG systems, yielded a qualitative interpretation of the recorded experimental data, providing information on the vibronic modes leading to the observed luminescence. In contrast, the investigation of Mn/RG solids was motivated by the absorption similarities between this transition metal atom and the simpler Hg/RG system, as both exhibit a ground ns2 electronic configuration and excited states derived from the ns1np1 configuration. However, the existence of low lying excited states in Mn, originating from electronic configurations other than the [Ar]3d54s4p configuration accessed in absorption, provides multiple radiative and non-radiative relaxation channels for excited state populations. The luminescence spectroscopy of Hg and Mn atoms isolated in rare gas matrices has shown that the solid state environment provides an ideal environment to study the solvation of ground and excited state metal atoms. It allows the extraction of information on long-lived electronic transitions (> 100 µsec) that cannot be observed in gas phase experiments. The experimental results obtained for the Mn/RG systems have shown that the site of isolation governs the excited state interactions with the host. This observation therefore would allow the extension of this work to investigate site selective excited state reactions with reagents such as CH4, CH3F, NH3 and H2doped rare gas matrices. Chapter I; Introduction 1 Chapter I The optical spectroscopy of metal atoms isolated in rare gas solids – Introduction I.1 Overview The following provides a synopsis of the ‘matrix-isolation’ technique with particular emphasis on its application to the spectroscopic interrogation of metal atoms isolated in rare gas solids (RG). First, a brief history of the development of the technique since its inception in the 1920’s is outlined. Second, some general remarks regarding the requirements of the apparatus necessary to realise a matrix-isolation experiment are given. Third, the structure of rare gas solids and the effect of the matrix environment on the dopant species are discussed in detail. Finally, a brief outline of the experimental and theoretical research conducted in this study is presented to provide a guide to the results which are presented in the Chapters following. This includes specific details of the metal atom systems, mercury and manganese, investigated. I.2 Matrix-isolation; History and development Matrix-isolation is a term, according to IUPAC, which refers to the isolation of a reactive or unstable species by dilution in a solid matrix such as argon, nitrogen or any other inert material. The matrix is usually condensed on a window or in an optical cell at low temperature, to preserve the structure of the reactive or unstable species for identification by spectroscopic means1. However, the expression is most commonly used in a narrower sense to refer to the technique of trapping atomic and molecular species in the solid rare gases. Occasionally reactive solids are used to investigate low temperature photochemistry. L. Vegard2,3,4 preformed the earliest reported experiments using low temperature matrix materials in the 1920’s and observed the luminescence resulting from the electron bombardment of rare gas and nitrogen solids prepared at liquid helium temperatures. In the 1940’s the first report of the isolation of molecular species in an optically transparent material for the purpose of spectroscopic investigation appeared. G. N. Lewis and co-workers5 in Berkeley studied the Chapter I; Introduction 8 centres in solids. The origin of Jahn-Teller effect which gives rise to structured broadband profiles observed for P ← S type electronic transitions is also discussed. I.4.I Multiple metal atom trapping sites As the M/RG solid condensation proceeds at low temperatures (4 to 35 K) metal atoms exhibit a preference for certain sites of isolation within the RG host as controlled by the M⋅RG ground state bond length. The Hg/Xe case mentioned in Section I.3 represents the ideal situation where the Hg(1S0)⋅Xe bond length is less than the ss site diameter of solid xenon. This is not always the case. If the ground state bond length is slightly larger than the substitutional site size available in the solid, isolation may occur in multiple sites. Thus isolation may occur in a larger site such as a tetravacancy but also in a substitutional site following an expansion. In addition, where the solid formation occurs at the lower end of the temperature range (T < 12 K) isolation in thermally unstable sites that contain defects such as lattice vacancies may result. These sites are subsequently removed by the formation of a more crystalline lattice by annealing, where the solid is gently heated and re-cooled to allow organisation of the lattice into its regular packing structure. Matrix annealing or high deposition temperatures are therefore required to assess the thermal stability of the sites of isolation occupied by dopant species in RG solids. Due to the larger site sizes available but also for rigidity reasons, solid Xe is known to be most efficient rare gas matrix for the isolation of atomic species. The rigidity is directly related to the magnitude of the RG⋅RG diatomic dissociation energy (De). Table I.2 presents the dissociation/binding energies for the rare gas dimers. A comparison of the De values for the Xe2 and Ne2 reveals the Xe⋅Xe binding energy is almost seven times larger than that of Ne⋅Ne. The difference in the RG dimer dissociation energies is manifest in how solid Xe achieves and maintains the crystalline fcc packing easier than Ne. Chapter I; Introduction 9 Table I.2 Lattice parameters (a) for the RG solids and the ground state bond lengths for the RG dimers presented in angstrom units. The binding / dissociation energies (De) for the RG dimers are presented in wavenumber units. RG Solid Lat. Parm. (a, Å)15 RG⋅RG Re (Å) De (cm-1) Ne Ar Kr Xe 4.462 5.312 5.644 6.131 Ne⋅Ne31 Ar⋅Ar32 Kr⋅Kr31 Xe⋅Xe31 3.091 3.756 4.017 4.363 29.4 99.5 138.4 196.2 I.4.II Gas Phase – Matrix transition frequency shifts It has been observed for several metal atom systems undergoing P ← S type electronic transitions, that the solid host strongly influences the spectroscopy. The differences from the free atom in the gas phase include the observation of a shift between absorption and emission energies. The magnitude of this shift is calculated as the difference in energy between the absorption and emission band maxima. It is generally referred to as the Stokes’ shift. In the gas phase absorption occurs from the ground state E0 to the excited state E1 corresponding to a transition energy ∆E1,0 = E1 – E0 and the emission occurs between the same levels. Therefore, as depicted on the left of Figure I.3 ∆E1,0 (Abs.) = ∆E1,0(Em.) in the gas phase. The introduction of the luminescent centre (the atom) to a crystalline solid (RG) results in a Stokes’ shift as ∆E1,0 (Abs.) > ∆E1,0(Em.). This is due to interaction with lattice phonons in both the ground E0 and excited E1 states. Electron-phonon coupling results in the population of E1(vn) from E0(v0) in absorption as shown in the middle panel (Atom A) of Figure I.3. Then fast non-radiative relaxation occurs from the excited state phonon levels accessed E1(vn) to E1(v0) followed by the radiative transition to the ground state phonon level E0(vn) where a second non-radiative relaxation occurs to E0(v0). As a result, the radiative transition energy is ∆E1,0 (Em.) = E1(v0) – E0(vn), and the Stokes’ shift observed is the sum of the non-radiative energies occurring in both the ground and excited states. This process is shown in Figure I.3, where the solid and dashed arrows represent radiative and non-radiative processes respectively. In Figure I.3 E1 ← E 0 represents the electronic transition for metal atoms undergoing P ← S type transitions in solid RG matrices. The matrix absorption energies are normally observed to blue shift by approximately 1% of the gas phase transition energy11 ∆E1,0. This effect is due, in part, to the electron-phonon coupling Chapter I; Introduction 10 and is dominated by the interaction of the excited state with the rare gas solid, occurring in the Frank-Condon region of the ground state potential energy surface. The gas-phase to matrix shift of electronic transition energies of M/RG solids arises from contributions from both the ground and excited state interactions. This is revealed by the linear correlation between the matrix shifts and the RG polarizability. This was shown by the analysis of the Zn/RG, Cd/RG and Hg/RG systems by Laursen and Cartland33. However, another factor leads to the observation of broad absorption features for P ← S transitions in RG solids. This results from the dynamic Jahn-Teller effect, which is discussed in the next section. In emission the relative positions of the ground and excited state potential energy curves are of great importance in determining the observed Stokes’ shift as the transition energies are restricted by the Frank-Condon approximation. If a large difference exists between the excited and the ground state minima (for the same configuration of the atoms), a large Stokes’ shift will be observed. If the excited state is stabilised greatly, the ground state may be very repulsive at the excited state minimum. However, the overall emission bandshape is determined by the electronphonon coupling strength, of which the Huang and Rhys factor34, S is a measure. The larger the value of S, the more Gaussian the observed emission band will be. In cases where interactions in the excited and ground states leads to a coincidence in the minima of both potential energy surfaces, radiative relaxation from E1(v0) to E0(v0), as shown on the right hand side of Figure I.3, allows the electronic transition to occur with minimal electron-phonon coupling (S ≈ 0). In this case the observed emission feature corresponds to the band origin, (ν0,0) and a narrow zero-phonon line (ZPL) is observed. These effects are of reference to Chapter III where the first evidence of ZPL’s for matrix-isolated metal atoms is presented for the 3P0 → 1S0 transition of atomic mercury. ZPL’s have, however been observed in the spectroscopy of matrixisolated molecules. Bondybey and Brus35 have discussed in detail the physical origin of electron-phonon lineshapes observed for matrix-isolated molecules such as Cl2 and C2. The material presented above indicates the profound changes that can occur on placing a metal atom into a solid-state environment. Therefore knowledge of M⋅RG systems in the gas and condensed (M/RG) phases provide powerful methods to Chapter I; Introduction 11 examine the interactions occurring in the condensed phase which contribute to the observed spectral band shapes. Figure I.3 A comparison between the interactions of luminescent metal atoms in the gas and condensed phase. E0 and E1 represent the ground and excited electronic states of the atom and v represent the vibronic interactions in the ground and excited states induced by electron-phonon coupling. The right hand panel in this figure depicts the case of a luminescent centre in the solid where weak electron-phonon coupling (S ≈ 0) exists. I.4.III Dynamic Jahn-Teller effect The optical absorption spectroscopy of matrix-isolated metal atoms undergoing P ← S transitions in the rare gas matrices often show a threefold split pattern13,18. Jahn and Teller36,37 showed that an electronically degenerate state of a non-linear complex is unstable with respect to some asymmetric nuclear displacement that lowers its energy by lowering the symmetry and thereby removing the electronic degeneracy. In the early literature, the presence of the structured absorption features on matrixisolated atoms was attributed to effects such as multiple site occupancy38, crystal field splittings39 and non-nearest neighbour M-M40, interactions in the solid. In the 1950’s, the JT effect was first detected in EPR spectra of paramagnetic ions isolated in crystals41. In the late 1970’s moment analyses using magnetic circular dichroism (MCD) studies of matrix-isolated Mg atoms42 in rare gas solids and comparison to the Energy E1 Ground State Gas Phase – Atom E0 Excited State Absorption Emission Ground State Condensed Phase Absorption Emission E1(vn) E1(v0) E0(vn) E0(v0) Atom (B) Absorption Emission E1(vn) E1(v0) E0(vn) E0(v0) ZPL (ν0,0) Atom (A) Electron – phonon coupling Weak electron – phonon coupling Ground State Energy E1 Ground State Gas Phase – Atom E0 Excited State Absorption Emission Ground State Condensed Phase Absorption Emission E1(vn) E1(v0) E0(vn) E0(v0) Atom (B) Absorption Emission E1(vn) E1(v0) E0(vn) E0(v0) ZPL (ν0,0) Atom (A) Electron – phonon coupling Weak electron – phonon coupling Ground State Chapter I; Introduction 12 observed absorption spectra, revealed for the first time substantial evidence that the threefold splitting observed for M(P ← S)/RG transitions was a consequence of the dynamic Jahn-Teller effect. It was concluded that the threefold absorption pattern resulted from a splitting of the Mg 1P excited state due to a quenching of the excited state orbital angular momentum for Mg atoms isolated in single site type in solid Ne, Ar, Kr and Xe. The quenching was believed to be as a result of mixing of the 3p Mg orbitals with orbitals from neighbouring host rare gas atoms. An analysis of higher MCD moments and absorption spectra, assuming octahedral site symmetry for the Mg atom showed a dominant non-cubic (JT active) vibronic mode contributed to the observed bandwidth in the rare gas hosts (Ar, Kr and Xe) where the threefold pattern was observed. A Jahn-Teller explanation of this splitting, via T1 × t2g coupling, was thus proposed for Mg/RG and assigned as the origin of other triplet splitting patterns observed for P ← S type transitions of matrix-isolated atoms. Analysis of the 2P ← 2S transition of matrix-isolated Li atoms43 in rare gas matrices using the MCD technique showed that strong Jahn-Teller coupling in the 2P state contributes to the absorption bands recorded in solid Kr and Xe. In addition, the application of simple crystal field models to these systems were unable to account for the threefold pattern. A theoretical analysis44 of the MCD and absorption measurements made for the Li/Xe system showed good agreement between simulation and experiment when the JT active modes, (eg and t2g) had equivalent frequencies and coupling strengths. The eg and t2g designations are the non-cubic vibronic modes of the rare gas lattice. The Na/Xe and Li/Xe systems were further investigated45 using a temperature dependent moment and theoretical lineshape analysis of the MCD spectra. This study conclusively assigned the threefold splitting pattern observed for the P ← S type electronic transition to the dynamic Jahn-Teller effect. Information on the actual site occupied by the guest metal atom in the rare gas solid was not available in the early ‘80’s. Since this time, accurate experimental and theoretical work has been presented on the 1:1 M⋅RG diatomics. This is particularly true of the Hg⋅RG diatomics and it is this system which is analysed first in the present work. Chapter I; Introduction 13 I.5 Luminescence spectroscopy of Hg/RG and Mn/RG solids The following sections provide a general outline of the experimental and theoretical (Hg only) analyses of the luminescence spectroscopy of atomic mercury and manganese isolated in rare gas solids (Ar, Kr and Xe) presented in this thesis. I.5.II Hg/RG The metal atom whose spectroscopy has been most extensively studied, both in 1:1 van der Waals (Hg⋅RG) complexes21 and isolated in solid rare gas matrices20 (Hg/RG), is mercury. This is particularly true of the 6p 3P1 ↔ 6s 1S0 transition46 which occurs in the gas phase at 39412.3 cm-1. Figure I.4 presents a schematic of the gas phase energy level diagram for atomic Hg46. As accurate Hg⋅RG pair-potentials, obtained from spectroscopy of the diatomic Hg⋅RG complexes stabilised in supersonic expansions are available in the literature21, spectral simulations of the matrix absorption and emission spectroscopies are undertaken in this study. The simulations completed are an extension of the pair-potentials approach the Maynooth group has implemented in the Zn/RG28 and Cd/RG47 matrix systems. The results of these calculations are presented in Chapter IV. Following the theoretical work, it was necessary to extend the experimental analysis of the emission spectroscopy to provide sufficient information for comparison with predictions. The aspects relating to the spectroscopy of the 6p 3P1 and 6p 3P0 states will be addressed in the luminescence spectroscopy reported in Chapter III. In Chapter V, the Hg/RG18 pair-potentials calculations, presented in Chapter IV for RG = Ar, Kr and Xe, are extended to model the luminescence of the atomic Hg 6p 3P1 ↔ 6s 1S0 transition in solid neon, reported by Chergui and co-workers48,49. In addition, the localised Hg/RG18 model was extended to simulate the emission spectroscopy of the atomic Hg 6p 3P0 → 6s 1S0 transition reported in Chapter III. Chapter I; Introduction 14 Figure I.4 Schematic of the energy levels of atomic Hg46 the arrows indicate the allowed transitions and there energies in wavenumber units above the ground 1S0 state of atomic Hg. The gas phase lifetimes50,51,52 of the transitions are also presented. I.5.II Mn/RG Like mercury17, manganese53 was one of the earliest metal atom systems to be investigated with the matrix-isolation technique. Since the first report by Schnepp53 on the absorption spectroscopy of atomic Mn/RG solids, further work has appeared in the literature, including a review of the UV/Vis absorption spectroscopy of Mn/RG solids18,54,55. In contrast, no reports of the Mn/RG luminescence spectroscopy of atomic manganese have appeared to date. Therefore, the experimental work presented in Chapters VI – IX report both the UV/Vis absorption and luminescence spectroscopy of Mn isolated in solid Ar, Kr and Xe. The motivation for this being the similarities between atomic manganese, which exhibits an ns2 ground state electronic configuration and M/RG systems such as Mg23,24 where the solid-state spectroscopy has been studied in detail. The similarities in the gas phase spectroscopy of Mg and Mn is evident on inspection of Figure I.5. The lowest energy electronic configuration of atomic Mn is [Ar]3d54s2 giving rise to the spherically symmetric a6S5/2 ground state. Chapter VI presents the UV/Vis 1S0 Energy 6s6p λ = 184.9 nm τ = 1.2 nsec λ = 253.6 nm τ = 113 / 125 nsec 1P1 54069 cm-1 Gas Phase – Atomic Hg 3P2 44043 cm-1 3P1 39412 cm-1 3P0 37645 cm-1 Energy 6s6p λ = 184.9 nm τ = 1.2 nsec λ = 253.6 nm τ = 113 / 125 nsec 1P1 54069 cm-1 Gas Phase – Atomic Hg 3P2 44043 cm-1 3P1 39412 cm-1 3P0 37645 cm-1 Hg: [Xe]4f143d106s2 Chapter I; Introduction 15 absorption spectroscopy of Mn/RG (RG = Ar, Kr and Xe) solids. The aim of this being to assign the of s → p electronic type transitions of Mn atoms from the ground a6S5/2 state to the excited Mn [Ar]3d54s4p states. The excited states are the ‘singletlike’ [3d5(6S)4s4p(1P°)] y6P and ‘triplet-like’ [3d5(6S)4s4p(3P°)] z6P states occurring in the gas phase56 at 35,725.85 cm-1 (279.91 nm) and 24,788.05 cm-1 (403.42 nm), as shown in Figure I.5. Following the identification of the fully allowed y6P ← a6S and z6P ← a6S transitions in Chapter VI, the subsequent Chapter presents the luminescence spectroscopy of Mn/RG solids following resonant (z6P5/2 ← a6S5/2) excitation with continuous and pulsed light sources. Excitation spectra are recorded to resolve multiple site occupancies, which are, convoluted in the absorption spectra and determine the matrix shifts of z6P5/2 ← a6S5/2 transition in each solid. Highresolution time-integrated and time-resolved emission spectra are recorded following resonance excitation to assign the observed emission features to radiative transitions from excited states of the Mn atom that occur to lower energy than the z6P state. Excited state lifetime measurements are reported to identify the excited state dynamics, inter-system crossing and inter-multiplet relaxation processes leading to the observed emission bands. Although the absorption spectra of atomic manganese and magnesium share many characteristics, the existence of several atomic states (a4D, z8P and a6D) that occur to lower energy than the z6P state means that the emission spectra of the former is complex. The spectral regions of the ‘forbidden’ 3d54s4p z8P5/2 ↔ 3d54s2 a6S5/2 and 3d64s a6D5/2 ↔ 3d54s2 a6S5/2 atomic transitions at 543.4 and 573.0 nm, (18402.46 and 17451.52 cm-1) were investigated with direct dye laser excitation. The intrinsic high-resolution of the laser coupled with the high intensity, allows excitation spectra for both of these forbidden transitions to be recorded. These spectra also provide a comparison of the absorption characteristics of Mn atoms undergoing P ← S or D ← S transitions. The emission spectroscopy recorded following direct laser excitation of the z8P and a6D ← a6S transitions provides more definitive state assignments of the emitting levels. Chapter I; Introduction 16 Figure I.5 Schematic representation of the energy levels of gas phase atomic manganese56 The allowed y6P5/2 ← a6S5/2 and z6P5/2 ← a6S5/2 transitions occur at 35726 cm-1 (279.91 nm) and 24788 cm-1 (403.42 nm) respectively, are indicated by arrows. Chapter IX provides a comparison of the excitation spectroscopy recorded for the y6P, z6P and z8P ← a6S transitions of atomic Mn. Allied with the similarities of the Mn/RG systems and those of Mg/RG systems, an attempt is made to identify the site or sites of Mn atom isolation in each RG system employing the polarizability model of Laursen and Cartland33. All of the experimental data reported in this thesis was recorded in the Low Temperature Laboratory in the Department of Chemistry, National University of Ireland – Maynooth, with the exception of the Hg(3P1)/Ne luminescence data which was provided by M. Chergui and co-workers48,49. The specifics of the experimental apparatus and spectroscopic techniques used to achieve these results are presented in Chapter II. a4DJ a6DJ 6S b4DJ a4PJ y6PJ λ =279.91 nm τ = 2.7 nsec z6PJ λ =403.42 nm τ = 52.0 nsec z4PJ z8P Energy λ =543.4 nm τ = 149.3 µ sec λ =586.43 nm τ = 3.4 sec Mn 3d54s4p 3d54s23d64s1 Mg 3s3p λ =285.3 nm τ = 2.0 nsec 3P1 λ =457.3 nm τ = 2.3 msec 1S0 1P1 Electronic Config. Mg: [Ne]3s2 s – block Mn: [Ar] 3d54s2 d – block a4DJ a6DJ 6S b4DJ a4PJ y6PJ λ =279.91 nm τ = 2.7 nsec z6PJ λ =403.42 nm τ = 52.0 nsec z4PJ z8P Energy λ =543.4 nm τ = 149.3 µ sec λ =586.43 nm τ = 3.4 sec Mn 3d54s4p 3d54s23d64s1 Mg 3s3p λ =285.3 nm τ = 2.0 nsec 3P1 λ =457.3 nm τ = 2.3 msec 1S0 1P1 Electronic Config. Mg: [Ne]3s2 s – block Mn: [Ar] 3d54s2 d – block Chapter I; Introduction 17 References 1 IUPAC Compendium of Chemical Terminology; 2nd Edition, (1997). Free text search for ‘matrix-isolation’ http://www.chemsoc.org/cgi-shell/empower.exe (Last accessed 7th June 2004) 2 L. Vegard, C. R. Acad. Sci., 176, 941 (1923). 3 L. Vegard, H. Kamerlingh-Onnes and W. H. Keesom, C. R. Acad. Sci., 180, 1084 (1925). 4 L. Vegard, Ann. Phys., 6, 487 (1930). 5 G. N. Lewis and D. Lipkin, J. Am. Chem. Soc., 64, 2801 (1942). 6 I. Norman and G. Porter, Nature, 174, 508 (1954). 7 I. Norman and G. Porter, Proc. Roy. Soc., A230, 399 (1955). 8 E. Whittle, D. A. Dows and G. C. Pimentel, J. Chem. Phys., 22, 1943 (1954). 9 E. D. Becker and G. C. Pimentel, J. Chem. Phys., 25, 224 (1956). 10 E. D. Becker, G. C. Pimentel and M. Van Thiel, J. Chem. Phys., 26, 195 (1957). 11 B. Meyer, Low Temperature Spectroscopy, American Elsevier Publishing Company Inc., New York, 1971. 12 D. W. Ball, Z. H. Kafafi, L. Fredin, R. H. Huage and J. L. Margrave, Rice University Houston, A Bibliography of Matrix Isolation Spectroscopy, 1954 – 1985, 1988. 13 M. Moskovits and G. A. Ozin, Cryochemistry, Wiley – Interscience, New York, 1976. 14 L. Andrews and M. Moskovits, Chemistry and Physics of Matrix – Isolated Species, North – Holland, Amsterdam, 1989. 15 H. E. Hallam, Vibrational Spectroscopy of Trapped Species, Wiley – Interscience, New York, 1973. 16 I. Dunkin, Matrix - Isolation Techniques, (A Practical Approach), Oxford University Press, 1998. 17 M. McCarthy and G. W. Robinson, Mol. Phys., 2, 415 (1959). 18 D. M. Gruen, Spectroscopic Identification and Characterization of Matrix Isolated Atoms in Cryochemistry, edited by M. Moskovits and G. A. Ozin, Wiley – Interscience, New York, 1976. 19 B. Meyer, Ber. Bunsenges. Phys. Chem., 82, 24 (1978). 20 C. Crepin-Gilbert and A. Tramer, Intl. Rev. Phys. Chem., 18, 485 (1999). 21 W.H. Breckenridge, C. Jouvet, B. Soep, in: M. Duncan (Ed.), Advances in Metal and Semiconductor Clusters, Vol. 3, JIA Press, Greenwich, CT, 1995. 22 M. Quigley, M.Sc. Thesis, National University of Ireland – Maynooth, Maynooth, Co. Kildare, Ireland, 2002, (unpublished results). Chapter II, Experimental 24 Table II.3 Experimental apparatus used to maintain and monitor the vacuum in the gas handling system. Component Manufacturer Model/Part Number Turbo – molecular pump8 Diaphragm pump9 Ionisation gauge11 Pfeiffer Balzers Vacuubrand Granville – Phillips TPU-180H MD-4T 307 II.4 M/RG sample preparation Solid M/RG samples were prepared by co-condensing metal vapour with the host gases listed in Table II.4 onto a CaF2 window at temperatures ranging up to approximately one quarter the melting point of the host gas depending on the sample characteristics under investigation. The method of metal vapour generation selected was specific to the physical properties of the metal. The two different techniques employed are discussed individually. Table II.4 Matrix host gases, purity and suppliers. The melting point of the host gases (Mp) in Kelvin. Host Gas Purity Supplier Mp (K)7 Argon, (Ar) Krypton, (Kr) Xenon, (Xe) 99.998 % 99.998 % 99.998 % BOC Gases Linde Gas UK Linde Gas UK 83.9 116.6 161.2 II.4.I Metal vapour generation, Mercury The deposition method used for mercury, exploits its high vapour pressure whereby metal ‘pick-up’ by the matrix gases flowing over a Hg reservoir at room temperature, (296 K) entrains sufficient Hg vapour to produce moderately absorbing samples (OD = 0.9). Mercury has a vapour pressure (v.p.) of 1 x 10-3 Torr14 at 16 °C. The reservoir consisted of a 15 cm long stainless steel tube, 1.5 cm in diameter containing 1 cm3 of mercury, connected by one quarter inch tubing to the gas inlet of the matrix shroud. Pick-up is controlled by the three-way valve arrangement shown in Figure II.4, where two valves isolate the reservoir from the matrix rig, except during Hg/RG sample formation, when they are opened and a third valve, directly connecting the Chapter II, Experimental 25 GHS to the matrix, is closed. Freeze-pump-thaw cycles were used to remove air from the reservoir after the initial fill with mercury. Figure II.4 A schematic of the three-way valve system used in the Hg ‘pick–up’ deposition. II.4.II Metal vapour generation, Manganese Manganese vapour was generated by electron bombardment (sputtering) of the bulk metal using an ultra-high vacuum Omicron evaporator, Model EFM3 (Evaporator with integral Flux Monitor)13 shown in Figure II.5. Manganese is a high temperature metal with a melting point of 1244 K and a vapour pressure 1 x10-3 torr14 at 1108 K. The electron beam from a tungsten filament is as shown in Figure II.6 focused on irregular manganese pieces15,16 (Goodfellow, Johnson Matthey; purity >99.5%) contained in a molybdenum crucible. The bombardment induces localised heating in the metal resulting in evaporation. The integrated flux monitor indicated the metal flux generated with this vaporisation. However, the absolute quantity of metal deposited in samples was ascertained from absorption spectroscopy, as irregular size Mn pieces were used the flux observed was dependent on the packing and subject to change as consistent loading of the crucible was not possible. Therefore, the flux monitored during sample deposition served only as a guide. The specifications of the EFM3, Mo crucible and the bulk metal are presented in Table II.5. Three – way valve system Granville Philips variable leak valve Type: 203 Hg Resevoir 298 K Nupro valves (JN Series) Hg vapor Hg/RG Sample Three – way valve system Granville Philips variable leak valve Type: 203 Hg Resevoir 298 K Nupro valves (JN Series) Hg vapor Hg/RG Sample Chapter II, Experimental 26 In CaF2 window shutter Out cooling water supply 'HV' supply flux monitor Exit shutter Figure II.5 A representation of the metal atom source Omnicron UHV EFM3 evaporator showing the orientation of the source with respect to the sample substrate CaF2 and the electrical and physical connections required for operation. Tungsten filament (0 - 3 A) cooling shroud ceramics 'HV' electrical supply (0 - 1000 V) Mn atom beam Molybdenum Crucible barrel connector Figure II.6 An illustration of the molybdenum crucible (containing the bulk manganese) and its immediate environs showing the relative positions of the filament and the exit path of the metal atom vapour within the Omnicron UHV EFM3 evaporator. This study centres on the atomic spectroscopy of metal atoms, therefore, low metal atom fluxes were used with an excess of the host gas to ensure atomic isolation dominated. Typically, for manganese, 700 V was applied to the EFM3 and filament currents (IFil) ranging from 1.3 to 1.5 A were used to achieve Mn/RG samples containing the maximum amount of isolated metal atoms while limiting the content of higher metal atom aggregates formed from metal nucleation during deposition. To ensure sample purity and limit the effect of contaminations from the evaporation, such as MnO2 adsorbed on the metal surface, the evaporant was degassed prior to depositing samples. This was done following each fill of the crucible. This procedure was completed in two stages. First, the filament was degassed by slowly increasing the filament current (IFil.) without high voltage. Second, the bulk metal was evaporated under the normal conditions (see above) with Chapter II, Experimental 27 the exit shutter (Figure II.5) closed until a stable flux could be maintained without any noticeable pressure rise in the sample chamber. This procedure was employed to remove the adsorbed species as the flux arising from these contaminants reduces over time whereas the remaining stable flux is due to the vaporisation of the bulk metal. Table II.5 Manufacturers of the equipment and the specific dimensions where appropriate for the UHV evaporator, molybdenum crucible and bulk manganese. Apparatus Manufacturer Model Specifications Evaporator Crucible Manganese Omicron13 Omicron Goodfellow15 Johnson Matthey16 UHV EFM3 Molybdenum - - IFil: 0 – 2.5 A Voltage: 0 – 1000 V Outer diameter: 5.0 mm Inner diameter: 3.5 mm Capacity: 75 mm3 Temperature (max): 2200 K Condition: Irregular pieces Purity: 99.5 % Condition: Irregular pieces Purity: 99.99 % Although the method by which the mercury and manganese vapours were produced are different, both present specific difficulties with respect to purification. However, once the vapour was achieved, the co-condensation with the host gas was completed as follows. Firstly, a layer of pure host gas (RG) was allowed to deposit on the CaF2 window to minimise metal nucleation and contaminant build up at the sample window. This precaution limited the amount of sample-to-sample cross contamination. Secondly, the metal vapour was admitted and co-condensed with the rare gas of interest. In the case of Hg, this was achieved by allowing metal ‘pick–up’ to proceed. Opening the EFM-3 exit shutter allowed the co-condensation of Mn/RG samples to begin. The samples reported here were deposited at a rate of 8-10 mmol/hr for a period of 30 minutes. The same M:RG ratios and sample thickness were achieved by varying the two factors which govern the deposition rate; a) the backing pressure (Pbk) in the GHS and b) the flow rate selected for the Granville – Phillips variable leak valve. Chapter II, Experimental 28 II.5 Luminescence measurements Following deposition of the M/RG thin films, the spectroscopic measurements reported in the following chapters were conducted using three optical arrangements, which can be considered in two classes depending on the excitation source employed. A) Continuous Lamp Excitation yielding steady–state excitation and emission spectra and B) Pulsed Laser Excitation allowing excitation and emission spectra and temporal measurements of the emission to be made. II.5.I Steady-state spectroscopy (continuous lamp excitation) A deuterium17 (Hamamatsu L6310 and a Cathodeon C713 power supply18) and/or tungsten lamp (30 W, GE Model DZA) were used as the light sources for the ultraviolet (UV, 180–500 nm) and UV/Vis (350–600 nm)19 spectral regions respectively to record both absorption and excitation spectra. An Acton Research Corporation (ARC) 0.30 m SpectraPro–300i monochromator20 fitted with a 1200 grooves/mm diffraction grating, blazed at 300 nm was used for wavelength selection. The absorption, excitation and emission spectra reported later employed the highest resolution gratings (1200 grooves/mm) fitted in both ARC SpectraPro monochromators. However, both monochromators were fitted with additional diffraction gratings, the specifications of which are listed in Table II.6. Figure II.7 presents a schematic of the spectrometer employed for continuous lamp absorption and luminescence spectroscopies. The monochromatic light transmitted through the thin film M/RG matrix samples, located on a vertical CaF2 window21, was focused onto a photomultiplier tube (Hamamatsu, 1P2822) by means of a quartz focusing lens (Fl1) (Figure II.7) mounted with the PMT on the sample chamber. Absorption spectra of the M/RG samples were obtained in the usual manner by rationing sample transmittance spectra with their corresponding blanks, i.e., pure RG films, (RGbl). The absorbance23 quantified in terms of the optical density (O.D.) is calculated using the Equation II.1, O.D. = - log10 (I/I0) Equation. (II.1) Assuming the same sample thickness for the M/RG sample and the corresponding rare gas blank RGbl, the absorbance was calculated using the following substitutions, I = I(M/RG) and I0 = I(RGbl). Chapter II, Experimental 29 An important consideration for optical measurements is the choice of the deposition substrate in this case calcium fluoride, CaF2. The substrate should be transparent in the regions of the electromagnetic spectrum where the measurements are preformed. Calcium fluoride is transparent over the range 129–1176 nm7. Also the matrix material, must be optically transparent in the region of interest. The rare gases are ideal as they are transparent over a very wide spectral range from the far infrared to the vacuum UV. The far IR absorptions are due to phonon absorptions, lattice vibrations of the atoms within the solid matrix crystal. The vacuum UV absorptions of the rare gas crystals correspond to Frenkel excitons, the lowest energy transition occurring for solid xenon at 150 nm23. Emission from the M/RG samples was monitored perpendicular to the excitation axis by focussing emitted light onto the entrance slits of an ARC 0.5 m SpectraPro-500i monochromator24 fitted with a 1200 g/mm grating, blazed at 300 nm. Photon detection was achieved using a Hamamatsu R928-P PMT25 maintained at –20 °C in a Products for Research cooled-housing (Photocool S600)26. This PMT was operated in photon counting mode by relaying its signal via an amplifier/ discriminator module27 (Electron Tubes Ltd, type AD6) to the ARC NCL data acquisition and controller unit. The specifications of the photo-multiplier tubes used are presented in Table II.7. The SpectraPro-500i emission monochromator was calibrated using the sodium D lines from a hollow cathode Na lamp, UV lines of molecular oxygen were used to calibrate the excitation monochromator SpectraPro300i. Table II.6 The specifications of the Acton Research Corporation (ARC) monochromators used during the course of this work. Note * indicates the specifications refer to the 1200 grooves/mm gratings. ARC Monochromator SpectraPro-300i20 SpectraPro-500i24 Focal length (mm) Wavelength range (nm) Gratings (grooves/mm) / Blaze (nm) Resolution (nm)* Dispersion (nm/mm)* Accuracy (nm)* 300 180 nm – far infrared 1200 / 300 300 / 300 0.1 @ 435.8 nm 2.7 ± 0.2 500 180 nm – far infrared 1200 / 300 600 / 600 150 / 300 0.05 @ 435.8 nm 1.7 ± 0.2 Chapter II, Experimental 30 Table II.7 The specifications of the Hamamatsu photo-multiplier tubes (PMT’s) employed during the course of this work. Note the IP28 was mounted on the sample chamber and used to monitor the radiation transmitted by the M/RG samples. Hamamatsu PMT IP2822 R928-P25 Range (nm) Peak wavelength (nm) Photo-cathode material Window material Cathode sensitivity (µA/lm) Anode sensitivity (A/lm) Response times – Rise time (ns) Electron transit time (ns) 185 – 650 340 Sb-Cs UV glass 40 200 2.2 22 185 – 900 400 Multialkali UV glass 200 2000 2.2 22 D2: Hamamatsu L6310 W: GE Model: DZA Emission Monochromator ARC SpectraPro - 500i ABS PMT D2 / W Lamp S.P.C. PMT SAMPLE Hamamatsu IP28 Excitation Monochromator ARC SpectraPro - 300i Hamamatsu R928-P Fl1 Cl1 Cl2 Fl1 - focusing lens (f.l.: 1 inch) Cl1 - collecting lens (f.l.: 1 inch) Cl2 - collecting lens (f.l.: 1 inch) Figure II.7 A Schematic of the luminescence spectrometer set–up used to record the steady–state (time–integrated) spectra. Chapter II, Experimental 31 II.5.II Time-resolved spectroscopy Emission spectra were also recorded with pulsed excitation using a Nd:YAG (Quantel YG 980E-10)28 laser normally operating at a repetition frequency of 10 Hz. A dye laser29 (Quantel TDL–90), pumped by either the second or third harmonics of the YAG, were used to produce tuneable laser radiation. The characteristics of the Quantel laser systems are presented in Table II.8. Table II.8 The specifications of the Quantel laser systems. Pump Laser System Nd:YAG, Quantel YG 980E Gain Medium Repetition rate (Hz) Energy (mJ) @ 1064, 532, 355 nm Pulse duration (ns) @ 1064 nm Flash-lamps Flash-lamp Voltage (V) Q – Switch pre-pulse (ns) Flash-lamp / Q – Switch delay (µs) Neodymium-doped crystal (Yttrium – Aluminium – Garnet) 10, 5, 2,1 850, 400, 165 6 SFL 611.09N/RX 1600 500 242 Dye Laser System Quantel TDL-90 Tuning range (nm) Linewidth (cm-1) @ 560 nm 220 – 750 0.08 The wavelengths required for mercury atom excitation around 250 nm were achieved by mixing the residual Nd: YAG fundamental at 1064 nm with the doubled output of the dye laser using DCM (4-Dicyanmethylene-2-methyl-6-(pdimethylaminostyryl)-4H-pyran) as the dye material. Manganese atom excitation in the UV at 280 nm was achieved by frequency doubling Rhodamine 590 (Benzoic Acid, 2-[6-(ethylamino)-3-(ethylimino)-2,7-dimethyl-3H-xanthen-9-yl]-ethyl ester, monohydrochloride). Manganese atom excitation in the 360-420 nm spectral region was produced by mixing the dye laser output with the Nd:YAG fundamental using DCM as the dye material. Table II.9 presents the details of the spectral characteristics of the dye materials used. KDP (Potassium Diphosphate) crystals, Quantel DCC2/3 and MCC1/2 were used to frequency double and mix respectively while quartz crystals, Quantel QCC1 and QCC2, were used to compensate for the walk of the resultant beams. The wavelength ranges accessible by frequency doubling and/or mixing the dye laser output are presented in Table II.10. Wavelength separation of Chapter II, Experimental 32 the final beam from residual beams was achieved with a Pellin-Broca prism, the UV output of which was trained onto the matrix sample (on a CaF2 window) without focusing optics. The experimental arrangement of the laser is shown in Figure II.8. The direct dye laser output was employed for pumping the forbidden a6D and z8P transitions Mn atom using Rhodamine 590 and Coumarin 500 as dye materials respectively. The former and the latter involved pumping with 2ω and 3ω of the Nd3+:YAG at 532 and 355 nm respectively. Table II.9 presents the spectral characteristics of the dye materials. The use of the tuneable dye laser output allowed the acquisition of laser excitation spectra, using the arrangement shown in Figure II.8. Wavelength separation was not required using the direct dye output of the TDL-90 so the Pellin-Broca prism was replaced by a right-angled prism. Excitation spectra were recorded by scanning the dye laser wavelength monitoring a given emission band maximum. The excitation spectra recorded in this manner were not corrected for the intensity distribution of the dye material. Table II.9 The specifications of the laser Dye materials employed for mercury and manganese atom excitation in RG solids. Dye Material Characteristic Specification DCM30, 31 (4-Dicyanmethylene-2-methyl6-(p-dimethylaminostyryl)-4Hpyran) C19H17N3 Manufacturer Solvent Pump Source: Nd: YAG (nm) Absorption maximum, (nm) Fluorescence maximum, (nm) Dye Laser Range, (nm) Exciton Ethanol 532 (2ω) 472 639 615–666 Rhodamine 59030, 31 (Benzoic Acid, 2-[6- (ethylamino)-3-(ethylimino)- 2,7-dimethyl-3H-xanthen-9-yl]- ethyl ester, monohydrochloride) C28H31N2O3Cl Manufacturer Solvent Pump Source: Nd: YAG (nm) Absorption maximum, (nm) Fluorescence maximum, (nm) Dye Laser Range, (nm) Exciton Ethanol 532 (2ω) 530 566 555–580 Coumarin 50030, 31 (7-Ethylamino-4trifluormethycoumarin) C12H10NO2F3 Manufacturer Solvent Pump Source: Nd: YAG (nm) Absorption maximum, (nm) Fluorescence maximum, (nm) Dye Laser Range, (nm) Exciton Ethanol 355 (3ω) 395 495 498–546 Chapter II, Experimental 33 Table II.10 Details of the technical processes used to achieve the laser frequencies29 required for UV mercury and UV-Vis. Manganese atom excitation. M/RG System Wavelength Range, nm Technical Process Crystals Dye Material Hg Mn (UV) Mn (Vis.) 231–272 267–325 360–420 Frequency Mixing after Doubling Frequency Doubling Frequency Mixing MCC2/QCC2 DCC2/QCC1 DCC3/QCC1 MCC1/QCC2 DCM Rhodamine 590 DCM Typical laser fluence of 20 µJ/mm2, measured with a Molectron power-max 500A meter and PM10V1 head, was achieved in the 250 nm spectral region using only the oscillator and pre–amplifier stages of the TDL-90 dye laser. The linewidth of the dye laser is 0.8 cm-1 at 560 nm. S.P.C. PMT ω2 SAMPLE Nd:YAG (900 mJ) Laser ω1 Dye Laser Mixing Doubling Hamamatsu R928-P ARC SpectraPro - 500i Quantel TDL-90 Quantel YG-980E Output Nd:YAG Laser Fundamental (ω1): 1064 nm Second harmonic (ω2): 532 nm Third harmonic (ω3): 355 nm Collecting lens (focal length 1") Figure II.8 A representation of the luminescence spectrometer and the interaction with the Quantel Nd:YAG (YG-980E) pumped Dye laser (TDL-90) used to record the time-resolved luminescence of M/RG samples following UV/Vis excitation at a repetition frequencies of 1 to10 Hz. Emission was monitored perpendicular to the laser beam and recorded in the photon counting manner described previously for the steady-state measurements, except that each data point in the spectrum was obtained by averaging ten laser shots. Recording emission spectra in the manner described above provided temporal Chapter II, Experimental 40 21 PMT mounted on vacuum shroud using the Acton Research Corp. Model PD-471 PMT Detector Housing with Integral High Voltage Power Supply. 22 Hamamatsu Data Sheet title ‘Side-On Photomultiplier Tubes’ 23 Mark Fox, Optical Properties of Solids, Oxford University Press, 2001. 24 Acton Research Corporation SpectraPro – 500i, 500i Manual V1097.1. 25 Hamamatsu Data Sheet title ‘Photomultiplier Tubes R928, R955’ 26 Products for Research, Inc., Photocool Series Power Supply, Instruction Manual, (Model PC177CE009 for R928). 27 Electron Tubes Limited, Photomultiplier Amplifier-Discriminator Type AD 6, DS_AD6, Issue 1, 11.02.97 28 Instruction Manual QUANTEL YG 980 Q-switched Nd:YAG laser, Doc. 980, Version #1, anglaise PM/DT (12.05.97). 29 Quantel – TDL 90, Instruction Manual – Issue 1. 30 Exciton, Laser Dyes Catalogue, 1992. 31 Ulrick Brackmann, Lambda Physik, Lambdachrome Laser Dyes Catalogue, July 1985. 32 The Andor iStar iCCD camera was mounted onto the ARC SpectraPro 500i using a monochromator specific flange, (Part Number: MFL-ARC-SPRO) to allow coincidence of the grating dispersed radiation and the detector focal plane. 33 D. V. O’Connor and D. Philips, Time – correlated Single Photon Counting, Academic Press, London, 1984. 34 Ortec, Model VT120 Fast Timing Pre-Amplifier, Operating and Service Manual, Part. No.760360, Revision B. 35 Ortec, Model 584, Constant – Fraction Discriminator, Operation and Service Manual, Ortec, Pt. No. 733550, Revision B. 36 Perkin Elmer Instruments, Model 661 Ratemeter, Operation and Service Manual, EG&G Ortec Pt. No. 740380, Revision B. 37 Fast ComTec GmbH, Model 7886, 2 GHz Fast Multiscalar, User Manual, Version 2.1, Aug. 1998. Chapter III, Hg(3PJ)/RG Luminescence 41 Chapter III Luminescence spectroscopy of 3P1 and 3P0 state atomic mercury isolated in solid Ar, Kr and Xe. III.1 Introduction Historically Hg 3P1 ↔ 1S0 was one of the first atomic systems studied with the matrix-isolation technique1,2,3,4. Despite the existence, for more than a decade now, of accurate Hg⋅RG pair-potentials, obtained from the spectroscopy of the diatomic Hg⋅RG complexes stabilised in supersonic expansions, no calculations have appeared in the literature of the corresponding atomic absorption or emission spectra in rare gas matrices. With the availability of accurate interaction potentials for the Hg⋅RG diatomics, spectral simulations of the matrix absorption and emission spectroscopies were extended to the Hg/RG system using the pair-potentials approach our group has implemented in the Zn5 and Cd6 matrix systems. However, with several emission pathways identified in the theoretical work, it was necessary to extend experimental analysis of the emission spectroscopy to provide sufficient information for comparison with predictions. Specifically, the temperature dependence of the matrix emission is examined, lineshape analysis is performed and excitation spectra are recorded. Details of the pair-potentials simulations and a comparison with the experimental data are presented in Chapter IV. The spectroscopy and reactivity of atomic mercury isolated in low temperature solids has been studied in greatest depth and scope by the Orsay group of Crepin and Tramer (C&T)7. As their work has been recently reviewed8, only aspects relating to the spectroscopy of the 6p 3P1 and 6p 3P0 states will be addressed here. Absorption recorded by C&T with a deuterium lamp yielded spectra in agreement with the earlier work9 showing a threefold split band for Hg/Xe in the vicinity of the gas phase 6p 3P1 ↔ 6s 1S0 transition of atomic mercury at 253.6 nm. Featureless bands, progressively blue-shifted from the gas phase transition, were observed in Kr and Ar matrices. Dye laser excitation of the 6p 3P1 state absorption produced multiple emission bands in the UV in all three rare gas matrices. Although excitation spectra were not presented in C&T’s work7, the most intense emission bands had the smallest Stokes’-shifts and were tentatively assigned to the occupancy of atomic mercury in substitutional sites. The Hg/Xe emission was quite different to that Chapter III, Hg(3PJ)/RG Luminescence 42 recorded for Hg/Kr and Hg/Ar in that the Stokes’ shift was very large and the emission bandwidth was much greater than the absorption bandwidth. C&T also identified narrow “atomic-like” features in the matrix emission spectra which they assigned to the forbidden 6p 3P0 → 6s 1S0 transition. The linewidth of these transitions decreased in the order Ar to Xe but the shift of the band positions was irregular with the emission in Ar located between that in Kr and Xe. Other work by C&T on matrix-isolated atomic Hg has involved an examination of the relaxation of the excited atomic 6p 1P1 state following resonance excitation with synchrotron radiation10,11 and pulsed laser excitation12,13. Population of all three triplet spin-orbit states (3P2,1,0) was observed as a result of 1P1 excitation. More recently, Chergui and co-workers14 have conducted spectroscopic studies in neon matrices, work which will be compared with theoretical calculations presented in Chapter V. This Chapter presents a study of the temperature dependence of the 3P1 → 1S0 and 3P0 → 1S0 transitions of atomic mercury isolated in the solid rare gases Ar, Kr and Xe, resulting from resonance excitation of the 3P1 excited state. This state is accessed with continuous lamp and pulsed laser excitation, facilitating a clear distinction of the latter forbidden transition and the former, nearly fully allowed transition. Excitation spectra are presented for the first time allowing identification of the origin of the multiple emission features observed. Lineshape analysis of high-resolution 3P0 → 1S0 emission spectra allow the strength of the electron-phonon coupling to be determined for this transition. III.2 Results III.2.I Hg 3P1 ← 1S0 absorption spectra Following Hg/RG matrix deposition, as outlined in Chapter II, absorption spectra were recorded with a deuterium lamp in the vicinity of the 3P1 ↔ 1S0 transition3 of atomic Hg at 253.6 nm. Figure III.1 shows the absorption spectra recorded at 12 K for atomic Hg isolated in Ar, Kr and Xe matrices deposited at 22, 25 and 35 K respectively. The Hg/RG samples were deposited at elevated temperatures to increase the matrix crystallinity and minimise thermally unstable sites of isolation. The spectra shown were obtained by rationing transmittance spectra recorded for Chapter III, Hg(3PJ)/RG Luminescence 43 Hg/RG thin films with corresponding pure RG films. The spectra observed are in good agreement with those presented in the study by C&T. 39.040.041.0 x103 E ner g y ( cm-1 ) 0.0 0.2 0.4 0.6 Hg/Xe 0.0 0.2 0.4 0.6 0.8 Absorbance Hg/Kr 0.0 0.2 0.4 0.6 Hg/Ar 3P1 _______ 1S0 Figure III.1 Hg/RG absorption spectra recorded at 12 K in the vicinity of the atomic Hg 3P1 ↔ 1S0 transition3 upon deposition at temperatures in excess of 12 K. Thus, as observed by C&T7 upon deposition at 14 K, a progressive red–shift of the Hg(3P1 ← 1S0)/RG band and decreasing linewidth of the matrix absorption band occurs from Ar to Xe. These spectra verified that the Hg/RG samples prepared by our group mirrored those reported in previous studies. III.2.II Hg 3P1 ↔ 1S0 excitation and emission spectra Emission spectra produced with continuous lamp excitation of the Hg 3P1 state are presented in Figure III.2 for Ar, Kr and Xe samples. The overall features of the Hg/RG spectra agree well with the nanosecond, time-resolved spectra presented previously by Crepin and Tramer7. One difference is the presence of a weak, resolved feature at 265.1 nm in Hg/Xe that is due to the long-lived 3P0 state emission of Hg. The main emission features are centered at 250.3, 254.1 and 273.0 nm in Ar, Kr and Chapter III, Hg(3PJ)/RG Luminescence 44 Xe respectively as listed in Table III.1. Excitation spectra recorded for these emission wavelengths are shown on the left in Figure III.2 and correspond to the dominant features in the previously reported7,9 absorption spectra. Table III.1 Photophysical characteristics of the triplet 6p 3P1 ↔ 6s 1S0 transition3 of matrix – isolated atomic mercury. λEx indicates the position of the central component of the three–fold split excitation spectrum and λEm indicates the emission bandcentre in nm units. The full-width at half-maximum intensity of the excitation/emission features is denoted by ∆ and the Stokes shift by SS - both in wavenumber (cm-1) units. Excitation Emission Hg/RG System λEx (nm) ∆ (cm-1) λEm (nm) ∆ (cm-1) SS (cm-1) Ar 245.9 484 250.3 399 715 Kr 248.9 397 254.1 465 816 Xe 253.6 344 273.0 1472 2802 34363840 x103E n er g y ( c m -1 ) Intensity (Arb. Units) 250 260 270 280 290 30 0 W ave l e n g t h ( n m ) Ar Kr Xe Hg(3P1)/RGEmissionExcitation Figure III.2 Emission spectra recorded at 12 K for the Hg/RG systems with lamp excitation of the Hg 3P1 ← 1S0 transition. The excitation spectra, recorded by monitoring emission at 250.4, 254.1 and 273.9 nm in Ar, Kr and Xe respectively, are shown on the left of the figure. Hg/Ar, Hg/Kr and Hg/Xe samples were deposited at 22, 25 and 35 K respectively. Chapter III, Hg(3PJ)/RG Luminescence 45 As observed by C&T, the emission band in Hg/Xe, centered at 273 nm, has a large bandwidth and Stokes’ shift compared with those in the Hg/Ar and Hg/Kr systems. Moreover, the Hg/Xe emission band exhibits a clear asymmetry. To investigate the origin of this asymmetry we have examined the temperature dependence of the Hg/Xe emission and conducted lineshape analyses at 12 K and elevated temperatures. The right hand panel in Figure III.3 provides a comparison of the Hg/Xe emission recorded at 12 and 42 K. As expected the emission bandwidth increases at elevated temperatures but contrary to expectation, the band-centre blue shifts. This effect is completely reversible because, although not shown in Figure III.3, the 12 K scan recorded after sample warming to 42 K is identical to the previous 12 K scan. Lineshape analysis of the time-integrated Hg/Xe emission spectrum is complicated by the presence of a small amount of 3P0 state emission in addition to 3P1 emission. The location of the former emission is revealed by overlaying, as shown on the left in Figure III.3, the 3P0 emission spectrum produced with pulsed laser excitation. As indicated in this comparison, the emission bands of these two states are quite different - the 3P0 state is very narrow while the 3P1 state is broad. The deconvolution of the fluorescent 3P1 and phosphorescent 3P0 components present in the time-integrated spectrum shown in the right panel Figure III.3 was also achieved temporally. The different lineshapes of the 3P1 and 3P0 excited state emission features were verified by time–gated emission spectra recorded with iCCD detection as shown in Figure III.4. It is evident from the time–gated spectra shown in Figure III.4 that employing no acquisition delay (td = 0 nsec) and a long gate width (95 msec) reproduced the time-integrated emission spectrum shown in the right panel of Figure III.3. Temporal separation of the 3P1 and 3P0 emission features was achieved using a delay time (td) of 1.0 µsec. As shown by the dotted trace in Figure III.4, this setting removes the broad 3P1 fluorescence while maintaining the narrow 3P0 emission. A long gate width of 95 msec was chosen to optimise the measurement time between excitation pulses from the Nd:YAG laser operating at 10 Hz. Chapter III, Hg(3PJ)/RG Luminescence 46 3.8 3.7 3.6 3.5 3.4 3.3 Energy (x 104 cm-1) Hg 3P03P1Xe 3.8 3.7 3.6 3.5 3.4 3.3 Energy (x 104 cm-1) Xe Temp. 12 K 42 K 3.8 3.7 3.6 3.5 3.4 3.3 Energy (x 104 cm-1) Hg 3P03P1Xe 3.8 3.7 3.6 3.5 3.4 3.3 Energy (x 104 cm-1) Xe Temp. 12 K 42 K Figure III.3 Details of Hg/Xe emission. The panel on the right shows a comparison of emission spectra recorded at 12 K and 42 K. In this comparison a reversible blue shift in the band maximum occurs with increasing temperature. The panel on the left shows a comparison of the emission spectra produced with pulsed and continuous excitation. The narrow feature at 265.1 nm was recorded with pulsed laser excitation and is gated to show only long-lived 3P0 emission. 343536373839 x103Energy (cm-1 ) 260 265 270 275 280 285 290 295 Wavelength (nm) Hg/Xe Intensity (Arb. Units) Delay 0.0 nsec Delay 1.0 µsec 343536373839 x103Energy (cm-1 ) 260 265 270 275 280 285 290 295 Wavelength (nm) Hg/Xe Intensity (Arb. Units) Delay 0.0 nsec Delay 1.0 µsec Figure III.4 Hg/Xe time-gated emission spectra recorded at 12 K using iCCD detection following pulsed laser excitation, λEx = 253.0 nm. The delay times (td) employed were 0 and 1.0 µsec with a constant gate width of 95 msec for the solid and dotted traces respectively. Chapter III, Hg(3PJ)/RG Luminescence 47 In the lineshape analysis conducted on the time-integrated Hg/Xe emission, provision was made for the presence of the narrow 3P0 state emission band. A satisfactory fit of the 12 K emission band, shown on the bottom left in Figure III.5, is obtained with three Gaussian functions (not counting the narrow 3P0 state emission). Details of the 12 K fits are presented in Table III.II, which gives the positions of the 3P1 emission components as 37535, 36619 and 35729 cm-1. Fitting the emission band profile to Gaussian functions is a realistic analysis for Hg/RG systems where the difference in the ground and excited state bond lengths ∆R is large. The excitation spectra recorded for all three emission components were identical. As the Hg/Xe samples were deposited at 35 K to minimize the formation of multiple trapping sites and annealed to approximately 60 K to remove any persistent unstable sites, we conclude the three components in the 3P1 state emission centered at 273 nm arise from the occupancy of Hg in a single site in xenon. The 42 K spectrum, shown on the right in Figure III.5, could also be fit adequately with three bands. An indication of the quality of the fits is provided in the upper panels, which show the residuals existing between data and fit. Also shown are the emission profiles generated in the fit. The comparison of the low and high temperature fits, shown on the bottom in Figure III.5, suggests that the origin of the unexpected blue shift occurring at higher temperatures in the emission band, arises from the increasing intensity of the unresolved blue component. However, when the spectra are plotted on absolute intensity, it becomes clear that it is the intensity of the central component, which is decreasing at higher temperatures. Lineshape analyses of the emission bands in the Hg/Ar and Hg/Kr systems were also conducted, the results of which are shown in Figure III.6 and collected in Table III.2. As in the Hg/Xe system, allowance had to be made in the Hg/Kr system for the presence of a weak 3P0 state emission. This component and a broad underlying feature, due to a thermally unstable site, occur to the red of the main 3P1 state emission in Kr. Adequate fits in Hg/Ar were only obtained when three components were allowed for, as shown on the left in Figure III.6. The temperature dependence in the Hg/Ar and Hg/Kr systems is simpler than in the Hg/Xe system in so far as the emission bands broaden and red shift with increasing temperature. Chapter III, Hg(3PJ)/RG Luminescence 48 3.8 3.6 3.4 Energy (x 104 cm-1) Hg/Xe T = 12 K 3.8 3.6 3.4 Energy (x 104 cm-1) T = 42 K 3.8 3.6 3.4 Energy (x 104 cm-1) Hg/Xe T = 12 K 3.8 3.6 3.4 Energy (x 104 cm-1) T = 42 K Figure III.5 Lineshape analysis of 3P1 emission in the Hg/Xe system. The panel on the left shows an acceptable fit of the spectra recorded at 12 K. In this fit, three broad Gaussian functions are required in addition to one, narrow function included for the 3P0 state emission. The panel on the right shows a fit of the high temperature emission. As in the 12 K fits, three broad and one narrow Gaussian functions provide an adequate fit. Energy (x 104 cm-1) Hg/Kr 12 K 3.95 3.9 3.85 3.8 3.75 28 K Energy (x 104 cm-1) Hg/Ar 12 K 4.0 3.95 3.9 22 K Energy (x 104 cm-1) Hg/Kr 12 K 3.95 3.9 3.85 3.8 3.75 28 K Energy (x 104 cm-1) Hg/Ar 12 K 4.0 3.95 3.9 4.0 3.95 3.9 22 K Figure III.6 Lineshape analyses of 3P1 state emission in the Hg/Ar and Hg/Kr systems. The panels on the top show acceptable fits of spectra recorded at 12 K, revealing the presence of multiple components in the Ar and Kr systems as observed in Hg/Xe. In contrast to the Hg/Xe system, little temperature dependence is exhibited as indicated in the panels on the bottom. The fourth Gaussian component used in the Hg/Kr system was required to account for a small amount of a red site present in this sample. Numerical values extracted in these fits are collected in Table III.2. Chapter III, Hg(3PJ)/RG Luminescence 49 Table III.2 Parameters extracted in Gaussian fits of the Hg/RG 12 K emission spectra produced with continuous lamp excitation. The band areas were determined by numerical integration of the fitted curves. Hg/RG Bandcentre ν0 (cm-1) Bandheight (counts) Bandwidth ∆ (cm-1) Integrated area (counts) Xe 37693.9 1658 84.3 1.488 x 105 37534.9 3151 710.5 2.383 x 106 36618.8 15229 1256.6 2.037 x 107 35728.9 1924 2455.8 5.020 x 106 Kr 39486.9 6883 299.0 2.190 x 106 39334.7 1249 371.7 4.944 x 106 39149.0 5122 503.0 2.742 x 106 38513.515 619 1650 1.050 x 106 Ar 40038.0 5482 252.5 1.473 x 106 39899.3 9309 329.4 3.263 x 106 39734.6 2967 467.2 1.475 x 106 Time-resolved emission spectra (TRES) produced with pulsed laser excitation and iCCD detection are shown in Figure III.7 for Hg/Kr. Such scans yielded short lived (nanosecond) decay times for the Hg(3P1)/RG emission features. 400 300 200 100 0 0 500000 1000000 1500000 2000000 250 252 254 256 258 260 262 Hg(3P1→1S0)/Kr λEx = 249 nm Wavelength (nm) Time (nsec) Raw Counts 400 300 200 100 0 0 500000 1000000 1500000 2000000 250 252 254 256 258 260 262 Hg(3P1→1S0)/Kr λEx = 249 nm Wavelength (nm) Time (nsec) Raw Counts Figure III.7 Hg/Kr time–resolved emission spectrum corresponding to the Hg 3P1 → 1S0 nanosecond fluorescence recorded at 12 K produced with pulsed laser excitation and iCCD detection. Chapter III, Hg(3PJ)/RG Luminescence 56 deviated from those recorded at 12 K for the Hg/Xe system only. Figure III.13 presents a comparison of the decay profiles recorded for the three emission features (263, 273 and 280 nm) at 42 K. It is evident in Figure III.13 that the decay profiles of the 263 and 273 nm features, represented by open triangles and stars respectively, are equivalent. This is reflected in the observed decay times ( τ obs) extracted at 46 K, collected in Table III.4, where the value of the emission (35.4 nsec) 263 nm is equal to that extracted for the 273 nm feature. 050 100 150 200 Ti m e ( nsec ) 104 106 108 Emission, λEx= 254.0 nm λEm = 263.0 nm λEm = 273.0 nm λEm = 280.0 nm Laser Hg/Xe Decay profiles of the emission components at 42 K Figure III.13 A comparison of the temporal decay profiles recorded for the three emission components identified in the lineshape analysis (Figure III.5) of the steady-state emission spectrum recorded for the Hg 3P1 → 1S0 transition in solid Xenon at 42 K. This observation reveals the Hg/Xe decay profile is temperature dependent consistent with the lineshape analysis completed which showed that the 273 nm feature is diminished at temperatures in excess of 30 K. Therefore the temperature dependence in the decay time extracted is reflecting the behaviour of the 266 and 273 nm components identified in the lineshape analysis shown in Figure III.5. Chapter III, Hg(3PJ)/RG Luminescence 57 III.2.III Hg 3P0 → 1S0 emission spectra In this section, details of the 3P0 state emission of mercury atoms resulting from intermultiplet relaxation (IMR) following pulsed laser excitation of the 3P1 level are presented. The combination of a high intensity, low-repetition excitation source (Nd:YAG laser) with photon counting detection is used for recording these spectra as it favours the long-lived 3P0 emission over the nanosecond 3P1 fluorescence. Hence the resulting spectra are free of the 3P1 state emission bands described in the preceding section. The excitation wavelengths chosen correspond to the Hg 3P1 ← 1S0 transition in solid Ar, Kr and Xe, at 245.9, 248.8 and 253.0 nm respectively. As shown in Figure III.14, the emission features resulting from pulsed laser excitation are centered at 258.9, 260.8 and 265.1 nm in Ar, Kr and Xe. The spectra shown in Figure III.14 exhibit a progressive red shift and decreasing linewidth from Ar to Xe. The red shift mirrors that of the 3P1 emission but the linewidth behaviour is the reverse of that shown in Figure III.2 for the 3P1 fluorescence. It should be noted that the progressive red shift evident in Figure III.14 was not present in C&T’s data7. In their spectra, the position of the Hg/Ar emission was intermediate between that of Kr and Xe. It is thought that the 3P0 data presented in the earlier Hg/Ar work, corresponds to Hg occupancy in a secondary site13 of argon. Chapter III, Hg(3PJ)/RG Luminescence 58 37.538.038.539.0 x103 E n e r g y ( c m -1 ) Emission Intensity (Arb. Units) 258 260 262 264 266 Wavelength (nm) Ar Kr Xe Hg(3P0)/RG Figure III.14 A summary of the long-lived emission features recorded in the Hg/RG systems at 12 K. These emission spectra were produced with pulsed laser excitation of the Hg atom 3P1 ← 1S0 transition and are gated to show only long-lived 3P0 emission. Note the increasing red-shift in the emission bands on going from Ar to Xe but the decreasing linewidth. The emission features in Kr and Xe matrices reveal fine-structure splitting when recorded under high resolution. As shown in Figure III.15, Hg/Xe exhibits a narrow line (fwhm = 10.5 cm-1) at 265.12 nm, (37718 cm-1) and a broader red feature (fwhm = 70 cm-1 at 265.54 nm, (37658 cm-1). High temperature scans, also shown in Figure III.15, indicate that the sharp feature is reversibly removed while the broad feature broadens and red-shifts. These lineshapes and their temperature dependence are characteristic of a zero phonon line (ZPL), for the narrow blue feature and a phonon sideband for the broader, red feature. Two features are also evident in the Hg/Kr system although not as well resolved as in Hg/Xe, but exhibiting the same temperature dependence. The main band of Hg/Ar at 258.9 nm exhibits little temperature dependence, except that the pair of weak side-bands at 261.0 and 263.7 nm are removed at high temperature. Chapter III, Hg(3PJ)/RG Luminescence 59 37.538.038.539.0 Emission Intensity (Arb. Units) 258.0 261.0 264.0 12 K 22 K Hg(3P0)/Ar, λEx. =245.9nm 38.138.238.338.4 x103Ene r g y ( c m -1 ) 260.4 261.1 261.8 262.5 Wavelength (nm) 12 K 25 K Hg(3P0)/Kr, λEx. = 248.8 nm 37.637.737.7 265.0 265.5 266.0 12 K 30 K Hg(3P0)/Xe, λEx. =253nm Figure III.15 High resolution scans of the Hg 3P0 state emission in Ar, Kr and Xe yielding resolved fine structure in the Hg/Kr and Hg/Xe systems. The solid traces were recorded at 12 K. High temperatures scans are shown by the dotted lines indicating reversible changes in the Hg/Kr and Hg/Xe spectra. The main band in the Hg/Ar system exhibits little temperature dependence but the weak pair of red bands, assigned to defect site occupancy, are quenched at elevated temperatures. To investigate the origin of the splitting observed at high resolution for the 3P0 state emission in Hg/Xe, a lineshape analysis was conducted using the Wp optical function. This function was originally derived by Huang and Rhys22. It is described in detail by Struck and Fonger23 and given by Erreur ! Equation (III.3) In this expression r = exp(−h, −ω /kT) and Ip(x) is a modified Bessel function of variable order p and at a given temperature T, of fixed argument θ = 2Sr½/(1 − r). The function W provides the distribution of intensity as a function of phonon number, p. It essentially provides Franck-Condon intensity factors for a single phonon mode of frequency h, −ω , coupling to the electronic transition with a strength S, a variable known as the Huang-Rhys factor. For large S, the Wp function approaches a Gaussian function. To maintain numerical accuracy in the weak electron-phonon characteristics of the Hg 3P0 state emission spectra, the sum form of the Wp function was used in our programming. The sum form is Erreur ! Equation (III.4) Chapter III, Hg(3PJ)/RG Luminescence 60 and its evaluation is truncated at θ m, the next integer greater than θ + 1. Clearly, the sum increases with the electron-phonon coupling strength S and the temperature. A fit with the Wp function allows identification of the position of the zerophonon line (ZPL or ν0,0), and the magnitude of the strength electron-phonon coupling S once the recorded spectrum has been transformed into phonon units, p. This initially involves estimating the magnitude of the phonon frequency, h, −ω , a task which was direct in the 12 K Hg/Xe spectrum due to the ZPL being resolved. A satisfactory fit, obtained with S = 1.3 and h, −ω = 21.0 cm-1, is shown on the left in Figure III.16. The fit verifies the ZPL is located at 37718 cm-1, the phonon side-band is, as expected, to the red of this. In addition it reveals the presence of weak, ‘hot’ emission at 37739 cm-1 to the blue of the ZPL. Using the 12 K fit parameters (S, h, −ω and ν0,0) the lineshape calculated with T = 30 K is compared on the right in Figure III.16 with the emission spectrum recorded at this temperature. Evident in this plot is the diminished intensity of the ZPL and the gain in the intensity of the phonon sideband now showing a maximum at 37673 cm-1. Similar fits were performed on the 3P0 emission in Hg/Kr and Hg/Ar. A satisfactory fit was obtained in Hg/Kr with S = 2.2, h, −ω = 28.0 cm-1 and ν0,0 = 38366 cm-1. The comparison shown on the right in Figure III.17 indicates that the ZPL is not resolved in Hg/Kr but is located as a blue shoulder on the partially resolved feature. The fits conducted in Hg/Ar are not as definitive as those in Hg/Xe or Hg/Kr because fine-structure splitting has not been resolved to provide an initial estimate of the phonon frequency, h, −ω . However, the lineshape generated with S = 3.3, h, −ω = 41.0 cm-1 and ν0,0 = 38740 cm-1 compares well, as shown on the left in Figure III.17 with the observed emission band. This fit indicates the ZPL is too weak to be resolved in the emission spectrum. The results of the Wp lineshape analyses are presented in Table III.5. Chapter III, Hg(3PJ)/RG Luminescence 61 37.60037.700 Emission Intensity 265.0 265.5 266.0 266.5 Data Wp Fit Hg/Xe T=12 K W a v elength (n m ) x103Photon Ene r g y ( c m -1 ) S= 1.3 hω=21.0 cm-1 v00=37718cm -1 ZPL * 37.60037.700 265.0 265.5 266.0 266. 5 T=30 K Figure III.16 The Wp lineshapes calculated with Equation III.4 for the 3P0 state emission in Hg/Xe recorded at 12 and 30 K and produced with pulsed laser excitation at 253 nm. The location of the zero-phonon line in the spectrum is indicated as ZPL and numerically as ν0,0. The presence of ‘hot’ emission in the 12 K spectrum indicated by the asterix is made evident in the comparison of fit and data, which reveals the existence of a line to the blue of the ZPL. 38.538.638.738.8 Emission Intensity 258.0 258.5 259.0 259.5 260.0 Data Wp Fit Hg/Ar T=12 K Wa v e l e n g t h ( n m) x103P h o t o n E n e r g y ( cm-1 ) S= 3.3 hω=41.0 cm-1 v00= 38740 cm-1 ZPL 38.138.238.338.4 260.5 261.0 261.5 262.0 Hg/Kr T=12 K S= 2.2 hω=28.0 cm-1 v00= 38366 cm-1 ZPL Figure III.17 The Wp lineshapes calculated for the 3P0 state emission in the Hg/Ar and Hg/Kr systems. The Hg/Ar emission was produced with laser excitation at 245.9 nm, the Hg/Kr spectrum with 248.8 nm excitation. Chapter III, Hg(3PJ)/RG Luminescence 62 Table III.5 The location of the ZPL extracted in the Wp lineshape function analysis conducted on the recorded atomic Hg 3P0 ↔ 1S0 emission in solid Ar, Kr and Xe. For comparison purposes, the results of a similar analysis on the central component in the Hg/Xe 3P1 ↔ 1S0 emission are also given. Hg/RG Transition ZPL, ν0,0 (cm-1) S h,− ω (cm-1) Hg/Ar Hg/Kr Hg/Xe 3P0 ↔ 1S0 3P0 ↔ 1S0 3P0 ↔ 1S0 38740 38366 37718 3.3 2.2 1.3 41 28 21 Hg/Xe 3P1 ↔ 1S0 ~39000 105 21 The two weak features in the Hg(3P0)/Ar emission spectrum at 261.17 and 263.75 nm (see Figure III.14 and Figure III.15), have always been present in the samples we have prepared. As shown in Figure III.15, both of these low energy features are absent in the high temperature scans but re-appear at 12 K. It has also been observed that the intensities of these two features are increased dramatically by the introduction of a third species to the Hg/Ar matrix. This is most evident in Ar samples containing less than 0.1% Xe, where the intensities of the red-pair increase together relative to the 258.9 nm band. Crepin et al.7 previously assigned the 263.75 nm feature to an Hg:H2O complex in the Ar matrix, however, we find this band always accompanies the 261.17 nm feature and both are present in Ar samples free of water. Thus, we assign this pair of bands to mercury atom occupancy in imperfect sites in the Ar lattice, for example, a substitutional site with one of the 12 nearest neighbour atoms missing. Long decay times, ranging from 8 – 530 ms, were recorded for the emission bands of Ar, Kr and Xe shown in Figure III.14 and Figure III.15, verifying that they correspond to the strongly forbidden Hg 3P0 → 1S0 transition. Least squares fits with a triple–exponential function were required to obtain adequate fits, as shown in Figure III.18 for the Hg/Xe system. This is indicative of complex decay kinetics, behaviour already reported7 by Crepin et al. It should be noted however, that by far the largest component in the decay curve has the shortest decay time ( τ = 8.8 ms) while the longest decay ( τ = 541 ms) is found only on the phonon sideband band and is the weakest component. An investigation of the temperature dependence in the decay times extracted monitoring the 265.3 nm emission up to 30 K allowed the assignment of the 8.8 ms component as the radiative lifetime for the 3P0 → 1S0 transition in solid Xe. Chapter III, Hg(3PJ)/RG Luminescence 63 0.0 0.3 0.5 Time (sec) 10 100 1000 10000 λem= 265.124 nm λem=265.3nm τ1=275.0msec A1=1618 τ2=093.0msec A2=1697 τ3=008.7msec A3=36439 τ1=541.0msec A1=2410 τ2=141.2msec A2=7925 τ3=008.9msec A3=23402 37.537.8 x103Ener g y ( cm-1 ) 265.0 265.5 266.0 266.5 267.0 267.5 Wa v elen g th (nm) 12 K λEx. = 253 nm Hg(3P0)/Xe Figure III.18 Hg/Xe high-resolution time-gated emission spectra produced with laser excitation at 253 nm. Inset results of the non-linear least squares analysis completed and the decay characteristics extracted for the resolved emission features at 265.124 and 265.3 nm respectively. III.3 Discussion III.3.I Hg 3P1 → 1S0 emission The gross spectral features recorded for the Hg 3P1 → 1S0 transition in the present study are in very good agreement with Crepin and Tramer’s7 earlier work. High resolution scans of the emission bands reveal complex profiles, requiring multiple emission components to yield adequate lineshape fits. Of the Hg/RG systems, the most complex behaviour is exhibited by the emission in Hg/Xe, which shows a blue shift with increasing temperature. Calculation of the Wp optical function for the 3P1 state emission in xenon yields a Gaussian curve with an electron-phonon coupling strength S of 105. This is almost two orders of magnitude greater than the value of 1.3 extracted for the 3P0 state emission and closely replicates the central Gaussian in the lineshape analysis of the 273 nm band. Gaussian fits done as a function of temperature reveal that the intensity of the central component is diminishing at high Chapter III, Hg(3PJ)/RG Luminescence 64 temperature. The origin of the multiple emission components in the Hg/RG systems is known from excitation scans not to be due to multiple site occupancy. Pairpotential simulations described in Chapter IV, present a model, which explains the origin of multiple emission features for atomic mercury isolated at a single site. These calculations also suggest a mechanism for the quenching of the central emission component in the 273 nm emission in the Hg/Xe system. Time-resolved emission spectra recorded allowed the assignment of the nanosecond radiative lifetimes τ rad of the emission features corresponding to atomic Hg 3P1 → 1S0 relaxation in solid Ar, Kr and Xe. The decay profiles extracted from the time-resolved emission spectra exhibited different decay characteristics for the emission components identified in the lineshape analysis completed for Hg/Xe system, providing further evidence for the multi-component nature of the Hg(3P1 → 1S0)/Xe emission. III.3.II Hg 3P0 → 1S0 emission The spectral and temporal characteristics of the Hg atom 3P1 and 3P0 excited state emissions are very different making them, as the comparison presented in Figure III.4 for Hg/Xe illustrates, easy to differentiate. Although the 3P0 state was easily observed with laser excitation in the three rare gas hosts used, the actual amount of it produced with 3P1 state excitation is rather small relative to emission of the 3P1 state. The inefficiency of the 3P1 → 3P0 intramultiplet relaxation is made clear in Figure III.2, where the only Hg/RG system showing a clear sign of the 3P0 state in continuous (time-integrated) emission is Hg/Xe. Even in Hg/Xe, the ratio of the 3P1 state radiative decay to the 3P1 → 3P0 intramultiplet relaxation is estimated as 200:1 from the lineshape analysis presented in Figure III.5 and the corresponding numerical data collected. Moreover, the efficiency of the intramultiplet relaxation is not enhanced with increasing temperature up to 42 K as shown in Figure III.5. Optical lineshapes generated with the Wp function indicate the 3P0 emission in the solid rare gases involves weak electron-phonon coupling. The S values extracted are 1.3, 2.2 and 3.3 for Xe, Kr and Ar respectively. In solid-state spectroscopy, weak electron-phonon coupling is indicative of unshifted potentials for the two states involved in the optical transition. While the ground and many of the excited state potentials of the Hg⋅RG complexes are accurately known, the ã30(3Σ) states of the Chapter III, Hg(3PJ)/RG Luminescence 65 Hg(3P0)⋅RG diatomics cannot be determined directly be spectroscopic means. However, in the limit of Hund’s case-c coupling, it can be extracted from the A31(3Π) and B30(3Σ) states24,25 of the Hg(3P1)⋅RG diatomics with Equation III.5 V(3P0) ã30 = 1/3[(VΣe + VΠe) + VΠ] Equation (III.5) presented by Duval et al26. In this expression VΠ is the spectroscopic A state, while VΣ is determined from the A and B states from the relationship VΣ = 2VB - VA. The potential energy curves extracted for the ã30(3Σ) states of the Hg(3P0)⋅RG diatomics from the spectroscopic A and B states are shown in Figure III.19 and their key parameters are collected in Table III.6. It is evident on inspection of the curves shown in Figure III.19 that the ground Hg(1S0)⋅RG X 10(1Σ) and excited Hg(3P0)⋅RG ã 30(3Σ) state potential energy curves are very similar. Conversely the form of the excited Hg(3P1)⋅RG A31(3Π) state potentials are very different to the ground states. This explains the very strong electron-phonon exhibited on the Hg(3P1) state emissions (e.g., Hg/Xe, S = 105) and the very weak coupling on the Hg(3P0) state emission (S = 1.3). 3.0 4.0 5.0 -1000 -500 0 Energy (cm-1) Hg/A r 3.0 4.0 5.0 R ( Å ) Hg (3P1)Π Hg (3P0)a Hg (1S0)X Hg/K r 3.0 4.0 5.0 Hg/Xe Figure III.19 Potentials of the Hg⋅RG diatomics of relevance to the Hg(3P0)/RG matrix emission spectra. The X and the A(Π) state potentials are obtained directly from spectroscopic data of the Hg⋅RG diatomics given in Ref.26 while the ã30state potential was obtained with Equation III.5 which assumes case-(c) coupling. Chapter IV, Hg(3P1)/RG Sims. 72 IV.2 Methods and Results The assumptions on which the localized cluster model is based are I) the photophysical properties (absorption and emission energies) of the metal atom chromophore are governed by its immediate environment within the rare gas solid. II) The interactions between the metal atom and the surrounding RG atoms can be described completely by the sum of diatomic pair potentials13. III) The atomic electronic angular momentum, Je, at the molecular asymptote is conserved within the cluster. The validity of the pair potentials approach has been examined by Beswick et al.2 for the triatomic Hg(3P1)⋅Ar2 complex by simulating the vibronic structure in the resonance two-photonionization (R2PI) spectra recorded for this cluster by Jouvet and co-workers3. A similar theoretical approach has also been used by Alexander and coworkers4 on the ground state B(2PJ)⋅Ar2 system. McCaffrey and Kerins1 adapted Beswick’s cluster method for the solid state in a simulation of the spectroscopy of 1P1 ↔ 1S0 transition of atomic zinc in the solid rare gases based on metal atom occupancy in a substitutional site. The most fundamental aspect of the solid-state calculations is the site occupied by the ground state metal atom in the rare gas lattices. From the ground state bond length (Re) data presented in Table IV.1 for the mercury atom-rare gas atom diatomics and the rare gas dimers, very good matches exist between the Xe2 and Kr2 systems and their Hg⋅RG counterparts. Very favourable matches also exist for the Hg⋅RG van der Waals bond lengths and the substitutional site (ss) sizes of the solid rare gases. Thus in solid Xe, ss is 4.334 Å, calculated from the lattice parameter14 a = 6.13 Å, while the Hg⋅Xe bond length is 4.25 Å and in solid Kr, ss is 3.991 Å while Re Hg⋅Kr is 4.07 Å. The match that exists for Hg in Ar is not quite as good, where ss is 3.756 Å and Re Hg⋅Ar is 3.98 Å. However, even in Hg/Ar, substitutional site occupancy is also expected15. Following identification of the site of isolation, the energy of a guest metal atom (M) occupying a substitutional site in a solid rare gas (M/RG) system is calculated for an M⋅RG18 cluster. The rare gas atoms in this cluster fall, as shown in Figure IV.1, into two categories based on whether they are located in the first or second sphere of host atoms surrounding the guest atom M. The first category has a cubo-octahedral arrangement of 12 host atoms located at a nearest neighbour (nn) distance of a/√2 from M. The other set, consisting of 6 atoms are located at a next Chapter IV, Hg(3P1)/RG Sims. 73 nearest neighbour (nnn) distance of the lattice parameter, a, from the guest atom and are arranged as a regular octahedron on the X, Y and Z axes. Table IV.1 Spectroscopic constants used to generate the Morse potential energy curves for the Hg⋅RG and RG2 diatomics. Data source are indicated by the references. Hg⋅RG State Morse Parameters Hg⋅Ar2 Hg⋅Kr10,11 Hg⋅Xe11 X 1Σ (10+) µHg-RG (amu) De(cm-1) ωe(cm-1) ωexe(cm-1) Re(Å) β(Å-1) 33.3614162 130.25 23.5 1.1 3.98 1.448348 59.2819820 178 20 0.54 4.07 1.40557 79.7926882 254 18.3 0.33 4.25 1.249072 A 3Π (30+) De(cm-1) ωe(cm-1) ωexe(cm-1) Re(Å) β(Å-1) 353.63 41.2 1.2 3.34 1.54104964 517, (628.7) 43.5, (40.63) 1.5, (0.691) 3.52, (3.35) 1.793813, (1.5193512) 1380.9 54.17 0.565 3.15 1.585736 B 3Σ (31) De(cm-1) ωe(cm-1) ωexe(cm-1) Re(Å) β(Å-1) 51.57 11.4 0.6 4.66 1.1166067 96, (104.8) 11.3, (11.1) 0.32, (0.301) 4.57, (4.58) 1.081374, (1.0166592) 187.6 9.71 0.215 4.47 0.77118 Ar⋅Ar16 Kr⋅Kr16 Xe⋅Xe16 X 1Σ De(cm-1) Re(Å)17 β(Å-1) 99.545 3.7565 1.40218 138.4 4.017 1.604 196.24 4.3634 1.509 The expectation of substitutional site occupancy of atomic Hg within solid Ar, Kr and Xe lattices is tested by comparison of the predicted and observed absorption energies. The absorption energy is calculated as the difference between the energy of the cluster in the ground and excited state within the Frank – Condon principle. The method of evaluating the cluster energies for the ground Hg(1S0)/RG18 and excited Hg(3P1)/RG18 states is outlined in the following sections. Chapter IV, Hg(3P1)/RG Sims. 74 Figure IV.1 The guest atom-based co-ordinate system used to calculate the energy of a metal atom, M in a substitutional site of an fcc lattice. The 12 nearest neighbour (nn) Rg atoms surrounding the guest atom located at the origin are shown as grey spheres on the edges of the cubic unit cell. The 6 next nearest neighbour (nnn) atoms in the second surrounding sphere are shown as the dark spheres on the X, Y, Z axes at the lattice parameter distance, a, from the guest metal atom. The axis system shown is co-incident with the three, fourfold (C4) symmetry axes of the cubo-octahedral fcc unit cell and is referred to in the text as 4-atom mode calculations. The image was generated by the gOpenMol18 programme. IV.2.I Ground 1S0 state The ground 1S0 electronic state of a closed shell ns2 metal atom within a M/RG18 cluster is simple to evaluate given the spherical symmetry of the atomic electronic angular momentum. The interactions between the ground 6s2 1S0 state mercury atom and the RG18 cluster is then simply a pair–wise sum of the interaction potentials at a specific distance (Rk). Because for Je = 0, there is no angle dependence. This is evident in Equation IV.1, used to evaluate the ground state energy of the cluster, in which the energy is obtained as a sum of the Hg⋅RG and RG2 pair-potentials. WX(R) = Erreur ! Equation (IV.1) Morse functions19 are used for the ground state potentials for the Hg⋅RG diatomics and for the rare gas dimers, RG2 (RG = Ar, Kr and Xe). The parameters used for these functions are listed in Table IV.1. IV.2.II Excited 3P1 state In order to calculate the mercury atom 6p 3P1 ↔ 6s 1S0 absorption and emission energies in the solid, it is necessary to obtain the energy of the electronically excited Chapter IV, Hg(3P1)/RG Sims. 75 3P1 state mercury atom in the Hg⋅RG18 cluster. This is more difficult than the ground X state due to the axial symmetry of the electronic angular momentum, Je = 1. The method and notation used by Beswick and coworkers2 was followed throughout. The Hg⋅RG molecular states arising from the approach of the Hg (3P1) to the closed shell rare gas atom with the internuclear axis defined as the Z–axis have pure Π and Σ arising from the projection of the atomic angular momentum onto the internuclear axis corresponding to px, py and pz orbital occupancy. The cluster potentials are generated by applying the Wigner rotation matrices using Hund’s case-(a) quantum numbers as the diatomic basis set. The general solution for Je = 1 corresponding to the three 6p-orbitals of the excited 3P1 state mercury atom give rise to the following cluster states, where n is number of metal-rare gas bonds in the cluster. (pz) W1(R) = Erreur ! Equation (IV.2) (px) W2(R) = Erreur ! Equation (IV.3) (py) W3(R) = Erreur ! Equation (IV.4) In contrast to the ground state potential, given by Equation IV.1, the excited state energy is not a simple sum of the VΠ(R) and VΣ(R) pair potentials, but now depends on the angle variables1 θ k and φ k. In a Cartesian co-ordinate system having the metal atom positioned at the origin, θ k is the angle subtended between a rare gas atom k and the Z-axis, while φ k is the angle obtained by projecting the vector connecting this atom and the origin onto the XY plane as shown in Figure IV.2. The VΠ(R) and VΣ(R) terms appearing in Equations IV.2-4 are the pure Π and pure Σ spatial state potentials and not the spectroscopic A and B states of the Hg⋅RG diatomics presented in Table IV.1. As the B(ΩA = ± 1) state is a linear combination of Π and Σ orientations, it is necessary to extract the pure Σ potential for use in Equations IV.2-4. This was achieved using the relationship20 V B = ½[VΣe + VΠe] which yields the spatial Σ component of the original B state. VΣ(R) = 2VB(R) – VA(R) Equation (IV.5) Since the A(ΩA = 0) state is of pure Π symmetry (VA = VΠe) this potential was used directly. The result of the deconvolution of the pure Σ component from the B state is shown in Figure IV.3 for the Hg⋅RG diatomics (RG = Ar, Kr and Xe). Chapter IV, Hg(3P1)/RG Sims. 76 RG(4) RG(1) Y re Y X RG(1) X re RG(2) Hg φ1 Z θ1 R1φ1 r RCM Hg r RG(2) RG(3) Figure IV.2 Diagram of the metal atom based coordinate system illustrating the origin of the variables used in Equations (IV. 2-4) for the calculation of the excited state energies. It is clear from Figure IV.3 (dashed line) that the Σ states deconvoluted with Equation IV.5 are not completely repulsive as they all show weakly bound regions at long range. This behaviour is consistent with a slightly attractive van der Waals interaction which exists between the metal atom Hg(pz) orbital and the rare gas atoms at long range. At short range, however, it was observed for all the Hg⋅RG systems, that the deconvoluted Σ state curves exhibited a non-physical minimum, instead of increasing exponentially like the B State. Fortunately, this non-physical behaviour of the Σ state (not shown in the plot) does not occur in the range of distances involved in the solid-state simulations. Even in Hg⋅Xe (the worst case of the Hg⋅RG diatomics) the Σ state becomes non-physical at distances less than 2.8 Å. As this is less than the shortest distance encountered for substitutional site occupancy (3.065 Å, half the lattice parameter of Xe), the raw deconvoluted Σ potential was used in all the excited state calculations. Chapter IV, Hg(3P1)/RG Sims. 77 3.0 4.0 5.0 6.0 7.0 -1500 -1000 -500 0 Energy (cm-1) Rmin=5.222 Å De= 58.464 cm-1 3Σstate parameters B( 31) A3Π(30+) 3Σ(deconvoluted) Hg(3P1)Ar 3.0 4.0 5.0 6.0 7.0 R ( Å ) Rmin=5.105 Å De= 96.484 cm-1 B( 31) A3Π(30+) 3Σ(deconvoluted) Hg(3P1)Kr 3.0 4.0 5.0 6.0 7. 0 Rmin=5.277 Å De=201.578 cm-1 B( 31) A3Π(30+) 3Σ(deconvoluted) Hg(3P1)Xe Figure IV.3 The Hg(3P1)⋅RG 3Σ states extracted with Equation IV.5 from the spectroscopic [A 30+ (3Π)] and [B 31] states whose constants are presented in Table IV.1. All of these states share the Hg atom 6p 3P1 asymptote at 39424.1 cm-1 but are shown dissociating to zero-energy for the purpose of comparison. Note that all the deconvoluted Hg⋅RG 3Σ states show a weakly bound region at long internuclear distance. These distances are listed as Rmin in the plots while the binding energies are indicated by De. Excited 3P1 state energies were calculated for the body and waist vibronic (2) modes for the three p-orbital (3) orientations. The energetics of the two modes and the three orbital orientations were determined for three co-ordinate systems based on the three symmetry poles (3) of the cubo-octahedron. The symmetry poles12 are fourfold, threefold and twofold symmetric and their calculations are referred to in this presentation as 4-atom, 6-atom and 2-atom modes respectively. The co-ordinate systems of the cubo-octahedron based on the three symmetry poles are shown in Figure IV.4. Thus for a given rare gas host, a total of 18 excited state potential energy curves were calculated for an Hg⋅RG18 cluster. Energies of the co-ordinate displacements were calculated at 0.001 Å intervals. As a check of the correctness of our code, absorption energies were compared for the three co-ordinate systems calculated. This can be used as a check since the three p-orbitals are degenerate at the centre of the cubo-octahedral substitutional site. Hence at R = 0, all modes must Chapter IV, Hg(3P1)/RG Sims. 78 produce identical absorption values irrespective of the vibronic mode, the symmetry co-ordinate system used or the orbital selected. Figure IV.4 Three co-ordinate systems based on the three symmetry poles of the cubooctahedron with the metal Hg atom positioned in a substitutional site within the RG18 cluster. In fourfold symmetry the Z-axis is aligned with one of the six fourfold symmetry poles of the cubo-octahedron (the centre of the cubic faces). In the threefold and twofold symmetries, the Z-axis is aligned with one of the eight threefold symmetry poles (eight vertices of the cube) and the twelve twofold symmetry poles (one of the twelve edges of the cube) respectively. In generating the solid state localized model, individual calculations were preformed for smaller clusters and the resulting energies were used to obtain the energy of the Hg(3P1)/RG18. These calculations simulated the motion of the excited state Hg atom along the symmetry poles of the cubo-octahedron and provided information on interactions involved in ‘body’ mode calculations, where the 3P1 Hg undergoes large amplitude motions within the rare gas lattice. The specific details of X Z Y Fourfold Symmetry X Z Y Fourfold Symmetry Z Y Z X Threefold Symmetry Twofold Symmetry Z Y Z X Threefold Symmetry Twofold Symmetry Chapter IV, Hg(3P1)/RG Sims. 79 the high symmetry cluster calculations completed are now presented. The results obtained and the connection between these smaller clusters (RGn, n ≤ 6) and the RG18 model are highlighted in the sections that follow. IV.2.II.I Tetragonal (4-atom) symmetry modes a. Q2 is the 4-atom ‘body’ mode presented in earlier work1 on the Zn/RG systems. In this contribution, the energetics of this mode are given not only for the pz orientation but also for the degenerate px/py set. These orientations correspond to the 3A and 3E states respectively of the excited Hg atom in a substitutional site. Details of the Hg⋅RG interactions considered in the calculation of the 4–atom ‘body’ mode are presented in Figure IV.5. The energetics of the Q2 mode i.e., motion of the metal atom along the Z-axis in the four-fold symmetry system (Figure IV.4 top panel), is constructed from the interaction of the excited state Hg atom with four smaller clusters, all of fourfold symmetry. The smaller clusters are shown in Figure IV.5, in which S.site1 and S.site2 represent two planar arrangements of four RG atoms positioned at the nearest neighbour distance (a/√2) and next nearest neighbour distance of the lattice parameter a respectively from the Hg atom positioned at the centre of the substitutional site. The second set of fourfold symmetry clusters Ioh1 and Ioh2 are two RG5 clusters forming square pyramids, whose base is located at the centre of the octahedral interstitial sites, at a distance of half the lattice parameter (a/2) from the centre of the substitutional site. Prior to completing the 4-atom ‘body’ mode calculations, the interaction of the excited Hg atom with these RG4 and RG5 clusters was examined. The highest symmetry motion modelled by these cluster calculations was that of the C4V approach of the Hg (3P1) atom towards a planar arrangement of four RG atoms1. This motion is representative of the motion of the metal atom from the centre of the substitutional site towards an octahedral interstitial site. This motion was restricted to the Z-axis perpendicular to the plane of the four RG atoms, (Figure IV.5) By choosing, the Cartesian co-ordinate system to coincide with the three 4fold symmetry axes of the cubo-octahedron (Figure IV.4, top panel), the general sum expressions given by Equations IV.2-4, reduce1 to the simpler product expressions which provide the energies of the 3A(pz) and doubly degenerate 3E(px, py) states of an excited 3P1 state metal atom in a substitutional site. W3A1(R) = 4[] cos2 θ V3Σ(R) + sin2 θ V3Π(R) Equation (IV.6) Chapter IV, Hg(3P1)/RG Sims. 80 W3E(R) = 2[] sin2 θ V3Σ(R) + (cos2 θ + 1)V3Π(R) Equation (IV.7) S.site 1 Ioh2 R4 R3 R2 R1 q2 θ2 θ3θ4 θ1 Four - fold, 'body' mode S.site 2 Ioh1 Hg atom 12 RG nn 6 RG nnn Extended RG lattice Figure IV.5 A diagram representing the distance and angle parameters used in the calculation of the Tetragonal fourfold symmetry, ‘body’ mode, Q2. The motion of the metal atom from the substitutional site towards the octahedral interstitial site Ioh1 along in the fourfold symmetry Z-axis is represented as q2. Shown are the 12 nearest neighbour (nn) RG atoms arranged as three RG4 square planar clusters in the XY plane. The six next nearest neighbour (nnn) atoms are also shown, four of these identified as S.site2 are positioned as a RG4 cluster in the XY plane with each RG atom at the lattice parameter distance from the Hg atom at the centre of the substitutional site. The final two nnn atoms are positioned on the Z-axis positioned in the apical position of RG5 square pyramidal clusters, the base of which corresponds to the octahedral interstitial site. Equations IV.6 and 7 were obtained upon summation of the expressions resulting after substituting values of φ k = kπ/4, k=1, 2, 3 and 4 in Equations IV.3 and 4. Plots of the 3A(pz) and 3E(px, py) states as a function of distance from the centre of mass Rcm of the Xe4 cluster are shown in Figure IV.6. The variables, R and θ in Equations IV.6 and 7 are obtained from the relations R = √(Rcm2 + r2) and θ = sin-1(r/R) where r = re/√2 is the distance of each Xe atom from the centre-of-mass of the Xe4 cluster. The distance between the Xe atoms arranged in the Xe4 (Figure IV.2) cluster is set at the equilibrium internuclear separation (re) of the ground state Xe dimer (see Table IV.1). It is observed in Figure IV.6 that the Hg(pz) 3A1 state is stabilised upon approaching the planar Xe4 cluster exhibiting a dissociation energy De = 5459.3 cm-1 at Rcm = 0 Å, corresponding to the centre of mass of the cluster. The doubly Chapter IV, Hg(3P1)/RG Sims. 81 degenerate 3E (px, py) states are also stabilized up to 2.38 Å from the centre of the cluster. However, further approach results in destabilisation. 0 1 2 3 4 5 6 RCM (Å) 3.4 3.6 3.8 4.0 4.2x104 Energy (cm-1) 3A1(pz) 3E(p x/py) 3A1(pz) Emin=33964.8 cm-1 Rmin=0.000 Å De=5459.3cm -1 3E(p x/p y) Emin=38030.8 cm-1 Rmin=2.380 Å De=1393.3cm -1 H g X e4C4v A p p roac h Figure IV.6 Potential energy curves of the 3A1 and doubly degenerate 3E states calculated using Equations IV.6 and 7 as a function of distance of the Hg atom from the centre of mass (Rcm) of the Xe4 cluster. The approach of the atomic Hg occurs along Z-axis perpendicular to the planar Xe4 cluster. The 4-atom ‘body’ mode, Q2 involves motion of the excited state metal atom along the Z-axis from the centre of the substitutional site towards the octahedral interstitial site located on the face of the unit cell as shown in the fourfold coordination system Figure IV.4, top panel. The octahedral interstitial site (Ioh) is positioned at the base of a square pyramid of RG atoms. Therefore, in simulating the approach of the excited state Hg atom to this site, an extension of the Xe4 cluster calculation to a C4v square pyramid of Xe5 atoms is required. This is achieved with the addition to the Hg⋅Xe4 expressions of a VΣ(R + r) for the 3A1 electronic state (Equation IV.6) and a VΠ(R + r) for the 3E state (Equation IV.7) as shown in Equations IV.8 and IV.9 respectively. W3A1(R) = 4[] cos2 θ V3Σ(R) + sin2 θ V3Π(R)+ V Σ(R + r) Equation (IV.8) W3E(R) = 2[] sin2 θ V3Σ(R) + (cos2 θ + 1)V3Π(R)+ V Π(R + r) Equation (IV.9) Chapter IV, Hg(3P1)/RG Sims. 88 0 1 2 3 4 5 6 RCM (Å) 3.70 3.75 3.80 3.85 3.90 3.95x104 Energy (cm-1) 3A1(pz) 3E(p x/py) 3A1(pz) Emin=37181.3 cm-1 Rmin=0.000 Å De=2242.8cm -1 3E(p x/p y) Emin=38255.3 cm-1 Rmin=1.220 Å De=1168.8cm -1 HgXe6C6v Approach Figure IV.12 Hg 3A1 and doubly degenerate 3E potential energy curves calculated for the C6v approach to a planar arrangement of six Xe atoms. Five categories of Hg⋅RG interactions are identified for the Q4 mode, the geometric details of which are illustrated in Figure IV.10. (I) S.site/W(R1, θ1), motion of the Hg atom away from the hexagon of 6 nn atoms on Plane B. (II) Tri.1/W(R2, θ2), motion towards the centre of the 3 nn atoms on Plane A (small, light shaded triangle). The separation between the close packed A, B and C planes, b, is √(2/3)ss. (III) Tri.2/W(R3, θ3), motion away from the 3 nn atoms on Plane C, (small, light triangle), i.e. the opposite of (II). (IV) Tri.3/W(R4, θ4), motion towards the 3 nnn atoms indicated by the large, dark shaded triangle on Plane A. These atoms are initially at a lattice parameter distance, a, from the guest metal atom in the substitutional site. (V) Tri.4/W(R5, θ5), motion away from 3 nnn atoms on Plane C, the opposite of (IV). The potential energy for each of the interactions was evaluated using Equations IV.2-4. The potential energy curves obtained by summing the five interactions, are shown by the solid traces in the top of Figure IV.13 for the Hg/Xe system. In contrast to the 4-atom body mode, the 3A1(pz) state in the 6-atom ‘body’ mode is not stabilised. The lack of stabilisation evident in Figure IV.13, arises from strong destabilisation occurring with movement of the metal (pz) orbital away from the hexagon (S.site) on Plane B and the initially repulsive interaction it experiences as it Chapter IV, Hg(3P1)/RG Sims. 89 approaches the 3 nn Xe atoms positioned as a triangle on planes A or C. The latter repulsive interaction is shown by the dash-dot line (Tri.1) in Figure IV.13, the former destabilisation by the dashed line (S.site). However, the 3E state of this mode, corresponding to the degenerate px/py orbital orientations of the excited 3P1 Hg atom, is slightly stabilised. The stabilisation arises from the approach of the Hg px/py orbital to the 3nn atoms arranged as a triangle (Tri. 1) on Plane A. This is nearly counteracted by the destabilisation incurred by movement away from the other nn triangle (Tri. 2) on Plane C, resulting in only a small net stabilisation. The slight stabilisation calculated for the Hg 3E state is that indicated by the C3v approach to the RG3 cluster shown earlier. The dominant interaction leading to the destabilisation of the 6-atom ‘body’ mode, (Q4) for the excited state Hg pz orbital orientation is the motion from the centre of the hexagon of Xe atoms forming S.site. This excited state destabilisation was predicted by the Xe6 cluster calculation, inspection of Figure IV.12 reveals that motion of the Hg 3A1 (pz) state from Rcm = 0 Å is a repulsive interaction. b. The 6-atom “waist” mode, Q5 involves, as shown on the right in Figure IV.10, in-phase contraction of 6 nn lattice atoms on the close-packed B plane towards the central metal atom. The total energies of the excited 3A1(pz) and 3E(px/py) states in the waist mode of the Hg(3P1)⋅RG18 cluster obtained from Equations IV.2-4 are given by the expressions W3A1(R) = 6[cos2 θ ΑV3Σ(R1) + sin2 θ ΑV3Π(R1)] Equation (IV.14) + 6[cos2 θ ΒV3Σ(R2) + sin2 θ ΒV3Π(R2)] + 6[cos2 θ CV3Σ(R3) + sin2 θ CV3Π(R3)] + Erreur ! W3E(R) = 3[sin2 θ ΑV3Σ(R1) + [cos2 θ Α + 1]V3Π(R1)] Equation (IV.15) + 3[sin2 θ ΑV3Σ(R2) + [cos2 θ Α + 1]V3Π(R2)] + 3[sin2 θ ΑV3Σ(R3) + [cos2 θ Α + 1]V3Π(R3)] + Erreur ! in which the angles θ Α , θ Β and θ C are defined with respect to the Z-axis and have values of π/5.104299, π/2 and π/3.288535 radians. R1 and R2 refer to the nearest neighbour distance (a/√2) and R3 refers to the next nearest neighbour (a) distance. In the calculation of the energetics of the 6-atom ‘waist’ mode, only the distance of the 6 Hg-RG interactions on Plane B (R1) is decreased, as shown on the right in Figure IV.10. Although the cluster size is restricted to an M⋅RG18 species, 24 Chapter IV, Hg(3P1)/RG Sims. 90 additional on-plane RG-RG interactions arising inside the 4th surrounding sphere are included. This term (m = 24 in Equation IV.14 and 15) is required21 to account for the strong lattice destabilisation that occurs on the close packed Plane B from contraction of the equilibrium rare gas distances. The results of the 6-atom “waist” mode calculations are shown on the bottom of Figure IV.13 for the Hg/Xe system. As indicated by the dashed lines in this figure, stabilization arises only for the pz orbital orientation with the contraction of the 6 Hg-RG bonds on Plane B. Strong destabilisation, shown by the broken grey line, comes from disruption of the nn RgRg distances on the close packed B plane of the lattice, greatly reducing the overall stabilisation (solid trace) of this mode. 0.0 1.0 2.0 3.0 3.6 3.7 3.8 3.9 4.0x104 Energy (cm-1) pz Q4body mode (Å) Hg(3P1)/Xe18 0.0 1.0 2.0 3.0 3.2 3.4 3.6 3.8 4.0x104 Energy (cm-1) Rmin=0.312 Å Eem= 37829 cm-1 Eabs=39270cm -1 pz Q5waist mode (Å) 0.0 1.0 2.0 3. 0 Rmi n=0.837 Å Eem= 38025 cm-1 Eabs= 39270 cm-1 px/p y 3 - fold symmetry 0.0 1.0 2. 0 S.site Tri. 1 Tri. 2 Tri. 3 Tri. 4 Lattice Sum px/p y Figure IV.13 Energetics calculated for the five specific interactions involved in the 6–atom ‘body’ (Q4) and ‘waist’ (Q5) mode for the Hg/Xe system are shown by the legend used. The total potential energy curves obtained by summing these five interactions and the lattice contribution are shown by the solid line. IV.2.II.III Two-fold (2-atom) symmetry modes Calculation of the 2–atom modes involves a 45° rotation of the Z and Y axes about the X–axis from the coordinate system used1 for the 4-atom modes, where the Cartesian axes were coincident with the three, four–fold symmetry axes. Figures IV.14 and IV.2 show the resulting arrangement of the 12 nearest neighbour (nn) RG atoms Chapter IV, Hg(3P1)/RG Sims. 91 around the guest metal atom in the substitutional site, (ss). When counted along the Z–axis, there is a 1, 4, 2, 4, 1 arrangement of the 12 nn RG atoms on this axis with the metal atom at the origin. It can be seen in Figure IV.14 that there are 2 nn on both the Z and Y axes with the remaining 8 nn atoms located in two rectangles at right angles to the Z-axis. Two types of 2-atom modes are identified in the excited state. One involves the in-phase contraction of the 2 nn rare gas atoms on the Y axis towards the central metal atom. The other is motion of the metal atom along the Z-axis directly towards one nn atom. q7 q6YZ X Z Y X 2 - fold symmetry 2-atom body mode, Q 62-atom waist mode, Q 7 Hg Hg Figure IV.14 A representation of the co-ordinate system used for the 2–atom ‘body’ and ‘waist’ mode Q6 and Q7 calculations respectively. The co-ordinate system has been chosen to coincide with the one of the twofold symmetry axes of the cubooctahedron. a). 2–atom ‘body’ mode, Q6, involves motion of the excited state metal atom on the Z–axis towards one of the 12 nn RG atoms positioned on this axis. It is illustrated on the left in Figure IV.14 and involves passage of the excited state metal atom through the 4 nn RG atoms arranged as a rectangle perpendicular to the Z-axis. With the length of the sides of this rectangle, the lattice parameter (a) and substitutional site size (ss), the distance of each of the four rare gas atoms to the centre of the rectangle is then Rcm = √⅜(a). Using the associated values of φ k in Equations IV.3 and 4, φ 1 = (½)cos-1(1/3), φ 2 = π - [(½)cos-1(-1/3)], φ 3 = [(½)cos-1(1/3)] + π and φ 4 = 2 π - [(½) cos-1(1/3)] the following expressions were obtained for the 3B1(px) and 3B2(py) states, Chapter IV, Hg(3P1)/RG Sims. 92 py ⇒ WB2(R) = 4/3[VΣ(R) sin2 θk + VΠ(R) (cos2 θk + 2)] Equation (IV.16) px ⇒ WB1(R) = 8/3[VΣ(R) sin2 θk + VΠ(R) (cos2 θk + 1/2)] Equation (IV.17) The following expression for the 3A1(pz) electronic state was obtained from Equation (IV.2). pz ⇒ WA1(R) = 4[VΣ(R) cos2 θk + VΠ(R) sin2 θk] Equation (IV.18) Calculation of the 2-atom modes for a Hg/RG18 cluster involves consideration of the eight interactions illustrated in Figure IV.14. They are as follows; (1) End 1/W(R1,θ1), motion of the Hg atom from the substitutional site towards one of the nearest neighbour nn atom on the Z–axis. R1 = ss – x where x represents displacement along the Z – axis. (2) End 2/W(R2,θ2), the motion of the metal atom away from the other nn on the Z–axis. This is the opposite of interaction (1) and the distance is R2 = ss + x. (3) Rect. 1/W(R3,θ3), this interaction involves approach to the rectangle of 4 nn atoms. Initially the Hg(pz) atom is at a distance ss/2 from the centre of mass of the rectangle. During the motion this distance becomes ss/2 – x. The distance from each of the RG atoms to the centre of mass of the rectangle is r = √⅜(a), so this interaction occurs at a distance R3 = [(ss/2 – x)2+(⅜ a)2]1/2 and the angle θ 3 = asin (r/R3). (4) Rect. 2/W(R4,θ4), this interaction is the motion of the excited state guest atom away from the rectangle of 4 nn. It is the opposite of interaction (3), therefore R4 = [(ss/2 + x)2+(⅜ a)2]1/2 and θ 4 = asin (r/R4). (5) 2 nn Y/W(R5,θ5), motion of the metal atom away from the 2 nn rare gas atoms on the Y–axis. This interaction takes place at R5 = [(ss)2 + (x)2]1/2 where θ 5 = asin (ss/R5). (6) 2 nnn X/W(R6,θ6), involves the interaction between the metal atom and the 2 nnn on the X–axis. Initially these 2 rare gas atoms are at the next nearest neighbour, (nnn) distance of the lattice parameter a from the metal atom. During the motion the distance becomes R6 = [(a)2 + (x)2]1/2 where θ 6 = asin (a/R6). (7) 2 nnn YZ/W(R7,θ7), this motion involves approach of the Hg atom to the 2 nnn RG atoms positioned on the YZ-plane on the diagonal initially at a distance a. R7 = [(ss - x)2 + (ss)2]1/2 where θ 7 = asin (ss/R7). (8) 2 nnn YZ/W(R8,θ8), this interaction is the opposite of (VII) and R8 = [(ss + x)2 + (ss)2](1/2) and θ 8= asin (ss/R8). The potential energy curves calculated for the interactions 1 to 8 involved in the 2-atom body mode, Q6 are shown for the Hg/Xe system in Figure IV.15. The potential energy curves for Q6 are obtained by the summation of the eight Chapter IV, Hg(3P1)/RG Sims. 93 interactions, and are shown by the solid traces in Figure IV.15. Only the py orbital orientation leads to stabilization22 for the Q6 mode in Hg/Xe. 0.0 0.5 1.0 1.5 3.5 3.6 3.7 3.8 3.9 4.0x104 Energy (cm-1) End 1 End 2 Rect. 1 Rect. 2 2 nnn X Axis Sum px Hg(3P1)/Xe18 0.0 0.5 1.0 1.5 3.5 3.6 3.7 3.8 3.9 4.0x104 Energy (cm-1) Rmin=0.57Å Eem= 38020 cm-1 Eabs= 39270 cm-1 px 0.0 0.5 1.0 1.5 Rmin=1.15Å Eem= 35919 cm-1 Eabs= 39270 cm-1 py 2 - fold Symmetry 0.0 0.5 1.0 1. 5 2nnYAxis 2nnnXZ 2nnnXZ pz 0.0 0.5 1.0 1.5 Lattice py 0.0 0.5 1.0 1. 5 Rmin=0.57Å Eem= 38020 cm-1 Eabs= 39270 cm-1 pz Q6body mode (Å) Q7waist mode (Å) Figure IV.15 Energetics calculated for the eight specific interactions involved in the 2–atom ‘body’ (Q6) and ‘waist’ (Q7) modes depicted in Figure IV.14. The potential energy curves obtained by summing these eight interactions for the three porbital orientation are shown by the solid lines for Hg/Xe. b). 2-atom ‘waist’ mode, Q7, involves, as shown on the right of Figure IV.14, the in–phase contraction of 2 nearest neighbour rare gas atoms on the Y-axis to the central metal atom. The overall energies of the excited 3A1(pz), 3B1(px) and 3B2(py) 3P1 states of the mercury atom in the M⋅RG18 cluster are obtained from Equations. IV.2-4. For the 3A1(pz) state the following expression is used W3A1(R) = 2[cos2 θ ΑV3Σ(R1) + sin2 θ ΑV3Π(R1)] + 8[cos2 θ ΒV3Σ(R1) + sin2 θ ΒV3Π(R1)] + 2[cos2 θ CV3Σ(R1) + sin2 θ CV3Π(R1)] + 4[cos2 θ DV3Σ(R2) + sin2 θ DV3Π(R2)] + 2[cos2 θ EV3Σ(R2) + sin2 θ EV3Π(R2)] + Erreur ! Equation (IV.18) In this equation the angles θ A, θ B, θ C, θ D and θ E are defined with respect to the Z–axis and have values of 0, π/3, π/2, 2π/3 and π radians respectively. R1 and R2 refer to the nearest neighbour distance and the next nearest neighbour distances respectively. Chapter IV, Hg(3P1)/RG Sims. 94 Results calculated for the Q7 mode are shown on the bottom in Figure IV.15 for Hg/Xe. Excited state stabilization occurs for both the px and pz orbital orientations, whereas py is strongly repulsive due to the pure Σ interaction with the two approaching rare gas atoms on the Y-axis. The px and pz orbitals, although arising from different electronic states of the excited state Hg 3P1 atom, show the same excited state minimum due to symmetry. Lattice destabilization must also be included in these calculations as the waist mode Q7 involves motion of two rare gas atoms with respect to their nearest neighbours. Since the RG atoms initially occupy equilibrium positions in the lattice, any displacement from these positions will destabilize the host lattice. 22 RG-RG interactions were considered, of which the motion of the two RG atoms on the Y–axis towards a rectangle of its nearest neighbours is the most important. This repulsive lattice interaction is represented by the grey line in the bottom panels in Figure IV.15, and reduces considerably the stabilization of these modes for the px and pz orbital orientations. IV.3 Discussion A summary of the excited state calculations conducted on the three Hg/RG systems is presented in Figure IV.16. In constructing this figure, identical calculations to those shown in detail for the Hg/Xe system were performed on the Hg/Ar and Hg/Kr systems. However, only the modes exhibiting stabilisation are shown, as they are the only ones that will lead to Stokes’-shifted emission. As indicated by the solid curves in Figure IV.16, the 4-atom waist mode, Q3, leads to excited state stabilisation for all the Hg/RG systems but only for the pz orbital orientation. In contrast, the 4-atom body mode, Q2, exhibits stabilisation only in the Hg/Xe system. The 3A1(pz) state in the 6-atom ‘body’ mode, Q4, is not stabilised in any of the solid rare gases. It is evident in the detailed Hg/Xe plot shown in Figure IV.13, that the reason for the lack of stabilisation is the repulsive interaction the metal (pz) orbital experiences as it approaches the 3 nn Xe atoms positioned as a triangle on planes A or C. This repulsive interaction (shown for Hg/Xe in Figure IV.13 by the dash-dot line, Tri.1), is much stronger in Hg/Kr and stronger still in Hg/Ar as the lattice parameters get smaller. The 3E(px,py) state of this mode shows a shallow Chapter IV, Hg(3P1)/RG Sims. 95 minimum in Hg/Xe. The 6-atom ‘waist’ mode, Q5, exhibits excited state minima in the pz orbital orientation for all the Hg/RG systems. 0.0 0.5 1.0 3.2 3.4 3.6 3.8 4.0x104 Energy (cm-1) Hg(3P1)/Ar18 0.0 0.5 1.0 Qn(Å) Hg(3P1)/Kr18 0.0 1.0 2.0 3. 0 Q2(pz) Q3(pz) Q4(px/py) Q5(pz) Q6(py) Q7(px/pz) Q1Hg(1S0) Hg(3P1)/Xe18 Figure IV.16 A comparison of the excited Hg(3P1) state potential energy curves of the vibronic modes exhibiting stabilisation in the Hg/RG systems. Shown also is the ground Hg(1S0) state potential energy curves calculated for the breathing mode, Q1. Particularly noteworthy is the crossing of this curve with the very strongly stabilized 4-atom modes in the Hg/Xe system. Stabilisation is not found for the 2-atom ‘body’ mode, Q6, in any of the orbital orientation for Hg isolated in Ar and Kr. However, the py orbital orientation exhibits a stabilisation for the body mode, Q6, in Hg/Xe. In contrast, the ‘waist’ mode of this symmetry, Q7, is stabilised for the px,py orbital orientations in all three rare gas systems. In the next section a comparison of the predicted absorption and emission energies is made with recorded matrix spectra presented in Chapter III. IV.3.I Absorption Energies The absorption energy of the guest mercury atom isolated in a solid rare gas lattice is calculated as the difference between the ground Hg(1S0)⋅RG18 and the excited state Hg(3P1)⋅RG18 cluster energies at the centre of a substitutional site, R = 0 Å. Within the Frank-Condon approximation no movement will occur between the Hg and the cluster atoms during the electronic transition, so the absorption energy is given by Chapter IV, Hg(3P1)/RG Sims. 96 Eabs = E[Hg(3P1)⋅RG18]Q(R=0) – E[Hg(1S0)⋅RG18]Q(R=0) Equation (IV.19) where Q(R = 0) represents zero displacement for a vibronic mode, Qn (corresponding to the centre of a substitutional site). Accordingly, for a given site occupancy, the calculated absorption energies must be identical for all vibronic modes. The level of agreement between the 4-atom modes and the new 6-atom and 2-atom mode calculations is evident for Hg/Xe in Figure IV.8, Figure IV.13 and Figure IV.15 by the identical ‘Eabs’ values (39270 cm-1) obtained for the Q3, Q5 and Q7 modes. Table IV.2 shows a comparison of the observed absorption wavelengths with those calculated for substitutional site occupancy. Table IV.2 A comparison of the observed absorption wavelengths (nm units) for the 3P1 ← 1S0 transition of matrix-isolated atomic mercury with the calculated absorption values. The difference between the observed band maxima and the predicted values are quoted as δObs-Cal in cm-1. For a given Hg/RG system, the quoted predicted value was found for the three symmetry systems used, the three porbital orientations and the body and waist vibronic modes. Hg/RG ECal λCalc λObs δObs-Cal Hg/Ar 40495 246.94 246.0 +155 Hg/Kr 39922 250.49 249.1 +227 Hg/Xe 39270 254.65 253.4 +192 The 246.94 nm absorption wavelength calculated for Hg/Ar, compares very well with the observed band centre at 246 nm. The recorded8 absorption band centre for Hg isolated in solid Kr is at 249.1 nm while the calculated value is 250.49 nm. Better agreement with observed data is achieved in Hg/Xe where the calculated value of 254.64 nm compares favourably with the observed absorption centered at 253.4 nm. From the comparison presented in Table IV.2, it is clear that the calculated absorptions match the red component of the threefold-split bands for all three Hg/RG systems. It thereby supports the assumption of substitutional site occupancy inherent in the pair-potential calculations conducted. It is not within the scope of the present calculations to examine the threefold absorption splitting effect because as indicated by Equation IV.19, the absorption values are determined only at the centre of the substitutional site i.e., at R = 0. Simulation23 of the Jahn-Teller structure on the absorption profiles requires displacement of the ground state metal atom from the centre of the substitutional site, a task difficult to implement in the code developed for the calculations presented. Chapter IV, Hg(3P1)/RG Sims. 97 IV.3.II Emission Energies The Hg(3P1 → 1S0) emission energies are calculated with the formula Eem = E[Hg(3P1)⋅RG18]Q(Rmin’) – E[Hg(1S0)⋅RG18]Q(Rmin’) Equation (IV.20) where Rmin’ represents the nuclear configuration of a given excited state vibronic mode, Q, at its energy minimum. In accordance with the Franck-Condon approximation, the energy of this vibronic mode on the ground state is obtained at the Rmin’ value identified in the excited state. The results calculated in this way for the vibronic modes exhibiting excited state stabilisation (shown in Figure IV.16) in the Hg/RG systems are collected in Table IV.3. Hg/Ar: The 4-atom (pz), 6-atom (pz) and 2-atom (px and pz) ‘waist’ modes exhibit excited state stabilisation in solid argon. The emission wavelengths calculated for these Q3, Q5 and Q7 modes are 256.14, 250.51 and 248.29 nm respectively. From the comparison made in Figure IV.16 of the three excited state vibronic modes, it is expected that Hg/Ar emission is dominated by the 6–atom ‘waist’mode, (Q5) as it exhibits more rapid stabilisation (i.e., a steeper gradient) than the more deeply bound 4–atom ‘waist’, (Q3) or the 2-atom ‘waist’ mode. As shown in Figure IV.17, the 250.51 nm emission calculated for the Q5 mode closely matches the deconvoluted central component at 250.69 nm in the observed band. The two other predicted emission bands lie to the blue and red of the two remaining deconvoluted emission components. Chapter V; I) Hg(3P1)/Ne, II) Hg(3P0)/RG Sims. 104 Chapter V A pair-potentials analysis of I) the Hg (3P1 ↔ 1S0)/Ne luminescence and II) Hg (3P0 → 1S0)/RG (RG = Ar, Kr and Xe) emission spectroscopy. V Introduction This Chapter consists of two parts both of which employ the M/RG18 localised model to simulate the observed luminescence of atomic mercury isolated in solid rare gases. In Part I, the Hg/RG18 pair-potentials approach, presented in Chapter IV for the rare gases Ar, Kr and Xe, is extended to model the luminescence of the atomic Hg 6p 3P1 ↔ 6s 1S0 transition isolated in solid neon. Part II focuses on Hg isolated in Ar, Kr and Xe in an attempt to simulate the emission spectroscopy of the Hg atom 3P0 → 1S0 transition reported in Chapter III. These two sections provide insights into 1) ground and excited state metal atom solvation, 2) the effect of local lattice perturbations caused by the dopant and 3) the importance of the site of isolation occupied by the metal atom in determining the observed luminescence. V.I A pair-potentials analysis of the Hg (3P1 ↔ 1S0)/Ne luminescence V.I.1 Introduction The localised pair-potentials approach1 is employed here to investigate the absorption and emission spectroscopy of the atomic 3P1 state Hg in solid neon. The spectroscopic studies of matrix-isolated2 atomic mercury focusing on the 6p 3P1 ↔ 6s 1S0 transition reported by the Orsay group3 and recently by our group at Maynooth4 outlined the luminescence in solid Ar, Kr and Xe as 12 K. This was the minimum temperature available with those experiments but recently Chergui and co-workers5,6 have conducted spectroscopic studies of atomic Hg in neon matrices at 4 K. The deposition temperature was achieved on a LiF window cooled by a liquid helium cryostat. The atomic Hg(3P1 ↔ 1S0)/Ne excitation and emission spectra reported by Chergui and co-workers5 are presented in the top panel of Figure V.1. The emission spectrum and fluorescence-excitation reported by Chergui were produced with UV laser excitation. Emission was detected using a UV-enhanced CCD camera or photomultiplier tube following dispersion by an Acton Research UV-Vis monochromator equipped with a 150-grooves/mm diffraction grating. Table V.1 presents a comparison of the photo-physical characteristics of the Hg(3P1 ↔ 1S0)/Ne Chapter V; I) Hg(3P1)/Ne, II) Hg(3P0)/RG Sims. 105 results5 and those recorded for Hg/Ar, Hg/Kr and Hg/Xe in this work and presented in Chapter III. 34363840 x103E n e r g y ( c m -1 ) Intensity (Arb. Units) 250 260 270 280 290 30 0 W a v e l e n g th ( n m ) Ne Ar Kr Xe Hg(3P1)/RGEmissionExcitation 3P1 ____ 1S0 Figure V.1 A comparison of the emission spectra recorded for Hg/Ne, Ar, Kr and Xe systems (shown right) produced with excitation of the Hg 3P1 ← 1S0 transition. The Hg/Ne emission spectrum shown (top right) was produced with laser excitation and recorded at 4 K using CCD detection, as reported by Chergui and co-workers5. The Hg/Ar, Kr and Xe data presented are time-integrated emission spectra recorded following deuterium lamp excitation at 12 K as outlined in Chapter III. The excitation spectra shown left were produced monitoring the Hg 3P1 → 1S0 fluorescence emission maximum for each Hg/RG system. Table V.1 Photophysical characteristics of the triplet 6p 3P1 ↔ 6s 1S0 transition of matrix – isolated atomic mercury. λEx indicates the position of the central component of the three–fold split excitation spectrum and λEm indicates the emission bandcentre in nm units. The full-width at half-maximum intensity of the excitation/emission features is denoted by ∆ and the Stokes’ shift by SS - both in wavenumber (cm-1) units. Excitation Emission Hg/RG System λEx (nm) ∆ (cm-1) λEm (nm) ∆ (cm-1) SS (cm-1) Ne5,7 247.8 470 252.5 ± 0.5 300-500 668 Ar 245.9 484 250.3 399 715 Kr 248.9 397 254.1 465 816 Chapter V; I) Hg(3P1)/Ne, II) Hg(3P0)/RG Sims. 106 It is evident from the comparison made in Figure V.1 and Table V.1 that the Hg(3P1)/Ne excitation and emission spectra do not follow the overall trend exhibited by the other Hg/RG systems, where a progressive blue-shift in the excitation and emission band maxima (λEx and λEm) is observed from Xe to Ar. However the Hg/Ne emission spectrum conforms to the trends evident for Hg/Ar and Kr where the Stokes’ shift (SS) and the emission linewidth (∆) increase from Ne to Xe. The expected blue-shift of the excitation band maximum (λEx) with decreasing rare gas polarizability for a given site of isolation of the metal atom within the RG matrix is not exhibited by the Hg/Ne system. A linear correlation between the gas phase – matrix absorption/excitation band maximum and rare gas polarizability was presented by Laursen and Cartland8 (L&C) for the M(1P1 and 3P1 ← 1S0)/RG transitions of the metal atoms Zn, Cd and Hg in Ar, Kr and Xe matrices. The polarizability model held true for the Hg (3P1 ← 1S0) transition in Ar, Kr and Xe matrices but until recently5,6, Hg/Ne results were not available for comparison. The Hg(3P1 ↔ 1S0)/Ne18 pairpotentials simulations presented in this Chapter allow an investigation of the deviations shown by the Hg(3P1)/Ne excitation and emission spectra from the trends exhibited by the other Hg/RG systems. The availability of the Hg(3P1)/Ne matrix data5 and diatomic Hg(3P1)⋅Ne [X 10+ (1Σ)], [A 30+ (3Π)] and [B 31] state potentials9,10 allowed the application of the Hg⋅RG18 model to solid Ne. In the sections which follow, an examination of the Hg 3P1 ↔ 1S0 Ne matrix spectroscopy is conducted using the Hg⋅RG18 cluster calculations. The tetragonal (4-fold symmetry), trigonal (3-fold symmetry) and 2fold symmetry calculations outlined in Chapter IV are performed for atomic Hg isolated in solid Ne matrices. The calculations presented in the previous Chapter are based on atomic mercury occupying unperturbed substitutional sites in solid Ar, Kr and Xe. The Hg/Ne simulations presented are based on substitutional site occupancy but allow the ground state Hg 6s2 1S0 atom to deform its immediate neon matrix environment. The original and deformed matrix environments are referred to in the text as “rigid” and “relaxed” lattice calculations respectively. Therefore, in addition to the calculations outlined in Chapter IV, details of a lattice expanding symmetric ‘breathing’ mode, (Q1) and modifications to the trigonal, 6-atom ‘waist’ mode, (Q5) are presented. The ‘breathing’ mode Q1 involves the symmetric expansion (or contraction) of the first solvation shell of Ne atoms surrounding the metal atom Chapter V; I) Hg(3P1)/Ne, II) Hg(3P0)/RG Sims. 107 thereby allowing relaxed substitutional site occupancy. Comparison of the calculated Hg(3P1 ← 1S0)/Ne absorption energies with the observed Hg/Ne spectroscopy allowed the assignment of relaxed substitutional site occupancy for atomic mercury in solid neon. The emission comparison allows an assessment of the effects of a ground state perturbation on the Hg(3P1) excited state dynamics. In solid neon a significantly Stokes’ shifted emission is predicted by the Hg/Ne18 cluster model only for relaxed lattice calculations. A combination of the re-establishment of the equilibrium neon lattice and stabilisation of the Hg⋅Ne interactions lead to energy values which compare well with observed emission. These conditions are fulfilled by ‘waist’ modes calculated where the Hg 3P1 state is stabilised by the trigonal 6-atom ‘waist + limited stretch’ mode, (Q8*) - a modification of the 6-atom ‘waist’ mode (Q5) presented in Chapter IV. V.I.2 Method and Results The pair-potentials analysis of the Hg(3P1 ↔ 1S0)/Ne luminescence spectroscopy outlined in this section employs the localized M/RG18 model1 presented in Chapter IV. Therefore only the modifications required to simulate the Hg/Ne spectroscopy are presented here. V.I.2.I Ground State Site occupancy The starting point for the Hg/Ne18 simulations is the selection of the site occupied by the ground state Hg atom within the host neon matrix. This is achieved by comparison of the Hg⋅Ne X (1Σ) ground state bond length with that of neon dimer Ne2. As presented in Table V.2, the Hg⋅Ne ground state bond length is 3.89 Å whereas the neon dimer bond length is 3.091 Å. The substitutional site (ss) size available in solid Ne is 3.155 Å11, calculated from the lattice parameter12 a = 4.462 Å, using the relationship ss = a/√2 Å. The comparison of the substitutional site size available and the Hg⋅Ne van der Waals bond length reveals a difference of 0.735 Å. This unfavourable comparison makes rigid substitutional site occupancy unlikely within solid neon. Therefore, an expansion of the substitutional site may be required to facilitate the isolation of the mercury atom. Whether Hg atom occupancy in a substitutional site leads to the deformation of the Chapter V; I) Hg(3P1)/Ne, II) Hg(3P0)/RG Sims. 108 original site can be determined by comparing the predicted and observed Hg(3P1 ← 1S0)/Ne absorption energies. The ground state energy of the cluster is evaluated with the method presented in Chapter IV, Section IV.2.I. The Morse parameters used to describe the Hg⋅Ne and Ne⋅Ne interactions are provided in Table V.2. Table V.2 Spectroscopic constants used to generate the Morse potential energy curves for the Hg⋅Ne and Ne2 diatomics. Data source are indicated by the references. Hg⋅Ne10 Morse Parameters X 1Σ (10+) A 3Π (30+) B 3Σ (31) Ne⋅Ne13 X 1Σ µHg-RG (amu) De(cm-1) ωe(cm-1) ωexe(cm-1) Re(Å) β(Å-1) 18.191701 42 17.2 1.77 3.89 1.378517 - 79 26.9 2.28 3.497 1.57198 - 13.3 7.7 1.12 4.71 1.09666 9.996219 29.4 29.1 - 3.091 2.090614 Ground State ‘breathing’ mode, (Q1) Due to the unfavourable match between the substitutional site size available in solid Ne and the Hg⋅Ne ground state bond length, the 12 nearest neighbour (nn) Ne atoms surrounding the Hg atom undergo a radial expansion. The ground state ‘breathing’ mode, (Q1) is akin to the previously described ‘waist’ mode as the motions of the lattice atoms are ‘in-phase’ relative to the fixed metal atom. Q1 lowers the energy of the Hg⋅Ne12 cluster as the Hg-Ne interaction distance is increased by +X Å from ss Ne = 3.155 Å to ss + X Å allowing the Hg-Ne distance to approach the Hg⋅Ne ground state bond length listed in Table V.2. The interactions involved in Q1 are shown in Figure V.2. The motion of the 12 nn Ne atoms results in a destabilisation of the Ne lattice as the Ne-Ne distances are displaced from their equilibrium value, (Re). The lattice destabilisation caused by the motion of the 1st sphere limits the amount the Hg⋅RG12 cluster can be stabilised. Therefore, the extent of the expansion occurring in the 1st sphere surrounding the metal atom is identified as the point where the Hg⋅Ne stabilisation and the Ne lattice destabilisation energies are equal. To identify this point, the energies at the equilibrium positions of the Hg⋅Ne12 (the Hg⋅Ne ground state bond length) and Ne lattice (where all Ne⋅Ne interactions occur at the distance Chapter V; I) Hg(3P1)/Ne, II) Hg(3P0)/RG Sims. 109 of the Ne2 equilibrium bond length) are both set to zero. Equation IV.I is applied with the number of Hg⋅Ne and Ne⋅Ne interactions occurring i.e., n = 12 and m = 120 respectively. Ground state ‘breathing’ mode Q1 Hg 1S0 3rd Sphere a(√1.5) 1st Sphere (a/√2) 4th Sphere a(√2) 2nd Sphere (a) 5th Sphere a(√2.5) 7th Sphere a(√3.5) q1 Ground state ‘breathing’ mode Q1 Hg 1S0 3rd Sphere a(√1.5) 1st Sphere (a/√2) 4th Sphere a(√2) 2nd Sphere (a) 5th Sphere a(√2.5) 7th Sphere a(√3.5) Hg 1S0 3rd Sphere a(√1.5) 1st Sphere (a/√2) 4th Sphere a(√2) 2nd Sphere (a) 5th Sphere a(√2.5) 7th Sphere a(√3.5) q1 Figure V.2 Representation of the interactions involved in the calculation of the energetics of the ground state ‘breathing’ mode (Q1). The symmetric expansion of the 12 nearest neighbour (nn) Ne atoms forming the 1st co-ordination sphere of the Hg atom isolated in a substitutional site, shown by q1. The relative positions of the 2nd and successive metal atom co-ordination spheres are indicated by shaded spheres and the radial distances to the Ne atoms forming the different shells are provided as a function of the lattice parameter of solid Ne (a = 4.462 Å). A total of 120 Ne⋅Ne interactions occurring within the 4th sphere are considered in the calculation of the lattice destabilisation due to the motion of the 12 nn. The Ne⋅Ne interactions considered are shown in Figure V.2 the geometric details of which are now presented, labelled I-V. I) The expansion of the 12 nn Ne atoms surrounding the Hg atom results in the contraction of the distance between the moving atoms and 12 RG atoms positioned in the 4th sphere (spotted circles), resulting in the contraction of 12 ‘on-axis’ RG interactions from the initial Chapter V; I) Hg(3P1)/Ne, II) Hg(3P0)/RG Sims. 110 substitutional site distance ss = a/√2 Å to ss – x Å. II) Each of the moving 12 nn Ne atoms approach a rectangle of Ne atoms located in the 3rd sphere by a distance axa 2 3 8 2 2 −     +. There are 48 of these Ne⋅Ne interactions. III) The 12 nn Ne atoms move away from the 6 next nearest neighbour (nnn) atoms located in the 2nd co-ordination sphere by the amount ax 2 2 2     +, contributing another 24 Ne⋅Ne interactions. IV) The radial expansion of the 12 nearest neighbour atoms on the surface of the cubo-octahedron results in the extension of 24 Ne⋅Ne distances from the substitutional site distance ss = a/√2 to ss + x. V) 12 Ne⋅Ne next nearest neighbour (nnn) interactions on the surface of the cubo-octahedron are extended from the lattice parameter, a to (a+√2x). The potential energy curves calculated for the expansion of the 1st sphere atoms from the Hg atom isolated in a substitutional site in solid Ne are shown in Figure V.3. Upon inspection of Figure V.3 it is evident that a lattice expansion of 0.293 Å occurs representing an increase in the substitutional site diameter of 9.29%. With this expansion the Hg⋅Ne12 cluster is stabilised by 353.0 cm-1 (Ne lattice destabilisation is -∆E cm-1) from the initial value (Esubs) for Hg atom isolated in a rigid site within the Ne12 cluster. The relaxed substitutional site size [ssRel = (3.155 + ∆Q)] is 3.448 Å. The stabilisation observed for Hg⋅Ne12 can be understood in terms of the Hg(1S0)⋅Ne ground X state potential shown in the right panel of Figure V.4, where the vertical lines crossing the potential energy curve indicate the rigid (NeSS) and relaxed (NeSS + ∆Q) substitutional site diameters. The ground state energy of the Hg(1S0)/Ne18 cluster is calculated for both rigid and relaxed substitutional site occupancy following the method detailed in Chapter III. Chapter V; I) Hg(3P1)/Ne, II) Hg(3P0)/RG Sims. 111 0.0 0.1 0.2 0.3 0.4 0. 5 Q1(Å) 0 500 1000 1500 Energy (cm-1) Ne-Ne Hg-Ne ∆Q = 0.2930 Å ∆E = 353.0 cm-1 Esubs = 1046.6 cm-1 H g / N eQ 1' b r eat h i n g ' m o d e Figure V.3 Potential energy curves for the Hg⋅Ne12 and 120 Ne⋅Ne interactions occurring for the symmetric ground state ‘breathing’ mode, (Q1). R = 0 Å, for the mode Q1 represents the undistorted substitutional site. The extent of the lattice expansion calculated is indicated as ∆Q = 0.293 Å. V.I.2.II Excited 3P1 state The excited state energy of the Hg(3P1)/Ne18 cluster was calculated as outlined in Chapter IV; Section IV.2.II. The known diatomic Hg(3P1)⋅Ne A [30+ (3Π)] and B (31) state potentials10 are used to construct the excited state Hg/Ne18 cluster following deconvolution of the 3Σ state from the experimentally observed B state using Equation IV.5. The A (3Π), B and deconvoluted 3Σ states for Hg⋅Ne are shown in the left panel of Figure V.4. The deconvoluted 3Σ state exhibits a long range minimum as observed for Hg⋅Ar, Kr and Xe in Chapter IV. Excited 3P1 state energetics were calculated for the body and waist vibronic modes in the three orbital orientations (px, py and pz) for the fourfold (4-atom), threefold (6-atom) and twofold (2-atom) co-ordinate systems shown Figure IV.4, Chapter IV. The calculations were completed for atomic mercury isolated in rigid and relaxed substitutional sites within the neon lattice. The absorption energies were compared for the three co-ordinate systems calculated and identical absorption energies were obtained for all modes given the condition of the site of isolation (rigid Chapter V; I) Hg(3P1)/Ne, II) Hg(3P0)/RG Sims. 112 or relaxed). As the absorption energy is calculated as the difference in the energy of the cluster in the ground and excited state within the Frank – Condon principle the absorption energies should be different for rigid and relaxed site occupancy. 3.0 4.0 5.0 6.0 7.0 R(Å) -80 -60 -40 -20 0 20 Energy (cm-1) Rmin= 5.265 Å De= 11.566 cm-1 3Σstate parameters B( 31) A3Π(30+) 3Σ(deconvoluted) Hg(3P1)Ne 3.0 4.0 5.0 6. 0 R ( Å ) 0 50 100 Energy (cm-1) X1Σ(10+) NeSS NeSS +∆Q Rmin=3.9Å De=42.0cm -1 X1Σstate Hg(1S0)Ne Figure V.4 The Hg(3P1)⋅Ne 3Σ state extracted with Equation IV.5 from the spectroscopic [A 30+ (3Π)] and [B 31] states (left panel) whose constants are presented in Table V.2. Note: the most recent spectroscopic data on the Hg⋅Ne 1:1 van der Waals complex presented in Ref. 10 was used to generate the potential energy curves presented. The 3Σ state shows a weakly bound region at long range corresponding to Rmin and the binding energy is indicated by De. All of the states share the Hg atom 6p 3P1 asymptote at 39424.1 cm-1 but are shown dissociating to zero-energy for the purpose of comparison with the potential energy curve for the Hg(1S0)⋅Ne X (1Σ) state shown right panel. Thus for Hg isolated in solid Ne, 18 excited state potential energy curves were calculated for both rigid and relaxed substitutional site occupancy. However, only the vibronic modes exhibiting stabilisation, are discussed as these are the only ones leading to emission. The tetragonal (4-atom) and the twofold (2-atom) ‘body’ and ‘waist’ modes Q2, Q3, Q6 and Q7 presented in the previous chapter were not stabilised for any p-orbital orientation. The trigonal (6-atom) ‘body’ mode Q4 was not stabilised for Hg/Ne. However the 6-atom ‘waist’ mode was stabilised and the results are now presented. The details of two new vibronic modes (Q8 and Q8*) based on the Chapter V; I) Hg(3P1)/Ne, II) Hg(3P0)/RG Sims. 113 6-atom ‘waist’ mode, (Q5) presented in Chapter IV are discussed in the following section. Hg/Ne18 Trigonal (6-atom) symmetry modes The co-ordinate system used to calculate the 6-atom ‘body’ (Q4) and ‘waist’ (Q5) vibronic modes is presented in Figure IV.4 where the Z-axis is co-incident with one of the four, threefold symmetry axes of the Hg⋅Ne12 cubo-octahedron. The details of the modes and the calculation of the excited state energetics are presented in detail in Chapter IV; Section IV.2.II.II. The potential energy curves calculated for the 6-atom modes for rigid and relaxed neon lattice are shown in Figure V.5 and Figure V.6 respectively. 0.0 1.0 2.0 3.9 4.0 4.1x104 Energy (cm-1) Rmin=0.000 Å Eem= 40339 cm-1 Eabs= 40339 cm-1 pz Q4body mode (Å) Hg(3P1)/Ne18 0.0 1.0 2.0 3.9 4.0 4.1 4.2x104 Energy (cm-1) S.site Tri. 1 Tri. 2 Rmin=0.000 Å Eem=40339cm -1 Eabs=40339cm -1 pz Q5waist mode (Å) 0.0 1.0 2. 0 Rmin=0.000 Å Eem=40339cm -1 Eabs= 40339 cm-1 px/p y Rigid Lattice ∆Q=0.00Å 3 - fold symmetry 0.0 1.0 2. 0 Tri. 3 Tri. 4 Lattice Sum px/p y Rigid Lattice ∆Q=0.00Å Figure V.5 Energetics calculated for the Hg/Ne18 trigonal (6-atom) ‘body’ and ‘waist’ modes, Q4 and Q5 respectively based on rigid substitutional site occupancy as indicated by ∆Q = 0.00 Å. The individual interactions (I–V) involved (the specific details of which are presented in Chapter IV) are shown by the legend. The total potential energy calculated for the mode is shown by the solid line obtained by summation of the individual interactions. Figure V.5 reveals that the excited state Hg(3P1)/Ne18 cluster is not stabilised by the 6-atom vibronic modes for Hg isolated in a rigid substitutional site in solid Ne. However the 6-atom ‘waist’ mode (Q5) does lead to stabilisation for the pz orbital Chapter V; I) Hg(3P1)/Ne, II) Hg(3P0)/RG Sims. 120 39.040.041.0 x103Ene r g y ( c m -1 ) Intensity (Arb. Units) 245 250 255 W ave l eng t h(n m ) Ne Ar Kr Xe Hg(3P1)/RG Figure V.12 A comparison of the observed excitation spectra and calculated atomic Hg (3P1 ← 1S0) absorption energies. The Hg/Ne spectrum (top) is that recorded at 4 K by Chergui and co-workers. The excitation spectra presented for Hg/Ar, Kr and Xe are those recorded at 12 K presented in Chapter III. The absorption energies calculated for atomic Hg isolated in rigid substitutional site in solid Ar, Kr and Xe reported in Chapter IV are indicated by the double dash dot (DD-D). The absorption energies calculated for Hg isolated in rigid and relaxed substitutional sites in solid Ne are indicated by DD-D and dashed lines respectively. Further evidence that atomic Hg occupies a relaxed substitutional site in solid neon is provided by the ground state lattice expansion of 9.29% required to accommodate the Hg atom and the observation that the shift of the excitation band maximum from the gas phase 3P1 ← 1S0 observed for Hg/Ne does not show the linear correlation with the rare gas polarizability15. Table V.5 presents the gas phase to matrix frequency shifts and the rare gas polarizabilities. A plot of the frequency shift (δ cm-1) of the Hg 3P1 ← 1S0 transition from the gas phase position to the Hg/RG matrix position (λEx) versus rare gas polarizability is shown in Figure V.13. Chapter V; I) Hg(3P1)/Ne, II) Hg(3P0)/RG Sims. 121 Table V.5 Gas phase to matrix frequency shifts of the atomic Hg 3P1 ← 1S0 transition, δ in wavenumber units. The rare gas polarizabilities and λEx the position of the Hg(3P1 ← 1S0) transition in the rare gas hosts. Hg/RG System λEx (nm) δ (cm-1) RG Polarization (Å3)15 Ne 247.8 + 942.83 0.400 Ar 245.9 + 1254.64 1.640 Kr 248.9 + 764.47 2.485 Xe 253.6 + 19.88 4.050 01234 Po l ( Å3 ) 0 250 500 750 1000 1250 δ(cm-1) Ne Ar Kr Xe Hg 3P1 ______ 1S0 Figure V.13 A plot of the gas phase to Hg/RG matrix frequency shifts (δ cm-1) observed for the 3P1 ← 1S0 transition of atomic mercury versus the polarizabilities of the host rare gas solids. The solid line shown highlights the linear correlation between the frequency shifts and rare gas polarizability observed by the Hg 3P1 ← 1S0 transition in solid Ar, Kr and Xe. The calculated emission energies (E cm-1) for atomic Hg isolated in an expanded substitutional site (i.e. relaxed lattice) are compared to the observed matrix emission spectra in Figure V.14. The best agreement is achieved using the 6-atom ‘waist + limited stretch’ mode, (Q8*) for the pz orbital orientation. The stabilization arises from the ‘in-phase’ contraction of the six Ne atoms on the close packed plane B towards the metal atom and in so doing, these Ne⋅Ne interactions approach their equilibrium lattice positions. The role of the metal atom is revealed as Q8* is stabilised for the pz orbital orientation thus maximising the pure Π Hg⋅Ne interaction. Chapter V; I) Hg(3P1)/Ne, II) Hg(3P0)/RG Sims. 122 The 6-atom ‘stretch’ contributes in reducing the overall energy of the Hg 3P1 excited state by reducing the Σ interaction. 38. 5 39.039.540.040.541.0 x103E n e r g y ( c m -1 ) Intensity (Arb. Units) 245.0 247.5 250.0 252.5 255.0 257.5 W a v e l e n g t h ( n m ) Absorption Q5(pz) Q8(pz) Q8*(p z) Hg(3P1)/Ne Figure V.14 A comparison of the observed and calculated Hg(3P1 ↔ 1S0)/Ne spectroscopy. The experimental spectra shown reported by Chergui and co-workers5,6,7 were recorded on deposition at 4 K. The vertical lines represent the absorption and emission energies calculated for Hg isolated in an expanded substitutional site in solid neon presented in Table V.3 and Table V.4 respectively. The vibronic modes leading to the stabilisation of the excited state are indicated in the legend. V.I.5 Conclusion The agreement between the calculated and observed5 absorption energies indicate that atomic Hg occupies distorted (relaxed) single substitutional sites in solid neon. This accounts for the observation that the Hg 6p 3P1 ← 6s 1S0 transition occurs to lower energy than that predicted by an extrapolation of the polarizability model8. However, as the expansion of the substitutional site produces essentially a different matrix environment than the rigid substitutional sites occupied by atomic Hg isolated in solid Ar, Kr and Xe, Hg/Ne will deviate from the linear behaviour. The observed emission spectroscopy is also best predicted by simulations based on relaxed substitutional site occupancy for Hg in neon. It is observed that the Chapter V; I) Hg(3P1)/Ne, II) Hg(3P0)/RG Sims. 123 excited state lattice reorganisation is both dependent on the lattice energy and the stabilisation of the metal atom excited state. Solid neon provided an ideal system to probe the effect of ground and excited state solvation as the cramped lattice site allowed an investigation of the lattice contribution to the overall excited state cluster stabilisation. This insight is not provided by the calculations presented in Chapter IV for atomic Hg isolated in rigid substitutional sites in solid Ar, Kr and Xe as the lattice interactions calculated for these systems, always showed a destabilising effect since these modes move the lattice atoms away from there equilibrium positions. Therefore the calculations presented for Hg/Ne show the importance of the equilibrium lattice restoration in producing the observed luminescence in cases where there is an unfavourable match between the site of isolation and the M⋅RG ground state bond length. They also show that for a cramped site of isolation, the ‘waist’ type vibronic modes are of greater importance than the ‘body’ modes in producing excited state stabilisation of a metal atom. This can be understood on the basis that the internal motion of the metal atom within such a cramped lattice is not feasible because of the immediate on-set of repulsive interactions. V.II Hg (3P0 → 1S0)/RG emission spectroscopy (RG = Ar, Kr and Xe) V.II.1 Introduction The remaining sections of this Chapter present pair-potential simulations of the emission spectroscopy of the 3P0 → 1S0 transition of atomic mercury isolated in solid Ar, Kr and Xe. The most recent experimental study of the Hg(3P0 → 1S0)/RG (RG = Ar, Kr and Xe) emission spectroscopy was conducted in the present study4. Chapter III presents the observed emission spectroscopy of the 3P0 state produced as a result of intermultiplet relaxation following pulsed laser excitation of the 3P1 ← 1S0 transition. The 3P0 state emission spectra exhibited a progressive red shift and decreasing linewidth from Ar to Xe. Recording the emission spectra at higher temperatures than 12 K suggested the presence of zero phonon lines and phonon sidebands for the Hg(3P0 → 1S0) emission in solid Kr and Xe. A lineshape analysis of the high-resolution emission spectra, using the Wp optical function allowing the identification of ZPL in Kr and Xe. Figure V.15 presents a summary of the Hg 3P0 → Chapter V; I) Hg(3P1)/Ne, II) Hg(3P0)/RG Sims. 124 1S0 emission spectroscopy reported in Chapter III. The spectral positions of the observed phonon sidebands and the calculated ZPL’s (ν0,0) are provided in Table V.6. 37. 5 38.038.539.0 x103 E n e r g y ( c m -1 ) Intensity (Arb. Units) 256 258 260 262 264 266 Wavelength (nm) Ar Kr Xe Hg(3P0)/RG v0,0 v0,0 v0,0 3P0 _______ 1S0 Figure V.15 A summary of the emission features assigned in Chapter III to the 6p 3P0 → 6s 1S0 transition of atomic Hg recorded in Ar, Kr and Xe at 12 K. These emission spectra were produced with pulsed laser excitation of the Hg atom 3P1 ← 1S0 transition. The calculated positions of the zero-phonon lines ν(0,0) are indicated by the solid lines. Table V.6 The location of the ZPL for the atomic Hg 3P0 ↔ 1S0 emission in solid Ar, Kr and Xe. The location of the phonon sideband and the difference in energy between the ZPL and the phonon sideband denoted by ∆ both of which are presented in wavenumber units. Hg/RG ZPL, ν(0,0) E (cm-1) / λ (nm) Phonon sideband EPSB (cm-1) / λ (nm) ∆ = ν(0,0) - EPSB (cm-1) Hg/Ar 38740 / 258.13 38625 / 258.9 + 115 Hg/Kr 38366 / 260.65 38314 / 261.0 + 52 Hg/Xe 37718 / 265.13 37672 / 265.4 + 46 The pair-potential calculations presented assume ground state rigid substitutional site occupancy for atomic mercury in Ar, Kr and Xe (identified in Chapter IV from a comparison of the calculated and observed Hg(3P1 ← 1S0) absorption energies). A pair-wise sum of the Hg(3P0)⋅RG (0) ã state potentials extracted from the experimental Hg(3P1)⋅RG [A 30+ (3Π)] and [B 31] states16,17,18 is used to examine the vibronic mode coupling with the excited 3P0 state metal atom Chapter V; I) Hg(3P1)/Ne, II) Hg(3P0)/RG Sims. 125 which lead to stabilization. Unlike the Hg 3P1 excited state which has axial symmetry, the Hg 3P0 state exhibits spherical symmetry as the atomic electronic angular momentum, Je is 0. A comparison of the calculated and observed emission energies lead to the identification of the excited state ‘breathing’ mode (Q1*) as the vibronic mode coupling with this excited state. The theoretical model predicts excited state stabilisation for the ‘breathing’ mode in Hg/Ar, Hg/Kr and Hg/Xe systems. The Hg(3P0 → 1S0) emission energies calculated are progressively red-shifted of the observed band maxima. V.II.2 Method and Results The localised pair-potential approach is applied to the simulation of the 6p 3P0 → 6s 1S0 transition of atomic Hg isolated in solid Ar, Kr and Xe. The simulations undertaken represent extensions to those presented in Part I of this chapter for the analysis of the Hg (3P1 ↔ 1S0)/RG luminescence for solid Ne and outlined in Chapter IV in simulating the observed luminescence in solid Ar, Kr and Xe. Therefore only the modifications necessary to complete the Hg(3P0)/RG calculations are presented. V.II.2.I Ground 1S0 and Excited 3P0 states Normally the simulated absorption energies are calculated and compared to the experimental data allowing the identification of site of occupancy. The Hg (3P0 ← 1S0) absorption cannot be observed experimentally, due to the negligible oscillator strength of the transition, making this comparison impossible. However, the pair potentials simulations presented in Chapter IV for the Hg 3P1 ↔ 1S0 transition concluded that Hg occupies rigid (unperturbed) substitutional sites (ss) only in solid Ar, Kr and Xe. Assuming exclusive substitutional site occupancy, a comparison of the experimental and calculated emission energies is completed, which provide insight into the vibronic modes coupling with the excited 3P0 state metal atom. Calculation of the solid-state Hg (3P0 ↔ 1S0) atom absorption and emission energies is achieved by calculating the energy for both the ground and excited states of the guest metal atom, (Hg) occupying a substitutional site in a Hg⋅RG18 cluster. The method employed to calculate the energy of the ground state for the Hg⋅RG18 cluster using a simple sum of the Hg(1S0)⋅RG and RG⋅RG pair potentials was described in Chapter IV. Chapter V; I) Hg(3P1)/Ne, II) Hg(3P0)/RG Sims. 126 Due to the spherical symmetry of the 3P0-excited state (Je = 0) the energy of the excited state cluster is also a simple summation of the Hg(3P0)⋅RG and RG⋅RG pair potentials, modifying Equation IV.1 for the excited state. The Hg(3P0)⋅RG diatomic potentials are not available experimentally for all the Hg⋅RG pairs (except Ar19). However the Hg[3P0 (0) ã]⋅RG potential can be calculated from the known Hg(3P1)⋅RG [A 3Π (30+)] and [B 3Σ (31)] state potentials presented. The A state is pure Π whereas the B state is a linear combination Π and Σ atomic orbitals. The pure Σ component was extracted from this B state, as outlined previously. The expressions for VΠ and VΣ were then used to obtain an expression for the diatomic Hg(3P0)⋅RG interaction using the following equation presented previously by Duval et al19. V( 3P0) ã30 = 1/3[(VΣe + VΠe) + VΠ] Equation (V.1) The Morse function parameters and the energy curves calculated for the Hg(3P0)⋅RG diatomics are presented in Figure V.16. 3.0 4.0 5.0 -1000 -500 0 Energy (cm-1) Hg (3P0)a(Exp.) a30 state parameters Rmin=4.260 Å De= 73.818 cm-1 a30 state parameters (Exp) Rmin=4.330 Å De=110.000 cm-1 Hg Ar 3.0 4.0 5.0 R ( Å ) Hg(3P1)Π Hg(3P0)a Hg(1S0)X a30 state parameters Rmin=4.130 Å De=136.317 cm-1 Hg Kr 3.0 4.0 5.0 a30 state parameters Rmin=3.600 Å De=351.275 cm-1 Hg Xe Figure V.16 Potentials of the Hg⋅RG diatomic the X and the A(Π) state potentials are obtained directly from spectroscopic data of the Hg⋅RG diatomics presented in Table IV.1, Chapter IV while the ã30state potential was obtained with Equation V.1 which assumes case-(c) coupling. The available spectroscopic data for the Hg(3P0)⋅Ar system reported by Duval et al19 is also presented. It is evident from the potential energy curves shown in Figure V.16 that the Hg (3P0) excited state is similar to the Hg⋅RG ground state interaction. By inspection Chapter V; I) Hg(3P1)/Ne, II) Hg(3P0)/RG Sims. 127 of Figure V.16, it is also clear that the Ar ground state bond length is less than that of the excited state, whereas in Xe, the reverse is the case. The ground and excited states for Hg⋅Kr show very little difference in bond length. The zero total angular momentum of the 3P0 state suggests that only a solidstate vibronic mode which conserves the spherical symmetry of the Hg(3P0)⋅RG18 cluster will couple with this excited state. The details of the interactions involved for the lattice expansion or contraction ground state ‘breathing’ mode (Q1) calculation are outlined in Section V.I.2.II. In this case the Q1* calculation pertains to the excited state of the metal atom where the same lattice interactions considered for Q1 are used to identify excited state stabilisation of the Hg 3P0 metal atom. The results of the excited state ‘breathing’ mode (Q1*) calculations are shown in Figure V.17. Inspection of the potential energy curves calculated for the excited state Q1* reveals that the preference for an expansion or contraction of the 12 RG nearest neighbour atoms is in line with the trends observed for the Hg(3P0)⋅RG and Hg(1S0)⋅RG states for the 1:1 van der Waals complexes. 0.0 0.2 0.4 0.6 0.8 3.4 3.5 3.6 3.7 3.8x104 Energy (cm-1) Rmin= 0.079 Å Eem= 38635.2 cm-1 Eabs= 38716.3 cm-1 Hg(3P0)/Ar18 0.0 0.2 0.4 0.6 0.8 Q1 *(Å) Sum 120 RG-RG 12 nn Hg-RG 6 nnn Hg-RG Rmin=0.009Å Eem= 38142.1 cm-1 Eabs= 38143.1 cm-1 Hg(3P0)/Kr18 -0. 6 -0.5-0.3-0.20.0 Rmin= -0.061 Å Eem= 37375.5 cm-1 Eabs= 37467.1 cm-1 Hg(3P0)/Xe18 Figure V.17 Energetics calculated for the Hg/RG18 (RG = Ar, Kr and Xe) symmetric excited state ‘breathing’ mode (Q1*) based on rigid substitutional site occupancy. The individual interactions involved are shown by the legend. The total potential energy calculated for the mode is shown by the solid line obtained by summation of the individual interactions. Chapter V; I) Hg(3P1)/Ne, II) Hg(3P0)/RG Sims. 128 Q1* involves a contraction in solid Xe as the Hg⋅Xe excited (3P0) and ground state bond lengths are 3.6 Å and 4.25 Å respectively. The Hg/Ar and Kr systems exhibit an expansion due to Q1* as the Hg⋅Ar and Hg⋅Kr ground state bond lengths are less than the excited state bond lengths as shown in Figure V.16. V.II.2.II Hg(3P0 ↔ 1S0)/RG18 Absorption and Emission Energies As stated earlier, neither absorption nor excitation spectra exist for the ‘forbidden’ 3P0 ↔ 1S0 transition of atomic Hg in solid Ar, Kr or Xe. However, assuming mirror symmetry between the 3P0 absorption and emission (equivalent full width at half maximum, (fwhm) and intensity) allows the creation of an estimated ‘best guess’ absorption spectra. The estimated absorption spectra presented in Figure V.18 were achieved by assuming mirror symmetry about the ZPL’s identified in Chapter III for the Hg(3P0 → 1S0)/RG, listed in Table V.6. Comparison of calculated absorption energies for the Hg 3P0 ← 1S0 transition in the RG solids are then made. 37.538.038.539.0 x103 E ne r g y ( c m -1 ) Intensity (Arb. Units) 256 258 260 262 264 266 26 8 Wavelength (nm) Q1 * Simulated Abs. ** Ar Kr Xe Hg(3P0)/RG Estimated absorption Emission Figure V.18 A comparison of the observed and calculated Hg(3P0 → 1S0)/RG emission spectroscopy recorded at 12 K (reported Chapter III), and the estimated Hg(3P0 ← 1S0)/RG absorption spectra (shown left). The vertical lines represent the Q1* absorption and emission energies calculated (Table V.7) for Hg isolated in substitutional sites in the solid RG’s. ** Indicates the calculated absorption energy for Q1* using the experimental Hg⋅Ar experimental data19. Chapter V; I) Hg(3P1)/Ne, II) Hg(3P0)/RG Sims. 129 Table V.7 A comparison of the observed emission wavelength (nm units) for the 3P0 → 1S0 transition of matrix-isolated atomic mercury with the calculated emission values for the excited state ‘breathing’ mode (Q1*) for Hg isolated in substitutional sites in solid Ar, Kr and Xe. The Hg/Ar** results presented were achieved using the spectroscopic available parameters19 reported. The difference between the observed band maxima and the predicted values are quoted as δObs-Cal in cm-1. Hg/RG ECal λCalc λObs δObs-Cal Hg/Ar 38635.2 258.83 258.9 - 10.2 Hg/Ar** 37965.0 263.40 258.9 + 660.0 Hg/Kr 38142.1 262.18 261.0 + 172.0 Hg/Xe 37375.5 267.55 265.4 + 296.5 The estimated absorption energy is obtained by subtraction of the difference in energy between the ZPL position and that of the phonon sideband labelled ∆ cm-1 in Table V.6. The estimated Hg(3P0)/RG absorption positions are 38855, 38418 and 37764 cm-1 for Ar Kr and Xe respectively, shown in Figure V.18 and collected in Table V.8. Table V.8 A comparison of the estimated experimental absorption wavelength (nm units) for the 3P0 ← 1S0 transition of atomic mercury with the calculated absorption values for Hg/Ar, Kr and Xe. The difference between the estimated band maxima and the calculated values are quoted as δObs-Cal in cm-1. ** Indicates the calculations were completed using the spectroscopic parameters for the Hg(3P0)⋅Ar system reported by Duval et al19. Hg/RG EAbs Cal λCalc λEstm δEstm-Cal Hg/Ar 38716.3 258.29 257.37 + 138.7 Hg/Ar** 37984.2 263.27 257.37 + 870.8 Hg/Kr 38143.1 262.17 260.29 + 274.9 Hg/Xe 37467.1 266.90 264.80 + 296.9 V.II.3 Discussion A comparison of the experimental data and the results of the pair potential simulations, shown in Figure V.18, reveals that the predicted emission is within the broad-band profile of the phonon sideband in solid Ar (using the extracted 3P0 potential). The calculations outlined in this report predict the position of the ZPL in absorption as the calculation assumes a pure Frank Condon transition. Table V.7 presents the calculated emission energy for the excited state ‘breathing’ mode of atomic Hg 3P0 excited state in solid Ar. Upon inspection it is observed that the Chapter VI; Mn/RG Absorption Spectroscopy 136 34.536.037.5 x103 E ner g y ( cm-1 ) 0.0 0.1 0.2 0.3 0.4 Absorbance 270 280 290 Wavelength (nm) Mn/Ar Td=12K low metal flux y6P____ 6S 34.536.037.5 x103 E ner g y ( cm-1 ) 270 280 290 Wavelength (nm) Td=12K TAn. =29K Figure VI.2 Mn/Ar absorption spectrum recorded at 12 K upon deposition (Td) at 12 K, in the vicinity of the atomic Mn y6P5/2 ↔ a6S5/2 transition, shown left. Sample deposition was completed at 12 K using a low manganese flux, and indicates the presence of sites on both the blue and red sides of the dominant feature centered at 278.1 nm. A comparison of the absorption recorded on deposition and following matrix annealing to 29 K (TAn.) is shown right. The absorption feature observed at 397.4 nm in solid Ar is assigned to the z6P5/2 ← a6S5/2 transition, based on its proximity to the location of this transition at 403.42 nm in the gas phase7. Inspection of the bottom panel of Figure VI.1 reveals the absorption band exhibits a progressive shift to lower energy with increased metal loading. This red-shift is attributed to the production of Mn2. The band maximum of this feature occurs at 402.3 nm but the high-energy blue wing, assigned to the atomic transition, is also evident, as shown in the top panel of Figure VI.1. The atomic z6P5/2 ← a6S5/2 transition of Mn is blue shifted in Ar by 375 cm-1 from its position in the gas phase. The very different intensities of the recorded y6P and z6P absorption bands of atomic Mn result from their singlet and triplet characteristics respectively of these two excited states. This is reflected in the reported Einstein Aki coefficients18, of the y6Po 3d5(6S)4s4p(1Po) ↔ 3d54s2 a 6S and z6Po 3d5(6S)4s4p(3Po) ↔ 3d54s2 a 6S transitions 3.7 x108 and 0.19 x108 sec-1 respectively. Inspection of the UV region shown in Figure VI.1, close to the gas phase x6P5/2 ↔ a6S5/2 transition energy reveals the presence of two features at 211.8 and Chapter VI; Mn/RG Absorption Spectroscopy 137 226.4 nm. The bottom panel shows that the 211.8 nm feature is present in very dilute Mn/Ar samples whereas the 226.4 nm19 feature gains in intensity relative to the dominant 278.1 nm band with increased metal loading. Therefore the 211.8 nm feature is assigned to the x6P5/2 ← a6S5/2 absorption in Ar, blue-shifted by 2058 cm-1 from the gas phase position at 221.45 nm, (45156 cm-1)7. The observed gas phase-Ar matrix shift is greatest for the x6P atomic absorption. This may arise because the excited electronic configuration ([Ar]3d64p) is reached by a two electron transition from the [Ar]3d54s2 ground configuration. Table VI.1 presents a summary of the absorption spectroscopy of atomic manganese isolated in solid Ar. Table VI.1 Spectral positions of the atomic absorption features assigned for atomic manganese isolated in solid Ar on deposition at 12 K. λabs indicates the position of the band centre or the central three-fold split component where possible. The gas phase transition energies for the Mn atom are also presented. The gas phase – Ar matrix shift is denoted by δ in wavenumber units. Mn/Ar Mn atom – Gas Phase7 Transition λAbs (nm) EAbs (cm-1) λ (nm) E (cm-1) δ (cm-1) z 6P5/2 ← a 6S5/2 y 6P5/2 ← a 6S5/2 x 6P5/2 ← a 6S5/2 397.4 278.1 (1°) 273.0 (2°) 211.8 25163 35958 36630 47214 403.42 279.91 211.45 24788 35725 45156 +375 +232 +904 +2058 The Mn/Ar samples prepared at 12 K using low metal fluxes show very weak absorption features at 254.4 and 311.6 nm. These features are observed to increase in intensity at medium metal fluxes, middle panel of Figure VI.1. Another 226.4 nm feature is present in the most concentrated samples19. In addition to the 226.4, 254.4 and 311.6 nm features, the red-shift of the band maximum of the 397.4 nm band to 402.3 nm is assigned to an increased contribution from a manganese dimer absorption. Therefore, the 254.4, 311.6 and 402.3 nm features are assigned to absorption transitions of manganese aggregates most probably Mn dimers in Ar matrices. Inspection of the high metal flux deposition shown in the top panel of Figure VI.1, for the most concentrated Mn/Ar samples prepared, show an additional set of features at 226.4 and 345.7 nm. These bands are tentatively assigned to Mnx where x > 2. The assignment to Mn2 species is rejected as Mn2 absorptions are identified even in the most dilute Mn/Ar samples at 254.4, 311.6 and 402.3 nm. The van der Waals Chapter VI; Mn/RG Absorption Spectroscopy 138 nature of Mn2 can result in the efficient formation of larger cluster species Mnx (x > 2) due to chemical bond formation in the higher metal aggregates. The 226.4 and 345.7 nm bands are assigned to absorptions of manganese aggregates with higher nuclearity than the dimer. This assignment is consistent with the trends evident in the concentration studies completed. The additional dimer features assigned from the concentration studies at 254.4, 311.6 and 402.3 nm showed no discernible temperature dependence. It is also evident from the top panel of Figure VI.1 that increasing the Mn loading increases the complexity of the absorption spectra recorded. This spectral congestion is most pronounced from 300 to 360 nm, where two weak bands appear at 330 and 345.7 nm. Additional features are also observed in the vicinity of the y6P5/2 and z6P5/2 ← a6S5/2 transitions at 290 and 438.5 nm respectively. The discussion and assignment of these features is postponed until the Mn/Kr absorption spectroscopy is presented. VI.2.II Discussion Mn/Ar absorption spectroscopy In the following section the absorption spectroscopy recorded for manganese isolated in solid Ar is discussed with reference to the previous Mn/Ar experimental work9,1013,17. The first reports of the absorption spectroscopy of Mn/Ar are those of Schnepp9 and Lee and Gutmacher10 who employed photographic detection methods to observe the atomic y6P and z6P ← a6S transitions in matrices deposited at 4.2 K. Overall the earlier absorption spectra showed good agreement with those presented in this Chapter in that both studies assigned the y6P ← a6S transition of atomic manganese in Ar as a triplet centered at 277.54 and 277.7 nm in Refs. 9 and 10 respectively. These results show close agreement with the centre of the threefold split absorption reported here, which occurs at 278.1 nm, Table VI.1. The weaker, absorption features to the blue of 278.1 nm were also observed and assigned9,10 to multiple site occupancy of Mn atoms in solid Ar. The z6P ← a6S transition in Ar was reported at 396.6 and 403.06 nm by Schnepp9 and at 392.77 and 402.25 nm by Lee and Gutmacher10. The higher energy features shows good agreement with the z6P ← a6S absorption reported here at 397.4 nm. The absorption at 402.3 nm was assigned to Mn2 in the previous section from Chapter VI; Mn/RG Absorption Spectroscopy 139 concentration study completed (Figure VI.1), as this feature is absent in dilute Mn/Ar samples. All the earlier samples9,10 exhibited strong absorption at 402/403 nm and represents the larger metal fluxes which accompany Mn vaporisation by bulk resistive heating. Shakhsemampour et al13 assigned two sites of isolation for atomic manganese in solid argon. The primary site (s1) was centered at approximately 279.5 and 393.5 nm with the secondary site (s2) observed at 275.25 and 381 nm corresponding to the y6P ← a6S and z6P ← a6S atomic transitions respectively. The spectral positions of the two sites reported for the y6P state are in good agreement with the red (1o) 278.1 nm band and the blue shoulder (2o) 273 nm identified in this Chapter. The report of the dominant site (s1) z6P ← a6S absorption occurring at 393.5 nm agrees with that assigned here at 397.4 nm. The presence of a secondary site (s2) at 381 nm13 is not immediately evident in the absorption spectra reported here. Its presence in Mn/Ar was identified however in luminescence excitation spectroscopy, the results of which are presented in Chapter VII. Lee et al10 reported an additional feature at 311.23 nm in solid Ar, which they assigned to the z4P ← a6S atomic transition, commenting that the absorption band appeared after annealing to 30 K. Mann and Broida11 also assigned the 311.5 nm band to the z4P ← a6S transition of atomic Mn occurring as imperfectly isolated Mn atoms. Given the weak oscillator strength of the z4P3/2 ← a6S5/2 transition (Aki = 0.0027 x 108 sec-1), observation of this spin-forbidden transition using absorption techniques seems unlikely. No evidence for this atomic absorption feature was observed upon deposition of dilute Mn/Ar samples (bottom panel Figure VI.1). However under medium/high metal loading conditions, absorption at 311.6 nm is observed. Consequently this band was assigned earlier in this Chapter to a Mn2 absorption. The observations of Lee et al10 that the intensity of the 311 nm absorption feature increased after matrix annealing and the assignment by Mann et al11 of the z4P ← a6S transition occurring within an imperfect matrix site are inconsistent, as annealing results in the production of a more crystalline matrix environment. Annealing should therefore weaken the z4P ← a6S atomic absorption if it occurs from an imperfect lattice site. However, both observations are consistent with the assignment of the 311 nm feature to a Mn2 absorption band whose presence on deposition is metal-loading-dependent and production after annealing resultant Chapter VI; Mn/RG Absorption Spectroscopy 140 from metal atom nucleation. Mn/Ar concentration studies reported by Ozin and coworker17 also lead to the assignment of the 311 nm feature to that of a manganese aggregate. Ozin and co-workers17 reported the absorption spectra for Mn/Ar in the vicinity of the x6P5/2 ↔ a6S5/2 transition7 of atomic manganese at 221.45 nm in the gas phase. They assign a band at 226 nm in solid Ar to the x6P5/2 ← a6S5/2 transition. The concentration study for Mn/Ar shown in Figure VI.1 shows that only samples prepared using high metal fluxes contain the 226 nm band. This absorption intensity is enhanced relative to the dominant 278.1 nm (y6P5/2 ↔ a6S5/2) atomic absorption feature under higher metal loading conditions. This behaviour is consistent with the assignment of the 226 nm feature to a Mn aggregate. In addition the absorption observed at 211.8 nm in dilute Mn/Ar samples (bottom, Figure VI.1) is in agreement with the assignment to the x6P5/2 ↔ a6S5/2 transition, blue shifted of the gas phase transition by 2058 cm-1. Ozin and co-workers17 observed a similar absorption feature at 211 nm under low metal loading conditions and assigned the feature to the w6P5/2 ↔ a6S5/2 transition. The assignment of the absorption feature to that of the w6P5/2 state produces an inconsistency in the observed matrix-shifts as this w6P5/2 absorption would exhibit a red-shift of 445 cm-1 from the gas phase position at 209.82 nm (47659 cm-1)7, while both the y6P5/2 and z6P5/2 ← a6S5/2 absorption features show blue-shifts. No data is available on the relative oscillator strengths of the w6P5/2 and x6P5/2 ↔ a6S5/2 transitions so assignment cannot be made from the observed intensities. However, from the matrix shifts exhibited and the observation that higher Mn atom concentration lead to the increase in the 226 nm band relative to the dominant y6P5/2 absorption feature, the aggregate assignment made in this Chapter is reinforced. In earlier work, the interpretation of the optical absorption spectroscopy of manganese species isolated in solid rare gases (especially Ar) has proven difficult due to the variety of species present on deposition. The concentration studies of Mn/Ar reported in this Chapter allowed the assignment of the features at 254.4, 311.6 and 402.3 nm to transitions of the manganese dimer. In the literature however, one definitive assignment of a Mn2 transition appears corresponding to the A ← X absorption reported at 650 nm in solid Ar14. Ozin and co-workers17 assigned two sets of small cluster species as Mnx and Mny from concentration studies completed. Mnx Chapter VI; Mn/RG Absorption Spectroscopy 141 they assigned to the binuclear molecule showing absorption features at 253, 312 and 400 nm. These conclusions are in good agreement with the Mn2 absorption species assigned in this Chapter at 254.4, 311.6 and 402.3 nm. The 226.4 nm absorption feature was identified here to a manganese aggregate of higher nuclearity than the dimer, however assignment to Mn2 cannot be rejected based on the absorption spectra reported earlier. In addition the persistence of this band in the most ‘dilute’ samples prepared by Ozin and co-workers17 lead to their assignment of the 226 nm feature to the x6P5/2 ← a6S5/2 atomic transition. However, from the concentration studies reported here this atomic assignment has been amended to an Mn2 absorption. VI.2.III Mn/Kr A comparison of the UV/Vis absorption spectra recorded at 12 K for manganese isolated in solid krypton using different metal fluxes is presented in Figure VI.3. All the spectra shown were recorded at 12 K (Ts) for samples deposited at 12 K, (Td). Examination of the concentration study presented for Mn/Kr in Figure VI.3 reveals that the dominant feature overlaps the gas phase7 y6P5/2 ↔ a6S5/2 transition of atomic manganese at 279.9 nm. The bottom panel of Figure VI.3 presents the absorption spectrum recorded for the most dilute Mn/Kr sample. Three pairs of bands are observed at 210 / 213.1, 279.3 / 284.9 and 385.5 / 401.9 nm. The dominant absorption feature occurring at 279.3 nm exhibits a threefold split pattern and is assigned to the y6P5/2 ← a6S5/2 transition, blue-shifted from the gas phase position by only 79 cm-1. The 279.3 nm absorption band in Mn/Kr exhibits a weak low-energy shoulder, shown Figure VI.4. This is the reverse of the Mn/Ar situation where the pronounced shoulder occurred to the blue of the threefold split absorption assigned to the y6P5/2 ← a6S5/2 transition. The shoulder occurs at 284.9 nm and is assigned to the y6P5/2 absorption occurring from a second site of isolation in solid Kr, red-shifted from the gas phase position at 279.9 nm7 by 626 cm-1. Chapter VI; Mn/RG Absorption Spectroscopy 142 25.030.035.040.045.0 x103 E ne r g y ( c m -1 ) Absorption Intensity 210 245 280 315 350 385 420 455 Wavelength (nm) High Medium Low Mn/Kr metal flux x6P____6S y6P____6S z6P____6S Figure VI.3 Mn/Kr UV/Vis absorption spectra recorded at 12 K following sample deposition at 12 K. The three spectra shown indicate the changes in the relative intensities of the observed bands with increased metal flux. Note the top panel presents the most concentrated sample prepared the dominant absorption feature located at ≈280 nm is fully absorbing. Annealing Mn/Kr matrices to 37 K resulted, as presented in the right panel of Figure VI.4, in the formation of a high-energy shoulder at 276 nm, increased resolution of the 284.9 nm feature and producing an additional band at 291.2 nm. The stability of these features was investigated by irradiation at the band maxima 284.9 and 291.2 nm for 15 minutes. Irradiation of the 291.2 nm feature lead to its removal and also the loss of the high-energy shoulder. No additional absorption bands were produced by the removal of these features. The production of the 276 and 290.9 nm bands by the annealing process is assigned to the production of a thermally induced site of atomic isolation. Mn′ is used to denote Mn atoms isolated in ‘thermally-induced’ sites. The 276 and 290.9 nm bands are assigned to the y6P5/2 and z6P5/2 ← a6S5/2 transitions occurring for Mn′. The annealing process had the effect of resolving the 284.9 nm feature resulting from the removal of a broad low energy component. Chapter VI; Mn/RG Absorption Spectroscopy 143 These effects indicate the presence of multiple thermally stable sites centered at 279.3 and 284.9 nm labelled blue (1°) and red (2°) sites respectively. The removal of the weak high-energy shoulder reveals the presence of at least one thermally unstable site of atomic isolation present in solid Kr on deposition at 13 K. 34.536.037.5 x103Ener g y ( cm-1 ) Intensity 270 280 290 Wavelength (nm) Mn/Kr Low metal flux Td=13K y6P____ 6S 34.536.037.5 x103Ener g y ( cm-1 ) 270 280 290 Wavelength (nm) λIrr. = 291.2 nm 34.536.037.5 x103Ener g y ( cm-1 ) 270 280 290 Wavelength (nm) TAn. =37K Figure VI.4 Mn/Kr absorption spectra recorded in the vicinity of the y6P5/2 ↔ a6S5/2 gas phase transitions of atomic manganese recorded at 12 K following sample deposition at Td (Kelvin) using low manganese atom concentrations. The absorption feature centered at 385.5 nm shows a resolved threefold splitting pattern and is assigned to z6P5/2 ← a6S5/2 transition exhibiting a blue matrix shift of 1152 cm-1. The 385.5 nm absorption feature is overlapped by the lower energy 401.9 nm band. The 401.9 nm band, like the 397.4 nm feature assigned to the z6P5/2 ← a6S5/2 transition in solid Ar, appears to red-shift at higher manganese concentrations. This occurs due to the appearance of a Mn2 band at 413.2 nm, top panel Figure VI.3. Accordingly the 401.9 nm feature is assigned to the z6P5/2 ← a6S5/2 transition of atomic Mn isolated in a secondary site (2o) in solid Kr. Chapter VI; Mn/RG Absorption Spectroscopy 144 Table VI.2 Spectral positions of the atomic absorption features assigned for atomic manganese isolated in solid Kr on deposition at 12 K. λabs indicates the position of the band centre or the central three-fold split component where possible. The dominant/primary (1o) and secondary (2o) site absorptions are labelled blue and red respectively to reflect their relative absorption energies. The gas phase transition energies for the Mn atom are also presented. The gas phase – Kr matrix shift is denoted by δ in wavenumber units. Mn/Kr Mn atom – Gas Phase Transition Site λAbs (nm) EAbs (cm-1) λ (nm) E (cm-1) δ (cm-1) z 6P5/2 ← a 6S5/2 y 6P5/2 ← a 6S5/2 x 6P5/2 ← a 6S5/2 Blue (1o) Red (2o) Blue (1o) Red (2o) Blue (1o) Red (2o) 385.5 401.9 279.3 284.9 210.0 213.1 25940 24882 35804 35100 47619 46926 403.42 279.91 211.45 24788 35725 45156 +1152 +94 +79 -626 +2463 +1770 The UV absorptions occurring at 210 and 213.1 nm, shown in Figure VI.3 are assigned to x6P5/2 ← a6S5/2 absorptions from a dominant blue (1o) and secondary red (2o) site blue-shifted by 2463 and 1770 cm-1 respectively from the gas phase position. Higher metal loading reveals a complex overlapping set of bands where the 210 nm feature dominates confirming its assignment as the primary site of Mn isolation. Table VI.2 presents a summary of the absorption spectroscopy of atomic manganese isolated in solid Kr matrices deposited at 12 K. Absorption spectra recorded for concentrated Mn/Kr samples formed at 12 K contain additional absorption features at 255.3, 317.7 and 413.2 nm evident from a comparison of the top and middle panels of Figure VI.3 recorded for high and medium metal concentrations respectively. The isolation of atomic manganese in Kr is much more efficient than in Ar matrices. This was manifest in the observation that producing samples with detectable amounts of dimer bands resulted in fully absorbing atomic transitions. The 229.1, 255.3, 317.7 and 413.2 nm absorption features observed are assigned to the Mn2 species. Further increasing the metal concentration in Kr results in additional absorption features at 330.7 and 349.1 nm, top panel of Figure VI.3. These features are also assigned to Mn2 absorptions, similar to the bands identified under high metal loading conditions in the 300-350 nm region in solid Ar by Vala and co-workers15. Chapter VI; Mn/RG Absorption Spectroscopy 145 VI.2.IV Discussion Mn/Kr absorption spectroscopy In this section the spectroscopy of manganese atoms and aggregates isolated in solid Kr are discussed in relation to the results and band assignments reported in earlier studies9,13,17. The absorption features assigned to the atomic absorption transitions of manganese in solid Kr are presented in Table VI.2. The absorption spectra reported in this Chapter reveal the presence of multiple sites of isolation for dilute Mn/Kr samples deposited at 12 K. Schnepp9 reported strong triplets of atomic Mn absorptions corresponding to the y6P5/2 and z6P5/2 ← a6S5/2 transitions centered at 279.1 and 395.1 nm respectively. Additional weaker absorption bands were observed at approximately 283.7 and 400.3 nm. The absorption features reported by Schnepp, using photographic detection methods, show good agreement with those identified during the course of this work (Table VI.2) leading to the assignment of multiple occupancy for Mn in solid Kr. The relative absorption strengths identified by Schnepp9 on deposition at 4.2 K mirrored those achieved here with matrix deposition at 12 K. Ozin and co-workers17 achieved similar absorption spectra for dilute Mn/Kr samples. However, Shakhsemampour13 observed absorption features only at 284.5 and 387 nm corresponding to the y6P5/2 and z6P5/2 ← a6S5/2 transitions respectively in solid Kr. The observed features match the red (2o) site for the y6P5/2 and blue (1o) site for the z6P5/2 identified from the results of concentration studies of Mn/Kr collected in Table VI.2. The observation of atomic absorptions correlated to different sites of isolation for the y6P5/2 and z6P5/2 ← a6S5/2 atomic transitions is difficult to resolve. In addition, as in solid Ar, Vala and co-workers13 assigned the z4P3/2 ← a6S5/2 transition of atomic Mn to occur at 310 nm from MCD measurements. This band was not present in the absorption spectra reported in the present study. Given the weak oscillator strength of the z4P3/2 ↔ a6S5/2 transition (Aki = 0.0027 x 108 sec-1), observation of this spin-forbidden using optical absorption techniques is unlikely. As in Ar matrices, the absorption spectroscopy of Mn clusters in Kr is difficult to assess due to the number of absorption bands observed for concentrated Mn/Kr samples. Mn2 absorption features at 229.1, 255.3, 317.7 and 413.2 nm were assigned from concentration studies of Mn deposited in solid Kr at 12 K, (Figure VI.3). These bands show good agreement with the 254, 317 and 410 nm absorption features assigned to Mnx (where x is most probably two) by Ozin and co-workers17. Chapter VIII; Mn(z8P and a6D)/RG Lum. 248 represents an apparent reversal of the site dominance observed in the z6P5/2 state absorption spectra reported in Chapter VI for Mn/Kr. However, the tentative assignment of the 587 nm and the 626.7 nm emissions to site-specific relaxation of the a6D9/2 state, where the a6D ↔ a6S transition is enhanced in one site, provides a plausible reason for this observation. Therefore, it is proposed that the broad 626.8 nm feature results from the emitting a6D level undergoing a stronger interaction with the matrix in a distinct site of isolation which results in broadening and shifting the a6D → a6S transition. If the site-specificity of the z6P state emission is maintained for the a6D state, the observation that the excitation spectra recorded for the 587 and 626 nm emission bands are in the same narrow spectral range indicates the site occupancy has only a small influence on the a6DJ ← a6S5/2 transition energy. These proposals are consistent with the unshifted positions of the excitation bands observed. An analysis of the excitation spectrum recorded monitoring the 587 nm emission allows the assignment of the observed features to electronic transitions between the ground a6S5/2 state and the spin-orbit levels of the a6DJ excited state atomic Mn for J = 1/2, 3/2, 5/2, and 7/2 respectively. This is revealed in Table VIII.4 by comparing the splittings recorded for the resolved matrix excitation features (∆Mn/Kr) at 566.18, 568.44, 572.54 and 578.27 nm with the gas phase spin-orbit splittings, (∆GP). The transitions identified in solid Kr occur to higher energy than the gas phase positions1 by 15 cm-1. This effect is attributed to a weak stabilisation of the a6S5/2 ground state of atomic Mn isolated in Kr. As presented for Mn/Xe in the previous section, a transition to the a6D9/2 level was not observed in excitation. On the basis of the conserved spin-orbit splittings, the a6D9/2 ↔ a6S5/2 transition is predicted to occur at 586.06 nm (17063 cm-1) in solid Kr. Emission spectra recorded at 12.5 K are shown in Figure VIII.9 resulting from a6D5/2 excitation at 572.46 nm following Mn/Kr sample deposition at 12.5 K and subsequent annealing to 34 K. The spectra presented show three emission features located at approximately 587, 604 and 628 nm. The spectra have been normalised, so a comparison of the relative intensities of the observed features, before and after matrix annealing, reveals the thermally instability of the 604 nm emission feature and the enhancement of the 587 nm feature by the annealing procedure. Chapter VIII; Mn(z8P and a6D)/RG Lum. 249 Table VIII.4 The transitions assigned and the photophysical characteristics of the resolved excitation features recorded by monitoring the emission at 587 and 626.8 nm. The spectral positions of the observed excitation features and the gas phase transition energies for the individual a6DJ ↔ a6S5/2 are indicated in nm and wavenumber units (cm-1) for the individual spin-orbit levels. ∆ indicates the splitting between successive spin-orbit levels in the gas phase (G.P.)1 and Kr matrix environment (Mn/Kr). The matrix-shift (δ) observed for the assigned transition is presented in wavenumber units. Mn Gas Phase Mn/Kr Excitation Transition (nm / cm-1) 1 ∆G.P. (cm-1)1 Assignment (nm) / (cm-1) ∆Mn/Kr (cm-1) δ (cm-1) a6D1/2 ↔ a6S5/2 566.98 / 17637 a6D3/2 ↔ a6S5/2 569.21 / 17568 a6D5/2 ↔ a6S5/2 573.03 / 17452 a6D7/2 ↔ a6S5/2 578.63 / 17282 a6D9/2 ↔ a6S5/2 586.43 / 17052 69 116 170 230 a6D1/2 ← a6S5/2 566.18 / 17662 a6D3/2 ← a6S5/2 568.44 / 17592 a6D5/2 ← a6S5/2 572.54 / 17466 a6D7/2 ← a6S5/2 578.27 / 17293 70 126 173 +25 +24 +14 +11 15.315.615.916.216.516.817.1 x103Ener g y ( cm-1 ) Intensity 590 600 610 620 630 640 650 Wavelength (nm) Td=13.0K TAn. =34.0K Mn/Kr - Emission λEx. =572.46nm T S=12.5K a6D9/2 ____a6S5/2 Figure VIII.9 Emission spectra recorded with dye laser excitation at 572.46 nm corresponding to the a6D5/2 ← a6S5/2 transition assigned. The dashed vertical line indicates the gas phase position1 of the a6D9/2 ↔ a6S5/2 transition of atomic Mn. High-resolution emission spectra recorded at 12.6 and 15 K with excitation at 568.4 nm (corresponding to the a6D3/2 ← a6S5/2 transition), are shown in Figure Chapter VIII; Mn(z8P and a6D)/RG Lum. 250 VIII.10. At 12.6 K the emission band shows a clear asymmetry and exhibits a band maximum at 585.75 nm (17072 cm-1). At a higher temperature (15 K) the intensity of the band maximum is reduced relative to the low energy wing. This temperature dependence was completely reversible allowing the assignment of the 585.75 nm (17072 cm-1) band as the band origin (ν0,0) of the pure a 6D9/2 → a6S5/2 electronic transition. Excitation of the remaining J levels in the four possible a6DJ ← a6S5/2 transitions produced the same 585.75 nm feature but with no additional emission bands. This behaviour indicates efficient IMR amongst the J levels, which populates the lowest energy J level. Given the slow radiative decay rate, the IMR rate must greater than 103 sec-1. A summary of the high-resolution excitation and emission spectroscopy corresponding to the electronic transitions to the a6DJ states is presented in Figure VIII.11. 16.9016.9517.0017.0517.10 x103 E ner g y ( cm-1 ) 0 200 400 600 Reaw Counts 586 588 590 59 2 Wavelength (nm) TS=12.6K TS=15.0K Mn/Kr-Em.,λEx . =568.4nm a6D9/2 ____a6S5/2 Figure VIII.10 High-resolution emission spectra recorded at TS (Kelvin) upon excitation at 568.4 nm (17593 cm-1) corresponding to the a6D3/2 ← a6S5/2 transition assigned. Chapter VIII; Mn(z8P and a6D)/RG Lum. 251 17.017.117.317.417.617.7 x103 E ner g y ( cm-1 ) Intensity 564 567 570 573 576 579 582 585 588 591 Wavelength (nm) Mn/Kr - Ex., λEm. =587.05nm Em., λEx. =568.39nm a6D1/2 _____ a6S5/2 a6D3/2 _____ a6S5/2 a6D5/2 _____ a6S5/2 a6D7/2 _____ a6S5/2 a6D9/2 _____ a6S5/2 ZPL ZPL ZPL ZPL Figure VIII.11 High resolution excitation spectrum recorded monitoring emission at 587.05 nm assigned to the a6D9/2 → a6S5/2, shown left, and emission recorded with excitation at 568.39 nm assigned to the a6D3/2 ← a6S5/2 transition of atomic Mn in solid Kr at 12.5 K. The dashed vertical lines indicate the gas phase positions of the a6DJ ↔ a6S5/2 transitions. The band origins, zero phonon lines (ZPL) assigned for the a6DJ ← a6S5/2 transitions observed in solid Kr are indicated. To further analyse the excitation spectrum shown on the left of Figure VIII.11 and identify the origin of the lineshapes assigned to the transitions to the individual spinorbit levels of the a6D excited state, excitation spectra were recorded at temperatures in excess of 12.6 K. Figure VIII.12 presents the temperature dependence recorded for the 566.18 nm (17662 cm-1), 568.44 nm (17592 cm-1) and 572.54 nm (17466 cm-1) excitation features assigned in Table VIII.4 to the a6D1/2; a6D3/2 and a6D5/2 ← a6S5/2 transitions of atomic Mn. Inspection of the right hand panel of Figure VIII.12 reveals evidence for the presence of a zero-phonon line in excitation as the relative intensity of the bands at 572.54 nm (17466 cm-1) and 572.05 nm (17481 cm-1) changes with increasing temperature. Although both features are diminished at higher temperatures (18 K) the rate at which this process occurs is different. Chapter VIII; Mn(z8P and a6D)/RG Lum. 252 17.6017.6517.70 x103 E ner g y ( cm-1 ) 0.0 0.3 0.6 0.9 1.2x104 Raw Counts 566 568 Wavelength (nm) Ex., λEm. = 587.05 nm 17.4 0 17.4517.5017.55 x103 E ner g y ( cm-1 ) 0 3000 6000 9000 570 572 574 Wavelength (nm) TS=12.6K* TS=15.0K TS=18.0K Figure VIII.12 Excitation spectra recorded at sample temperatures TS (Kelvin) monitoring the emission at 587.05 nm assigned to the a6D9/2 → a6S5/2 transition of atomic Mn isolated in solid Kr. The excitation features assigned to the a6D1/2 and a6D3/2 ← a6S5/2 transitions at 566.18 nm (17662 cm-1) and 568.44 nm (17592 cm-1) are shown left. The feature assigned to the a6D5/2 ← a6S5/2 at 572.54 nm (17466 cm-1) and the resolved but unassigned 572.05 nm (17481 cm-1) feature, shown right. Increasing the sample temperature from 12.6 K (solid) to 15 K (dashed) results in an immediate reduction in the intensity of the sharp 572.54 nm feature while the intensity of the 572.05 nm feature is only slightly reduced. This allows the assignment of the 572.54 nm (17466 cm-1) feature as the band origin of the a6D5/2 ↔ a6S5/2 transition (ν0,0). The broader band at 572.05 nm (17481 cm-1) is then assigned to the phonon sideband. In addition to this effect, all the excitation features manifest a decrease in intensity and a shift to lower energy. Following the identification of the zero-phonon line (ν0,0) for the a6D5/2 ↔ a6S5/2 transition at 17481 cm-1, the matrix shift (δ) for the transition is then +15 cm-1. Subtracting this matrix shift from the recorded excitation spectrum monitoring the a6D9/2 → a6S5/2 emission, and the high-resolution emission spectrum allows a critical assessment of the transitions assigned. Comparison of the location of the resolved excitation features corrected for the matrix-shift (δ = + 15 cm-1) with the gas phase positions in Figure VIII.13 reveals how good the agreement is. Slight deviation in the location of the features assigned to the a6D1/2 and a6D3/2 ← a6S5/2 transitions at Chapter VIII; Mn(z8P and a6D)/RG Lum. 253 566.18 nm and 568.44 nm (17662 and 17592 cm-1) is also revealed. The linewidths of the a6DJ ← a6S5/2 bands at 566.18, 568.44, 572.54 and 578.27 nm assigned to the J = 1/2, 3/2, 5/2 and 7/2 levels increase, as is evident upon inspection of Figure VIII.13, with increasing J. This behaviour may arise from the partial removal of the level degeneracy leading to band profiles broadening as 2J+1. 17.017.117.317.417.617.7 x103Ener g y ( cm-1 ) Intensity 567 570 573 576 579 582 585 588 591 Wavelength (nm) Mn/Kr - Ex., λEm. = 587.05 nm δEx. =-15.0cm -1 Em., λEx. =568.39nm δEm. =-15.0cm -1 a6D1/2 a6D3/2 a6D5/2 a6D7/2 a6S5/2 a6D9/2 a6S5/2 Figure VIII.13 High resolution excitation spectrum recorded monitoring, λEm. = 587.05 nm corrected by subtraction of the matrix-shift (δ) of 15 cm-1 calculated from the position of the ZPL assigned to the a6D5/2 ← a6S5/2 transition, shown left. The emission spectrum recorded with excitation at 568.39 nm assigned to the a6D3/2 ← a6S5/2 transition of atomic Mn in solid Kr at 12.5 K following correction for the matrix-shift (shown right). The dashed vertical lines indicate the gas phase positions of the a6DJ ↔ a6S5/2 transitions to the excited state spin-orbit levels1. From its spectral position, the single emission band observed for atomic Mn isolated in solid Kr at 585.75 nm (17072 cm-1) is assigned to the a6D9/2 → a6S5/2 transition. The correction of the emission spectrum for the matrix-shift identified in excitation also succeeds in accounting for the ZPL assigned in emission. A decay profile recorded with the TCSPC technique by monitoring the 585.7 nm emission feature at 12.5 K is presented in Figure VIII.14. An adequate fit was achieved employing a double exponential trial function. Decay times of 227 and 26.5 µsec were extracted. The longer microsecond component dominated the decay profile Chapter VIII; Mn(z8P and a6D)/RG Lum. 254 recorded at 12.5 K. Recording the decay profile at higher temperatures showed only minor temperature dependence, Table VIII.5. Therefore the observed 227 µsec lifetime at 12.5 K is assigned as the radiative lifetime of the electric quadrupole a6D9/2 → a6S5/2 transition of atomic manganese in solid Kr. Table VIII.5 Decay characteristics, components and amplitudes (A) extracted from double exponential fits of the temporal profiles recorded monitoring 585.7 nm emission feature at different temperatures, (Ts) following pulsed laser excitation at 567.55 nm. Note the dominant decay contribution is presented in bold. Ts. (K) Fit Range (msec) A1 τ 1 (µsec) A2 τ 2 (µsec) 12 14 0.005 – 0.8 0.005 – 0.8 380 217 227 224 172 111 26.5 22.6 0.0 0.2 0.4 0.6 0.8 1. 0 T i m e ( m sec ) 102 103 104 105 106 Emission, λEx.=567.55nm Data Fit (Range 0.005 to 0.8 msec) τ1= 0.226679 msec τ2= 0.026525 msec A1= 379.51 A2= 172.01 Mn/Kr Decay profile λEm.= 585.7 nm @ 12.5 K, (Td=13.0K,T An. =33.0K) -100 -50 0 50 100 Residuals Figure VIII.14 Decay profile recorded by monitoring the 585.7 nm emission feature at 12 K using TCSPC following pulsed laser excitation at 567.55 nm. The resolved features assigned to the transitions to the a6D state spin-orbit levels are also present in the excitation spectra recorded monitoring the emission feature centered at 626.7 nm, Figure VIII.8. Analysis of the excitation spectra recorded monitoring the 585.7 and 626.7 nm features shows the spectral overlap of the bands. However, the different linewidths indicate that the site specificity in emission is maintained with direct a6D excitation but less than that observed with z6P5/2 excitation. Chapter VIII; Mn(z8P and a6D)/RG Lum. 255 A time-resolved emission spectrum (TRES) recorded monitoring the emission features at 626.7 nm and 585.7 nm is presented in Figure VIII.15 following pulsed laser excitation of the a6D5/2 ← a6S5/2 transition at 572 nm. This figure shows the intensity of the 585.7 nm feature, assigned to the a6D9/2 → a6S5/2 phosphorescence, decreasing on a microsecond timescale. This behaviour is in agreement with the TCSPC measurements presented earlier in this section, in which the observed radiative lifetime of the a6D9/2 state in the matrix that was identified as 227 µsec. In contrast, the intensity of the 626.7 nm feature is observed, (as shown in Figure VIII.15) to increase on this timescale. The TRES spectrum in Figure VIII.15 also reveals the presence of an emission feature centered at 603 nm corresponding to the thermally unstable red emission feature that is, as reported in Chapter VII, produced with z6P5/2 3° site excitation. The broad emission is not completely removed by the annealing procedure and therefore may be the origin of the 23.5 µsec component present in the decay time recorded monitoring the 585.7 nm emission feature presented in Table VIII.5. This allows a correlation of this band to the 656 nm feature observed in solid Xe and strengthening the argument that the Mn/Xe 25 µsec component results from the z8P5/2 → a6S transition. 0 60 120 180 210 0.5 1.0 1.5 2.0 2.5 3.0 x10 5 580 590 600 610 620 630 640 650 660 Raw Counts Wavelength (nm) Time ( µsec) Mn/Kr Emission; λEx. = 572.0 nm 0 60 120 180 210 0.5 1.0 1.5 2.0 2.5 3.0 x10 5 580 590 600 610 620 630 640 650 660 Raw Counts Wavelength (nm) Time ( µsec) Mn/Kr Emission; λEx 0 60 120 180 210 0.5 1.0 1.5 2.0 2.5 3.0 x10 5 580 590 600 610 620 630 640 650 660 Raw Counts Wavelength (nm) Time ( µsec) Mn/Kr Emission; λEx. = 572.0 nm 0 60 120 180 210 0.5 1.0 1.5 2.0 2.5 3.0 x10 5 580 590 600 610 620 630 640 650 660 Raw Counts Wavelength (nm) Time ( µsec) Mn/Kr Emission; λEx Figure VIII.15 Time-resolved emission spectra recorded monitoring the red features (at 12 K) following pulsed dye laser excitation at 572 nm (a6D5/2 ← a6S5/2). The temporal step and width used was 30 µsec with a delay time of td = 0.0 nsec. Mn/Kr sample deposition was completed at 12.5 K and subsequently annealed. Chapter VIII; Mn(z8P and a6D)/RG Lum. 256 Figure VIII.16 presents a decay profile recorded at 12.5 K for the 626.7 nm emission feature using TCSPC. First inspection of the recorded decay profile reveals a rising portion at short time, indicating the presence of the feeding step. This was also observed in the time-resolved emission spectra, (TRES) presented in Figure VIII.15 in which the 626.7 nm intensity increased on the microsecond timescale shown. A trial function consisting of three exponential components allowed an adequate fit of the temporal profile as shown by the residuals. Two decay components 2.19 and 1.61 msec and a rising component of 108 µsec were extracted. 0 2 4 6 8 1 0 Ti m e ( m sec ) 102 104 106 Emission, λEx.=572.0nm Data Fit (Range 0.012 to 10.0 msec) τ1= 2.187645 msec τ2= 1.613220 msec τ3= 0.107957 msec A1= 553.136 A2= 1184.360 A3= -566.533 Mn/Kr Decay profile - λEm. = 626.0 nm @ 12.5 K, (Td=13.0K,T An. =33K) -200 -100 0 100 200 Residuals Figure VIII.16 Decay profile of the 626 nm emission feature recorded at 12.5 K using TCSPC following pulsed laser excitation of the a6D5/2 ← a6S5/2 transition identified in Kr to occur at 567.55 nm. The residuals present the difference between the triple exponential fit completed and the decay recorded. Note the negative amplitude A3 indicates the presence of a feeding step or rising portion in the decay profile recorded. Analysis of the decay profiles was also done at 14 K. The decay characteristics extracted from the analysis of the decay profiles at 12 and 14 K are presented in Table VIII.6. Comparison of the excited state lifetime components extracted at each temperature reveals that all the components are temperature sensitive over the small temperature range. Therefore the radiative lifetime for the 626.8 nm feature has not been observed. The relaxation mechanism involved and the Chapter VIII; Mn(z8P and a6D)/RG Lum. 257 state assignment of the 626.7 nm emission are discussed in detail at the end of this Chapter following a presentation of the luminescence spectroscopy recorded with direct laser excitation of the z8P excited state. Table VIII.6 Decay characteristics, components and amplitudes (A) extracted from nonlinear least squares analysis of the temporal profiles recorded monitoring emission at 626 nm at different temperatures, (Ts) following pulsed laser excitation at 572 nm. Note the dominant decay contribution is presented in bold. Ts. (K) Fit Range (msec) A1 τ 1 (msec) A2 τ 2 (msec) A3 τ 3 (µsec) 12 14 0.012 – 10.0 0.005 – 8.0 553 605 2.19 2.09 1184 542 1.61 1.29 -567 -228 108 124 Table VIII.7 Photophysical characteristics and excited state assignments of the emission feature produced following excitation of the 3d64s a6DJ ← 3d54s2 a6S5/2 transitions of matrix – isolated atomic manganese isolated in solid Kr. λEm, indicates the emission band-centre in nm and wavenumber units. The matrix shift (δ) for the transitions are presented in wavenumber (cm-1) units. The positions of the band origins (ZPL) identified are presented in wavenumber units. Mn Gas Phase1 Mn/Kr Matrix – Emission Transition Nm / cm-1 Assignment λEm. (nm) / (cm-1) δ (cm-1) Decay Characteristic a6D9/2 ↔ a6S5/2 586.43 / 17052 a6D9/2 → a6S5/2 585.75 / 17072 ZPL = 17072 +20 τ Rad = 227.0 µsec z8P5/2 ↔ a6S5/2 543.4 / 18402 a6D9/2 ↔ a6S5/2 586.43 / 17052 (?) 628.7 / 15906 -2496 -1146 τ Obs. = 1.61 msec ( τ Rise = 108 µsec) VIII.2.III Mn(a6D)/Ar The luminescence spectroscopy of the z6P5/2 excited state of atomic Mn isolated in solid Ar, presented in Chapter VII, identified three site specific emission features at 590, 625 and 604 nm produced with 1° red, 2° blue and 3° site excitation respectively. The 590 nm emission feature resulting from red (1°) site excitation, shown Figure VII.41, was tentatively assigned to the a6D9/2 → a6S5/2 transition. Table VII.6 presents the photophysical characteristics of this emission feature. Blue (2°) site excitation produced emission at 625 nm, which has not been state assigned. Finally, careful annealing experiments showed the 604 nm feature to be a thermally unstable site for Mn isolated in solid Ar. Chapter VIII; Mn(z8P and a6D)/RG Lum. 264 VIII.3 Discussion - Mn(a6D)/RG luminescence In this section, the observed luminescence of the a6DJ ↔ a6S5/2 transition of atomic Mn isolated in solid Ar, Kr and Xe is summarised and some trends evident in the recorded excitation and emission spectroscopy are presented. VIII.3.I Mn(a6D)/RG Excitation spectroscopy The excitation spectra recorded in the region of the gas phase a6D5/2 ↔ a6S5/2 transition, monitoring the red emission features tentatively assigned in Chapter VII to the a6D9/2 state of atomic Mn isolated in solid Ar, Kr and Xe, are presented in Figure VIII.23. All the excitation spectra exhibit resolved features, some with very narrow linewidths, and on the basis on the splitting exhibited and their proximity to the gas phase energies are assigned to transitions to the individual spin-orbit levels of the a6D state from the ground a6S5/2 state. The assignments made to the individual resolved features are presented in Table VIII.1, Table VIII.4 and Table VIII.8 for Mn/Ar, Mn/Kr and Mn/Xe systems respectively. The excitation bands recorded for different sites in a given rare gas matrix overlap indicating that the energy of the a6D state does not shift significantly with different site occupancy in the matrices from the gas phase. Hence, the site of isolation plays only a minor role in the observed excitation spectroscopy. In the next section the role of host in determining the excitation spectroscopy is discussed. As evidenced in Figure VIII.23, a progressive shift to higher energy is observed for the a6D levels from Ar to Xe matrices. The matrix-shift was calculated to be -23, +15 and +120 cm-1 for Ar, Kr and Xe respectively. As the relative splitting between spin-orbit levels is maintained in all three solids, the observed matrix shift is ascribed to the manifestation of the extent of ground state stabilisation for the Mn atom isolated in a given site of isolation within a particular host solid. The comparison of the Mn/Kr high-resolution excitation spectra to the gas phase transition energies and the temperature dependence observed, allowed the assignment of the ZPL’s for the a6DJ (J = 1/2, 3/2, 5/2 and 7/2) ← a6S5/2 transitions in solid Kr. It is significant that the excitation spectra recorded here do not conserve the site dominance identified in the previous Chapter with resonance z6P excitation. In Chapter VII the excitation spectra allowed the identification of specific sites of isolation (1°, 2°, red, blue etc) in the vicinity of the resonance z6P5/2 ↔ a6S5/2 gas Chapter VIII; Mn(z8P and a6D)/RG Lum. 265 phase transition of atomic Mn. However, in the a6D case the luminescence data suggests that although the site-specific emission features, identified in Chapter VII, are present, the emission features produced with blue site (z6P5/2) of isolation at 625 and 627 nm are much weaker with direct a6D excitation. This effect was most evident in solid Xe where the emission intensity for the thermally unstable 656 nm feature persisted after annealing and was observed to have an equivalent intensity as the 620 nm band. As the 620 nm band is the only Mn/Xe thermally stable emission feature reported to date, the weak emission intensity observed with direct a6D excitation must be attributed to the site of isolation. The sites of isolation are discussed in greater detail in Chapter IV where the blue sites are assigned to single substitutional site occupancy in Ar, Kr and Xe. However, the blue site only allows the a6D ← a6S transition to occur efficiently in Xe, whereas the red sites in Ar and Kr allow an enhancement of the transition. This effect may owe its origin to either the site size or the site symmetry. 17.117.417.718.0 x103 E ner g y ( cm-1 ) Intensity 555 560 565 570 575 580 Wavelength (nm) 590 625 17.117.417.718.0 x103 E ner g y ( cm-1 ) Intensity 555 560 565 570 575 580 Wavelength (nm) 588 626 17.117.417.718.0 x103 E ner g y ( cm-1 ) Intensity 555 560 565 570 575 580 Wavelength (nm) 620* a6D5/2 ____ a6S5/ 2 Mn(a6D)/RG - Excitation Ar λEm. (nm) Kr Xe Figure VIII.23 Dye laser excitation spectra recorded monitoring the red emission features as indicated in wavelength units at 12 K for all the Mn/RG systems, in the vicinity of the Mn a6DJ ← a6S5/2 transition. All spectra were recorded following Mn/RG sample deposition at 12 K and matrix annealing. The dashed vertical line indicates the position of the a6D5/2 ← a6S5/2 gas phase transition of atomic Mn at 573.03 nm (17452 cm-1) 1. Chapter VIII; Mn(z8P and a6D)/RG Lum. 266 VIII.3.II Mn(a6D)/RG Emission spectroscopy The emission spectroscopy reported in the previous sections following excitation of the a6D ↔ a6S transitions is summarised in Figure VIII.24. The transitions assigned and photophysical properties of the observed emission features are presented in Table VIII.10, Table VIII.7 and Table VIII.3 for Mn/Ar, Mn/Kr and Mn/Xe respectively. The emission spectroscopy reported here allowed a more definitive assignment of the emission features observed at 590, 587 and 620 nm (presented in Chapter VII following z6P5/2 excitation) in Ar, Kr and Xe to the a6D9/2 → a6S5/2 phosphorescent transition of atomic Mn. The most definitive assignment was achieved for the 587 nm band in the Mn/Kr system. Temperature dependence observed in the emission intensity of the 587 nm band of Mn/Kr allowed the assignment of the ZPL for the band to be at 585.75 nm (17072 cm-1) blue shifted from the gas phase a6D9/2 ↔ a6S5/2 position1 by +20 cm-1. The blue shift observed correlates well with the matrix-shift of +15 cm-1 extracted from excitation of the a6DJ (J = 1/2; 3/2; 5/2;and 7/2) ← a6S5/2 transitions. Inspection of the middle panel of Figure VIII.24 shows the agreement achieved and the effect of increased resolution on the emission band profile observed. The emission features assigned to the a6D9/2 → a6S5/2 transition of Mn isolated in solid Ar, Kr and Xe at 590, 585.75 and 620 nm show uncharacteristic matrix-shifts of –103, +20 and –923 cm-1 respectively. This effect is attributed to the Frank Condon accessible region in excitation for certain site types in the rare gases. In the previous Chapter, Mn isolated in the red sites (identified on the z6P5/2 transition) in Ar and Kr lead to the production of the 590 and 585.75 nm features which are now assigned to the emission of the a6D9/2 state (shown Figure VII.65). However the 620 nm feature assigned to the same electronic transition was observed to occur from the blue site equivalent. It is suggested that the site size available in Kr allows access to the minimum of the excited state potential energy surface, by allowing the observation of the ZPL for the a6D9/2 → a6S5/2 transition. In solid Xe, the Mn atom occupies a different site type and the site size does not allow coincidence of the minima of the ground and excited state. Subsequent interaction with the matrix host results in the broadening of the emission lineshapes and in the matrix-shifts observed. Chapter VIII; Mn(z8P and a6D)/RG Lum. 267 15.016.017.018.0 x103 E ner g y ( cm-1 ) 555 570 585 600 615 630 645 660 675 Wavelength (nm) 15.016.017.018.0 x103 E ner g y ( cm-1 ) 555 570 585 600 615 630 645 660 675 Wavelength (nm) 15.016.017.018.0 x103 E ner g y ( cm-1 ) 555 570 585 600 615 630 645 660 675 Wavelength (nm) 15.016.017.018.0 x103 E ner g y ( cm-1 ) 555 570 585 600 615 630 645 660 675 Wavelength (nm) 15.016.017.018.0 x103 E ner g y ( cm-1 ) 555 570 585 600 615 630 645 660 675 Wavelength (nm) 15.016.017.018.0 x103 E ner g y ( cm-1 ) 555 570 585 600 615 630 645 660 675 Wavelength (nm) a6D9/2 ___ _ a6S a6D5/2 __ __ a6S * Mn(a6D)/RG - Excitation Emission λEx. (nm) Ar λEm.(nm) 590 569.9 Kr 588 568.4 Xe 620* 574 Figure VIII.24 Emission spectra (shown right) recorded at 12 K for all the Mn/RG systems investigated produced with laser excitation in the vicinity of the Mn a6D ← a6S5/2 transition. The excitation wavelengths used are shown (right) as λEx. (nm). The excitation spectra shown (left), were recorded by monitoring emission bands as indicated by λEm. in wavelength units. All spectra were recorded following Mn/RG sample deposition at 12 K and matrix annealing. The dashed vertical lines show the spectral locations of the gas phase transitions of atomic Mn. The asterix indicates that the Mn/Kr emission presented was recorded at lower spectral resolution over this range. VIII.4 Conclusion Mn(a6D)/RG luminescence The excitation spectra recorded monitoring the emission features in the region of the gas phase a6D5/2 ↔ a6S5/2 transition are presented in Figure VIII.23 for Mn atoms isolated in solid Ar, Kr and Xe. The observation of the individual spin-orbit levels in excitation is attributed to weak coupling of the a6D excited state of Mn to the solidstate environment provided by the site of isolation. This contrasts with the excitation spectra recorded monitoring the same emission features, in the vicinity of the gas phase z6P5/2 ↔ a6S5/2 and z8P5/2 ↔ a6S5/2transitions, in the previous Chapter VII and later in this Chapter. The P ← S excitation spectra show characteristic Jahn-Teller threefold splitting pattern and large shifts from the gas phase positions. The comparison indicates the stronger interaction of the z6P5/2 excited state with the Chapter VIII; Mn(z8P and a6D)/RG Lum. 268 matrix environment, highlighted by the observation of different matrix shifts for particular sites of atomic isolation. The conservation of the gas phase spin-orbit splittings on the a6D state in the solid allows the identification of the matrix shift observed for the a6D5/2 ↔ a6S5/2 transition in all RG hosts, to the extent of ground state stabilisation for the Mn atom isolated in a given site of isolation within a particular host. The temperature dependence in the recorded spectra allowed the first observation of a ZPL for a matrix-isolated metal atom in excitation. This is best illustrated for the a6D5/2 ← a6S5/2 transition for the Mn/Kr system. The spectroscopy recorded following direct a6DJ excitation allowed a definitive assignment of the 585 nm emission to the a6D9/2 → a6S5/2 transition in the Mn/Kr system. In Ar and Xe such a definitive assignment was not possible but the 590 and 620 nm emission bands in these solids are attributed to the a a6D9/2 state. The emission features observed at 625, 626.7 and 656 nm in Ar, Kr and Xe matrices remain unassigned. VIII.5 Results Mn(z8P)/RG luminescence The following sections present the luminescence spectroscopy produced with direct laser excitation of the z8P5/2 excited state of atomic Mn isolated in solid Ar, Kr and Xe. Excitation spectra recorded in the vicinity of the gas phase position1 of the 3d54s4p z8P5/2 ↔ 3d54s2 a6S5/2 transition at 543.4 nm (18402 cm-1) allow the identification of the z8P5/2 ← a6S5/2 transition in each RG host. Time-resolved emission spectra and excited state lifetime measurements following pulsed laser excitation are used to attempt state assignments of the red emission bands reported at 625 and 627 nm in solid Ar and Kr respectively. These measurements also serve as a probe of the excited state relaxation mechanisms leading to the observed emission. This secondary role is of great importance in assessing the relaxation paths, which produce the emission features assigned to the a6D9/2 → a6S5/2 transitions produced from the higher lying z6P5/2 state, the luminescence spectroscopy of which was presented in Chapter VII. The dye material used for direct z8P state laser excitation was Coumarin 500. This is tuneable over the spectral range 485 to 535 nm (Chapter II, Table II.IX, Chapter VIII; Mn(z8P and a6D)/RG Lum. 269 bottom panel) and exhibits a fluorescence maximum at 504 nm when pumped with the third harmonic of the Nd:YAG laser at 355 nm. VIII.5.I Mn(z8P)/Xe In Section VIII.2.I the emission at 620 nm produced with a6D7/2 ← a6S5/2 excitation at 574.65 nm was cautiously (Figure VIII.6) assigned to the relaxation of the a6D9/2 state. Figure VIII.3 showed that the 620 nm emission was separable from the thermally unstable 656 nm feature, identified in Chapter VII, Figure VII.2. Figure VIII.25 now presents high-resolution excitation spectra recorded in the vicinity of the gas phase z8P5/2 ↔ a6S5/2 transition by monitoring the 621 (solid line) and 650 nm (dotted line) emissions. Inspection of the 621 nm excitation spectrum reveals two well-resolved features at 521.24 and 527.31 nm and a third weaker feature at 533.21 nm. The overall band shape is characteristic of Jahn-Teller threefold splitting of the z8P5/2 ← a6S5/2 transition of atomic Mn, exhibiting a blue matrix-shift of 562 cm-1 from the gas phase position1 at 543.4 nm, (18402 cm-1). The average linewidth observed for the threefold split components is 192 cm-1. The weak intensity of the lowest energy threefold component is due to the intensity drop-off of the Coumarin 500 dye curve at 535 nm. 18.418.819.219.620.0 x103 E ner g y ( cm-1 ) Intensity 500 510 520 530 540 Wavelength (nm) λEm . =621.0nm λEm . =650.0nm z8P5/2 ____ a6S5/2 Mn/Xe - Laser Excitation Figure VIII.25 Mn/Xe dye laser excitation spectra recorded at 12 K, in the vicinity of the z8P5/2 ↔ a6S5/2 gas phase transition (dashed vertical line). Chapter VIII; Mn(z8P and a6D)/RG Lum. 270 The most striking feature evident in the spectra presented in Figure VIII.25 is the presence of the threefold splitting pattern. This is attributed to the P ← S nature of the electronic transition involved and indicates a strong interaction of the excited state Mn atom with the matrix environment. Earlier in this Chapter the excitation spectra recorded for the a6D excited state, allowed the identification of electronic transitions to the individual spin-orbit (J) levels. This occurred due to the weak guest-host interaction in the a6D excited state of atomic Mn. In the z8P case, the gas phase splitting between the spin-orbit levels1 (5/2 ↔ 7/2 and 7/2 ↔ 9/2) is 130 and 173 cm-1. The observed splittings between the threefold components identified in the solid are 221 and 210 cm-1. From this comparison it is evident that the observed matrix splittings exceed the gas phase spin-orbit splittings1. Moreover, the broad linewidth and the constant splitting further reinforces the assignment of the three observed bands as arising from Jahn-Teller interaction between the degenerate z8P state and the matrix environment. Following the identification of the z8P state in excitation, emission spectra were recorded with z8P ← a6S excitation at 529.4 nm. The emission spectrum recorded is shown in Figure VIII.26 – it exhibits no additional features to those observed previously. 15.016.518.019.5 x103 E ner g y ( cm-1 ) Intensity 510 540 570 600 630 660 69 0 Wavelength (nm) Mn a6D5/2 ____ a6S5/2 Mn z8P5/2 ____ a6S5/2 Mn/Xe Excitation, λEm. = 621.0 nm Emission, λEx . = 529.4 nm Figure VIII.26 Emission spectrum (shown right) recorded at 12.5 K following pulsed laser excitation at 529.4 nm, corresponding to the central threefold split component identified in excitation (shown left). Sample deposition was completed at 12.5 K and subsequently annealed to 37 K. Chapter VIII; Mn(z8P and a6D)/RG Lum. 271 Table VIII.11 Photophysical characteristics of excitation and emission bands assigned to the 3d54s4p z8P5/2 ← 3d54s2 a6S5/2 transition and emission features produced following z8P5/2 ← a6S5/2 excitation of atomic manganese isolated in solid Xenon. λEx / λEm indicates the position of an individual threefold excitation feature and emission band-centre in nm units respectively. The full-width at half-maximum intensity of the excitation/emission features is denoted by ∆ and the gas phase to matrix frequency shifts (δ) are presented for the assigned absorption and emission features are presented in wavenumber (cm-1) units. The decay characteristics extracted for the observed emission feature at 12.7 K are also presented. Mn Gas Phase Mn/Xe Matrix – Excitation Transition1 nm / cm-1 Assignment λEx. (nm) / ∆ (cm-1) δ (cm-1) z8P5/2 ↔ a6S5/2 543.4 / 18402 z8P5/2 ← a6S5/2 521.24 527.31 / ≈ 192 533.21 +562 Mn/Xe Matrix – Emission λEm. (nm) / ∆ (cm-1) Decay (msec) δ (cm-1) a6D9/2 ↔ a6S5/2 586.43 / 17052 a6D9/2 → a6S5/2 620.0 / ≈ 240 τ Obs = 1.75 -926 Figure VIII.27 presents the decay profiles recorded for the 620 nm emission band at 12.7, 14, 22 and 27 K. It is clear from Figure VIII.27 that the radiative decay of the excited state has not been identified, as the decay profile observed is sensitive to temperature, even over the range 12.7 to 14 K. Figure VIII.28 presents a fit of the decay profile recorded by monitoring the 620 nm emission at 12.7 K. Therefore, the observed lifetime of the 620 nm emission is identified as 1.75 msec. The results of the fit of the decay profiles recorded at various temperatures (Table VIII.12) show that the relative contributions of the three temporal components were constant over the range investigated. However, all three decay characteristics decrease with increasing temperature. The temperature dependence observed is consistent with the assignment of the emission feature to the a6D9/2 → a6S5/2 transition of atomic Mn, as the observed excited state lifetime τ Obs (λEx = z8P5/2) of 1.75 msec is longer than that observed with direct z6D5/2 excitation where a microsecond excited state lifetime dominated, as shown in Table VIII.3. Therefore, the temperature dependence exhibited by the decay profiles presented in this section is a manifestation of the efficiency of the z8P ⇒ a6D inter system crossing (ISC) process leading to the observed emission. Accordingly, no z8P5/2 state emission features are identified in the spectra resulting from direct z8P5/2 excitation in solid Xe. Chapter VIII; Mn(z8P and a6D)/RG Lum. 272 0.0 0.5 1.0 1.5 2.0 2.5 Time ( msec ) 102 103 104 Emission, λEx.=528.25nm Mn/Xe - Decay profile λ Em. = 620.0 nm 12.7 K 14.0 K 22.0 K 28.0 K Figure VIII.27 Decay profiles recorded monitoring the 620 nm emission recorded at various temperatures as indicated, following pulsed laser excitation at 528.25 nm. 0 2 4 6 8 Ti m e ( m sec ) 103 104 105 106 107 Emission, λEx.= 528.25 nm Data Fit (Range 0.01 to 7.5 msec) τ1= 1.746429 msec τ2= 0.725057 msec τ3= 0.128203 msec A1= 1059.732 A2= 866.835 A3= 349.999 Mn/Xe Decay profile - λ Em. = 620.0 nm @ 12.7 K, (Td=12.5K,T An. =37K) -200 -100 0 100 200 Residuals Figure VIII.28 Decay profile of the 620 nm emission recorded at 12.7 K using TCSPC following pulsed laser excitation at 528.25 nm. The residuals presented allow an assessment of the fit quality as they represent the difference between the triple exponential fit completed and the decay recorded. Chapter VIII; Mn(z8P and a6D)/RG Lum. 273 Table VIII.12 Decay characteristics, components and amplitudes (A) extracted from nonlinear least squares analysis of the temporal profiles recorded monitoring emission at 620 nm at different temperatures, (Ts) following pulsed laser excitation at 528.25 nm. Note the dominant decay contribution is presented in bold. Ts. (K) Fit Range (msec) A1 τ 1 (msec) A2 τ 2 (µsec) A3 τ 3 (µsec) 12.7 14.0 18.0 22.0 25.0 28.0 0.01 – 7.5 0.01 – 7.5 0.01 – 7.5 0.01 – 6.0 0.01 – 6.0 0.01 – 4.7 1060 876 869 1263 1010 994 1.75 1.67 1.59 1.39 1.34 1.21 867 677 810 784 733 503 725 632 724 569 592 460 350 261 322 342 283 233 128 72.0 124 85.6 81.2 75.3 VIII.5.II Mn(z8P)/Kr The luminescence spectroscopies resulting from z6P state excitation (presented in Chapter VII), and resulting from a6D state excitation, presented earlier in this Chapter, have shown the complexity of the Mn/Kr system, where multiple sites of atomic isolation and emission features have been identified. The emission feature produced to lower energy than the z8P5/2 ↔ a6S5/2 gas phase transition1 of atomic Mn at 585.75 has been definitively assigned to the a6D9/2 → a6S5/2 transition. However the 626.8 nm emission band has not been state assigned. The production of the broad 626.8 nm emission was only observed with blue (1°) site excitation of the z6P5/2 state, Figure VII.63. The temperature dependence observed in the steady-state and decay time measurements indicated a complex inter-system crossing (ISC) process leads to the production of the emission feature. The complexity of the relaxation process involved serve to make direct z8P5/2 resonance excitation important in deciphering the relaxation processes leading to the observed emission. High-resolution excitation spectra recorded with a dye laser monitoring the thermally stable 628.3 and 589.4 nm and unstable 605.2 nm emission features are presented in Figure VIII.29. These excitation restore the site-specific nature of the emission features previously observed for the z6P5/2 excited state luminescence, Chapter VII; Figure VII.43. The dominant 1° site is located to higher energy and exhibits resolved threefold splitting. The excitation spectrum obtained monitoring the 589.4 nm emission is located at lower energy (2° site) and manifests two resolved features and a weaker, low energy feature8. The excitation spectrum recorded Chapter VIII; Mn(z8P and a6D)/RG Lum. 280 Table VIII.14 Photophysical characteristics and excited state assignments of the emission feature produced following site-specific 3d54s4p z8P5/2 excitation. λEm. indicates the emission band-centre in nm and wavenumber units. The matrix shift (δ) for the transitions are indicated in wavenumber (cm-1) units. Mn Gas Phase Mn/Kr Matrix – Emission (1° Site) Transition1 nm / cm-1 Assignment λEm. (nm) / (cm-1) δ (cm-1) Decay Characteristic z8P5/2 ↔ a6S5/2 543.40 / 18402 z8P5/2 ↔ a6S5/2 a6D9/2 ↔ a6S5/2 z8P5/2 → a6S5/2 (?) 565.2 / 17693 627.4 / 15938 -709 -2464 -1114 τ Obs. = 43.4 µsec τ Obs. = 1.61 msec τ Rise = 108 µsec Mn/Kr Matrix – Emission (2° Site) a6D9/2 ↔ a6S5/2 586.43 / 17052 a6D9/2 → a6S5/2 586.75 / 17043 -9 VIII.5.III Mn(z8P)/Ar In this section the luminescence spectroscopy of the z8P5/2 ↔ a6S5/2 transition is presented for Mn isolated in solid Ar. Chapter VII; Section VII.2.II and Section VIII.2.III of this chapter presented the luminescence of the z6P5/2 and a6D excited states of atomic Mn. The three emission features reported at 590, 605 and 625 nm, occurred to lower energy than the z8P5/2 ↔ a6S5/2 gas phase position1 at 543.4 nm (18402 cm-1). The Mn(z6P5/2)/Ar spectroscopy identified multiple sites of isolation, and subsequently attributed the emission features at 590 and 625 nm as resulting from distinct sites of isolation. Results from direct a6D state excitation allowed the 590 nm emission band to be assigned to the a6D9/2 → a6S5/2 transition, Figure VII.41. However, the 625 nm emission feature has not been assigned to either the a6D9/2 → a6S5/2 or z8P5/2 → a6S5/2 transitions of atomic Mn in solid Ar. Dye laser excitation spectra recorded monitoring the thermally stable 590 and 625 nm and unstable 605 nm emission features are presented in Figure VIII.35. These excitation spectra exhibit the same site-specific characteristics previously observed in excitation spectra of the z6P5/2 state, in Chapter VII. The low energy, red site dominates (1° site) and exhibits very well resolved threefold splitting and a blue matrix shift (δ) of 636 cm-1 from the gas phase transition1. Resolved threefold splitting was also recorded monitoring the 625 nm feature. This corresponds to the blue (2°) site of isolation identified for the z6P5/2 transition. The excitation spectrum Chapter VIII; Mn(z8P and a6D)/RG Lum. 281 recorded monitoring the thermally unstable emission at 605 nm shows a broad lineshape consistent with that obtained for the z6P5/2 state, Chapter VII; Figure VII.12. The photophysical characteristics of the excitation spectra are shown in Figure VIII.35 and presented in Table VIII.13. 18.418.819.219.620.020.420.8 x103 E ner g y ( cm-1 ) Intensity 490 500 510 520 530 540 Wavelength (nm) 590 625 605 1osite 2osite 3osite Mn z8P5/2 ____ a6S5/2 Mn/Ar - Excitation λEm. (nm) Figure VIII.35 Mn/Ar high-resolution excitation spectra recorded at 12.5 K by monitoring the emission features as indicated. The dashed vertical lines indicates the spectral location of the z8P5/2 ↔ a6S5/2 gas phase transition1. Table VIII.15 Photophysical characteristics of the 1°, 2° and 3° sites of isolation 3d54s4p z8P5/2 ↔ 3d54s2 a6S5/2 transition of atomic manganese. The spectral position and average full width at half maximum (fwhm) denoted as ∆AV of the three components identified in Gaussian lineshape analyses for the three–fold split excitation spectra are presented in wavenumber units. Gas phase to matrix frequency shifts (δ) are presented for the atomic Mn z8P5/2 ← a6S5/2 transition1 (G.P.: 18402 cm-1) in wavenumber units. Note the frequency shifts are calculated with respect to the central feature of the observed three-fold pattern. Mn/Ar Site E (cm-1) ∆AV (cm-1) δ (cm-1) 1° 19354 19038 18798 ≈ 193 + 636 2° 19989 19617 19315 ≈ 320 + 1215 3° 19773 19346 18982 ≈ 428 + 944 Chapter VIII; Mn(z8P and a6D)/RG Lum. 282 Emission spectra recorded with site-specific excitation at 528.8 nm (1°); 508.8 nm (2°) and 515.8 nm (3°) are shown in Figure VIII.36. The emission spectra shown reveal the red (1°) site excitation at 528.8 nm leads to the production of the 590 nm emission feature previously assigned to the a6D9/2 → a6S5/2 transition. Blue (2°) site excitation at 508.8 nm produced the 625 nm emission previously identified. However, unlike the 565 nm band in the Mn(z8P)/Kr system, no emission feature with a small Stokes’ shift is observed in Ar with direct z8P5/2 ← a6S5/2 excitation. The sections that follow present a detailed analysis of the site-specific emission spectroscopy observed following z8P5/2 excitation, focussing on the excited state lifetime measurements conducted in an attempt to identify the excited state dynamics leading to the observed emission features. 15.516.016.517.017.518.018.519.0 x103Ener g y ( cm-1 ) Intensity 540 560 580 600 620 640 Wavelength (nm) 528.8 1o 508.8 2o 515.8 3o x2.3 Mn a6D9/2 ____ a6S5/2 Mn z8P5/2 ____ a6S5/2 Mn/Ar - Emission λEx. (nm) Site Figure VIII.36 Mn/Ar emission spectra recorded at 12.5 K produced with site-selective laser excitation of the 1°, 2° and 3° sites assigned to the Mn z8P5/2 ← a6S5/2 transition on deposition at 12.5 K. The spectra have been normalised and the scaling factors are shown. The dashed vertical lines indicate the positions of the z8P5/2 ↔ a6S5/2 and a6D9/2 ↔ a6S5/2 gas phase transitions1. VIII.5.III.I Mn(z8P)/Ar - 1° site luminescence As shown in Figure VIII.36, laser irradiation at 528.8 nm corresponding to excitation of the red (1°) site, leads to the production of the 590 nm emission feature assigned in Chapter VIII; Mn(z8P and a6D)/RG Lum. 283 Section VIII.2.III to the a6D9/2 → a6S5/2 transition. Figure VIII.37 presents the decay profile recorded monitoring the 590.5 nm feature at 12 K. The decay times extracted using a trial triple exponential function are dominated by two long lived components of 1.28 and 0.51 msec. These long-lived millisecond components dominate for all temperatures in the range 12 to 24 K. 0 2 4 6 T i m e ( m sec ) 102 104 106 108 Emission, λEx.= 527.7 nm Data Fit (Range 0.01 to 7.0 msec) τ1= 1.2848 msec τ2= 0.5121 msec τ3= 0.0890 msec A1= 2331.56 A2= 2188.21 A3= 1445.36 Mn/Ar Decay profile λEm. =590.5nm@12K(T d=12K,T An =30K) -200 0 200 Residuals Figure VIII.37 Decay profile of the 590.5 nm emission feature recorded at 12 K using TCSPC following pulsed laser excitation at 527.7 nm. The residuals present the difference between the triple exponential non-linear least squares fit completed and the decay recorded. Comparison of the decay times extracted at all temperatures following pulsed laser excitation of the z8P5/2 ← a6S5/2 transition in Table VIII.16 can be made with those presented in Table VIII.9 following a6D5/2 ← a6S5/2 excitation. This reveals that the excited state decay characteristics are slightly longer with z8P5/2 excitation than with a6DJ. This therefore suggests that the relaxation occurs by an efficient z8P ⇒ a6D ISC process which is 100% efficient with red (1°) site excitation as no rise times (non-radiative feeding rates) are evident in the decay profiles recorded. Chapter VIII; Mn(z8P and a6D)/RG Lum. 284 Table VIII.16 Decay characteristics, components (τ) and amplitudes (A) extracted from nonlinear least squares analysis of the temporal profiles recorded by monitoring emission at 590 nm at different temperatures, (Ts) following pulsed laser excitation at 528.8 nm. Note the dominant decay contribution is presented in bold. Ts. (K) Fit Range (msec) A1 τ 1 (msec) A2 τ 2 (µsec) A3 τ 3 (µsec) 12.7 15.0 18.0 21.0 24.0 0.01 – 7.0 0.01 – 5.0 0.01 – 5.0 0.01 – 5.0 0.01 – 5.0 2332 1054 1482 1414 1781 1.28 1.08 0.99 0.91 0.83 2188 641 916 908 1201 512 330 288 240 216 1445 464 581 574 740 89.0 55.2 41.9 21.8 21.9 VIII.5.III.II Mn(z8P)/Ar - 2° site luminescence Emission spectra recorded with blue (2°) site excitation at 508.8 nm yielded the single thermally stable emission centered at 625 nm. Unlike Mn/Kr, presented in the previous section, no feature in the 560 nm region was observed in solid Ar, Figure VIII.36. However, the strong temperature dependence observed for the 565.2 nm band in Kr, suggests that the minimum temperature accessible of 12 K, may not allow the production of an emission feature due to fast non-radiative relaxation of the z8P excited state. Therefore experiments below 12 K are suggested but cannot be realised using the current cryogenic apparatus. Decay profiles recorded monitoring the 625 nm emission band show no indications of a rise time component unlike that present in the 627 nm feature in solid Kr. Analysis of the decay profiles recorded monitoring the 625 nm emission required a triple exponential function, Figure VIII.38 at 12 K. The dominant millisecond decay components extracted at all temperatures are presented in Table VIII.17. The conflicting spectral and temporal characteristics of the 625 nm emission feature preclude a definitive assignment to either the z8P5/2 → a6S5/2 or the a6D9/2 → a6S5/2 transitions of atomic Mn isolated in Ar. Specifically, the broad symmetric lineshape is indicative of a P → S type transition, while the long decay time (980 µsec) and the spectral location suggest an assignment to the a6D9/2 → a6S5/2 transition. The assignment of this emission band is discussed further at the end of this Chapter. Chapter VIII; Mn(z8P and a6D)/RG Lum. 285 0 1 2 3 4 5 Ti m e ( m sec ) 102 104 106 108 Emission, λEx.= 511.2 nm Data Fit (Range 0.001 to 4.5 msec) τ1=0.9802msec τ2=0.3069msec τ3=0.0597msec A1= 1981.65 A2= 1632.07 A3=956.13 Mn/Ar Decay profile λEm. =625.0nm@12K(T d=12K,T An =30K) -400 -200 0 200 400 Residuals Figure VIII.38 Decay profile of the 625 nm emission feature recorded at 12 K using TCSPC following pulsed laser excitation at 511.2 nm. The residuals present the difference between the triple exponential non-linear least squares fit completed and the decay recorded. Table VIII.17 Decay characteristics, components and amplitudes (A) extracted from nonlinear least squares analysis of the temporal profiles recorded monitoring λEm. = 625 nm at different temperatures, (Ts) following pulsed laser excitation at 511.2 nm. Note the dominant decay contribution is presented in bold. Ts. (K) Fit Range (msec) A1 τ 1 (µsec) A2 τ 2 (µsec) A3 τ 3 (µsec) 12.7 15.0 18.0 25.0 0.001 – 4.5 0.001 – 4.0 0.001 – 4.0 0.001 – 4.0 1982 711 909 1032 980 961 960 854 1632 641 852 827 307 290 324 270 956 369 470 564 59.7 39.7 63.5 40.9 The site-specific luminescence spectroscopy of the z8P5/2 ↔ a6S5/2 transition of atomic Mn isolated in solid Ar is summarised in Figure VIII.39. The photophysical characteristics and the transitions assigned to the observed emission features are presented in Table VIII.18. Chapter VIII; Mn(z8P and a6D)/RG Lum. 286 19.020.0 x103 E ner g y ( cm-1 ) Intensity 495 510 525 540 Wavelength (nm) λEm . =590.0nm(1 o) λEm . =625.0nm(2 o) Mn z8P5/ 2 ____ a6S5/2 Mn/Ar - Excitation 15.516.016.517.017.5 x103 E ner g y ( cm-1 ) 585 600 615 630 645 Wavelength (nm) λEx. = 528.8 nm λEx. = 508.8 nm Emission Mn a6D9/ 2 __ __ a6S5/2 Figure VIII.39 Emission spectra recorded at 12.5 K with site-selective pulsed dye laser excitation of the Mn z8P5/2 ← a6S5/2 transition (right). The excitation spectra (1° and 2° site) were recorded by monitoring emission at 590 (solid trace) and 625 nm (dotted trace), shown left. Sample deposition was completed at 12 K and subsequently annealed to 28 K. The dashed vertical lines show the spectral positions of the relevant gas phase transitions1 of atomic Mn. The 600 nm emission band results from excitation of the thermally unstable 3° site. The photophysical characteristics of the excitation and emission features are presented in Table VIII.15 and Table VIII.18 respectively. Table VIII.18 Photophysical characteristics and excited state assignments of the emission features produced following site-specific 3d54s4p z8P5/2 excitation. λEm, indicates the emission band-centre in nm and wavenumber units. The matrix shift for the transition is indicated δ in wavenumber (cm-1) units. The observed excited state lifetimes are also presented as τ Obs. at 12 K. Note additional decay times extracted of substantial amplitude are also presented. Mn Gas Phase Mn/Ar Matrix – Emission (1° Site) Transition1 nm / cm-1 Assignment λEm. (nm) / (cm-1) δ (cm-1) Decay Characteristic a6D9/2 ↔ a6S5/2 586.43 / 17052 a6D9/2 → a6S5/2 590 / 16949 -103 τ Obs. = 1.28 msec 512 µsec Mn/Ar Matrix – Emission (2° Site) z8P5/2 ↔ a6S5/2 543.40 / 18402 a6D9/2 ↔ a6S5/2 586.43 / 17052 (?) 625 / 16000 -2402 -1052 τ Obs. = 1.03 msec 360 µsec Chapter VIII; Mn(z8P and a6D)/RG Lum. 287 VIII.6 Discussion Mn(z8P)/RG luminescence In this section, the observed luminescence spectroscopy of the z8P5/2 ↔ a6S5/2 transition of atomic Mn isolated in solid Ar, Kr and Xe is summarised and some trends evident in the excitation and emission spectroscopy are presented. VIII.6.I Mn(z8P)/RG Excitation spectroscopy A summary of the excitation spectra recorded by monitoring the red emission features of Mn isolated in solid Ar, Kr and Xe in the region of the gas phase z8P5/2 ↔ a6S5/2 transition is presented in Figure VIII.40. In all matrices and for all the sites, the excitation spectra exhibit resolved threefold split excitation patterns. The photophysical characteristics of the sites of isolation identified from the excitation spectra are collected in Table VIII.15, Table VIII.13 and Table VIII.11 for Mn/Ar, Mn/Kr and Mn/Xe respectively. Of particular note in the excitation spectra is the occurance of threefold splitting patterns indicating the strong interaction of the z8P5/2 excited state with its matrix environment. This contrasts the excitation spectra recorded for the a6D ← a6S5/2 transition (Section VIII.3.I) in which the individual spin-orbit levels were identified reflecting the weak coupling of the a6D excited state to the local solid-state environment. The sites of isolation identified with z8P state are spectrally well separated like those of the z6P, exhibiting large matrix shifts (100’s of cm-1). It will be remembered that the excitation spectra of the a6D state all occurred within the same narrow spectral range. Both of these observations support the conclusion that the interaction of the Mn atom within the site is much stronger for the z8P state than the a6D state. In the next section the emission spectroscopy resulting from z8P excitation is discussed, highlighting the role of the site of isolation in determining excited state relaxation processes. Chapter VIII; Mn(z8P and a6D)/RG Lum. 288 18.418.819.219.620.020.420.8 x103 E ner g y ( cm-1 ) Intensity 480 490 500 510 520 530 540 Wavelength (nm) 590 nm 625 nm 18.418.819.219.620.020.420.8 x103 E ner g y ( cm-1 ) Intensity 480 490 500 510 520 530 540 Wavelength (nm) 565/626nm 588 nm 18.418.819.219.620.020.420.8 x103 E ner g y ( cm-1 ) Intensity 480 490 500 510 520 530 540 Wavelength (nm) 620 nm z8P5/2 ____ a6S5/2 2osite 1osite 1osite 2osite 1osite Mn(z8P)/RG - Excitation Ar λEm. Kr Xe Figure VIII.40 Mn/RG excitation spectra recorded at 12 K for all the Mn/RG systems investigated produced with laser excitation in the vicinity of the Mn z8P5/2 ← a6S5/2 transition. The spectra shown were recorded monitoring emission bands as indicated by λEm. (left) in wavelength units. All spectra were recorded following Mn/RG sample deposition at 12 K and matrix annealing. The dashed vertical line indicates the position of the z8P5/2 ← a6S5/2 gas phase transition1 of atomic Mn at 543.3 nm. VIII.6.II Mn(z8P)/RG Emission spectroscopy The emission spectroscopy presented in the previous sections with excitation of the z8P5/2 ← a6S5/2 transition is summarised in Figure VIII.41. The transitions assigned and photophysical properties of the observed emission features are collected in Table VIII.18, Table VIII.14 and Table VIII.11 for Mn/Ar, Mn/Kr and Mn/Xe respectively. The Mn/Kr system provided an additional emission feature (shown centre, Figure VIII.41) at 565.2 nm (17693 cm-1) with blue (1°) site excitation at 513.8 nm. This band represents the only emission feature observed which has been definitively assigned to the phosphorescence of the z8P5/2 excited state. This assignment was made on the basis of the excitation spectra recorded, the excited state lifetime measured and the spectral location of the emission. This band was observed earlier in Moskovits’ work11 on Mn/Kr in which a fixed frequency Ar ion laser at 514.8 nm Chapter VIII; Mn(z8P and a6D)/RG Lum. 289 was used as the excitation source. Based on the presence of dimer bands at 667 and 690 nm in the absorption spectra it was assumed in this earlier work that the 565 nm band was also related to Mn2. The data recorded for this band in the present study clearly indicates the atomic origin of the 565 nm emission in Mn/Kr. 16.017.018.019.020.0 x103 E ner g y ( cm-1 ) 480 510 540 570 600 630 Wavelength (nm) 590 625 16.017.018.019.020.0 x103 E ner g y ( cm-1 ) 480 510 540 570 600 630 Wavelength (nm) 527.8 508.8 16.017.018.019.020.0 x103 E ner g y ( cm-1 ) 480 510 540 570 600 630 Wavelength (nm) 565 / 626 588 16.017.018.019.020.0 x103 E ner g y ( cm-1 ) 480 510 540 570 600 630 Wavelength (nm) 513.8 531.8 16.017.018.019.020.0 x103 E ner g y ( cm-1 ) 480 510 540 570 600 630 Wavelength (nm) 620 16.017.018.019.020.0 x103 E ner g y ( cm-1 ) 480 510 540 570 600 630 Wavelength (nm) 527.0 z8P5/2 ____ a6S5/2 a6D9/2 ____ a6S5/2 Mn(z8P)/RG - Excitation Emission λEx. (nm)Ar λEm. (nm) Kr Xe Figure VIII.41 Mn/RG site-specific emission spectra (shown right) recorded at 12 K for all the Mn/RG systems investigated produced with laser excitation corresponding to the z8P5/2 ← a6S5/2 transition. The excitation wavelengths used are shown (centre) as λEx. (nm). The excitation spectra shown (left), were recorded by monitoring emission bands as indicated (left) by λEm. in wavelength units. All spectra were recorded following Mn/RG sample deposition at 12 K and matrix annealing. The dashed vertical lines show the spectral positions of the gas phase transitions1 of atomic Mn. Excited state lifetime measurements and the temperature dependence observed in the emission spectra recorded, have not allowed a definitive assignment of the 627 nm or the 625 nm emissions in Kr and Ar respectively. The possible state assignments of these emission features are discussed at the end of this Chapter. The spectroscopy recorded with specific 1° site excitation in Ar and Xe and 2° site excitation in solid Kr revealed the emission features at 590, 620 and 585.8 nm respectively, shown in Figure VIII.41. These features were assigned to the a6D9/2 → a6S5/2 transition of atomic Mn earlier in this chapter. These assignments are Chapter VIII; Mn(z8P and a6D)/RG Lum. 296 References 1 N.I.S.T. Atomic Spectra Database, Website: http://physics.nist.gov/cgibin/AtData/display.ksh?/XXE0qMnqIXXP-15XXT2XXS, (Last accessed 4th February 2004). 2 R. Schnabel, A. Bard and M. Kock, Zeitschrift für Physik D, 34, 223, 1995. 3 G. A. Martin, J. R. Fuhr and W. L. Wiese, Atomic Transition Probabilities Scandium through Manganese, J. Phys. Chem. Ref. Data, Vol. 17, No. 3, 1, 1988. 4 A. A. Radzig and B. M. Smirnov, Reference Data on Atoms, Molecules and Ions, Springer-Verlag, Berlin, 1985. 5 A. Corney, Atomic and Laser Spectroscopy, Clarendon Press and Oxford University Press, 1977. 6 W. C. Martin and W. L. Wiese, Atomic Spectroscopy, A Compendium of Basic Ideas, Notation and Formulas in Atomic, Molecular, and Optical Physics Handbook, National Institute of Standards and Technology, Gaithersburg, MD, Website: http://physics.nist.gov/Pubs/AtSpec/ (Last accessed 12th July 2004). 7 The narrow emission features centered at 600 and 700 nm (shown in Figure VIII.6) were produced with fixed wavelength (λEx. = 576.45 nm) pulsed dye laser excitation. These bands are assigned to sample contaminants (non atomic species) as they are not present in the emission spectra recorded using CCD detection following excitation at various wavelengths in the vicinity of the a6D ↔ a6S gas phase transition as shown in Figure VIII.3. 8 The excitation band-profile is due to the weak emission intensity of λEm. 589.4 nm, confirmed by the marked baseline to higher energy, due to the Coumarin 500 dye response. 9 The index of refractive used for solid Kr at 241 nm is 1.428 at 12 K, (P. Gürtler, unpublished results, 1996). 10 The excitation spectrum recorded by monitoring the 587 nm emission band and the emission spectra presented have been normalised. The additional emission bands evident in Figure VIII.34 at 603 and 626 nm are due to spectral overlap of the 1° and 3° sites with the 2° site, which produces the 587 nm emission. 11 A. D. Kirkwood, K. D. Bier, J. K. Thompson, T. L. Haslett, A. S. Huber and M. Moskovits, J. Phys. Chem., 95, 2644 (1991). Chapter IX; Mn/RG Site Occupancy 297 Chapter IX Sites of manganese atom isolation in RG solids, RG = Ar, Kr and Xe IX.1 Introduction The absorption and luminescence spectroscopy of atomic Mn isolated in solid Ar, Kr and Xe, reported in Chapters VI to VIII, identified multiple trapping sites and highlighted the importance of metal atom site occupancy in determining the excited state luminescence. In addition emission bands recorded for the resonance z6P5/2 ↔ a6S5/2, the ‘forbidden’ z8P5/2 ↔ a6S5/2 and the quadrupole a6D ↔ a6S5/2 transitions for the Mn/Ar, Mn/Kr and Mn/Xe systems yielded high-resolution excitation spectra that permitted the extraction of the photophysical properties of the trapping sites present but not resolved in absorption. Although, the spectral and temporal characteristics observed allowed the assignment of many of the features to the emission from specific atomic levels, a complete analysis has not been possible due to the presence of multiple intersystem crossing (ISC) and intermultiplet relaxation pathways (IMR) in each Mn/RG system. Therefore, knowledge of the site of isolation occupied would provide great insight into the interactions occurring between the matrix cage and the excited state Mn atom that allow the radiative and non-radiative processes identified. This Chapter collects the information extracted from the observed absorption and luminescence excitation spectroscopy and presents an analysis of these results to assign the sites occupied by Mn atoms in the RG solids. Overall, the two thermally stable blue and red sites of isolation of atomic Mn in solid Ar and Kr are assigned to substitutional site and multi-vacancy (tetra-vacancy) sites respectively. The single thermally stable site identified for Mn isolated in Xe is assigned to a substitutional site. The assignments completed are based on the application of the polarizability model of Laursen and Cartland1 (L&C) to the z6P5/2 ← a6S5/2; z8P5/2 ← a6S5/2 and y6P5/2 ← a6S5/2 electronic transitions of atomic Mn. This model allows the association of certain site types occupied by metal atoms in the rare gas solids from an analysis of the gas phase to RG matrix frequency shifts observed for P ← S type electronic transitions. The required condition being a linear correlation of the matrix shifts with rare gas polarizability for those metal atoms ‘trapped’ in a particular site type. As discussed in Chapter I, (Introduction), atomic Mn has a spherically Chapter IX; Mn/RG Site Occupancy 298 symmetric a6S ground state and therefore will favour isolation in spherical sites of the lattice. Trends, such as the preference for Mn occupancy in certain sites of isolation in the rare gases (Ar, Kr and Xe), as evident from the absorption spectroscopy, are used to assess the trapping environment. These trends, coupled with polarizability arguments form a major part of the sections which follow, as currently no information is available on the simpler Mn⋅RG 1:1 complexes from gas phase or ab initio studies. Instead, the similarities between atomic manganese, which exhibits an ns2 ground state electronic configuration and the M/RG systems (M = Zn2, Cd3, Hg4,5 and Mg6) where the solid-state and gas phase7 spectroscopy has been studied in detail are exploited. The spectroscopic parameters such as the ground state bond lengths for these known systems are used to tentatively assess the sites occupied by atomic Mn. The comparison between these M(P ← S)/RG (M = Zn, Cd and Hg) and the Mn/RG systems allows the extraction of finer details regarding the ground and excited state interactions of Mn with the host matrix. To achieve these goals this Chapter has the following structure. Firstly, the absorption results and luminescence excitation results obtained for the P ← S transitions presented in the previous Chapters are reviewed and details specific to site occupancy are highlighted. Secondly, the L&C polarizability model is applied to the excitation spectra recorded in the vicinity of the z6P5/2 ↔ a6S5/2 gas phase transition. Thirdly, as UV absorption spectra recorded for Mn/RG samples provided information on the multiple sites of isolation on the y6P5/2 ↔ a6S5/2 transition, excitation spectra recorded by monitoring the atomic emission features assigned in Chapter VII are presented for the first time. This therefore allows a comparison of the matrix shifts observed for the ‘singlet – like’ y6P5/2 ↔ a6S5/2 and ‘triplet – like’ z6P5/2 ↔ a6S5/2 transitions and an assessment of the extent of the Mn atom matrix interaction for these different excited states. The results of this comparison are directly comparable to the ns2 metal atom 1P1 and 3P1 ← 1S0 transitions discussed by L&C1. The analysis is extended by showing the relationship of the different ‘triplet – like’ excited state interactions that are manifest in the different matrix shifts observed for the z6P5/2 ← a6S5/2 and z8P5/2 ← a6S5/2 transitions. Overall the application of the polarizability model to the P ← S type transitions of atomic Mn in RG solids allowed the correlation of the high-energy blue sites in Ar and Kr with the single site in solid Xe and the subsequent assignment of this site to Chapter IX; Mn/RG Site Occupancy 299 Mn atoms in single substitutional sites. The analysis also allowed the grouping of the low energy red sites in Ar and Kr and their assignment to Mn atoms isolated in matrix tetra-vacancies. This was achieved numerically using a comparison of the Mg⋅RG ground state bond lengths assuming the transference of the known Mg⋅RG parameters to the Mn⋅RG systems. The site dominance showed that Mn atoms in solid Ar exhibit a preference for trapping in tetra-vacancy sites whereas single substitutional site occupancy is preferred in Kr, while this site is the single thermally stable site in solid Xe. IX.2 Site Analysis Mn(z6P ← a6S)/RG The UV/Vis absorption spectroscopy recorded near the gas phase z6P5/2 ↔ a6S5/2 transition8 of atomic Mn isolated in Ar, Kr and Xe (see Chapter VI, Figure VI.8) allowed the identification of multiple thermally stable features to the z6P5/2 transition in solid Kr only. Table VI.2 presented the spectral positions of the sites assigned as the blue (1°) and red (2°) sites. In solid Ar the occurrence of multiple site occupancy was indicated by the apparent loss of the linear correlation between the observed matrix shifts in absorption. Consideration of this effect predicted the existence of a secondary site of atomic isolation in solid Ar. The results of the Mn(z6P)/Ar excitation spectroscopy reported in Chapter VII, revealed the presence of the weak blue (2°) site. Therefore two thermally stable sites of isolation were identified in solid Ar and Kr but with intensity reversals. In solid Xe absorption spectra recorded following sample deposition at various temperatures and/or matrix annealing following deposition at 12 K allowed the identification of only a single thermally stable trapping site. The high resolution excitation spectra recorded by monitoring the atomic emission features produced with steady-state excitation of the z6P5/2 ← a6S5/2 transition for all Mn/RG systems investigated are presented in Figure IX.1. The excitation spectra shown, reveal the presence of threefold split excitation patterns for each of the sites identified in all the RG gas solids. The splitting observed is attributed to the Jahn Teller effect indicative of Mn atom occupancy in highly symmetric matrix environments. Chapter IX; Mn/RG Site Occupancy 300 24.025.026.027.0 x103 E ne r g y ( c m -1 ) Intensity 360 370 380 390 400 410 420 Wavelength (nm) Mn(z6P)/RG - Excitation Ar λEm. (nm) Kr Xe 413 438 625 λEm. (nm) 428 590 416 440 627 428 586 620 1osite 2osite 2osite 1osite 1osite z6P5/2 ____ a6S5/2 Figure IX.1 Excitation spectra recorded by monitoring the emission features as indicated reported in Chapter VII (at 12 K) resulting from z6P5/2 ← a6S5/2 excitation for the atomic Mn/Ar, Mn/Kr and Mn/Xe systems. The excitation spectra shown were recorded following Mn/RG sample deposition at Td= 12 K and matrix annealing. The dashed vertical line shows the spectral location of the gas phase transition. Because of its spatial symmetry, the a6S ground state of atomic manganese will favour isolation in spherical sites of isolation. Therefore, as discussed in the Chapter I, (Introduction) only spherically symmetric trapping sites within the RG fcc lattice are considered for atomic Mn isolation. The simplest matrix system, with respect to site occupancy, is Mn/Xe as only a single thermally stable site was identified in absorption and excitation spectra of annealed samples. This provides the starting point for this site analysis as Xe represents the ideal matrix host for atomic isolation due to the increased site sizes available as shown by the numerical analysis presented in Table I.1, Chapter I. The similarity between the energetics of the 1P and 3P ↔ 1S gas phase transitions of atomic Mg and those of the y6P and z6P ↔ a6S transitions of Mn is clearly evident from the diagram presented in Figure I.5. The Mg⋅RG (RG = Ar and Xe) diatomic ground state bond lengths, known from 1:1 complexes prepared in supersonic expansions7, are believed to be a good approximation to those of the Mn⋅RG systems. The Mg atom exhibits a 3s2 ground state electronic configuration Chapter IX; Mn/RG Site Occupancy 301 while that of Mn is 3d54s2. The presence of the half filled 3d5 shell and the small difference in energy to the 4s orbital makes the correlation with the 3s2 Mg ground state feasible. The known Mg⋅RG (RG = Ar and Xe) ground state bond lengths (re) are presented in Table IX.1. Comparison of the ground state bond lengths with the available site sizes reveals either isolation of Mg (and therefore Mn) in deformed substitutional sites or tetra-vacancy sites in solid Xe. The presence of a single thermally stable site of isolation in solid Xe shows the preference for the Mn atoms for a particular site type. However, an assignment is not possible based solely on application of the Mg⋅Xe ground state bond length to the Mn system, as the comparison reveals the possibility of multiple site occupancy. Table IX.1 Site sizes9 in angstrom units (Å) for specific spherically symmetric site types in the solid rare gases. The details of these sites was presented in Chapter I, Introduction. In addition, the polarizability of the solid rare gases and the known Mg(1S0)⋅RG diatomic ground state bond lengths are also presented. RG Solid ss (Å) TVac (Å) RG Polarization (Å3)10 Mg⋅RG, re (Å)7 Ar Kr Xe 3.756 3.991 4.335 4.404 4.679 5.083 1.640 2.485 4.050 4.49 4.56 Therefore, trends shown by Mn atoms isolated in solid Ar and Kr are used to strengthen any assignment of site occupancy. In solid Ar, the red site dominates the z6P state excitation spectra. However, the dominant site of isolation in solid Kr is as shown in Figure IX.1 the blue site. This difference represents a reversal of the dominance of a particular site type from Mn/Ar to Mn/Kr. To identify the trends in the site occupancy and the relationship of thermally stable sites of isolation in each RG solid, the polarizability model1 is employed. This is achieved by plotting the gas phase to matrix frequency shifts for the z6P5/2 ← a6S5/2 transition, calculated from the central threefold split component for each of the thermally stable sites in the solid RG’s (observed in the excitation spectroscopy and shown in Figure IX.1), against host RG polarizability data given in Table IX.1. The polarizability analysis is presented in Figure IX.2 and it is evident that a linear correlation exists between RG polarizability and the matrix shifts (δ, cm-1) observed for the high-energy sites identified in Ar and Kr and the single site Xe. An extrapolation of the red site Ar and Chapter IX; Mn/RG Site Occupancy 302 Kr data (triangles) in Figure IX.2 clearly does not include the single site present in the Xe system. 2 3 4 Pol ( Å3 ) 0 500 1000 1500 δ(cm-1) Blue Site Red Site 2 3 4 Pol ( Å3 ) 0 500 1000 1500 δ(cm-1) Ar Kr Xe Mn z6P5/2 ______ a6S5/2 Figure IX.2 A plot of the gas phase to Mn/RG matrix frequency shifts (δ cm-1) observed for the blue and red sites identified for the z6P5/2 ← a6S5/2 transition of atomic manganese versus the RG host polarizabilities. The squares (connected by the solid line) highlights the linear correlation between the frequency shifts and rare gas polarizability observed for the Mn z6P5/2 ← a6S5/2 transition occurring within the blue sites of isolation. The red site, which dominates the Mn/Ar solid-state spectroscopy, is correlated with the red but minor site identified in Kr. The site dominance is reversed from red to blue from Mn/Ar to Mn/Kr and there is a correlation between the blue sites identified in all three rare gas hosts. The red/blue site dominance is attributed to the preference for a different site type in the heavier RG solids. A comparison of the Mg⋅Ar ground state bond length (4.49 Å) with the site size available for the tetravacancy in solid Ar (4.404 Å) and the substitutional site (3.756 Å) reveals a favourable match with the former site. Therefore the red sites of atomic Mn isolation identified in solid Ar and Kr are assigned Mn atom trapping in tetra-vacancy sites. The preference for a single site in solid Xe, and the correlation of the matrix-shifts observed for the blue sites, allows the assignment of the blue sites to the trapping of Mn atoms in single substitutional sites of these matrices. In the following section the results of excitation spectroscopy recorded in the vicinity of the y6P5/2 ← a6S5/2 transition are presented for the first time and the Chapter IX; Mn/RG Site Occupancy 303 polarizability model is applied to these results and those achieved in Chapter VIII for the z8P5/2 ← a6S5/2 transition, to check the trends with respect to site occupancy evident for the z6P5/2 ← a6S5/2 transition. In addition some trends in the photophysical characteristics of the excitation bands observed for the P ← S type transitions related to the site occupancy are discussed. The ‘singlet’ vs. ‘triplet’ nature of the excited state transitions are discussed with respect to the excited state matrix interactions leading to the observed matrix shifts. IX.3 Site Analysis Mn(y6P and z8P ← a6S)/RG IX.3.I Mn y6P ← a6S Excitation spectroscopy The UV absorption features recorded for Mn/RG solids in the vicinity of the y6P5/2 ↔ a6S5/2 gas phase transition8 and presented in Chapter VI provided more direct information on the atomic trapping sites than the corresponding z6P5/2 ↔ a6S5/2 transition. This is in part due to the increased oscillator strength of the ‘singlet’ like y6P5/2 ← a6S5/2 over the ‘triplet’ like z6P5/2 transition but also the better separation on the former transition. This is clear upon inspection of the Mn/RG absorption spectra shown in Figure VI.7 where the y6P5/2 absorption bands dominate the spectra. In solid Xe, the ratio of the absorption intensity for the y6P5/2 and z6P5/2 ← a6S5/2 IAbs(y6P) : IAbs(z6P) was found to be 14:1 providing a measure of the relative oscillator strengths. The absorption bands assigned to the y6P5/2 state exhibited a red-shift of the band maximum from Ar to Kr to Xe which deviate from linearity, consistent with the reversal of red dominant / blue minor sites of isolation from Ar to Xe observed on the z6P5/2 ← a6S5/2 transition. In solid Ar, the two thermally stable y6P5/2 ← a6S5/2 absorption features were identified occurring to higher energy than the gas phase transition at 273 and 278.1 nm. These bands were assigned to the blue (2°) and red (1°) sites of isolation respectively. In Kr the 1o Mn/Kr absorption feature at 279.9 nm overlaps the gas phase transition and the 2° site occurred to lower energy. In solid Xe high temperature deposition and matrix annealing experiments allowed the definitive identification of the band located at 288.2 nm to a single site. Tables VI.I to VI.III present the details of the transition energies for the sites of isolation identified on the y6P5/2 ← a6S5/2 transition of atomic Mn occurring in solid Ar, Kr and Xe respectively. Chapter IX; Mn/RG Site Occupancy 304 High-resolution excitation spectra recorded in the vicinity of the y6P5/2 ← a6S5/2 transition monitoring the site-specific atomic emission features produced with z6P5/2 excitation in the Chapter VII allowed the identification of the sites of isolation. Figure IX.3 presents the UV excitation spectra recorded. These all show resolved threefold split patterns indicative of Mn occupancy in high symmetry matrix sites for all the Mn/RG systems. The photophysical characteristics of the 1° and 2° sites extracted from the excitation spectra shown in Figure IX.3 are presented in Table IX.2. 34.435.236.036.837.6 x103 E ner g y ( cm-1 ) Intensity 270 275 280 285 290 295 Wavelength (nm) y6P5/2 ____ a6S5/2 2osite 1osite 1osite 2osite 1osite Mn(y6P)/RG - Excitation Ar λEm. (nm) λEm. (nm) Kr Xe 413 438 428 590 416 440 626 428 620 Figure IX.3 Excitation spectra recorded in the vicinity of the y6P5/2 ↔ a6S5/2 gas phase transition by monitoring the emission bands reported in Chapter’s VII and VIII (at 12 K) to result from z6P; z8P and a6D excitation for each of the Mn/RG systems investigated. Mn/RG sample deposition was completed at 12 K and matrix annealing. The dashed vertical lines show the spectral position of the gas phase transition. The photophysical properties of the sites of isolation identified are presented in Table IX.2. As observed in the absorption spectroscopy, the y6P5/2 ← a6S5/2 transition occurring for Mn atoms isolated in the 1° site of solid Kr overlaps the gas phase transition (indicated by the vertical line in Figure IX.3), while the Mn/Ar and Mn/Xe the bands occur at higher and lower energies respectively. It is noteworthy that the matrix Chapter IX; Mn/RG Site Occupancy 305 excitation spectra of the z6P5/2 state are blue of the gas phase transition for all three hosts. This observation is attributed to the spin ‘singlet’ characteristic of the y6P5/2 state and ‘triplet’ nature of the z6P5/2 state. Comparison of the y6P5/2 state excitation spectra presented in Figure IX.3 reveals the same site specificity as observed with excitation of the z6P5/2 state shown in Figure IX.2. Table IX.2 Photophysical characteristics of the sites of isolation (1° and 2°) revealed in the excitation spectra of the 3d54s4p y6P5/2 ↔ 3d54s2 a6S5/2 transition of atomic manganese. Where possible the spectral position and average linewidth (full width at half maximum, fwhm) is denoted as ∆AV of the three components identified in Gaussian lineshape analyses for the threefold split excitation spectra are presented in wavenumber units. Gas phase to matrix frequency shifts are presented for the atomic Mn y6P5/2 ← a6S5/2 transition8 (G.P.: 35726 cm-1), δ in wavenumber units. Note the frequency shifts are calculated with respect to the central feature of the observed threefold pattern. Mn/RG Site Component E (cm-1) ∆AV (cm-1) δ (cm-1) Argon Red (1°) Blue (2°) 1 2 3 1 2 3 36138 35978 35778 36879 36676 36438 ≈ 215 ≈ 265 +252 +950 Krypton Blue (1°) Red (2°) 1 2 3 1 2 3 36030 35851 35656 - 35386 35231 ≈ 200 ≈ 115* +125 -340 Xenon (1°) 1 2 3 34841 34692 34540 ≈ 206 -1034 IX.3.II Mn z8P ← a6S Excitation spectroscopy A summary of the site-specific excitation spectra recorded by monitoring the red emission features of Mn isolated in solid Ar, Kr and Xe in the region of the gas phase z8P5/2 ↔ a6S5/2 transition is presented in Figure IX.4. The excitation spectra presented were recorded monitoring atomic emission bands produced with direct dye laser excitation. It is evident in Figure IX.4 that all the sites exhibit excitation spectra with resolved threefold split excitation patterns in all matrices. Chapter VIII presents the specific details of each system and comments on the effect of the intensity Chapter IX; Mn/RG Site Occupancy 312 References 1 S. L. Laursen and H. E. Cartland, J. Chem. Phys., 95, 4751, (1991). 2 V. A. Braken, P. Gürtler and J. G. McCaffrey, J. Chem. Phys., 107, 5290 (1997). 3 B. Healy and J. G. McCaffrey, J. Chem. Phys., 110, 3903 (1999). 4 C. Crepin and A. Tramer, J. Chem. Phys., 97, 4772 (1992). 5 M. A. Collier and J. G. McCaffrey, J. Chem. Phys., 119, 11878, (2003). 6 J. G. McCaffrey and G. A. Ozin, J. Chem. Phys., 101, 10354 (1994). 7 W. H. Breckenridge, C. Jouvet and B. Soep, Advances in Metal and Semiconductor Clusters, edited by M. A. Duncan (JAI, Greenwich, 1995), Vol. III. 8 N.I.S.T. Atomic Spectra Database, Website: http://physics.nist.gov/cgibin/AtData/display.ksh?/XXE0qMnqIXXP-15XXT2XXS, (Last accessed 4th February 2004). 9 H. E. Hallam, Vibrational Spectroscopy of Trapped Species, Wiley – Interscience, New York, 1973. 10 T. M. Miller, Handbook of Chemistry and Physics, edited by D. R. Lide, 73rd ed. (CRC, Boca Raton, 1991-1993). Chapter X; Conclusion 313 Chapter X Conclusions IX.1 Hg/RG Hg/RG absorption and excitation spectra recorded in the vicinity of the gas phase 6s16p1 3P1 ← 6s2 1S0 transition revealed bands at 245.9, 248.9 and 253.6 nm in Ar, Kr and Xe respectively. Spectra were obtained with a deuterium lamp following matrix deposition at 22, 25 and 35 K for Ar, Kr and Xe. The progressive red-shift of the excitation bands from Ar to Xe was accompanied with decreasing linewidths but with better resolved threefold splitting. Excitation into the recorded absorption bands produced emission features centered at 250.3, 254.1 and 273 nm in Ar, Kr and Xe respectively. These features all showed nanosecond emission lifetimes and from temperature dependent studies of the recorded decay curves, are identified as the matrix radiative lifetimes of the Hg atom 3P1 → 1S0 transition. The emission bands exhibited an increased Stokes’ shift and linewidth from Ar to Xe - results in good agreement with those reported earlier by Crepin and Tramer. At higher temperatures the linewidth of the Hg(3P1 → 1S0)/Xe band increases but the band maximum blue shifts – an effect which is completely reversible. Gaussian lineshape analysis showed three components are required to reproduce the emission bands in Ar, Kr and Xe. They also revealed that the reduced intensity of central, 273 nm component is the origin of the blue shift observed in the Mn/Xe system at high temperature. The multicomponent nature of the 3P1 state emission is shown from excitation spectroscopy not to arise from solid-state effects such as multiple site trapping. Its origin is examined with a pair-potential method in which the energetics of excited state vibronic modes are calculated for Hg(3P1)/RG18 clusters. The multi-component emission bands were identified as arising from the stabilisation of several ‘waist’ and ‘body’ type vibronic interactions occurring for the excited 3P1 state Hg atom isolated at single substitutional sites in each rare gas host. Time-gated emission spectra recorded following Hg 3P1 ← 1S0 excitation revealed the presence of weak, narrow linewidth features at 258.9, 260.8 and 265.1 nm respectively in Ar, Kr and Xe. The millisecond decay times measured for these emission bands allowed their assignment to the forbidden 3P0 → 1S0 transition of Chapter X; Conclusion 314 atomic mercury. The maximum contribution of this band to the time integrated emission intensity is found in Xe where it is only 0.5%, indicating the inefficiency of intramultiplet 3P1 → 3P0 relaxation compared with radiative decay of the 3P1 excited state. The efficiency of the intramultiplet relaxation increases only very slightly at higher temperatures. The presence of resolved fine structure on the Hg 3P0 → 1S0 emission in Xe (and partly in Kr) matrices and the temperature dependence exhibited, allowed the identification of a resolved zero phonon line and a phonon side band. This assignment is confirmed in the lineshape simulation conducted with the Wp function yielding small S values (1.3 and 2.2 in Xe and Kr respectively), which represent weak electron-phonon coupling. The close match between the excited Hg⋅RG(3P0) ã30-(3Σ) and ground X Hg⋅RG(1S0) 10+(1Σ) state bond lengths is identified as the origin of the very weak electron-phonon coupling in the Hg(3P0)/RG matrix system. This represents the first observation of a resolved zero-phonon line for an electronic transition of a matrix isolated metal-atom. From the close agreement found with observed absorption energies, simulations based on the localised, pair-potentials approach developed by the Maynooth Group to Hg(3P1 ↔ 1S0)/RG18 indicated that atomic mercury occupies essentially undistorted substitutional sites in solid Ar, Kr and Xe. The Hg⋅RG18 calculations predict the spectral location of the pure 3P1 ← 1S0 electronic transition to occur red of the observed band maxima but within the observed band profile in all cases. These calculations also succeeded in predicting multiple emission energies for Hg/Ar and Hg/Kr revealing that three vibronic modes lead to emission in these matrices. Of the stabilised modes, the 6-atom ‘waist’ mode, Q5, is expected to dominate the low temperature spectra as it has the steepest stabilization gradient. In Hg/Ar matrices, this mode predicts emission in good agreement with the observed bands, but in Hg/Kr, it is slightly to the red of the observed band. Excited state calculations showed the importance of the site of isolation in determining the solidstate luminescence, as Stokes’ shifted emission is only predicted for ‘waist’ vibronic modes. In the waist mode the RG atoms move with respect to the fixed Hg atom at the centre of a substitutional site. The ‘body’ modes, involving internal motion of the excited state Hg atom from the substitutional site, do not lead to an excited state minimum in Ar and Kr. However, for Hg/Xe calculations indicate that emission can arise from both ‘body’ and ‘waist’ modes to give a total of six stabilised modes. The Chapter X; Conclusion 315 2–atom ‘body’ Q6(py) mode leads to emission which most closely matches the observed band centre at 273 nm. This mode involves motion of the Hg(py) atom to one of the 12 nearest neighbour Xe atoms and corresponds to excimer type behaviour that Crepin and Tramer proposed was the origin of the emission in Hg/Xe. The solidstate calculations show, however, that to achieve stabilization, the excimer type interaction is specific to one orbital orientation. Tetragonal calculations, namely the 4–atom ‘waist’ (Q3) and the 4–atom ‘body’ modes (Q2), predict emission at 377.73 and 358.74 nm respectively. However, no Hg/Xe emission features have been observed in the 350-400 nm spectral region. A mechanism by which these modes are quenched was identified in the calculations, involving the crossing of these strongly bound excited state vibronic states by the repulsive ground state potential. This crossing does not occur in the Hg/Ar and Hg/Kr systems. Hg(3P1)⋅Ne18 pair-potential simulations showed good agreement between the calculated and observed absorption energies for atomic Hg occupying distorted (relaxed) single substitutional sites in solid neon. This accounts for the observation that the Hg 6p 3P1 ← 6s 1S0 transition occurs to lower energy than that predicted by an extrapolation of the polarizability model. A radial expansion of 0.293 Å of the 1st co-ordination sphere surrounding the Hg atom occurs to allow the isolation of the ground state Hg atom in a single substitutional site. The metal atom induced site deformation, calculated with the ground state ‘breathing’ mode (Q1) for the Hg(1S0)⋅Ne18 cluster, represents a 9.29% expansion of the nearest neighbour distance. The observed emission spectroscopy can only be predicted with simulations based on relaxed substitutional site occupancy for Hg in neon as no excited state stabilisation giving rise to substantially Stokes’ shifted emission is found for rigid site occupancy. It is observed that the excited state lattice reorganisation is dependent on I) the lattice energy and II) the stabilisation of the metal atom excited state. Solid neon provided an ideal system to probe the effect of ground and excited state solvation as the cramped lattice site allowed an investigation of the lattice contribution to the overall excited state cluster stabilisation. In emission the best agreement is achieved using the 6-atom ‘waist + limited stretch’ mode, (Q8*) for the pz orbital orientation. The stabilization arises from the ‘in-phase’ contraction of the six close packed Ne atoms towards the metal atom, a motion which allows half the Ne⋅Ne interactions to regain their equilibrium lattice positions. Therefore, the simulations highlight the Chapter X; Conclusion 316 importance of restoring the equilibrium lattice in producing the observed luminescence for metal atoms isolated in cramped sites. The role of the metal atom is revealed as Q8* is stabilised for the pz orbital orientation, thus maximising the attractive pure-Π Hg⋅Ne interaction. The 6-atom ‘stretch’ contributes to the overall energy of the Hg 3P1 excited state by reducing the Σ interaction. Overall, ‘waist’ type vibronic modes are more important than ‘body’ modes in producing excited state stabilisation of metal atoms located in cramped lattice sites, as internal motion of the metal atom is not feasible under these circumstances. The pair-potential approach was also used to simulate the recorded Hg(3P0 → 1S0)/RG spectroscopy. The simulations predict excited state absorption and emission energies for an excited state ‘breathing’ mode (Q1*), with small Stokes’ shifts consistent with the observation of zero-phonon lines for the transition in Ar, Kr and Xe. The calculated emission energies do not show good agreement with the observed matrix emission bands. A comparison reveals the emission energies calculated are progressively red-shifted of the observed band maxima. The red shift discrepancy is attributed to a breakdown in the pure Hund’s Case–c coupling of the Hg (3P0) state in the solid. In addition, the Hg(3P0)⋅RG diatomic potentials are not available experimentally for any of the Hg⋅RG pairs except Ar. As such the [3P0 (0) ã] potentials used were calculated from the known [A 3Π (30+)] and [B 3Σ (31)] states of the Hg(3P1)⋅RG diatomics. Overall the pair-potential calculations conducted during the course of this work demonstrate that given accurate Metal (M)⋅Rare Gas (RG) interaction potentials, available from spectroscopic studies or ab initio calculations, the M⋅RG18 localised cluster model succeeds in identifying the sites of metal atom isolation and provides insight into the ground and excited state interactions which lead to the observed solid-state luminescence. IX.2 Mn/RG Absorption spectra were recorded in the vicinity of the gas phase 3d54s4p y6P ↔ 3d54s2 a6S; 3d54s4p z6P ↔ 3d54s2 6S and 3d64s1 x6P ↔ 3d54s2 6S transitions of atomic manganese at 279.91, 403.42 and 221.45 nm respectively, with deuterium and tungsten lamps. The simplicity of the absorption spectra reported and the spectral location of the bands observed allowed easy assignment of the matrix equivalent of Chapter X; Conclusion 317 these atomic P ← S type transitions. The Mn/RG samples prepared during the course of this work were more ‘atomic’ than those reported in previous studies. This was due to the increased control of the Mn vaporisation afforded by localised heating of Mn via electron bombardment compared to the bulk heating of Knudsen cells used in previous works. As a result, the samples prepared under low Mn concentrations showed only small amounts of Mn aggregate absorptions and the spectra recorded in this study allowed easy distinction of Mn atom and Mn2 transitions. Multiple thermally stable sites of atomic isolation were identified in solid Ar and Kr whereas single site occupancy is indicated in solid Xe. The Mn(y6P5/2 ← a6S5/2)/RG transitions were observed to occur at 273 / 278.1; 279.3 / 284.9 and 288.2 nm in Ar, Kr and Xe respectively. A progressive red shift was observed from Ar to Xe with respect to the gas phase y6P ↔ a6S transition energy for the Mn atom. The red site at 278.1 nm dominated only in Mn/Ar. The features assigned to the z6P5/2 ← a6S5/2 transition occur at 397.4, 385.5 / 401.9 and 395.5 nm in Ar, Kr and Xe respectively all of which are blue of the gas phase transition. A progressive red-shift from Ar to Xe is not observed for the z6P5/2 ← a6S5/2 transition as the absorption band in Mn/Kr appears to higher energy than the Mn/Ar band. The presence of a secondary site (corresponding to the 273 nm band on the y6P), blue-shifted from the assigned 397.4 nm absorption in Ar, (identified in excitation spectra) rationalized the behaviour exhibited by the matrix shifts for the z6P5/2 ← a6S5/2 absorption. This allowed a linear correlation between the gas phase-matrix shift and host matrix polarizability. The x6P5/2 ← a6S5/2 transition was also identified in absorption spectra and occured blue-shifted of the gas phase transition energy by approximately 2000 cm-1 for all Mn/RG systems investigated. The luminescence spectroscopy observed for the z6P excited state of matrixisolated atomic Mn confirmed the presence of multiple sites of isolation of Mn atoms in Ar and Kr and a single site in solid Xe. The excitation-emission spectra recorded for each Mn/RG system allowed the deconvolution of badly overlapped absorption bands into separate site contributions. The steady-state and time-resolved emission spectroscopy (TRES) allowed the identification of two very site-specific relaxation channels in the Mn/Ar and Mn/Kr systems giving rise to z6P → a6S fluorescence and a4D → a6S phosphorescence from the blue and red sites respectively. The a4D state emission occurs by z6P ⇒ a4D intersystem crossing a process which is 100% Chapter X; Conclusion 318 efficient. Excited state lifetime measurements definitively assigned z6P → a6S fluorescence at 413 and 416 nm in Ar and Kr respectively. In addition, the matrix lifetime for the forbidden a4D → a6S transition is reported for the first time to be approximately 25 msec in both matrices. Currently no experimental gas phase lifetime data is available for this transition which is both electric-dipole and electricquadrupole forbidden. The results of lineshape analyses of the very narrow emission features assigned to the a4D → a6S transition allowed the assignment of zero phonon lines and phonon sidebands with distinct lifetimes of 25 and 10 msec. Also the requirement of two sets of phonon frequencies allowed the identification of a small (10 cm-1) crystal field splitting parameter for the a4D state of Mn atoms isolated in the red sites in Ar and Kr. In solid Xe, no emission features were observed in the 400 – 600 nm spectral range so neither the z6P → a6S nor the a4D → a6S transitions of atomic Mn occur in this host. High-resolution dye laser excitation spectra were recorded in the region of the gas phase a6D5/2 ↔ a6S5/2 transition by monitoring Mn emission features present in the red spectral region for solid Ar, Kr and Xe with z6P excitation but not definitively assigned. The excitation spectra revealed that the gas phase spin-orbit splittings were maintained in all three solids. The temperature dependence exhibited allowed the assignment of zero phonon lines for the excited state J = 1/2, 3/2, 5/2 and 7/2 levels of the a6DJ ← a6S5/2 electric-quadrupole transition in Mn/Kr. This represents the first report of zero phonon lines observed in excitation for metal atoms isolated in RG solids. Their occurrence is attributed to weak coupling of the a6D excited state of Mn to the solid-state environment provided by the site of isolation. In addition, an enhancement of the transition was observed to occur for Mn atoms in the larger, red sites of isolation. However, different excitation lineshapes were observed monitoring Mn atoms located in high symmetry sites in Ar and Kr but the excitation spectra recorded by monitoring site specific emission features showed no relative shifts. This behaviour, different to P ← S transitions, occurs due to the weak coupling of the a6D state with different environments within the same matrix host. The emission spectroscopy recorded following a6D state excitation allowed the definitive assignment of emission to the relaxation of the a6D9/2 state in the Mn/Kr system. A comparison of the matrix radiative lifetime (227 µsec) to the calculated gas phase lifetime for the a6D ↔ a6S transition (3.4 sec) shows the extent of the enhancement Chapter X; Conclusion 319 of this forbidden transition in the matrix environment. Selective excitation of the individual levels assigned to the a6D state in Kr resulted in the production of only the a6D9/2 emission due to efficient intermultiplet relaxation. The a6D9/2 → a6S5/2 transition occurs with direct a6D red site excitation in Ar at 590 nm. In this case the electronic transition occurs to lower energy than in solid Kr with a substantially longer lifetime (1.09 msec). These observations are attributed to the smaller site size available in solid Ar. In solid Ar and Kr, no rise times are observed in the decay profiles for the a6D emission bands suggesting direct feeding processes lead to the observed emission bands. Blue site a6D excitation also produces direct phosphorescence emission features at 625, 626.7 and 620 nm in Ar, Kr and Xe matrices. The emission bandbands are Gaussian in Ar and Kr and asymmetric in Xe, but linewidths of approximately 260 cm-1 are observed in all cases. These bands were assigned to the a6D9/2 → a6S5/2 transitions from the long lifetimes recorded, (1 msec) and a comparison of the gas phase splittings between the z8P/a6D pair that is maintained in emission for Mn atoms isolated in cramped matrix environments. The lifetimes recorded (1 msec) are substantially longer than the radiative lifetime extracted for the a6D state emission at 586 nm in solid Kr following red site excitation. The long decay times and the substantial Stokes’ shifts observed for these bands arise due to size and symmetry reasons in the cramped substitutional site which for does not allow the enhancement of the transition like the red tetravacancy site. The excitation bands recorded for the forbidden z8P5/2 ← a6S5/2 transition of Mn atoms isolated in solid Ar, Kr and Xe showed well resolved threefold splittings, behaviour evident in the Mn z6P state absorption and excitation spectra. Moreover, the z8P5/2 ← a6S5/2 transition reveals multiple Mn atom trapping sites consistent with the z6P and y6P states. Emission spectra recorded with z8P5/2 ← a6S5/2 excitation allowed the identification of z8P5/2 state emission at 565 nm in solid Kr only. Moskovits and co-workers previously observed this emission feature with fixed wavelength excitation at 514 nm using an Ar+ laser but assigned it to Mn2. The excitation spectroscopy and the lifetime measurements recorded here indicate, however, that the 565 nm band is emission of the z8P excited state of atomic manganese. As observed with z6P excitation, the smaller, blue site in Mn/Kr leads to direct z8P state emission whereas the larger, red site leads to the a6D emission via an z8P ⇒ a6D ISC process of 100% efficiency. Overall, the luminescence and excited Chapter X; Conclusion 320 state lifetime measurements reported allowed the definitive assignment of emission from all the excited states of Mn, which occur below the z6P state. Mn atoms are isolated in single substitutional and tetra-vacancy sites in solid Ar, Kr and Xe, with a preference for the larger site increasing from the heaviest rare gas host, Xe to the lightest, Ar. The application of the polarizability model to the y6P ← a6S, z6P ← a6S and z8P ← a6S transitions of atomic Mn showed the same overall trends as the ns2 metal atom systems investigated by Laursen and Cartland insofar as the behaviour evident on the y6P ← a6S and z6P ← a6S matrix shifts mirrored those reported for the 1P ← 1S and 3P ← 1S transitions of Zn, Cd and Hg. This highlighted the differences in the relative contributions from the Π and Σ M⋅RG excited state interactions in the Frank Condon accessible regions of the excited states. The importance of the excited state spin multiplicity on the observed matrix shifts was evident in comparison of the ‘singlet’ like y6P ← a6S and ‘triplet’ z6P ← a6S and z8P ← a6S transitions. In addition, the analysis of the z6P and z8P excited state energies reflected the spin triplet nature of these two states. IX.3 Summary The luminescence spectroscopy reported here for Hg and Mn isolated in rare gas matrices have shown that the solid state environment provides an ideal environment to study the interactions of the ground and excited state metal atoms. It allows the extraction of information on long-lived electronic transitions (> 100 µsec) which cannot be observed in gas phase experiments. The solid state simulations have shown that given accurate potentials, the interactions between the metal atom and the solid RG environment can provide information on the vibronic modes leading to the observed luminescence. The results of the Mn/RG experimental work have shown that the site of isolation critically governs the excited state guest/host interactions. This observation therefore would allow the extension of this work to investigate site selective excited state reactions with reagents such as CH4, CH3F, NH3 and H2 doped RG matrices. In addition studies of Mn/G matrices (G = CH4, CF4 and N2) are suggested. However, the most pertinent information to allow a complete analysis of the Mn/RG luminescence reported would be the simulation of the solid state using diatomic pair–potentials which are unavailable at present for the Mn⋅RG 1:1 van der Waals complexes from either spectroscopic studies or ab initio calculations.