MSU LIBRARIES m \— RETURNING MATERIALS: Place in book drop to remove this checkout from your record. FINES wi11 be charged if book is returned after the date stamped below. THE GAS PHASE CHEMISTRY OF TRANSITION METAL-CONTAINING ANIONS WITH ORGANIC MOLECULES By Stephen Ward McElvany A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1985 ABSTRACT THE GAS PHASE CHEMISTRY OF TRANSITION METAL-CONTAINING ANIONS WITH ORGANIC MOLECULES By Stephen Ward McElvany The gas-phase chemistry of metal-containing anions with l—chloro-n-alkanes, 1-hydroxy-n—alkanes, l-bromo-n-chloro-alkanes, l-chloro-n-alcohols (n=1 to 6), l-nitro-n-alkanes (n = 1 to 4), and n-butyl nitrite is presented. The metal-containing anions studied (Fe(CO)3,4", Cr(CO)3-5', Co(CO)2,3', and CoNO(CO)1,2") are formed by low energy electron impact on the corresponding metal carbonyls, M(CO)n. The results suggest a common mechanism in which the metal inserts into the C-functional group bond and the charge is transferred (delocalized) to the electronegative group (e.g., Cl, OH, and N02). Metal insertion into C-C and 0-H bonds which is seen in the corresponding positive metal ion reactions is not observed in the reactions of metal-containing anions. The metal anions also attack bonds within the functional group in reactions with n-alcohols (0-H bond) and n-nitroalkanes (N-O bond). Following formation of the metal insertion/charge transfer intermediate, rearrangement of the ion may occur. Evidence is presented for a 8-H shift process in reactions of metal-containing anions, which is also common in the corresponding positive metal ion reactions. The products from the reaction Stephen McElvany of nitroalkanes are shown to result from rearrangements of the metal insertion/charge transfer intermediate. The excess energy which remains in these complexes is then lost through a competitive ligand loss process. This ligand loss process is quite different from the analogous process for positive metal ions. The exothermicity of the charge transfer process appears to determine the products which are observed. The reaction trends of the various metal-containing anions suggests the following order of electron affinities of the metal-containing species: E.A.(Cr(CO)3) < E.A.(Co(CO)2) < E.A.(CoCONO) < E.A.(Co(CO)3) < E.A.(Fe(CO)3). Several ligand effects observed in the reactions of metal-containing anions are discussed. These include the decrease in reactivity as the number of ligands present on the metal increases and the possible participation of carbonyl ligands in the metal insertion and rearrangement processes. ACKNOWLEDGMENTS I would like to thank John Allison for his help and guidance during ‘my research at M.S.U. I would like to give special thanks to my parents for their help and guidance throughout my educational career. I would like to thank Dr. Watson for serving as my second reader and the M.S.U./N.I.H. Mass Spectrometry Facility for financial support. I would like to thank everyone in the Allison group (especially the ICR group (Tony and Barb)) and all the friends I have met at M.S.U. for making my stay here enjoyable. Finally, I would like to thank all those who made MAC tours a success, including the participants (D.B.), Wherehouse Records, and the east coast connection (Bob and Parge). ii TABLE OF CONTENTS LIST OF TABLES ................................................................................. LIST OF FIGURES ............................................................................... LIST OF SCHEMES .............................................................................. CHAPTER 1 - INTRODUCTION ........................................................... CHAPTER 2 - EXPERIMENTAL I .............. . ............................................ A. The Ion Cyclotron Resonance Technique .................................. 1. The ICR Cell .................................................................. 2. Detection of Ions ............................................................ 3. Ion Cyclotron Double Resonance ............... ................ B. Experimental Parameters for Negative Ion Studies ................... ‘ CHAPTER 3 - REACTIONS OF N-CHLOROALKANES ........................... A. Parent Substitution ................................................................ B. C1 Abstraction ....................................................................... C. HCl Abstraction .................................................................... D. Reaction Trends and Mechanisms .. .......................................... CHAPTER 4 - REACTIONS OF N-ALCOHOLS ...................................... A. Reactions Resulting From Metal Insertion into the C-OH Bond .................. . ............................................. B. Reactions Resulting From Metal Insertion into the 0-H Bond .................................................................. CHAPTER 5 - REACTIONS OF BIFUNCTIONAL ORGANIC MOLECULES ................................................. A. Reactions of l,n-Bromochloroalkanes ...................................... B. Reactions of l,n—Chloroalcohols .............................................. iii vi vii ll 13 15 23 23 28 28 30 35 35 39 45 45 59 CHAPTER 6 - REACTIONS OF NITROALKANES .................................. 67 A. Reactions of n-Nitroalkanes ................................................... 68 B Reactions of 2-Methyl-2-Nitropropane .................................... 84 C. Reactions of n-Butyl Nitrite ................................................... 85 D Further Insights into the Metal Insertion/Charge Transfer Mechanism .............................................................. 91 CHAPTER 7 - CONCLUSIONS ............................................................. 94 APPENDIX A - PERTINENT ELECTRON AFFINITIES ........................... 98 APPENDIX B - GAS-PHASE REACTIONS OF Co+ AND Co(ligand)n+ WITH NITROALKANES ...................... 100 APPENDIX C - ICR STUDIES OF TRIMETHYLALUMINUM .................... 111 APPENDIX D - UNSUCCESSFUL ICR STUDIES ..................................... 132 LIST OF FOOTNOTES AND REFERENCES ........................................... 135 iv LIST OF TABLES Table 1 Metal-Containing Anions Formed by Low Energy Electron Impact .......... .................. ................ .. 2 Reactions of Chromium, Iron, and Cobalt-Containing Anions with n-Chloroalkanes ..... . ........... . ....... 3 Reactions of Chromium, Iron, and Cobalt-Containing Anions With D'AICOhOIS 000 0000000000 00000000000000000 000000 0 0000000000000000000000000000 4 Reactions of Iron and Chromium-Containing Anions with 1,n-Bromochloroa1kanes ..... ...... ............................ 5 Reactions of Cobalt-Containing Anions with l’n-BromOChloroalkanes COO. OOOOOOOOOOOOOO .00...0.0.0..00.000000000000000000000000....O 6 The Average Branching Ratios of Cl and Br-Containing Products and HCl and HBr Abstraction Products for the 1,n-Bromochloroalkanes (n = 2 to 6) ...... ............................. 7 Reactions of Iron and Cobalt-Containing Anions with 1,n-Chloroalcohols .. ............... . .............................................. 8 Reactions of Chromium-Containing Anions . with 1,n-Chloroalcohols ............... . ....................... . ........................ 9 Reactions of Iron-Containing Anions with n—Nitroalkanes and Z'MCIDYI'Z‘NIII'OPI'OPHHC 000000000000000000000000000000000000000000 0000000 0 00000 0 10 Reactions of Cobalt-Containing Anions with n-Nitroalkanes and 2-Methyl-2—Nitropr0pane .......... . ..... . ............. ........ ....... ll Reactions of Chromium-Containing Anions with n-Nitroalkanes and 2-Methyl-2-Nitropropane ..... . ............ . ........... .......... ...... 12 Reactions of Iron, Cobalt, and Chromium-Containing Anions with n-Butyl Nitrite (C4H90NO) .............. . ........... ...... ............ 13 Pertinent Electron Affinities (E.A.) (kcal/mole) .............................. Page 18 24 36 46 48 52 60 62 69 70 71 86 98 LIST OF FIGURES Figure Page 1 Three-region ICR cell ............. . ...................................................... 8 2 Vector contributions resulting in the drift motion .......................... 10 3 Simplified circuit diagram of the marginal oscillator detector ....................................................................... 11 4 Block diagram of the ICR spectrometer ......................................... 15 5 Energy dependence of negative ions formed from Fe(CO)5 .............. 17 6 Low energy («:1 eV) electron impact negative ion mass spectrum of Fe(CO)5 .............................. . ..................................... 20 7 Low energy («4.5 eV) electron impact negative ion mass spectrum of Cr(CO)5 .................................................................... 21 8 Low energy («4.5 eV) electron impact negative ion mass spectrum of Co(CO)3NO ............................................................... 22 LIST OF SCHEMES Scheme I ............................................................................................ Scheme II ........................................................................................... Scheme III .......................................................................................... Scheme IV .......................................................................................... Scheme V ........................................................................................... Scheme VI .......................................................................................... Scheme VII ......................................................................................... Scheme VIII ........................................................................................ Scheme IX .......................................................................................... Scheme X .............................. Scheme XI .......................................................................................... Scheme XII ......................................................................................... Scheme XIII ................................................................................... Scheme XIV ........................................................................................ Scheme XV ......................................................................................... Scheme XVI ........................................................................................ vii Page 31 32 37 41 43 51 56 58 64 65 75 76 77 78 89 90 CHAPTER 1 INTRODUCTION In recent years there has been a growing interest in the study of the gas-phase reactions of atomic metal and metal-containing positive ions with organic molecules. Studies of these reactions using ion cyclotron resonance (ICR)1'4 and ion beam techniquess’8 yield information on the activation of bonds in organic molecules by metal ions in the absence of complicating solvent effects. Thermodynamic, kinetic, and mechanistic information concerning the organometallic and coordination chemistry of metal ions can be obtained from these studies. An important area of interest has been the study of the interaction of metal ions with organic molecules containing specific functional groups. Metal ions appear to insert into polar bonds such as the C-Cl bond in isopropyl chloride9. The metal ion (Co+) inserts into the C-Cl bond as shown in reaction 1. Co++ >—c1 —) 2—00+—Cl fit) Juncéecm (1) H 2 In this intermediate structure, one carbon is a to the metal and two carbons are 8 to the metal. A hydrogen which is on a B-carbon (i.e. a 6-H) is observed to shift onto the metal as shown in reaction 1. This 8-H shift mechanism is common in both solution10 and gas-phase positive metal ion/molecule reactions“. Following the 3-H shift, rearrangement may occur producing an HCl ligand on the metal as shown in reaction 1. The metal insertion/ 3 -H shift mechanism produces two molecules (HCl and propene) from one. These two molecules then compete as ligands on the metal ("competitive ligand loss") as shown in reaction 2. Jun Co+(-CIH ——> J---- Co+ 4' HCI ' (2) > Co+ecm + y The proton affinities of ligands have been shown to parallel the strength of the metal-ligand interaction”. Ligands which have low proton affinities are lost preferentially to ligands with higher proton affinities in the competitive ligand loss process. Ligands which are I -donors (e.g. olefins) also interact strongly with positive metal ions and frequently are preferentially retained in the competitive ligand loss process. In addition to studies of the reaction of metal ions with alkyl chlorideszi9’11, the reactions of transition metal ions with organic species such as alkane5516113‘15, alkenes3’6117, alcohols9, amines13, aldehydeslgizo, ketonesl9’20, carboxylic acids”, esters19, etherszo, sulfides”, and nitroalkanes22 have also been reported. The major reaction pathway for the majority of these compounds involves insertion of the metal ion into the 3 relatively weak C-functional group bond as the first mechanistic step. This may be followed by the 8-H shift and competitive ligand loss process similar to that of the alkyl chlorides. Metal insertion has been observed not only into C-functional group bonds but also C-C bonds, C-H bonds (e.g. alkanes5’6913‘15) and even into bonds within functional groups (e.g. nitroalkanes”). The products which are observed in the positive metal ion reactions can usually be explained by the general mechanism: metal insertion] 8-H shift/competitive ligand loss“. The ability to explain and even predict the products observed in metal ion/molecule reactions with neutral organic molecules containing various functional groups has led to the utilization of these metal and metal-containing ions as chemical ionization (CI) reagentsl9123'26. Metal and metal-containing ion/molecule reactions have been shown to provide molecular weight, functional group, and structural information about the neutral organic reactant species. The metal ions used in these studies may be formed in several ways. The two most common sources of metal ions utilized have been laser ionization of metal foils27 which generates the bare metal ion, M+, and 70 eV electron impact on metal carbonyls (M(CO)n)28. One advantage of the use of metal carbonyls is that metal ions with various ligands present (e.g. M(CO)x+(x=1 to n)) are formed in addition to the bare metal ion M". This allows the study of ligand effect51912’13’22125’29 (i.e. changes in reactivity as the number of ligands on the metal is varied). The majority of the gas-phase organometallic ion/molecule reactions studied have dealt with metal and metal-containing positive ions. These previous studies have reported, changes in reactivity as the type of metal, number of ligands on the metal (ligand effects), and the neutral organic molecule are varied. Relatively few studies have been performed on the corresponding 4 metal and metal-containing anions; i.e., none has determined changes in the chemistry of an isolated metal "center" due to a change in the charge on the metal species. Early mass spectrometric studies have shown that M(CO)n-1‘ is the predominant anion formed by 70 eV electron impact on transition metal carbonyls, M(CO)n, with a small percentage of M(C0)n.2' also being formed30'32. The first metal anion/molecule reactions studied with ICR were performed by Dunbar33 and Beauchamp34. The stable l7-electron species, M(CO)n.1', were found to be generally unreactive in the gas phase, whereas the M(CO)n.é' anions (ls-electron species) were observed to react with the neutral M(CO)"33 as shown in reaction 3. Fe(CO)3" + Fe(CO)5 ~—->-Fe2(co)6‘ + 200 (3) Recently Wronka and Ridge4 have observed sequential anion/molecule reactions in Fe(CO)5 up to Fe4(CO)13'. Reactions of these metal-containing anions with organic molecules have been studied using ICR and flowing afterglow techniques. Weddle and Ridge35 reported the chemistry of Fe(CO)3' and Fe(CO)4‘ with a series of thirteen organic molecules. Ligand substitution of the neutral organic molecule for one or two carbonyl ligands was observed for 9 of the 13 compounds studied. Only one rearrangement-type reaction was reported, ion-induced decarbonylation _ of maleic anhydride. The predominance of ligand substitution reactions and lack of (apparent) bond—breaking and rearrangement reactions for Fe(CO)3' and Fe(CO)4' is not surprising since, in the reactions of asitive metal ions, as the number of ligands on the metal increases ligand substitution processes dominate the products observedlil2113122925129. Freiser et. al.36 were able 5 to study the chemistry of the bare metal anion Cr" with a series of Br'énsted acids. Only proton abstraction reactions were observed. The bare metal anion Cr" was formed during collision induced dissociation on the Cr(CO)5' anion which was produced by electron impact on Cr(CO)6. Squires et. al.37 have formed metallocarboxylic anions in the gas phase (Fe(CO)4COOH') by reacting OH‘ with neutral Fe(CO)5 in a flowing afterglow apparatus, in an attempt to study the water gas shift reaction. McDonald et. al.33’39 have also used a flowing afterglow apparatus to study the reaction of anions formed by electron impact on Fe(CO)5. The anion Fe(CO)4" reacts with a series of halo-methanes to yield halogen atom transfer products and . in some cases ligand substitution with loss of one or two carbonyl ligands”. The reactions of Fe(CO)3' were also studied39 in which adduct products with the neutral molecule were observed due to the termolecular collisional stabilization with the He/CH4 buffer gas, in addition to ligand substitution of the neutral molecule for carbonyl ligands. The reaction of Fe(CO)3' with ~ CilgBr yields a ligand substitution product of CligBr for two carbonyl ligands and abstraction of a bromine atom from CHgBr (reactions 4 and 5). Fe(CO)3" + CH3Br Fe(CO)(CH3Br)‘ + 200 (4) t Fe(CO)gBr‘ + 0113- (5) The mechanism proposed was the formation of an ion/radical collision complex produced by the bromine atom transfer, although oxidative addition is still possible since Fe(CO)3’ has two available coordination sites. It is suggested that the ion/radical complex could dissociate to produce the bromine abstraction product or the methyl radical could form a bond with the metal and displace two carbonyl ligands to yield the substitution product of the neutral CH3Br in the form: To date, there has been no attempt to formulate a general mechanism to describe metal-containing anion/organic molecule interactions. The studies described in this dissertation were initiated in an attempt to obtain an understanding of the reactions and mechanisms of these metal-containing anions. Low energy (0 to 5 eV) electron impact was used to produce a greater percentage of the more reactive M(CO)n_2‘ and M(CO)n.3‘ species from Co(CO)3NO, Fe(CO)5, and Cr(CO)630'32. The reactions of these metal-containing anions with a series of n-chloroalkanes, n-alcohols, n-nitroalkanes, and several bifunctional compounds were studied in order to propose a general mechanism for the reaction of metal anions with polar organic molecules. Based on the current literature, then, what similarities and differences would be expected in the chemistry of organometallic anions and cations? Metal anions may be expected to insert into bonds in a manner similar to that of cations. There is n_o evidence in the literature which suggests that 3-H shifts may occur for anions, however organometallic anions containing M-H 7 bonds have been reported“. If a rearrangement such as 8 -H M8 + C3H7Cl —’C3H7-Ma-Cl ——-)-(03H6)M3(HCI) shift occurs when a is (+), ligand loss correlates with the ligands' proton affinities. When a is H, ligand loss appears to correlate with Lewis acidity“. Thus, we may expect both similarities and differences in anion and cation reactions. The previous studies mentioned have not addressed the question concerning the location of the negative charge in these anions. In the corresponding positive ion reactions, the positive charge remains predominantly on the metal due to the lower ionization potential of the metal compared to that of the ligands present. In the discussion of the anion reactions, the electron affinities of the metal species and ligands must be considered when determining the location of the negative charge. The electron affinities relevant to this work which are available are listed in Appendix A. The amount of electron affinity data available on metal carbonyl species is small and incomplete but proves invaluable in explaining the anion/molecule reactions observed. For example, the electron affinities of all three bare metals (Fe, Cr, and Co) have been determined, but the various carbonyl-containing species of only one metal (Fe) have been determined. It will be shown that the location of the negative charge and the types of ligands present play an important role in explaining the reactions and mechanisms observed. CHAPTER 2 EXPERIMENTAL A. The Ion Cyclotron Resonance Technique 1. The ICR Cell Several reviews have been published on the ion cyclotron resonance (ICR) technique42‘45. The technique is extremely powerful in the study of bimolecular gas-phase ion/molecule reactions. A typical three-region ICR cell is shown in Figure l. "radiating minaret: Drift vol”; 9 r m reg-on - Drm voltages 9 Source region - - Trooping vamp o— \ Figure 1. Three-region ICR cell43. 9 The cell is placed between the poles of an electromagnet with the magnetic field B (H) directed as Shown. Electron impact is utilized to ionize the sample molecules in the source region of the cell. Electrons are emitted from a hot rhenium filament (collimated by B) located outside the cell, pass through the cell (if not scattered by collisions) and strike the collector. The motion of a charged particle in a uniform magnetic field (B) is constrained to a circular orbit of angular frequency we in a plane normal to B and is unrestricted along the axis parallel to B. The potential applied to the trapping plates (<.5 V) prevents the ions from drifting to the sides of the cell. The polarity of the trapping plate potential is made positive to trap I positive ions and negative to trap negative ions. The force on an ion in the plane perpindicular to B is given by F=ma=evB ~ For an ion of mass m, charge e, and velocity v normal to B, an acceleration a will occur. The force, F, is normal to both B and v. Circular motion results for sufficiently large values of B with a = v2/r where r is the radius of the ion path, resulting in the equation of force: F = ma = mv2/r = evB If (no is the angular frequency of the ion, then the basic cyclotron equation (1) may be derived: mv/r = mmc = eB 10 w c = eB/m (radians/second) (1) or “c = eB/2 11m (cycles/second) The cyclotron equation may be rewritten as m/e = B/21Tvc (2) which shows that at constant frequency, m/e varies linearly with B. Therefore, a linear mass scale for resonant ions may be produced by varying B. The ions are made to drift from the source region to the resonance (analyzer) region of the cell by applying a potential difference between the plates above and below the electron beam (drift plates). The resulting crossed electric and magnetic fields cause ions to drift in a direction perpendicular to both fields. The drift velocity is given by Vdrift = 6 drift/3 (3) where Edrift is the electric field intensity. The relationship between the three vectors (B, adrift, Vdrift) involved is shown in Figure 2. ea Figure 2. Vector contributions resulting in the drift motion“. 11 The drift velocity for a typical cell (8.6 cm. in length) under normal conditions is 1150 m/sec. Under these conditions the ion spends NZ X 10‘3 seconds in the cell. This relatively long ion residence time (compared to conventional mass spectrometric techniques) results in the formation of ion/molecule reaction _ products due to collisions of the reactant ions with a neutral species. 2. Detection of Ions If a radiofrequency (rf) field is applied to the drift plates of the ICR cell, an ion in resonance with this field (i.e., the cyclotron frequency of the ion and the frequency of the applied field are equal) absorbs energy which results in an increase in the velocity of the ion. Equation 1 shows that vc is independent of v and r separately but is dependent on their ratio v/r. Thus, an ion remains in resonance by increasing its radius in proportion to the increase in velocity. The absorption of power from an rf field is the basis for the detection of ions in ICR. The most common form of detector utilized in ICR is the marginal oscillator (MO) shown in Figure 3. Amplitude C Detector I) 31 I I9 Figure 3. Simplified circuit diagram of the marginal oscillator detector“. The oscillator may be considered to be a constant current generator which drives a parallel LC tank circuit at its resonance frequency. The two analyzer 12 drift plates are a capacitive element of the marginal oscillator. Ions'are detected when their cyclotron frequency (equation 2) is equal to that of the oscillator. The detector response reflects the change in rf voltage across the tank circuit. When a signal is observed, a current is generated in the circuit due to the power absorbed by the ions. The frequency of the marginal oscillator is usually chosen such that a change of 100 gauss in B corresponds to a change of one atomic mass unit (11). Using equation 2 this frequency corresponds to 153.57 kHz: 1.6021 x 10-19 c x 0.01 T vc = _2;-——- = 153.57 kHz 6.2832 x 10. x 1.6604x10 kg./u. The force (field intensity times charge) exerted on the ion which is in resonance with the rf field will be F = ma = (eff/2k where Erf is the rf electric field intensity. The velocity of the ion which has been in resonance for time t is given by: t v = I adt = (erfet)/2m 0 The product of the force on the ion and its velocity yields the equation for instantaneous power absorption: 13 A(t) = Fv = (cl-{2 e2 t)/4m The average power absorbed by an ion from the rf field while in the analyzer region for a time (0 < t < T) is T A(r) = A(t) = i new: = (erf2e21)/8m <4) 0 Note that in equation 4 the power absorbed by an ion depends inversely on its mass. It has been shown42 that in marginal oscillator detection the signal intensity must be divided by the mass of the ion in order to be proportional to the relative ionic abundance. This is due to the fact that the power absorbed by the ions is proportional to the time spent in the analyzer region and that the drift velocity is inversely proportional to the magnetic field (equation 3) (i.e., higher mass ions move slower and thus absorb power from the marginal oscillator more time than the lower mass ions). The signal/noise ratio is greatly enhanced by signal modulation and phase-sensitive detection. Common modulation methods include modulation of the magnetic field (by a few gauss), the electron beam voltage or current, or the trapping voltage. Modulation of the trapping voltage (15 to 25 Hz) by applying a square wave (+Vt to -Vt) to one trapping plate is the most common modulation technique. 3. Ion Cyclotron Double Resonance A distinguishing feature of ICR spectrometry is the correlation that can be made between a product ion and the reactant ion in an ion/molecule reaction. 14 This technique is known as ion cyclotron double resonance (ICDR). This is accomplished by applying an rf signal of variable frequency and amplitude to the drift plates of the cell. Thereby adding energy to possible reactant ions while leaving the product ion in resonance with the detector. If the amplitude of the rf field on the drift plates is large enough, it is possible for the ions to absorb enough energy to cause them to collide with the drift plate (i.e., their radius exceeds the dimensions of the cell). Suppose A+ reacts with B to yield C+ and D as seen in reaction 6. A+ + B ——1-» 0* + D (6) It can be seen from equation 2 that me vc is a constant when the product ion (Cl) is in resonance at the required value of B. Possible reactant ions (mx') can be brought into resonance at that value of B by supplying the frequency v ' such that mcvc = mx'vx'. Since A+ (mg) is reacting to form C“ (the), increasing the velocity (or rate of ejection) of A+ will cause a change in the amount of C+ which is observed by the detecting oscillator. If reaction 6 is exothermic, the double resonance ejection of A+ will result in a decrease in the C+ signal observed. If reaction 6 is endothermic, the acceleration of the A"' ions by the rf field may drive the reaction faster and cause an increase in the C"' intensity prior to the ejection of A+. A block diagram of a basic ICR, capable of performing all of the operations . discussed previously, is shown in Figure 4. 15 Drift and .- Um MW " Marginal l—-l Amplifier l-C- oscillator m (Drift and.) Modulation ,7 rat (Mag. told mod.) on“ oacillaror mm doubla—reaonama mod.) "in” l m... pm... oaoillator smiths XY m “'l m "l *‘L Figure 4. Block diagram of the ICR spectrometer“. B. Experimental Parameters for Negative Ion Studies The ion cyclotron resonance mass spectrometer used to perform all of the ion/molecule reacton studies in this dissertation was built at Michigan State University from conventional design. The dimensions of the three-section ICR cell are 0.88 in. X 0.88 in. X 6.25 in. The source and analyzer regions are 2.00 in. and 3.75 in. long respectively. In normal operation the electron filament is emission regulated. The filament controller and plate voltage controller for the ICR cell were designed and constructed at M.S.U. The data presented here were obtained under normal drift-mode conditions by using trapping voltage modulation and phase-sensitive detection. The marginal oscillator detector was constructed at M.S.U. based on the design of Warnick, Anders, and Sharp“. A Wavetek Model 144 sweep generator was used as the secondary oscillator for double resonance experiments. The ICR cell is housed in a stainless steel vacuum system and is situated between the polecaps of a Varian 12 in. electromagnet (1.5 in. gap). The 16 . electromagnet is controlled by a Varian V-7800, 13 kW power supply and a Fieldial Mark I Magnetic Field Regulator. The instrument is pumped by a 4 in. diffusion pump with a liquid nitrogen cold trap and an Ultek 20 US ion pump. Samples are admitted from a dual inlet with Varian 951-5106 precision leak valves. Approximate pressures are measured with a Veeco RG 1000 ionization guage. All chemicals used in this work were high-purity commercial samples which were used as supplied except for multiple freeze-pump-thaw cycles to remove noncondensible gases. Although ICR is most often used to study positive ion/molecule reactions, it is extremely simple to convert the system to study negative ions. The only- experimental condition which must be changed to detect negative ions is the polarity of the trapping plates (negative potential to trap negative ions). All other experimental conditions remain the same with the exception of "re-tuning" the plate voltages for optimum peak height and peak shape. The early mass spectrometric studies of negative ions formed by electron impact on metal carbonyls (M(CO)n) report that the major ion produced by 70 eV electron impact is the M(CO)n-1‘ anion (loss of one CO) with a small amount of M(C0)n-2" also being formed30’32. Recall that the previous metal anion/molecule reaction studies found the 17- electron M(CO)n_1 species to be unreactive. The early mass spectrometric studies of negative metal ions also reported the energy dependence of the negative ions formed by collision of slow electrons (0 to 10 eV) with metal carbonyls. The energy dependence - of negative ions formed by dissociative electron capture of slow electrons with Fe(CO)5 is shown in Figure 5. Note that it is possible to maximize the relative amount of each fragment anion by varying the ionizing electron energy. 17 RICO); (:10) ION CURRENT (arbitrary units) V l L O I 2 3 4 5 B 7 O ELECTM ENERGY (0V) Figure 5. Energy dependence of negative ions formed from Fe(CO)532. Since the M(CO)n-1" anions were reported to be unreactive, this study utilized low energy electron impact to maximize the relative concentrations of species with fewer ligands present (M(CO)n-2", M(CO)n_3'). The ionizing energies utilized for this study of anions from Fe(CO)5, Cr(CO)5, and Co(CO)3NO are listed in Table 1 along with the anions which are formed. The M(CO)n-1" anion is still the major ion formed under these conditions, but a sufficient amount of the lower fragment anions are produced to allow the study of their ion/molecule reactions. In Figure 5 a maximum for the formation of Fe(CO)2‘ is observed at «4.5 eV, but note that its intensity is magnified 10X compared to that of Fe(CO)4'. Although it would be desirable to form these lower fragments, even at the suggested energy for maximum production of Fe(CO)2‘ a sufficient amount of Fe(CO)2' is not formed which prevents the study of their ion/molecule reactions. Typical low energy electron impact negative 18 Table 1. Metal-Containing Anions Formed by Low Energy Electron Impact Neutral Io Ene Anions Formed Fe(CO)5 1 eV Fe(CO)4" Fe(CO)3" Cr(CO)6 4.5 eV Cr(CO)5‘ Cr(CO)4" Cr(CO)3' Co(C0)3N0 1.5 eV - Co(CO)2NO‘ ' CO(CO)3- CoCONO‘ C0(CO)2' 19 ion mass spectra for the three metals (Fe(CO)5, Cr(CO)6, and Co(CO)3NO) at pressures of approximately 2 X 10 ‘6 torr are shown in Figures 6 through 8. In a typical experiment, low pressure 70 eV positive ion mass spectra of both the metal carbonyl and organic compound were taken to determine the purity of the samples. The ICR was then configured for negative ion studies (i.e., the trapping voltage and ionizing electron energy were changed). Spectra were then recorded of a 1:1 mixture (by pressure)lof the metal carbonyl to organic compound at a total pressure of approximately 1 X 10"5 torr. Spectra were always taken to masses greater than the sum of the molecular weight of the metal carbonyl and the organic compound. The anion/molecule reaction products were identified and precursors of each determined using double resonance techniques. The branching ratios listed for the reactant ions are accurate to within 110%. Although in some cases empirical formulas other than those listed may also be possible for the products, those listed are believed to be the most reasonable based on observed reaction trends and the reactions of the other metal anions. 20 .maoovom «o 5:36on 6.2:: cc. 033mm: “page: c2320 3o :3 >928 so,— 6 0.59m 32.. an: -2093. SEE I‘lllll N): 1% - - -209: 8.3).. $.— Rusuaju; 3:3 2 c 14111. l -28.": 2:13... 21 369.6 2:03... $89.5 .3 5:30on 9.3:. :2 950mg Loon—E aches—o go. we); huge—.0 30.— 4. 952m 3:.— an: T a): L. -2095 S 7!... guzz— 3: 4 -2815 «an»? 4——— Kiisuaiui 22 62200.00 .0 50.5000... 80:. :0. 02.0mm: .0095 00500.0 30 «.3. .8300 30.— .a 0.5»... 3.... an: 7 -:0030 5...»): 02:00.60 2.0.... All) 8... 1020030 F — nukes—— -2023 2.7.3... $80.8 2.3.... .02.. 3: ( E N.02....00.30 -£0z.00«\o0 us can...» E «a . 85:00.30 .20.... - 2:00.30 «ans... 02:00.30 80!... - 05:00.30 guns}: <1—-——— flusuaiui CHAPTER 3 REACTIONS OF N-CHLOROALKANES The ion/molecule reaction products and their branching ratios for the reactions of chromium, iron, and cobalt-containing anions with the series of l-chloro-n-alkanes (n-chloroalkanes) (n = 1 to 6) are listed in Table 2. Three types of reactions are observed for the metal-containing anions with n-chloroalkanes: ligand substitution by the chloroalkane molecule for two or three carbonyl ligands (Parent substitution); abstraction of chlorine from the chloroalkane (Cl abstraction); and abstraction of chlorine and hydrogen from the chloroalkane (HCl abstraction). A. Parent Substitution The general reaction for substitution of the chloroalkane molecule for carbonyl ligands is shown in reaction 7. M(CO)X- + CnH2n+lCl ‘—_’ M(CO)x—a(CnH2n+1Cl)- + 800 (7) Parent substitution reactions by the chloroalkanes are observed for the metal-containing anions Cr(CO)3", Fe(CO)3', and Co(CO)2‘. The reaction of Fe(CO)3‘ and Co(CO)2" are accompanied by a loss of two carbonyl ligands 23 24 N_. am. am. ma. «N. cu. ac. ow. an. me. am. N.. .m... me. _o. mv. av. 5.. mm. mm. 0.. Amm.v an. m_. c—. ——. .b. mp. as. c.. N.. a.. o.. .m... m.. .v. me. n.. c—. n_. .m~. .N. n..... .N. mmm mmm mmm 8...:— $22.9... 3.... z. 2. $0.. .2. mp. .34 n—. mm. mm. mm. 0300:2000 « 030,—. 5....0 + -.0=~.00.60 In 00 + 5....0 + .82008 III-I 8. $5.20 + -.0~.00.60 A S. 00 + .+§....0 + -.00060 17'] S. 00. + -.0.+..~:..060 Ir 8.25.1.0 + -N.00.oo mz‘ulull-Gremzeo + -1086... 00 + $.10 + -.0..~.00.6.. Al 2... .25....0 + -5200}. III) 00. + -.0.+..~:..0006... III.- l- .0.+..~=..0 + -3005... mz Iallll 87.5.10 + -....00...0 602 TI 5.2.31.0 00 + 5....0 + -.0=.-..00..0 I) + -1085 N... .25....0 + -.0...00...0-A S. 00 + .+......=..0 + -.o~.00...0 ‘IJ a... 00.. + -.0.+§=..0._0 I 6.25:..0 + .2025 find 603000: 85562356 5... 86.5. 22.859580 0.... .32. 6.2525 .6 22.88. .« oz... 25 3.. mm. 2.. mm. nu: mm. 3. mm. 2... WI... mu... 803. 0:205... 3. mm. 0.00. .0000... 0.20000. .0. 0030.. 002000.... 0.0 0...... 0.: 0.... 00.000.00.00 0. 000—0.. 0 000.300.. >00 09000.. .00 0:. 00. 05 00.00.00. ”.2 o mzil-ll 5.2.31.0 + 02200.60 02 lull-ll 8.2.31.0 +.-...00.60 00 + 52:0 + 20:0260UI .25....0 + 20020060 5.20.1.0 + -020060 I 003000: .5.— 39003000 N 309—. 26 (a=2 in reaction 7) while Cr(CO)3" loses all three carbonyl ligands (a = 3). Two possible product ion structures are suggested, structures '1' and '2' (where x = 0forM=Cr,Co;x=lforM=Fe). (CO)" I (CO)x-—M'<—C|CH 'Cl—M—CH n 211+! 11 2n+t .l - 2 ~ Structure "1 results from ligand substitution by the chloroalkane for two or three carbonyl ligands with the chloroalkane molecule remaining intact as a ligand on the metal anion. The negative charge must remain on the metal species in structure ‘1' due to the large negative electron affinity of the chloroalkane ligand (e.g. E.A.(CH3C1) = -79.6 kcal/mole). While an electron pair on the chlorine may form a dative bond to the metal, there will certainly be a strong ion-dipole repulsive interaction in this intermediate. The magnitude of this repulsion may be estimated using equation 549 which describes the interaction between a point charge 0 and dipole u in the gas phase which are separated by a distance r at an angle 9 . 00) = 5:52- cos 6 (5) Assuming a distance of 2 A between the ion and dipole and 6: 0°, the magnitude of this repulsion is estimated to be 35 kcal/mole for l-chlorobutane and 29 kcal/mole for 1-butanol50. Similar calculations for a distance of 5 X yields 27 ion-dipole repulsions of approximately 6 kcal/mole and 5 kcal/mole for l-chlorobutane and l—butanol respectively. Comparison of the electron affinities for the metal-containing species Fe(CO)3 (41.5 kcal/mole) and CI (83.3 kcal/mole) suggests the possibility that, once the metal anion-chloroalkane. complex is formed, it is thermodynamically favorable for the charge to be transferred from the metal anion to a chlorine atom. Structure 2 results from metal insertion into the C-Cl bond (which occurs in positive metal ion reactions) with transfer of the electron to the chlorine atom. Presumably the metal anion insertion/charge transfer process is sufficiently exothermic such that excess energy results in the loss of two or three carbonyl ligands. If we are actually observing metal insertion into the C-Cl bond, a metal-carbon a-bond is also formed. Reactions of aliphatic halides with NazFe(CO)4 in solution54 also proceed through intermediates which have a metal-carbon o-bond from the alkyl group to the metal anion Fe(CO)4‘. The electron affinities of these alkyl ligands (15 to 26 kcal/mole) indicate that the presence of the alkyl ligand may help delocalize the negative charge in structure 2'. Therefore, structure 2 seems to be the most reasonable structure for the Parent substitution product ion from the electron affinity data and the mechanisms predicted from condensed phase reactions. Note that the Parent substitution product ion, structure 2, does not consist of an intact parent organic molecule (predicted from positive ion reactions), but instead is a metal insertion structure which contains the organic molecule as two separate ligands . on the metal. The amount of excess energy released in the metal anion insertion/charge transfer process (reaction 8) is dependent upon both the exothermicity of the metal inSertion (i.e. the bonds which are broken and formed) and the exothermicity of the charge transfer process (i.e. the difference in the electron 28 affinities of the metal species and the ligand). (0.0),.El M(CO)X' + Cl—CnH2n+1 —-> ‘Cl—M-CnH2n+1 + aCO (8) Assuming that the differences in the metal-ligand bond strengths are small for the three metals, the differences in the exothermicity for the metal insertion/charge. transfer process for the metal species is an indication of the difference in the electron affinities for the metal species. Since Fe(CO)3' reacts by loss of only two carbonyl ligands and Cr(CO)3" reacts by losing all three carbonyl ligands (and assuming that this reflects the fact that the metal insertion/charge transfer process is more exothermic for Cr(CO)3"), then presumably the electron affinity of Cr(CO)3 is less than the electron affinity of Fe(CO)3 (E.A.(Cr(CO)3) < E.A.(Fe(CO)3) = 41.5 kcal/mole) B. Cl Abstraction Once the metal insertion/charge transfer intermediate «.2 is formed, the excess energy may not only result in loss of carbonyl ligands (Parent substitution) but may also lead to loss of the alkyl ligand (reaction 9). This reaction is observed for the anions Cr(CO)3', Fe(CO)3", Co(CO)2', and CoCONO' with the n—chloroalkanes for n = 1 to 6. Competitive ligand loss . of the alkyl ligand with concurrent loss of zero or one carbonyl ligands leads to the chlorine abstraction products (M(CO)x.bCI’) observed in Table 2. C. HCI Abstraction The third reaction type observed for n-chloroalkanes (n >, 2) is abstraction of I-ICl as shown in reaction 10. 29 M(CO)x‘ + an2n+101 ——)>M(CO)X-CHC1‘ + CnHZn + cCO (10) This product may also result from metal insertion into the C-Cl bond as a first step. A common mechanism in both solution and gas-phase organometallic reactions involves the shift of a B-hydrogen atom onto the metal centerloill. This apparently occurs in the metal-containing anion reactions as well; In the case of chloromethane, the intermediate resulting from metal insertion into the C-Cl bond does not have any B-H's to shift onto the metal. As a result, the HCl abstraction product is 32?. observed for chloromethane (Table 2). For the larger n-chloroalkanes (n = 2 to 6) however, B-H's are available to shift onto the metal. Once the 8-H shift occurs, the alkene ligand produced (CnHZn) is lost, yielding the HCl abstraction products (M(CO)X.CHCI’). The two possible structures for this product ion are shown in structures ,3 and ,4. (C Oh I H—M—cr' 3 4 ~ ~ (c O)»:— M '—- (H C I) Structure ,3 represents the structure if HCl exists as a single ligand on the metal. The electron affinities of H (17.4 kcal/mole) and CI (83.3 kcal/mole) suggest that structure 4, where H and C1 are separate ligands on the metal, - may be the more stable structure. The addition of the hydrogen atom as a ligand, which has a positive electron affinity, may help delocalize the negative charge. This effect is also seen in the comparison of the electron affinities of Fe (3.8 kcal/mole) and FeH (21.5 kcal/mole). The alkene ligand which is formed following metal insertion and 13-11 shift 30 in the corresponding positive metal ion reactions is retained preferentially over the HCl ligand9. In the metal anion reactions however, the alkene ligand is never retained and is always lost in the competitive ligand loss process. This is not unexpected if the electron affinity of the alkene is considered and structure 4' is assumed to be the HCl abstraction product. The electron affinity of ethylene (-35.7 kcal/mole) suggests that the alkene ligand would be lost preferentially relative to ligands which possess a positive electron affinity, i.e., those which help to delocalize the negative charge. The difference in metal-ligand bonding in positive and negative metal ions has been studied by Corderman and Beauchamp‘l1 suggesting that the competitive ligand loss process I for positive and negative metal ions may be quite different. The metal-ligand bond energy in positive ions is determined largely by the I-donor ability of the ligand while back-bonding effects are less important. In the negative ions however, the bond energy is predominantly dependent upon the t-acceptor ability of the ligand with the ‘l-(IOI'IOI' ability playing a less important role. - Thus, the difference in the metal-ligand bonding in positive and negative ions may be a factor in the retention or loss of the alkene in the positive and negative ion reactions. D. Reaction Trenth and Mechanisms The general mechanism for the reactions of the metal-containing anions with n-chloroalkanes (n = 1 to 6) is shown in Scheme 1. Note that all three types of products observed (Parent substitution, C1 abstraction, and HCl abstraction) proceed through intermediate structure 2. 31 Scheme l M(CO)' 4- anzM.CI (can (anZnMr-M-c'- .2. co 'bco a q -an,,,,, sum / (CO) x-o (c'o),,c (CnHznoi)_M—C‘ (c°)fi'M—Cl "_v_c" I (Parent substitution) (Cl abstraction) Hzcéf-cn-IHZn-l (n-Itos) (n-ttoG) H -cCO -an2n (C'O)'.c H—M—CI‘ 4 (HCI obstruction) (n I 2 to 6) 32 Two possible mechanisms for forming structure 2 are presented in Scheme 11. Note that structure 23 is equivalent to structure 2 for X = Cl. Scheme ll (co)n (CF), - '. RX 4» M(CO)_| ‘_—=-5_ R—g—X —-—> R—M—X' ~ 6 a»: (con-m}. 7 ‘R In the first mechanism, the metal anion initially inserts into the R-X bond (analogous to positive metal ion insertion) to give structure ,é. If the charge transfer from the metal species to X is exothermic, structure Q (the metal anion insertion/charge transfer intermediate) is formed; If the charge transfer is not exothermic then 4? may reform the metal anion and RX. In the second mechanism (Scheme 11), the metal anion interacts with both X and the alkyl R as seen in structure 1. This type of initial complexation may reduce the ion-dipole repulsive interaction discussed previously. If the charge transfer is exothermic, an M-X" bond is‘ formed along with the cleavage of the R-X bond and formation of the M-R bond yielding structure 2. Regardless of the mechanism, it appears that all reactions observed proceed through the metal insertion/charge transfer intermediate Q and the driving force for the reactions appears to be the exothermicity of the charge transfer process. Additional information may be obtained through trends observed in the branching ratios of product ions as the alkyl chain length (n) increases (Table 33 2). For all three metal-containing anions, Parent substitution and Cl abstraction decrease and HCI abstraction increases as n increases. These trends support a mechanism in which all products observed proceed through a common intermediate (structure 2). One possible explanation for the increase in the amount of HCl abstraction observed may be the thermodynamics of the reaction. Thermodynamic calculations indicate that less energy is required to form an alkene from the corresponding alkyl chloride as the length of the alkyl chain increases. For example, if the reactions of l-chloroethane and l-chloropentane are considered (reactions 11 and 12), thermodynamic calculations indicate that A H12 (csnllcn is 4.3 kcal/mole less than A H11 (021150053 C2H5Cl -—’HCI 4' C2H4 (ll) C5H1101 -—>HCI + 051110 (12) The effects that the number and types of ligands present on the metal have on reactivity and mechanisms in positive metal ion reactions have been described previously1912,13,22,25,”. Several ligand effects are also observed in the reaction of metal-containing anions. The carbonyl ligand is lost preferentially to the nitrosyl ligand as seen in the reaction of CoCONO‘ with n-chloroalkanes (Table 2). This effect is also observed in the positive metal ion reactions and is due in part to the fact that the nitrosyl ligand is a three electron donor and the carbonyl ligand is a two electron donor. Another ligand ‘ effect observed for metal anions is that the reactivity of the metal-containing anions decreases as the number of ligands present increase. As expected, the stable 17 electron species, M(C0)n-1' is unreactive towards the n-chloroalkanes. In summary, the reactions of the metal-containing anions with chloroalkanes 34 proceed through a mechanism in which the metal inserts into the C-Cl bond with transfer of the electron to the chlorine, due to the higher electron affinity of Cl. This complex may then undergo rearrangements ( B-H shift) and ligand loss processes to yield all products which are observed. In the corresponding positive metal ion reactions with n-chloroalkanes, products resulting from metal insertion into C-C and C-H bonds are also seen in addition to C-Cl insertion products”. All reactions for metal-containing anions appear to result from interaction with chlorine which has a' relatively high electron affinity, and thus no products resulting from C-C or C-H insertion are observed. This may reflect _the importance of the charge transfer step, since, if Fe(CO)3' inserted into a C-C bond, the charge would remain on the metal (E.A.(Fe(CO)3) = 41.5 kcal/mole, E.A.(CnH2n+1) = ~20 kcal/mole), and no chemistry from such intermediates is observed. CHAPTER 4 REACTIONS OF N—ALCOHOLS A. Reactions Resulting From Metal Insertion into the C—OH Bond The ion/molecule reaction products and branching ratios for the reactions of chromium, iron, and cobalt-containing anions with the series of l-hydroxy-n-alkanes (n-alcohols) (n = 1 to 6) are listed in Table 3. Parent substitution reactions are observed in the reaction of Cr(CO)3' with the series of n—alcohols for n = l to 6 and are accompanied by the loss of one or two carbonyl ligands (reaction 13) Cr(CO)3‘ + cnnzmlon ——-)-Cr(CO)3-anH2n+10H' + xCO (13) If metal insertion into the C-OH bond is assumed in the Parent substitution process (analogous to C-Cl insertion in the chloroalkanes), the charge would be transferred to the OH ligand due to it's relatively high electron affinity (42.2 kcal/mole) as shown in structure 2 (Scheme III). Clearly, the charge will be distributed over a number of atoms, however, we will continue wherever possible, to put the charge on the most electronegative species in a given structure, to parallel what is done for the positive ion analogs. This intermediate 35 93:23.. 3:: oEovco «o: Eu :3 05 330.9: m2: :2 ill 5152.0 + 62.2088 :2 lullnl :22“:an + -2830 . 35...... oo + N: + 65250on ‘ulll. =o_+§==o + .9308 8;"... Nz + -ogzsofoovono AIIIIII :orfizco + -3085 mz ‘IIII morszso + -zoozm 3-. u 5 00 + a: + -ogzcorogom ‘IIII. :o...§==o + -2088 «2 I sorcazso + -rooto smz ‘l..|l =o~+§==o + H8076 36 8. 8. S. 8. 2.. $5.20 + 50588.6 AIIJ 08 + «E + baamzsoooso All 08 + N= + -oézcoooso ‘Illl 8. 8. 3. 2. 08 + hats—.5096 hill! 2. 2.. :. 00 + s; + -o~-§==o£oorolmlll 2. 3. 2. an. 3. 00 + N= + -05....03035 ‘III 2. 8. S. 3. 2.. 00 + -=o_+§==o£ooro Allllfi =or§:=o + -2095 H: . V": M": N".— u H: Envmdnvsg Bags:— 3282;. 5:. 23.2 23359.38 o5 .8... .5336 .o 93:83. .a 03:. Q 37 ~ Scheme Ill Cr(CO); + CnH2n+10H J, (c0)3 (an2n+t)— Cr— 0H- 9. -xCO 'CnHZnH (60):“x (CnH2n+1)—Cr—-OH- (CO)3——Cr——OH- (Parent substitution) (OH obstruction) (x I 1.2) 38 may undergo a competitive ligand loss process to lose the excess energy produced in the metal insertion/charge transfer process. Loss of carbonyl ligands yield Parent substitution products (Cr(CO)3-anH2n+10H‘) and loss of the alkyl radical (CnH2n+1) yields the OH abstraction product (Cr(CO)30H‘) as seen. in Scheme III. The formation of the OH abstraction product is analogous to the Cl abstraction product observed in chloroalkanes. Note that the OH abstraction products are observed only for Cr(CO)3’ (Table 3) suggesting that the C-OH insertion intermediate Q (Scheme III) occurs exclusively for the reaction of Cr(CO)3‘. The loss of only one or two carbonyl ligands in the Parent substitution of n-alcohols, in contrast to three carbonyls displaced in the chloroalkane reactions, may be due to a decrease in the energy released in the charge transfer step, which would arise due to the lower electron affinity of OH (42.2 kcal/mole) compared to CI (83.3 kcal/mole). Parent substitution reactions are no_t_ observed for either cobalt or iron-containing anions presumably due to the comparable electron affinity of OH (42.2 kcal/mole) and the metal species (e.g. E.A.(Fe(CO)3) = 41.5 kcal/mole). Thus, the charge transfer process is not favorable in this case resulting in the absence of Parent substitution and 0H abstraction products. Since the cobalt anion Co(CO)2' does not react with the alcohols by Parent substitution or CH abstraction, the difference in the electron affinity of Co(CO)2 and OH must not be large enough to cause these reactions to occur. Therefore, we may infer (making the same assumptions as in the chloroalkane discussion) that the electron affinity of Cr(CO)3 is less than the electron affinity of Co(CO)2 (E.A.(Cr(CO)3) < E.A.(Co(CO)2)). Products indicative of H20 abstraction may be expected in the Cr(CO)3‘ reactions (analogous to HCl abstraction in chloroalkanes) since 8-H atoms are present after the metal inserts into the C-OH bond of all the n-alcohols 39 (n >, 2). Thermodynamic calculations for the production of l-butene from the corresponding 1-chlorobutane and l-butanol (reactions 14 and 15) show that less energy is required to form the alkene from n-alcohols than from the analogous n-chloroalkanes58. C4chl -—->C4H3 + 1101 all =+13.31 kcal/mole (14) C4H90H ——-’>C4H3 + H20 aH = +8.67 kcal/mole (15) The H20 abstraction product, however, is flo_t observed for the reaction of Cr(CO)3'. Apparently sufficient energy is released in the electron transfer to the Cl ligand to result in rearrangement (B-H shift) and loss of the alkene ligand in the chloroalkane reactions, but not enough energy is released in the electron transfer to the OH ligand (due to it's lower electron affinity) to result in rearrangement and loss of the alkene in the n-alcohol reactions. Therefore, a barrier to the B-H shift process appears to exist. B. Reactions Resultilg Prom Metal Insertion into the O—H Bond The only reaction observed for the cobalt and iron-containing anions with n-alcohols is the elimination of H2 (n = 1 to 6) with Co(CO)2', CoCONO", and Fe(CO)3" as shown in reaction 16. This hydrogen elimination product is also observed for the chromium anion Cr(CO)3'. The hydrogen elimination product is n_ot expected to proceed through metal insertion into the C-OH bond since products indicative of C-OH insertion (Parent substitution and OH abstraction) are not observed for the cobalt or iron-containing anions. A possible mechanism for the elimination of H2 following 40 metal insertion into the C-OH bond is shown below. (CIO)x ananOH +M(CO)x —> CnH2M1—M—OH- \Lfl-H shift (00)“, (CO)x (Cn Hzn)---- M — 0‘ (— (cn Hzn)---- ila— ('3‘ H H + H2+0CO This mechanism however, does not explain the observance of this product for methanol (n=1) since no B-H atoms are present following insertion into the C-OH bond. Therefore, a more comprehensive mechanism must be proposed. Scheme IV illustrates the formation of an intermediate in which the metal inserts into the 0-H bond. This intermediate is predicted to be stable due to the relatively high electron affinities of the alkoxy group («.40 kcal/mole) and the atomic hydrogen (17.4 kcal/mole) ligands on the metal following O-H insertion. The charge may then be transferred to the alkoxy ligand due to its relatively high electron affinity. Note that a 8-H is available to shift onto the metal in this intermediate even for methanol (n = 1). Following metal - insertion into the O-H bond and charge transfer, the 8-H shift and loss of H2 may lead to two different structures (3 and 1,9) in Scheme IV. In structure 2, the alkoxy ligand is converted into an aldehyde (t-donor) ligand following the 8-H shift with the negative charge being transferred back to the metal. In positive metal ion reactions, these I-donor ligands are found to form strong 41 Scheme IV M(C0)x + CnHanOH (0'0)x H—M—0_ l H—f—H cn-lHZn-I 5-H _ shift 1 (61°31 .. ((2.0),, H—“Ei‘3' H- 7-49 ' H/M (5-H Cn-iHZn—i cn-tHzn-i 3..” shift \L—oCO (cl0)xlo" H—M—C—H I | (00) M 9 H Cad-*2": — ---I x-o C H -oC0 cn-tHZn-i -Hz 2. u°t (CO);-°M—?—H cn_‘ H2n_1 10 42 bonds with the metal. Recall that in the metal anion reactions with chloroalkanes however, the I-donor alkene ligand produced following the 8-H shift to yield HCl was lost preferentially to the other ligands on the metal. In contrast to the positive ion studies, these t-donor ligands do _no_t appear to form strong bonds to the metal in these negative ion studies. The alkyl ligand (metal-carbon bond) appears to be a relatively strongly bound ligand in the metal anion reactions, since the alkyl ligand is in some cases retained during the competitive ligand loss process (e.g. Parent substitution) for chloroalkanes and alcohols. Therefore, structure 19 (Scheme IV) is the more probable structure for the product ion resulting from H2 elimination from alcohols. Following the 8-H shift, the metal may form a bond to the carbon which is B to the metal resulting in the metal-alkoxy structure 1,0. In this structure, the negative charge remains on an alkoxy ligand which has a high electron affinity. The anion Cr(CO)3' reacts further with the alcohols to eliminate a second molecule of hydrogen. This product is only observed for Cr(CO)3‘ presumably due to its lower electron affinity than those of other metal species. Elimination of two hydrogen molecules does not occur for methanol and ethanol, but does occur for n >,3. A general mechanism for the elimination of the second molecule of hydrogen is shown in Scheme V. The mechanism begins with structure 1,9 (following elimination of one H2) which may proceed through a 8-H shift to produce structure {1 which contains two metal-carbon bonds. Another 8-H ‘ is now available to shift from the C3 carbon to yield structure {2. Note that, in the proposed mechanism, the charge remains on the alkoxy oxygen (the most electronegative site) and that the mechanism accounts for the formation of this product ion for n >, 3 only. In summary, the metal anion attacks exclusively at the functional group Scheme V 0.. (CO)? C:i;—(:2— H h-l‘c— H cn-z Han-3 1 O 43 C-H fl-H / (cog—[ck] H C—H V/ shift H—C-H cn--3 HZn-s 1 1 5-1-1 shift 44' (no evidence for C-C or C-H insertion) due to the relatively high electron affinity of OH. In comparison to the chloroalkane reactions, however, the metal appears to insert into not one but M bonds (C—OH and O—H) in the reactions of n-alcohols. Further evidence is seen in the n-alcohol reactions for the formation of a metal anion insertion/charge transfer intermediate followed by possible rearrangements ( 8-H shifts) and competitive ligand loss. The enhanced reactivity of the C,r(CO)3‘ anion with n-alcohols confirms the inference from' the n-chloroalkane reactions that 'Cr(CO)3 has the lowest electron affinity of the metal species studied. The ligand effects observed in the reactions of chloroalkanes also were observed for alcohols: the nitrosyl ligand is retained in the reaction of CoCONO‘ and the stable 17 electron metal anion species are unreactive. "‘1 If: CHAPTER 5 REACTIONS OF BIFUNCTIONAL ORGANIC MOLECULES The reaction of positive metal ions with bifunctional organic molecules (e.g. 1-chloro-2-ethanol) have been studied previouslylzizsi26 and compared to the reactions of the corresponding monofunctional compounds in an attempt to determine the utility of metal ions as chemical ionization (CI) reagent ions. The reactions observed for the bifunctional molecules include: products indicative of both functional groups; products only typical of one functional group (indicating the metal shows a preference for one functional group over another); and products unique to the particular combination of functional groups. The reaction of metal-containing anions with two series of bifunctional compounds (1,n-bromochloroalkanes and 1,n-chloroalcohols) were studied to determine the similarities and differences from the corresponding monofunctional compounds (n-chloroalkanes and n-alcohols). A. Reactions of 1,n-Bromochloroalkanes The ion/molecule reaction products and branching ratios for the reactions of 1-bromo-n-chloro-alkanes (1,n-bromochloroalkanes) (n = 2 to 6) and 1,1-bromochloroethane with iron and chromium-containing anions, and with cobalt-containing anions are listed in Tables 4 and 5, respectively. Parent 45 If“ 46 ma. mo. 2. ms. nu. ma. we. 2”. on. an. «N. we. 2.. N_. 2. 2. mo. me. an. 2. an. as. we. 2.. ma. 3. a. an. an. ma. mm. an. «5. mm. .l-m I. Z. 5. 5. 2. 2. z. 2. 5. 5. 5. S. 2.. 2. 5. 5. 2.. ... 2. a... m NM 8.21255... as. «a. on. on. «N. an. a. 3.. 2.. 00 + 8.-52:0 + 22.20025 1...! 00 + .2.-:~2:0 + -.02200.._0ATI 8:.520. + 12200.5 ‘J 00 + 8:.«20. + (5200.541, 2:220. + -.0200..0-qu 00 + 5:220. + -.0200..0‘!l 00: + 752:0 + (£82.04!- 00: + 52:0 + 128.0 All am;— 00 + 8.-52:0 + 122200.»... All. 8:220. + 22200.0... All 00 + 8:220. + 22200.9. 1.1 2:120. + -.0200.o2 .ll. 00 + 2.2520. + -8200}. All 00: + 52:0 + -._2.0s...lll 005 + -.0:.N20.._200s2 AIL cocoa—gguoEéJ 5.8 3.2.3 23380-525950 tel 69... .0 93.88.: 4 8:63:00 v and? 10:38.... + .2085 8:38.... + -200.s2 . 82520.... + -2002... 8.68.. 936,—. 47 220.820.. 2.2 0300:: 8: ED :0. 05 2320.05 :2 0 029.895.200.295- _ . ~ 2 221A||8:.~20.2 + -200.:0 .2. 8. 3. S. 8. 2.. 00 + 8:220. + -2200...0UI .2... 5. 3. 2.. S. 2. 00 + 2:38. + -.0200...0 82520.2 .. 200.5 23 mn: . win-m .nunle- wul: 21—..«2. 8.803. no.8: 92029.: 23:35: . 0.2:... 48 3. 3. mm. Nu. me. an. S. we. 3. ms. mm. mm. 9.. on. 3. ma. ma. mm. :. ma. we. we. NM. my. an. an. ¢~. :. we. a. .6. mo. 3. 2. A3. cu: 2. on. em. 2. 2.. 2. Z. 2. 2.. Va. 3:3— 330:8: an. 5.. 2. we. 3. Z". 2. 2. mm. «Inn.- on. 2.. 2. ma. 2. on. S. a. mg. no. 2.1"": 8 + 8755:“. + -5828 all 00 + 5752:". + -8828 AT 8:38. + -50288 ‘1 8 + 8:38. + 59.8 All 5:38. + Loozoooonll on + 5:38. + -8oz8 AIL 875::0 + -5:£8.8 All 8 + 875::0 + -5808 ‘1 8 + 575::0 + -8=88 AT 8:28. + 5:8.8 ‘1 8 + 8:38. + 5881.1 5:38. + -8283“. All 00 + 5:38. + -888‘.! 8: + 5::0 + -5881 8... + -8:%8.58‘.L Eggs-BE.— 5.’ gnu-I gin-00!“; .0 93:08»— om 03GB. 89.58 m 03-... . 8538.5 + -ozoo8 r8:385 + -2088 8.83: 49 £2 3. an. m2 alum m2 2... vu: amz 2.4 I i.- ' 82.5 «5:055 2.; .54 co.“ 2.4 an _ 7N": 5:288.— 28 098:: 8: Eu :0. 05 8.8.05 m2 5 2358.3— :ooEoB- _ . . a 8:38. + -502£8.8|A8£N8.5 + 6228.8 8“ + 875::0 + -5=88 8 + 8755:“. + -5=£8.8 8 + 8:38. + -5£8.8 8:185 + -2088 8.38: 38.3: m. 03-... 50 substitution (to form products of the type M(CO)a-bClCnH2nBr‘), bromine or chlorine abstraction (to form M(CO)a.cCl‘ and M(CO)a-cBr‘), and HCl or HBr abstraction products (to form M(CO)a-dHCl' and M(CO)a-dHBr‘) are observed in the reactions of 1,n-bromochloroalkanes. These products may- also be predicted from the (monofunctional) chloroalkane reactions. A product corresponding to abstraction of bromine 93g chlorine (to form MClBr‘ with loss of all carbonyl ligands present) is observed and is unique to the reaction of the 1,n—bromochloroalkanes (i.e., is not expected from the chloroalkane reactions). The same mechanism utilized in the chloroalkane reactions may also be used to explain most of the 1,n-bromochloroalkane reactions as shown in Scheme VI. The metal anion inserts into the C-X bond (X = C1 or Br) with transfer of the electron to the halogen .(metal insertion/charge transfer). This intermediate may undergo either loss of carbonyl ligands to yield Parent substitution products or loss of carbonyl and alkyl (CnHZnY) ligands to yield halogen abstraction products. A 6-H. shift rearrangement may occur resulting in the formation of H0] or HBr abstraction products. The presence of two halogen atoms on the neutral molecule is seen to effect the distribution of the observed products when compared to the chloroalkane products. Parent substitution products are E observed in the reactions of Cr(CO)3‘. Recall that all three carbonyl ligands were lost in the Parent substitution reactions of Cr(CO)3' with the chloroalkanes. Apparently the metal insertion/charge transfer process for the bromochloroalkanes is more exothermic than for the chloroalkanes resulting in only X and BK abstraction and no Parent substitution. The reactions of 1,n-bromochloroalkanes allow the study of the preference for the site of attack (Br or C1) by the metal-containing anions. Table 6 lists 51 Scheme VI M (c a); + x (cu-1,); (CIO)° YCnHfi-M—X' -bCO Jl-anZnY shift (CID) a_ b ((‘:|O)‘I chuzfi-M—x (c0);cM—x" ch Fir—x" H (Parent substitution) (X obstruction) -dCO (c|°)o-d' H—M—x" (Hx obstruction) 52 2.80.8 2: : 2: c S 3 S a. -0808 2. 3 :0 mm -~80.o0 2. 3. 2.80.8 3 m: m: 2. $80.8 8 2: 8 c m. a: a 2. 2.80.0: 3:083: 5: 5.8.53: 8: 30:02: 8:02: :o.:< mam—.8280 u .5 «55580 I Ru G 3 N u 5 aggzoofioum-cJ 05 :8 80:02.. 5:838 5: 0:: 8: 0:: 8:095 8:285 0:: 8 :0 8:3. 8:055 00:55. 2F .o 036% 53 the average branching ratios for products in which either the bromine or chlorine is exclusively attacked (e.g. Parent substitution is not included) by various metal—containing anions with the 1,n-bromochloroalkanes. These values are calculated by summing the branching ratios for all the 1,n-bromochloroalkanes (n = 2 to 6) in which bromine or chlorine is exclusively attacked and calculating the average value. Also listed in Table 6 is a comparison of the average amount of HCl and HBr abstraction observed for the 1,n—bromochloroalkanes. The preferential attack at the chlorine on the molecule by the metal anion [may be expected since the electron affinity of chlorine is 6 kcal/mole greater than that for bromine. The opposite trend, preference for attack at the bromine, is actually observed as seen in Table 6. This trend may reflect the thermodynamics of the metal insertion into the C-X bond. The C-Br bond energy is 14 kcal/mole less than the C-Cl bond energy“. Metal-containing species having relatively high electron affinities such as Fe(CO)3, Co(CO)2, and Co(CO)3 (predicted from previous reaction trends) exhibit a preference for attack at the weaker C-Br bond as seen in Table 6. In contrast, Cr(CO)3’ which has been shown to have a low electron affinity exhibits virtually no preference for reaction at the Cl or Br end of the molecule due to the large amount of energy available from electron transfer to _e_i_t_l_1_e£ Br or Cl. Therefore, it appears that both the strength of the C-X bond which must be broken, and the difference in electron affinity between the metal species and the halogen (i.e., the energy available from the charge transfer process) determine the ’ preference of attack by the metal anion. The difference in the amount of HCl and HBr abstraction observed also reflects the corresponding electron affinity of the reactant metal anions. Virtually no preference is seen for metal species with a relatively low electron affinity (Cr(CO)3), but HBr abstraction occurs exclusively for metal species with high electron affinities (Fe(CO)3 54 and Co(CO)3). The following order of electron affinities of the metal species is suggested by the results in Table 6: E.A.(Cr(CO)3) < E.A.(Co(CO)2) < E.A.(CoCONO) AH+ + B CH+ + B -——-1-BH+ + C are observed, then the following must be true: P.A.(C) < P.A.(B) < P.A.(A). If the proton affinities of A and C are known, upper and lower limits on the unknown P.A.(B) may be deduced. Double resonance experiments may also be performed to verify these reactions. A similar process was applied to the negative ions in which electron transfer reactions were observed. If the ~ following reaction A+B-——*A'+B is observed, then the following assumption may be made: E.A.(B) < E.A.(A). 55 Species A may then be varied to determine upper and lower limits of the E.A.(B) (metal species in this case). In order to test this technique, the metal-containing anions from Fe(CO)5 (since these electron affinities are known) were reacted with Brz. Reaction 1? would be expected to occur since E.A.(Fe(CO)3) = 41.5- kcal/mole < E.A.(Fe(CO)4) = 55.3 kcal/mole < E.A.(Brz) = 57.7 kcal/mole. Fe(CO)3‘ (or Fe(CO)4") + Br2 ——> Brz' + Fe(CO)3 (or Fe(CO)4) (17) When this experiment was performed however, no Brg‘ was observed. It was thought that the Br2' formed may have dissociated to Br" (already present from electron impact on Brg) and Br'. Double resonance on Br' however showed no response from Fe(CO)3' or Fe(CO)4‘. Many ion/molecule reaction products besides electron transfer were observed between the iron-containing anions and Brz. Many other,anions were present in the mass spectrum due to electron impact on products from the neutral-neutral reaction of Fe(CO)5 and Brz (no double resonance responses observed). Apparently these processes interfere and prohibit the observance of reaction 17. Even if this technique had been successful, the applicability would have been limited since very few neutral molecules exist which have known high electron affinities. The product resulting from abstraction of _b_ofl1_ bromine and chlorine (MClBr‘) from the 1,n-bromochloroalkanes is observed for Co(CO)2'. Cr(CO)3‘,and Fe(CO)3’ (for n = 2 to 6) and is accompanied by the loss of all ‘ carbonyl ligands on the metal. In the reaction of 1,2-bromochloroethane, the metal may complex with the chlorine and bromine simultaneously (Scheme VII) resulting in abstraction of both Br and Cl and loss of a neutral ethylene molecule. A similar product is observed in the reaction of Co+ with 1,2-bromochloroethane25 but is not observed for the larger 56 Scheme VII - + . W00)CI . Br(CzH4) on B r ”I \ H mQ—M '2 0 \s H \ / 2 Cl \L— o C O 5‘ 5' H H 57 1,n-bromochloroalkanes (n >, 3). The two possible mechanisms for the reaction of the larger 1,n-bromochloroalkanes (n = 3 to 6) with metal-containing anions are shown in Scheme VIII. The first mechanism is similar to Scheme VII for 1,2-bromochloroethane except that the neutral lost is the corresponding cycloalkane (as opposed to the loss of a biradical species). The second mechanism involves insertion of the metal $13 one of the carbonyl ligands into a C-X (X = Cl or Br) bond. The other halogen (Y) is then transferred to the metal through a cyclic intermediate. The carbonyl ligand is now incorporated in the neutral product which is in the form of a cyclic ketone. In gas-phase positive metal ion reactions, evidence for active participation of carbonyl ligands in insertion processes has occasionally been observed12:22:29. To this point, it has not been necessary to invoke active participation of ligands on the metal anion to explain the products observed. Reaction 18 shows the formation of Br, Cl, and the cycloalkane from the 1,n-bromochloroalkane. Reaction 19 shows the incorporation of CO with the 1,n—bromochloroalkane to form Br, Cl, and a cyclic ketone. Cl(CnH2n)Br ——)— Cl + Br + cyclo-CnHZn (18) Cl(CnH2n)Br + CO ——> Cl + Br + cyclo-Cn+1H2nO (19) The difference in AB for these two reactions is given by: A Hdiff = Ang - AH19 = AH(cyc1o-CnH2n) - AH(cyclo-Cn+1H2n0) + A H(CO) Thermodynamic calculations58 indicate that the loss of the cyclic ketone is favored over the loss of the cycloalkane in the reactions of '. 58 Scheme VIII M (c 0);" + x (c H2)nY ’X'—.CH2 -x (CO)O" ( > " < ) \L CO ‘M' CH2 . a ‘ "‘2 I O ‘- / Y’ \cé HzC cu, (CHzln-z l-OCO l—(o-lfliO 0 g 6" 6" H —-CH 6" 6‘ / \ x—M_Y + 2C\ I 2 x-M—Y '5' Hzc CH2 (CH2)n-2 \ / (0:2)“ 59 1,n-bromochloroalkanes by 26.2 kcal/mole for n = 4, 9.23 kcal/mole for n = 5, and 3.27 kcal/mole for n = 6. As the alkyl chain length increases, the thermodynamic advantage of forming the cyclic ketone over the cycloalkane decreases due to the decreased difference in ring strain for the two structures. The same ligand effects are observed in the 1,n-bromochloroalkane reactions as seen in the chloroalkanes and alcohols: retention of NO in reactions of CoCONO‘, and unreactivity of the 17 electron metal anions. Also, the same trends in the branching ratios are observed as in the chloroalkane reactions; i.e., an increase in HX abstraction and decrease in X abstraction as the alkyl chain length (n) increases. B. Reactions of 1,n-Chloroalcohols The products and branching ratios for the reactions of 1-chloro-n-alcohols (1,n-chloroalcohols) (n = 2 to 6) with iron and cobalt-containing anions and chromium-containing anions are listed in Tables 7 and 8, respectively. These studies allow the determination of the preference for attack at Cl vs. OH by the metal-containing anions. Parent substitution (MCICnHZnOH') is observed for the three metal anions Cr(CO)3‘, Fe(CO)3‘, and Co(CO)2" with the loss of all carbonyl ligands. Since Parent substitution was not observed in the reactions of Co(CO)2‘ and Fe(CO)3‘ with alcohols, but was seen for all three metal anions with chloroalkanes, the mechanism for Parent substitution in reactions of 1,n-chloroalcohols presumably involves metal insertion/charge transfer into the C-Cl bond (and not the C-OH bond) with competitive ligand ‘ loss of the carbonyl ligands similar to Scheme I for the chloroalkanes. If the metal inserts into the C-Cl bond in the 1,n-chloroalcohols, other products typical of chloroalkane reactions would be expected. The abstraction of Cl and 1101 from the chloroalcohols by all three metal anions are observed as seen in Tables 7 and 8. These products also result from metal insertion/charge 60 m~. mv. ~c. vm. —~. n~. mu. N~. mm. an. we. am. am. he. on. pa. no. no. mo. mm. am. 0:. mm. mm. m—. an. m_. v~. up. nu. c~. ¢~. .vlum-nlufl 8:3. 05555 mm. bfi. on. «a. 0:. ac. we. vw. va. 808:8 p :38. 00: + 2. . 8: + -0-:N=:08 lull-J. 00: + 8: + -0:~::08 5": 00: + 5::0 + -888 lull. 05::0 + -8z2088 TL. 00 + 05::0 + -8808 T. 00: + 05::0 + .880 All 8:38. + L02088 ‘ll-I. 00 + 8:38. + -8008 5|L 00: . 50:38.88 ‘lll-I :0:38.8 + -2088 :5. lllllll 8:3.88 + -2080... 00 + 8: + 85.102085 00 + 05::0 + -_0:200.:... 8:38. + -.0200.:: 00: + 50:38.85 20:38.8 . -2085 9288: 5200—0520:.— 5... 22:: 28:00-58: 0:: :2. .0 280:3. 0. 03:... 61 an. we. Qin- av. pm. «Mm 808$...555 228:9. 2:: 0300:: 8: 20 :0. 05 3.8.0:. mz : :2 ‘l :0:%:88 + 82200.8 :2 I'll :0:.~:0.8 + $888 00 + 05::0 + -8:0z8‘|I|.-| 8.. 8:18. + -802008 Alllll :0:.N:0.8 + -02008 «n: 638:3— :o::..:8 p :39: 20.88.. :5. 0m 80:: 8: 20 :0. 05 8.8.0:. 5.. 0 :8. 8.60.: 0:88. :0. no.2... 820:9... .0 :5: 05 9.: 3858.80 :_ 83.; : :mzflll 853.88 + -2088 62 mm. 2.. vs. 2.. 2.. pm. mo. «c. we. ma. no. av. cw. 2.. we. no. a. 2.. no. 2.. z. a. a. :0:.N:8 + -8:08._0.A 8.. m... 8.. 00 + :0:.~:0. + -8200.8‘||L| 8.22.8.0 + 208.0 3. 00.. + :~::0 + -888 Au-II! 8. 2.. 8. 8:.«8. + -:0200.._0 lull! :. :.... S. 00 + 05::0 + -8:208._0 ‘1'! 2.. S. 00.. + 05::0... -8:._0 AIIIL 2. 2. a... 8.2.8. + -.0208.0|:l| 2.. :. E. 00 + 8:28. + -82088 [all], 8. 00 + 8: + -0§::0200.._0 ATIJ, 00 + N5 + -0..-:~::08208._0‘I||-_ 00 + N: + -0.-:~::0.0200.._0‘||| 00 + -:0:.~8.8208.0 all! 8. 00.. + «:N + -0...-:N::08._0‘||L 2. :2... a... 00.. + N: + -0.-:N::08.0 LTIIL a. 2. m... 00... + -:0:.~:88._0A-|-|| 8...”:88 + -2088 WI: :8 Mul: flow-am.- 333— Hafiz—95 29.00%; 52.- 8 95:380-335....8 no 8308.— .0 036,—. 63 transfer into the C-Cl bond followed by a 8-H shift and competitive ligand loss similar to Scheme 1. The chromium anion Cr(CO)3‘ reacts with the 1,n-chloroalcohols to yield products expected from the n-alcohol reactions. These include OH abstraction and elimination of one or two hydrogen molecules. The mechanisms for these reactions are similar to those for the reactions of n-alcohols in Schemes III, IV, and V. Since both OH abstraction and Parent substitution products proceed through the same intermediate (C-OH insertion) in the n—alcohol reactions (Scheme III), the Parent substitution ion for 1,n-chloroalcohols (with Cr(CO)3‘) probably consists of both insertion into the C-Cl and C-OH bonds. Products. predicted from the reactions of n-alcohols (loss of H2) are not observed for iron and cobalt-containing anions with 1,n-chloroalcohols. Products which are unique in the reactions of 1,n-chloroalcohols include the abstraction of both Cl and OH (MClOH‘) and the elimination of HCl (M(CO)x.aC2H4O‘). The abstraction of both Cl and OH is observed for the ‘ anions Co(CO)2‘ and Cr(CO)3' with 1,2-chloroethanol and is accompanied by the loss of all carbonyl ligands on the metal. This product is not observed for the larger 1,n-chloroalcohols (n >, 3). The mechanism for abstraction of Cl and OH is shown in Scheme IX. The metal does not insert into the C-Cl or C-OH bonds but complexes with both Cl and 0H simultaneously. Abstraction of Cl and OH with electron transfer to the ligands with higher electron affinity is accompanied by the loss of a neutral ethylene molecule similar to Br and Cl abstraction in the 1,n-bromochloroalkanes. The elimination of HCl from 1,2-chloroethanol is observed for the anions Fe(CO)3' and Cr(C0)3". This product is also not observed for the larger 1,n-chloroalcohols (n >, 3). Following metal insertion into the C-Cl bond and electron transfer to the chlorine (Scheme X), a six-member ring intermediate may form (only for n = 2). Cleavage of 64 Scheme IX M 00" + ( )x CI(C2H4)OH 6.— Cl\ I” H2 (cm-M. 'H. x iio’ 2 \L—xCO 6'.’ 6" H\ /H Cl—M-—0H + =C\ 65 Scheme X M (00)x + c: (CZH‘) 0H (CO)—M—-C—-H + HCI 66 the C-Cl bond and electron pair transfer results in loss of HC1 and structure R3 in which the charge has migrated to the alkoxy ligand. The reactions of 1,n—chloroalcohols indicate that the preference for attack (Cl or CH) by the metal anions is at the chlorine end of the molecule. The preference for chlorine is due to the higher electron affinity of CI (83.3 kcal/mole) compared to OH (42.2 kcal/mole). The thermodynamics of the metal insertion may also play a role since the C-Cl bond strength is 10 kcal/mole less than the C-OH bond strength“. The same ligand effects are observed in the reactions of 1,n-chloroa1cohols that were seen in the previous reactions. The reactions of the bifunctional molecules have shown that the reactions observed and the general mechanism proposed for the corresponding monofunctional compounds may be used in explaining products indicative of one functional group. The preference for attack at a particular functional group may be explained in terms of the difference in electron affinities of the metal-containing anion and the functional group and also the C-functional group bond energy. The trends observed for the preference of attack were used to propose an ordering of the electron affinities of the metal-containing species. Several unique reactions (not predicted from the monofunctional reactions) involving both functional groups were also observed. CHAPTER 6 REACTIONS OF NITROALKANES The ion/molecule reactions of metal-containing anions formed by low energy electron impact on Fe(CO)5, Co(CO)3NO, and Cr(CO)6 with a series of nitroalkanes and n-butyl nitrite were studied as a continuation of the previous work with alkyl halides and alcohols. The general mechanism proposed previously invoked the initial formation of a metal anion insertion/charge transfer - intermediate with insertion occurring at the C-functional group bond only. The exothermicity of the charge transfer process appears to determine the products which are observed. Also, the electron affinity data available enables one to predict the types of ligands which may be formed or lost (e.g., Cl has a high electron affinity and always remains on the metal in the alkyl halide reactions). This general mechanism may now be applied to the reactions of metal-containing anions with nitroalkanes. The corresponding positive ion reactions of Co+ and Co(ligand)n+ with nitroalkanes and alkyl nitrites was studied previously22 and is included in Appendix B. The nitroalkanes were extremely reactive compared to the reactivity of other monofunctional organic molecules. The products observed for Co+ with nitroalkanes resulted from metal insertion into C-H, C-C, C-N, 67 68 and N-O bonds. Many nitroalkane products were best explained via a "nitrite-like" intermediate - possibly indicating a metal-induced nitro-to—nitrite (RN02 to RONO) isomerization. Although metal anion insertion into the C-H and C-C bonds is not expected (from the results of the previous study of metal anion reactions), metal anion insertion into the C—N and N-O bonds in the nitroalkanes is possible. The relatively high electron affinities of N02 (53 kcal/mole) and O (33.7 kcal/mole) suggest that the nitroalkanes may be reactive with the metal-containing anions by insertion into the R-NOz and N-O bonds. A. Reactions of n—Nitroalkanes The ion/molecule reaction products and their branching ratios for the reactions of iron, cobalt, and chromium-containing anions with the series of l-nitro-n-alkanes (n-nitroalkanes) (n = 1 to 4) are listed in Tables 9 to 11, respectively. The number of products observed for Fe(CO)3‘, Co(CO)2’, and Cr(CO)3" (8, 9, and 17 products, respectively) correlates with the electron affinities predicted for these metal-containing species from the alkyl halide and alcohol reactions (E.A.(Cr(CO)3) < E.A.(CO(CO)2) < E.A.(Fe(CO)3)). A greater amount of energy is released in the charge transfer process for the metal species with lower electron affinities and thus a larger variety of products is observed. Substitution of the neutral nitroalkane for l to 3 CO ligands (Parent substitution) is observed for the reactive (i.e. non-17 electron) metal-containing anions for all three metals studied (reaction 20). M(CO)X- + CnH2n+1NOZ —+M(CO)x.aCnH2n+lN02- + 8C0 (20) The nitroalkanes displace 1 or 2 CO ligands in the reaction of Fe(CO)3’ and Co(CO)2", but all three CO ligands are lost in the Parent substitution products 69 68.800. :5. 0:80.... .0: 0.0 :0. 05 «0.00.0... .m.z : .20. 80008 0.800.... ..0. 00.»... 3.50:2: .0 .5... 0:. 0.3 3005:0880 :. 003.; a :.:.z AI-Il N0275220 .. -208... . 00 + :0 . -0z5::0208:.. .6 v . 02...5::0 + 8208:... , .... N.00 . 02 + m.5::0208o..A._ol-J 5 2 a. E .. 005 + 05: . -0z.-5::0.08£ ill, .... 00 + 500 + 02 + .-. 5::0883. 3 .. . 00.. + 02. + -0z.-5::0:...A|-l 5... 5 2.. .... 00 . 500 + 5::0 + 82:88:... All; v .5 00 . N00 + 75::0 + -0z.00.:..A||l ._ 00... + ...5::0 + -5029. .a S .. .. E 005 .. 0275::0 + 82:08:: All a. .. a. E 00.. + 0275::0 + -0~:o....f|., .. a a. 2 .... ... 00 . N00 + -0z..5::0.08:..A|-l 8 5 .5. :3. 5.. .... ... 005 + 500 + 8:25.102. All. a .. a. 2 00 + -Noz.+5::0208:... All; a .. 2 2 a 005 + -«0z..5::0.08o.. Allrroz..5::0 . -208»... 108228.. I III . :u: :u: an: «n: .u: 0.5.0.5: 8008.. 8 3.35.... ESZ-fiigoluu and ill-9.3.2:.— 5.3 g :58ng .0 I333 . a 03-h. 70 am. am. 20:02.8: 0": 6.8.88. 2:: 0280:: .0: 0.0 :0. 05 8.8.0:. ....2 0 62200.00 .0 3:058... .08.... 8.808 :8... 00:08.88. 0. 0:0 03830 0. 030000.. 005208.. 03:00 0: ...... 60.0008 0..- n:0. 8:008 000:... n 9N Na 2 ..N ..— ON N.” a. : 0n ..— Z N. a. ... NN fl“: 803.8 «N av N. A... a. ..n I 0.... 0— an N. n— n. on N.” n. 6:0. 8:008 0.8000. ..0. 00...... 3.88.: .0 :5... 0:. 0.... 008588: :. 8:.0> 0 0...: All. N02.+:..::0 + 82200.8 0...: All N0205::0 . -2008 = 152:0 + -.02.02..00.8 All. >. 02 + 00 + 2.5::0 + -50.:8 All . 02 + 02.-5::0 + 85:80.8 .5 . 00 + 0275::0 + 8.028 All . 02.+5::0 + -.0.02008 A-l 00 + -.02..5::0.0z.8 All- N02.5.0.0 . -02008 .... N00 + 02 + -.+5::0.00.8 8 . 00. + 0.: + -02.-5::08 AI! = 00 . 5::0 + -~02:.00.8 All. .. 00. + 5::0 + -.02:8 Al .... 00 2.00 + 5::0 + -02:8 All! ... 00. + 02...5::0 + 85:81. .... ... 00 + N00 +-02.+5::08A|l, . 02.+5::0 + ..02088AI 00 + -.02.+5::0.00.8 _ 005 + -.02.+5::08UI N021528 + -2088 0.5.250 8.88.. 2:28.53... Eggs-$.23?” :5 8.2-: 5... 38:2 55:09.8 .o 80:8. .2 :.::... 71 av NN 25020228.. v": . n N .. n q o .. a 8 v ... n n e ... 3. _ 5 n N 3. ... N n 2.. 5 2 0 v 2. N H".- 8:-u 8 oN 0.m. mN : N": 3 ON a. n— .5 = = = >— >. >— >— >. ... .... ._= ... E .— 05.0.50 00 . .00 + 02 . -.+52:0.00.8 ..o 00.. + 0.2 + 02752208 All! 00.. + n-52..0 + -502028 AIL 00.. + 752:0 + -502528 All, 00.. + 5::0 + -8228 All 00 + ~+52:0 + -~.02200.8 All.- 005 + $5220 + -~02.00.8 All 00.. + 2752.8 + -5288 All 00.. + 252:0 + -2508 AIIL 00 + 2.+52=0 + 30200.8 I 005. + 2.25220 + 30.00.81 00.. + 275220 + -508 Al. 00. + 02 + 815220.088 AI... 00 - N00 + T52=0 + 02200.8 All, 00. + N00 + 82:52:08 4|! 025220 + 20200.81, 00 + 0252.8 + -20200...0A.|L 0215220 + 0200.811 00 + 02:52.8 + -0200...0 All! 00.. + -«0z_+52._0._0 AL 30.0228...— §.IINI.>50IIN i 83.2.... 5.! 28.:< 53.808.230.35 .0 0.3.88.— . 203.5. nusaaucoo —~.0223F II N0275220 + -2008 €8.08..— 72 cc. 250202.20: cu: NN : co. an a 7-: an: 9N a. an av 88.800.— 2:0 00.00:: .0: 0.0 :0. 0:. 8.8.0:. ....z a .20. 8300.5 0.20000. ..0. 00.8.. 0:20:02: .0 Es... 0:. 0.8 0000:8080 :. 002.; 0 .... ..— 0.5.035 08.003085 2.2.2 All N02:52.8 + 200.8 00. + N00 + 02:52:08 All, 00 + 025220 + 20200.8 All 00 + 0225220 + 8200.8 All, 00.. + -~.02.+52=0.00.8 All 00.. + -~0z.+52._08 AIILI- N027.5220 + 200.8 €3.08.— 00552—00 : 03-h. 73 with Cr(CO)3‘. This trend correlates well with the electron affinities (and thus exothermicity of the charge transfer processs) of the three metal species mentioned previously. This suggests that a metal insertion/charge transfer mechanism may account for the Parent substitution products. In the chloroalkane and alcohol reactions, the Parent substitution product structure presumably does 2g consist of an intact organic molecule as a ligand, but was predicted to result from metal insertion/charge transfer into the C-functional group bond. This was proposed due to the‘negative electron affinity of the neutral molecule and due to the fact that all the products observed could be formed through this common intermediate. The electron affinity of the neutral nitroalkanes, however, is positive (e.g. E.A.(CH3N02) = 9.2 kcal/mole) Suggesting that the Parent substitution product structure may be either an intact nitroalkane on the metal or a metal insertion type structure. Two metal insertion structures may be predicted for the reactions of nitroalkanes with metal-containing anions. Intermediate I results from metal insertion into the N-O bond. This structure is predicted to be stable due to the relatively high electron affinity of O (33.7 kcal/mole). The electron affinity of CnH2n+1NO, however, is not known. The electron affinities of N02 (53 kcal/mole) and CnH2n+1 (~20 kcal/mole) suggest that intermediate II (C-N insertion) would also be stable. ‘3 ‘3 cnHznw ”N‘Y-‘O CnHzml' [Yr-N'O (c0), (cc)x 74 In the nitroalkane discussion, structures will be depicted with the negative charge on the metal center. However, recall that charge transfer/charge delocalization over the various ligands present still occurs as in the alkyl halide and alcohol reactions. Metal insertion into C-C and C-H bonds are {E predicted for the nitroalkane reactions since they were not observed in the alkyl halide and alcohol reactions. . The proposed intermediate structures which lead to the various product ions are also listed in Tables 9 to 11. The mechanisms for the formation of product ions from intermediates I and II are shown in Schemes XI and XI], respectively. A large number of products, however, do not result from intermediates I and II. These products result from intermediates III and IV as shown in Schemes XIII and XIV, respectively. ('3 o - I CnHznu-k'd -NO+ anznfl-N-n'r-o + WO’x-a (C0)“, 1:: III In intermediate III, the nitroalkane exists as three ligands (CnHZnH: NO, and 0) on the metal. Intermediate III may result from rearrangement of intermediate I (N-O insertion) through an alkyl migration as shown in Scheme ‘ XIII. An analogous process has been observed in the reactions of Fe+ and ketones20 (reaction 21). 75 H \ .00. _ ...x 0026. - «8+ 3.807-22- 252:0 Ale-0--.). 2-2.220 IIJ _. 0 o o 3.2%. of . 0.28.32 202.82: 0.Lll ENE-220275220. Adam-fl TCNITCW ......n: ...... :0 + 0.x.ou.-zn..fl.__,__ 751720 - _ 0 a 2.0”... w... 02 52 c0 + :0--2 - Z60. 00.. + 20 ”25.2 .... ....m In .6”... H 000 +02 152:0 + 9.2 -..-:8. 0--.). -...- .Euxco ill- 0 .«ozzfi 2 c 0 + .80.- -2 .2 05200 76 :N C can o I 3. # 2 Tamruéuw 9.x 3. 2 000+ OIZILZ......= __ . Us... 0 I _ I : O .+CN C I I IOIX £5 00 + I 0+0 a -.2 A00. rum 0 j, H :8. q _ 92 -..2 .2525 __ o Nozzcmzco + 52 ..-s. 5 255m 821.-.... - N....8. .512...» . N... am. ...... CNICU ¢ 02-02 .- N-IOU. ‘I GO + OZI- ..... I‘ll-Ill... . u-: :-q x z ... 8 ...-5.252.... 62-2-8.1 ...: 8..-... . oz . z8......- Eazcu fl ‘ .-...ow. ..x.o.._.. 8 ..-...52- -..... - ..cuzau ..l. ~852 --.2- Zara”. .l. o2--2- :52... .J . 0.1000. ANOU. 3.8. 3.8. ... 7 of .02.. ...-.75.. .61] oz-- 2-... 52.5. 7 8"... a. n- ...—KIN .c 9:8. ......8. 2 ..-...8. w . . . u- .1 .....u 75:20.13: - - -..... ...... . 1oz}... .. 2.6. 2.5- : oz J. ”.....-zIIWQ... 10 IO 10 oz .2522”. . -.2 -..-28. i ..u E ow 38¢. Out-0.x. + 02 + IQ.‘ Izcmzcutlouo «OZI-e.‘ ...-.5130 ill-III O O H ...8. II - 1 o I E I..CNI CU O H ...8. .ICNICU ‘11 . IZI.$. . “w 0 EX 052.0% 78 .ICNI—ICU . 2.-.;ch . :0--.2 -o:-.|I...o--<.,....“u2__ :o 2.... ...-m 2.525 . :0 --.2 -0...on .. Io --.... 2.5-7522... z :5: ..w + o .....-o-Jou. o ...... :-m. H H a-x.o.u. .8“... OUU+OII§IZIL+CNICU All Ollfi I-op—NICU O 2% 059.8 79 o o + c g cnzca ——> CH Fe+ ('3 Fe + H _ - 3 3- _ 3 'L’\CHzCH3 ./ (21) CHJ-Fe+-CO + ' I -—> FoCO + cup-120.4, CHzCH3 Intermediate III may also be formed from intermediate II as a result of the oxygen shifting onto the metal center (Scheme XIII). Several products observed in the reactions of nitroalkanes result from an intermediate in which both oxygens are present on the metal as shown in intermediate IV. Intermediate IV may be formed from intermediate I as shown in Scheme XIV. A similar intermediate was observed in the reactions of Co+ with nitroalkanes“. The reaction products observed from the four intermediates will now be discussed in detail. Once intermediate I is formed (Scheme XI), it may undergo a ligand loss process to form M(CO)x.aO‘. This product is observed for metal-containing anions from all three metals studied. The retention of the 0 ligand is. expected since it has a relatively high electron affinity (33.7 kcal/mole)(E.A.(CnH2n+1NO) is unknown). This product may also be formed through intermediate III (Scheme XIII) with the loss of the N0 and alkyl (CnH2n+1) ligands. Note that the electron affinities of the two ligands which are lost (0.5 kcal/mole and ~20 kcal/mole respectively) are less than the electron affinity of O (33.7 kcal/mole) which is retained. In some cases however, the formation of M(CO)x.aO‘ M proceed through intermediate III due to the larger number of ligands which would be present on the metal in intermediate III compared to intermediate I, i.e., in intermediate III the nitroalkane is broken 80 into three ligands on the metal but in intermediate I it exists as only two ligands. For example, Cr(CO)20" may proceed through intermediate III since only five ligands would be present at any one time. The product Cr(CO)3O‘ cannot result from intermediate III since it requires six ligands to be present on the metal which violates the 18 electron rule. The corresponding product ion M(CO)x.aCnH2n+1NO’ (i.e., _lgs_s_ of O) is observed for Fe(CO)3‘, Co(CO)2', Cr(CO)3‘, and Cr(CO)4' (n = 1 to 4). This product is not predicted to result from intermediate I since it would require loss of the oxygen ligand which has a relatively high electron affinity. The active participation of carbonyl ligands in the metal insertion process has“ been suggested in both positive and negative metal ion reactions. If the metal anion and one of the carbonyl ligands inserts into the N-O bond, intermediate 1' (Scheme XI) may be formed. In this structure, the oxygen is never a ligand on the metal but instead exists as a 002 ligand on the metal. The 002 ligand is then lost due to its negative electron affinity (-13.8 kcal/mole). This product . ion may also result from an intermediate similar to intermediate III. If the oxygen in intermediate 11 shifts onto a CO ligand instead of the metal then a C02 ligand is formed. The COZ ligand is then lost to yield intermediate III' as seen in Scheme XIII. A a—H may shift onto the metal in intermediate I to yield structure k4 (Scheme XI). The O and H ligands on the metal may exist as a single OH ligand (E.A.(O) = 33.7 kcal/mole, E.A.(H) = 17.4 kcal/mole, and E.A.(OH) = 42.2 » kcal/mole) as shown in structure 1‘51. The electron affinities suggest that O and H may exist as either one or two ligands on the metal. Structure 1"4 may proceed through a ligand loss process to yield M(CO)x.aOH‘ (retention of OH, for Cr(CO)3") or M(CO)x.aCnH2nNO' (loss of OH, for Fe(CO)3"). Since these products (formation of an OH ligand) are observed for nitromethane (n = 1), 81 they gan_no_t_ proceed through intermediate III since 92 s-H’s are present when n = 1. A second hydrogen may shift onto the metal in structure R4 to produce an H20 ligand (E.A.(HZO) = 0.0 kcal/mole) which is lost to yield the product ion M(CO)x_aCnH2n-1NO'. This product is observed for metal-containing anions from all three metal carbonyl compounds as seen in Tables 9 to 11. The formation and subsequent loss of H20 may suggest that O and H are present as a single OH ligand on the metal. Since this ion is observed for nitromethane (n = 1), then it must proceed through intermediate I and 1121 from intermediate III (similar to the above argument). In one case, Fe(CO)CnH2n_1NO“, this. product ion is flit. observed for n = l but only for n = 2 to 4. Therefore, this ion may result from either intermediate I or III (Scheme XIII), but note that the ion Fe(CO)2CnH2n+1‘ (loss of N 02) is isobaric with this product ion. The product ions M(CO)x.aN02" and M(CO)x.aHNOZ' are formed through intermediate 11 (C-N insertion) as seen in Scheme XII. Once intermediate . II is formed, the alkyl ligand may be lost to form M(CO)x.aN02' since the electron affinity of CnH2n+1 (~20 kcal/mole) is less than the electron affinity of N02 (53 kcal/mole). This ion may also be formed through intermediate 111 but N02 then exists as two ligands (O and NO) on the metal. The electron affinities (E.A.(O) = 33.7 kcal/mole and E.A.(NO) = 0.5 kcal/mole) suggest that the product ion resulting from intermediate 11 delocalizes the negative charge more than in intermediate III, thus M(CO)x-aN02‘ presumably is formed . through intermediate 11. Note that FeNOz" is observed for n = 2 to 4 but not for n = 1 (Table 9). There is no apparent reason for this product ion not to occur for n = 1. Also note that all three carbonyls are lost in the formation of FeNOZ" but only one or two carbonyls are lost in the formation of Cr(CO)x..aN02' (Table 11). The loss of all three carbonyl ligands is unexpected 82 in the formation of FeNOz" since the reactions of Cr(CO)3‘ are always more exothermic than Fe(CO)3' (the charge transfer/delocalization process is more exothermic since E.A.(Cr(CO)3) < E.A.(Fe(CO)3)). Therefore, the ion FeNOz‘ is probably not the correct product assignment with the isobaric ion FeCO(H20)‘ being the more probable product ion structure. The product ion M(CO)x_aHN02' is observed for Co(CO)2' and Cr(CO)3‘ with the n—nitroalkanes (n >, 2). This suggests a mechanism in which a 8-H shifts onto the metal in intermediate II to yield structure k5 (Scheme XII). The alkene formed is then lost due to it's negative electron affinity (e.g. E.A.(CZH4) = -35.7 kcal/mole). The two remaining ligands (H and N02) both have positive electron affinities (17.4 and 53 kcal/mole respectively) to help delocalize the negative charge. The product ions M(CO)CnH2n+10' and M(CO)x.aO" may be formed through intermediate III by loss of NO or NO and CnH2n+1 ligands respectively as seen in Scheme XIII. The ligands which are lost (NO (0.5 kcal/mole) and CnH2n+1 (~20 kcal/mole)) both have lower electron affinities than the O ligand which remains on the metal (33.7 kcal/mole). Two H-shifts may occur in intermediate III to yield either structure lg or R7 in Scheme XIII. Structure “5 loses the CnHZn-l ligand to yield the product ion M(CO)x.aH20" which is observed for Fe(CO)3' and Co(CO)2' with the nitroalkanes (n >, 2). The electron affinity data suggests that H20 (0.0 kcal/mole) exists as OH (42.2 kcal/mole) and H (17.4 kcal/mole) ligands in order to delocalize the charge as in structure {6. - Structure k7 results from the formation of an H20 ligand which is then lost to yield M(CO)x.aCnH2n-1NO‘ which was discussed previously. The CnH2n-1 ligand has a positive electron affinity (e.g. E.A.(CgHs) = 12.5 kcal/mole) to help delocalize the charge. Several products result from intermediate [11' (i.e. following the loss of 83 O as C02) as seen in Scheme XIII. The loss of the N O ligand from intermediate III' yields M(CO)x_aCnH2n+1" which is observed for metal-containing anions from all three metals studied. This corresponds to a net loss of N02 from the nitroalkane. Although it is more straightforward to lose N02 from, intermediate 11 (C-N insertion), this process is £01; predicted since E.A.(NOZ) > E.A.(CnH2n+1), i.e. the ligand with the higher electron affinity would have to be lost. If this product is formed through 111', however, the N02 is lost as 002 (-13.8 kcal/mole)‘ and NO (0.5 kcal/mole) which have low or negative electron affinities. The loss of CnH2n+1 from intermediate III' yields the product ion M(CO)NO" which is observed for Cr(CO)3' and Fe(CO)3'. When n >, 2 for the n-nitroalkanes, a 8-H may shift onto the metal in intermediate III' to yield M(CO)x.2HNO‘ with loss of the corresponding alkene (E.A.(C2H4) = -35.7 kcal/mole) which is formed as seen 'in Scheme XIII. Note that HNO may exist as one ligand on the metal (E.A.(HNO) = 7.8 kcal/mole) or as H (E.A.(H) = 17.4 kcal/mole) and NO (E.A.(NO) = 0.5 kcal/mole). All of the reaction products observed with Fe(CO)3‘ and Co(CO)2' may be explained through intermediates I, II, or III. Approximately 4596 of the products from Cr(CO)3' result from intermediate IV (Scheme XIV) in which two oxygen ligands are present on the metal. The electron affinities of the ligands present in intermediate IV (E.A.(O) = 33.7 kcal/mole and E.A.(CZH5N) = 43 kcal/mole) indicate that this structure should be stable. Loss of the CnH2n+1N ligand yields the product M(CO)x_aOZ'. This is not expected if . the electron affinities of the two ligands are considered (i.e. E.A.(CnH2n+1N) > E.A.(O)). If the electron affinity of the Cr02 species (55.3 kcal/mole) is considered then the CnH2n+1N ligand _ig expected to be lost since it has a lower electron affinity. The product ions MOgH' and MOZHZ‘ result from one and two 8-H shifts from intermediate IV respectively as seen in Scheme XIV. It 84 is interesting that products resulting from intermediate IV occur almost exclusively for Cr(CO)3". Since Cr(CO)3" is predicted to have the lowest electron affinity of the metal-containing anions studied, the largest amount of energy released in the charge transfer (delocalization) process occurs for Cr(CO)3". Therefore, a large amount of energy appears to be required to form intermediate IV from intermediate I since it only occurs for Cr(CO)3". B. Reactions of 2—Methyl—2-Nitropropane The reactions of the metal-containing anions from Fe(CO)5, Co(CO)3NO, and Cr(CO)6 with 2-methyl-2—nitropropane are also included in Tables 9 to 11 respectively. The 2-methyl—2-nitropropane appears to be less reactive than the corresponding l-nitrobutane with only two reactions observed with Fe(CO)3" and Co(CO)2' and eight reactions with Cr(CO)3’. The only product ion observed for Fe(CO)3' besides the Parent substitution product is FeCnH2n+1NO‘ (loss of 0) resulting from either intermediate 1' or 111'. If this product is formed through intermediate III', the formation of Fe(CO)HNO‘ would be expected since B-H's are present and it is the largest product observed for l-nitrobutane. This product is not formed suggesting that FeCnH2n+1NO‘ is formed through intermediate 1'. The other products expected from intermediates I and I' are only minor products for l-nitrobutane and are not observed for 2-methyl—2-nitropropane. The two products formed from the reaction of Co(CO)2' with 2-methyl—2-nitropropane are Co(CO)20' from intermediate I and Co(CO)HN02' from intermediate 11. This is reasonable since these are the only two products observed from these two intermediates for the n-nitroalkanes. Also, there are fi-H's available to shift in intermediate II for 2-methyl—2-nitropropane to produce Co(CO)HN02". The reaction of Cr(CO)3‘ with 2-methyl-2-nitropropane yields the product 85 ions Cr(CO)20" and Cr(CO)3O" from intermediate I and CrCnH2n+1N'O‘ from intermediate 1'. The product ion Cr(CO)x_aOH' seen in the reaction of l-nitrobutane is M expected in the 2-methyl-2-nitropropane reactions since there are no B-H's present in intermediate I with 2-methyl—2-nitropropane. A small amount of Cr(CO)3OH" is observed, however, which may proceed by a H-shift through a six-member ring intermediate as shown in structure 133. 0 ll "' CO 053/N—¥_( )x .001 0 CH3 IC-H" H 1.8 The largest product formed through intermediate IV in the l-nitrobutane reactions is CrOsz‘. In the 2-methyl-2-nitropropane reactions, however, there are E B-H's present in intermediate IV and thus, the only product resulting from intermediate IV is Cr(CO)202‘. The product ion Cr(CO)x_aN02‘ is formed through intermediate 11 in the reactions of 2-methyl—2-nitropropane. Although B-H's are available to shift 8 to produce Cr(CO)x_aHN02‘, these B-H's are terminal hydrogens for 2-methyl—2-nitropropane. The lack of terminal B-H shift is not unexpected since the terminal B-H's in the reaction of nitroethane also do M shift to yield Cr(CO)x.aHN02' (Table 11). C. Reactions of n-Butyl Nitrite The ion/molecule reaction products and branching ratios for the reactions of n-butyl nitrite (C4H90NO) with iron, cobalt, and.chromium—containing anions are listed in Table 12. In the corresponding positive ion reactions of n-butyl 86 cc— 5.00:0 @— @— n— X” N. mm n. m— _N -2085 ...: S ... 3 -2088 -0208 .«.00.8 -2003. 8:3. 2.5.3.0 ......0 . -~0z...8.2 00 + 02 + N... + 022.028.... 00 . 02: + N2 + 002.028.: 00 + 02 + N: + 0.2.00.8... 00 . 0.2.0 + 022.00.: 8 + 02.. + 002.6022 00 + 02.. . 002.0300... 00.7... + 02 + 0.2.0209: 00..-... + 02 + 0.5.0.8.: 0.2.0 + -0228... 00 + ......0 + -3228... mm<<<<<<<<N oEosum 90 Scheme XVI M- - (C0)x + C4 HSONO (C0)x l C4H9- M- - ONO B 1 ¢ [ta-H (c0); M’-ONO+ c4149; shift H I (<|:o>,‘._cJ ll --- M"- ONO + 0C0 H-C l l H Cszl (C0)“, l H-M'-ONO+C4H8 91 (e.g., l7 electron species are unreactive) including further evidence for the active participation of the carbonyl ligands in the reaction mechanisms. The metal-induced nitro—to—nitrite isomerization which was observed in the positive metal ion reactions of Co+ with nitroalkanes is 1191 observed in the reactions of metal-containing anions. D. Further Insights into the Metal Insertion/Charge Transfer Mechanism There are two possible explanations for the lack of C-C and C-H insertion in the reaction of metal-containing anions with organic molecules. The first involves the stability of the three metal insertion intermediates following insertion into C-C, C—H, and C-X bonds as shown in structures 3’4 to 16 respectively. R'- MT—R"x '—R"X R—M‘Z—x 24 26 N Structure %6 (C-X insertion) is predicted to be more stable than structures a4 and as due to the relatively high electron affinity of the functional group and thus greater charge delocalization. (This analysis contains a number of assumptions, e.g., without charge delocalization, all of the metal-ligand 0 bonds in 2"} to ?»6 are of approximately the same energy.) The other explanation may be that metal insertion occurs via complexation/charge transfer/fragmentation as shown below (i.e., charge transfer occurs prior to metal insertion). 92 M"+l—>M--l —9 m: In order to differentiate between these two processes, one must determine when in the mechanistic sequence the charge transfer process occurs. The latter process can be considered analogous to the dissociative electron capture process which occurs during low energy electron impact. The electron impact negative ion mass spectra of the organic molecules studied here are consistent with a mechanism in which the charge transfer process occurs prior to the metal insertion process. For example, the negative ion mass spectra of alkyl chlorides" at high electron energies (> 50 eV) include Cl” and several alkyl fragment ions (e.g. CZH", Cg“, and CH”). At lower electron energies, however, the only ion present in the mass spectrum is Cl‘, i.e., electron attachment leads to the formation of Cl‘ and R-. This parallels the metal insertion/charge transfer process in which the R-Cl bond is broken and M—Cl‘ and M-R bonds are formed. The major negative ions present in the 90 eV electron impact mass spectra of alkyl alcohols78 are (M-H)‘, OH", and O“. The formation of the ion/radical pairs ((M-H)‘ and H-) and (OH‘ and R-) parallel metal insertion/charge transfer into the O-H and C-OI-I bonds of n-alcohols, respectively. The predominant negative ions in the low energy (4.5 eV) electron impact mass spectra of nitroalkanes79 are N02", 0', and CN". At higher electron energies more alkyl and rearrangement type ions are present. The formation of the ion/radical pairs (NOg‘ and R-) and (0' and RNO-) in the electron impact process again suggests that the charge transfer process may occur prior to the metal insertion into the R-NOZ and N-O bonds of nitroalkanes. 93 Thermodynamic calculations58 for the dissociative electron capture process (reaction 22) yield values for A H22 of N-6 kcal/mole for X = C1, M15 kcal/mole for X = Br, and~+48 kcal/mole for X = OH. RX + e- ———a»R- + x- (22) These results correlate well with the preference for site of attack which was observed in the chemistry of the bifunctional organic molecules. For example, metal insertion/charge transfer into the R-Br bond was preferred over the C-Cl bond in 1,n-bromochloroalkanes, and C-Cl insertion occurs preferentially over C-OH insertion in 1,n-chloroalcohols. These trends may be predicted by the exothermicity of the dissociative electron capture process in reaction 22. A similar analysis of the formation of RO“ and H- from n-alcohols, i.e., a process similar to metal insertion/charge transfer into the 0-H bond, yields a value of AH =4: +60 kcal/mole which is approximately 12 kcal/mole more endothermic than for the formation of OH‘ and R-. This would suggest preference for metal insertion/charge transfer into the R-OH bond over the 0-H bond of n-alcohols. The opposite trend, however, was observed in the chemistry of n-alcohols with metal-containing anions. This may reflect the effect which the formation of different bonds (e.g., R-M-OH and RO-M-H) has on the preference of attack by the metal anions in the metal insertion/charge transfer process. CHAPTER 7 CONCLUSIONS A general mechanism has been proposed for the reaction of metal-containing anions with polar organic molecules. Low energy electron impact on Fe(CO)5, Cr(CO)6, and Co(CO)3NO was utilized as the source of the metal-containing anions in order to obtain a significant amount of the more reactive (non- l7-electron) species. The reactions of alkyl halides and alcohols with these metal-containing anions suggest that the initial step is the formation of a metal insertion/charge transfeddelocalization) intermediate into the C-functional group bond. The negative charge may be transferred to a ligand (e.g., Cl) which has a relatively high electron affinity. This is followed by possible rearrangement (e.g., s—H shift for CnH2n+1X when n >, 2) and competitive ligand loss similar to that observed for positive metal ions. Evidence for the occurrence of 8-H shifts in the reactions of metal—containing ' anions has been reported here for the first time. The ligand loss process observed for metal anions is quite different from that for the corresponding positive ions. For example, ligands which are I-donors (e.g., olefins) are strongly bound to the metal in the reactions of positive ions but are always lost in the reactions of metal-containing anions. 94 95 The exothermicity of the charge transfer/delocalization process appears to determine the products which are formed. The trends in reactivity of the various metal-containing anions results in a possible ordering of the electron affinities of the metal-containing species: E.A.(Cr(CO)3) < E.A.(Co(CO)2) < E.A.(CoCONO) < E.A.(Co(CO)3) < E.A.(Fe(CO)3) = 41.5 kcal/mole. This ordering of electron affinities is consistent with all of the reactions observed in this dissertation. For example, Cr(CO)3' is always the most reactive species since it is predicted to have the lowest electron affinity and thus the greatest amount of energy is released in the charge transfer process. The general mechanism which was proposedcan be successfully applied in explaining the reactions of several bifunctional organic molecules and nitroalkanes. The products observed in the chemistry of bifunctional organic molecules (e.g., 1,n—bromochloroalkanes) indicates the preference of attack at one functional group over another. The preference for attack appears to depend on both the electron affinity of the functional group (and thus the energy available in the charge transfer process) and the bond energy of the C-functional group bond which must be broken on insertion. Products are also observed which are indicative of metal insertion with m functional groups. The reactions of n-alcohols and nitroalkanes are unique in that bonds within the functional group are also attacked by the metal anions. Products are observed in the reaction of nitroalkanes which result from rearrangement following formation of the metal insertion/charge transfer intermediate. In ‘ contrast to the corresponding positive metal ion reactions, there is no evidence in the chemistry of metal-containing anions for metal insertion into C-C or C-H bonds. The lack of C-C and C-H insertion may be expected if the metal insertion/charge transfer process is compared to .the processes which occur in the dissociative electron capture negative ion mass spectra of the organic 96 molecules. The similarities in the metal insertion/charge transfer process and the dissociative electron capture process suggest that the charge transfer may occur prior to the metal insertion. Several ligand effects are observed in the reactions of metal-containing anions. The reactivity of the metal decreases as the number of ligands increase. The l7-electron species (M(CO)n-1’) from electron impact on a metal carbonyl containing n CO ligands are generally unreactive as reported in earlier studies. Metal species with fewer ligands however are very reactive. Products are observed which indicate the possibility that the carbonyl ligands which are present may actively participate in the metal insertion and rearrangement. processes. This has also been suggested in the reactions of positive metal-containing ions. The original premise for the application of these metal-containing anions as chemical ionization (Cl) reagent ions appears to be somewhat limited. The Parent substitution products observed may provide information on the molecular weight of the sample molecule. Functional group information can be obtained through the products observed from metal insertion into the C-functional group bond (e.g., M(CO)xCl"). The only structural information which is available from the reactions of metal-containing anions is the presence or absence of products resulting from a 8-H shift rearrangement depending on the length of the alkyl chain. The reaction of the corresponding positive metal ions results in a larger variety of product ions and thus more information on the analyte molecule. One example is the presence of products resulting from metal insertion into C-C and C-H bonds which are M observed in the anion reactions. Based on this work, the anion which would provide the most information in a CI experiment would be Cr(CO)3‘ since the largest variety of products are usually formed with this metal-containing anion. 97 The results presented in this dissertation suggest several possible future projects on the reactivity of gas-phase metal anions. If the electron affinity of the metal-containing species could be experimentally determined, they could be compared to the ordering which is suggested in this study. The experimentally determined electron affinities would also be beneficial in predicting the products observed in the reactions of metal-containing anions with other organic molecules. The determination of the reactivity of the bag metal anions would also be very interesting. The use of collision induced dissociation and ion ejection techniques with Fourier transform mass spectrometry (FTMS) allows one to form the bare metal anion, M', from the corresponding metal carbonyl, M(C0)n. The reactivity of the bare metal anions with the organic molecules studied in this dissertation could be compared to the reactivity of the metal-containing anions, M(CO)x‘. The effect that the presence of the ligands have on the reactivity could then be determined. Squires80 has reported that the bare metal anion Cr‘ is unreactive towards propene, but Cr(CO)3H' reacts to eliminate H2. The presence of the ligands in the reactions of metal-containing anions may be necessary either to remove the excess energy from the charge transfer process (since carbonyl ligands are frequently lost) or to actively participate in the metal insertion or rearrangement process as illustrated in this study. APPENDICES APPENDIX A Appendix A Table 13. Pertinent Electron Affinities (E.A.) (kcal/mole)ll Anion (M') E.A. (M), kcal/mole Reference H" 17.4 56 Cl' 83.3 53 Br" 77.6 53 131-2“ 57.7 53 Cr‘ 15.4 74 Cr02" 55.3 ' 72 Co' 15.3 75 Fe" 3.8 52 Fe(CO)‘ 29.1 52 Fe(CO)2’ 28.1 52 Fe(CO)3' 41.5 52 Fe(CO)4' 55.3 52 FeH“ 21.5 40 CO" _ 31.6 76 002- -13.8 66 0" 33.7 64 HO" 42.2 53 H20' 0.00 67 CH3o- 36.7 60 021150” 39.8 61 n-03H70‘ 41.2 61 Table 13 continues 98 99 Table 13 continued n-C4H90' 43.8 62 CH3C1' -79.6 48 CH3' 26 55 C2H5' 23 55 C3117" 16 55 C4H9' 15 55 C2H4' -35.7 57 C3H5' 12.5 69 CH3N02' 9.2 65 NO" 0.5 68 HNO' 7.8 70 N02' 53 63 C2H5N' 43 71 8Electron affinity is defined as All for the reaction M' —-)- M + e“ APPENDIX B 100 rlnted from the Journal of the American Chemical Society. "84. 106. 6125. Rep Copyright 0 I984 by the American Chemical Society and reprinted by permission of the copyrkht owner. Gas-Phase Reactions of Co+ and Co(ligand),,+ with Nitroalkanes C. J. Cassady.’ B. S. Freisar." S. W. Meflraey.‘ and J. Aflsoe“ Com-mm from :10 Department of Chemistry. Purdue University. West Lafayette. Indiana 47907. and the Departure. of Chemistry. Michigan State University. East Lamtng. Michigan 48824-1322. Received Mart-II 28. I984 Ahulncc‘lhegss-phasecbun'stryofCa°andianofthotypeCo(ligaad). withssariesafnitraalkanesisprmented. Also. lucamparhaareactionsolmethylniuiteandl-batylniuiteaminduded. TheCo’ionresctswithnitroalkanesbytnserting intoC-hLC-C.C—N. andN-Obands. Manymumlhnepadtmanbebmtexplamedviaa‘mhhn'intuundiate-prnibly indicatingametal-induced nitrotonitritc isomerintion. Collisianinduceddbaciation analysisof(i)primary reaction products and(fi)pmducuofsubsaquontreacuannusedwmggestprodnctionsmum Withligandoptesentonthemetal dramatic changes of reactivity are abound. While nitroalkane react by simple ligand displacement with Co(ligand).’ typical of manafunctianalalhanQalkylaitritn ashibitamttch richarchemiatry. maretypicelaf maltifonctionalorganicmaleculco. lnrecentyearstherelnsbeenagrowiaginterestinthestndy of thegas-phsse reactionsof atomic metal and metal-containing ionswithorganicmoleculm. Studinafthmereaaiansusingia cyclatmnrnamnaaflcmspwtrunetry‘aadionbamtechniqan' yieldinfaruntianantheactintianofbondsinuganicmolemla bymetalionsintheabsenceafcomplicatingsolventeffects. Thumadymntiekinedeandmechnisticinfarmatiancanoerning theintrinsicarganometallicandaaordinstionchemistryafmetal ionacanbeobtainedfromthesesttidies. Also. thespeciftc chemistryof metalionswitharganicmalaculcsisthe buis far saewappraachtochemical ianintionmasss Animpartantareaofinteresthasbeenthestudyoftheuin- teractionafmotaliamwitharpnicmolecttlascantainingspedfic functional groups. In addition to studies of the reactions of a varietyof metal ions with alkanes.” the reactions of transition- metalitmwithargam'cspeamsuchasalkyllnlides.“alkeneo.“' a mi “mun mind W mind eaters.“° ethers.’-'° sulfides.”"’ and mercaptans" have also been studied. Aaanesampleofthetypeaofreactiansthathavcbeen oboerved.thercsctiansofCo’ with several C,H,X speciesare given'tnTable l. Withtheexceptional’ethylamine. the major reaction pathway farall of these compounds involves insertion of themetalioninto the relatively weak C-X bondasthe first mechanistic step. Afi-H shift framthealkyl ligand maythen occur. forming C0(C;H.)(HX)* which then dissociates by com- petitiveligandlossasseeniareactionl. Forthereactiansof c-’ + cm -- Mon‘ - Calla-00’4“ I one». + to: at.” + Calla (ll Ca’withaamfluhncfioulmpsachascarboayficacids. Was-sum. 'wmw. ‘l’ahlsl. ListiagolthaNoetrslsaadlanlunltingfromthe WofCo’withVarianMannamEthnm Co’ + C,H,X -0 A34 neutral X A’ out!!! rel 1» ref H no reaction 5d CH, CoC3H.’ CH. Ill 5d CoC,H.’ H, 69 l CaC,H.’ HI 7! 6 Col’ CzH, I I CoHl’ C,H. I I OH CoH,O‘ C,H. I) 6 CoC1H.’ H30 87 SH CoC3H.’ H,S 70 I2 CoSH’ C,H, 7 CoH,S‘ C,H. 23 NH, CoC,H,N’ H, I S 8 CaCH,N° CH. 26 CzflgN. CO" 59 o CoCO’ C,H. 8 IO CC". mgflg. C,H.O 2‘ CaC,H.O‘ C,H. $2 CaC,H.O‘ CH. I I CoCJhO’ H, 5 9 CaHp’ who 36 IO ‘°“ Cocm.’ cup, 31 CaCH,O,° C,H. 5 CoC,H.0’ H,O 2I coclfldol. Hi 1 ? Wylie cup; 3‘ I0 0‘“ CoCH10,‘ Cane 64 insertionintotheC—Xbandisstillamajarreactianpsthwsy.bst aamacleangeafbantwithinthefunctionalgronpalsaoccan’ 101 6126 1. Ant. Client. Soc. Vol. 106. No. 2!. I984 Carsady er al. Table II. Listing of Neutrals and Ions Resulting from the Primary Reactions of Co' with Nitromethlnc and ”CUM Nitrite Co’ + reactant -° CoR’ + neutral(s) _ CH,NO, CH,ONO CaR" neutral(s) rel S intermediate structure rel ‘5 intermediate structure CoHNO,‘ CH, 2 I 0 CoH’ CH,ONO 2 III. E I5 C. E C003 HCN. H,O I7 IV. V 0 CoOH’ CH,NO 2I IV. V 0 Co(OH),’ HCN 6 IV. V 0 CoCH,NO‘ OH 3 IV. V 0 CoCHNO‘ H,O 3 IV. V 0 CoNO‘ CH,O 2 E S E CoOCH, NO 40 E 70 E CaOCH, HNO 2 E 5 E CaOCH H,. NO 2 E 5 E Tehhlll. ListingofloaseadNeutralsRosuItiagiromthaPriiury ReactionsaiCo’withNitroethene C0. '9‘ CzflgN03 -. COR. ‘9' neutral“) Colt' neutraKs) intermedis‘ to structure rel ‘6 CoHNO,’ C,H. I 12 CoC,H.’ HNO, 3 CoC,H,° H,. NO, I. III I4 CoCH,’ CH,NO, II 5 CoCH,NO,’ CH, II 7 CoCH,NO,’ cu, II 2 C007 CH,CN. H,O IV. v 2 C00". C,H.NO IV. V 20 Co(OH),’ CH,CN IV. V 4 CoC,H.NO‘ . OH IV. V 3 CoC,H,NO" “,0 IV. V 4 CoNO’ C,H,O E 3 CaC,H,0‘ NO 3 4 CoC,H.O’ HNO 3 I0 CoC,H,O‘ H,. N0 2 s AlogicaleatensionofthiseImwlu‘chhasbo-iteen’vingrecmt interesListhestudyoftheeffeaofvariouingandsonmetelion reactivity. Although thisworkisstillinits infancy. preliminary studieshaveshawnthatthepresanaeofsligandmayleedtanew reactionmechanismsaswellesenhanceordeactivatethebere metalionJM" Asecantinuationofaurworkinthisemtheresultsafastudy afthereactionsofCo’withsomealipheticnitrocompaundsaie prurntedhere. Inadditian.reactiotlinnlvingCoL*genersted commandmemaainiat-ermiadudaths followug: (a) Huang. S.K.;All'noa.1. lies no.2. II]. (b) loam. W.:Staloy. H. . D. 3.;Proiser. 3. 8.16“. 790,103.14“ (c)Wri-h.1.; WDUP Did. I“ [06 67 (2)(aimllclnmhplLtAnOei-tSaelflle. 734. (b) ArmonmP. 3.:3nnchamp.1. I... I‘ll. III. 703. “'23. (c) HelIs.LF.; MRNMJWLW I. ”(3)”. R. Cclylio. D: Fm. 3.8.4“. Chart. In”. I64I. ”(‘)MM.3 AlliaJ. In. J. UmSmlonPhys. taster. 2'(iHIIByNLG. D.; hrm’u..:R.C FHWISJMCMISarm 104. 3565. (b)3yld.0 D.:Frn'u.3.S.lbidlfl1104.s944 (cIJaaohm. D. 3.; Freiser. 3. S. (bid. I”; 101. 736. (d) Jacobson. D. 3.; Freiser. 3. S. M. m 101. Sl9‘l. (6) Alien. 1; Frees. R. 34 My. D. P. (bid. I979. 101. I332. (6) All'nae. 1.: Ridge. D. P. 1. Ant. Chem. Soc. I979. 101. 4999. (7) (a) Armaetroot. P. 3.; Hello. L. F; In Sec. III. 103. 6624. (b) Allison. 1.; Ridge. D. P. I (8) Redechi. 3.; Allison. 1.). Ant. Cheat. Soc. I'd. 106. 946. (9) 3uraier. R. C.; Dyrd. O. D.; Print. 3. S. I. An. Gent. Sec. III. 103. 4360. ,.(I0) lyrd. O. D. Ph.D. Thum.‘ Purdue Univuut' ' y. Nut Lafayette. IN. I 2. (II)CarliaT..-1 Null 3.; PM. Immoral”! ”“2743 02le 1. PhD. MPurduaUuinrn‘ty. West LafayetteIN. I96 (IJ)(a)AIIhae.1.; HuaegS.K.; LambershiM; Mcflnaywsw- nae-ens AeiuimaSaa'otyfuMauSpactrnotry mrll. .(bH’mrbopeuImA; “listlwurs Tim". Listiagol’NeutraIsendIonsRmuItingiromthePrimary ReactionsafCa’ with Nitropropene Co’ + C,H,NO, -° CaR’ + neutral(s) C0. + C)H1N03 " R" ‘P neutral(s) rel % intumadiets I- 2- CoR’ or R" neutral(s) structure C,H,NO, C,H,N0, C,H,‘ CoNO, I I2 I7 CaHNO,’ C,H. I 21 I6 CaC,I-I.‘ HNO, I 4 3 CoC,H,’ H,. NO, I. III 10 6 CoCH,‘ C,H.. NO, II 2 2 CoC,H.NO,‘ CH, II 6 6 CoC,H,N0,’ CH, II 2 J CaCH,NO,’ C,H, II 2 2 CoCH,N0,° C,H. II I 2 CoOH' C,H,NO IV. V IS 16 CoNO’ C,H,O E 2 2 CoC,H10’ NO E I 3 CaC,H.0’ HNO 3 6 7 CaC,H,O‘ H,. NO 3 I4 II esprimerypoductsfromtheinteractiaiofCo’withnitroelkaam as well as Co(CO),’ (x I 0-2) and Co(CO),NO" (x I 0—3) generated by electron impact on Co(CO),NO are prorated. input-anal Section Theeapuimantsinvolviegthebaremstalion.Co’.wuamrriedaut byusingaprototypeNicolot FTMS-IOthichlnsbaeapreviaust described in detail." The ma- spectrometer is equipped with a 5.2-cm cubic trapping cell situated between the poles at a Varian IS-in. elm- tromagnet maintaind at 0.9 T. The cell has a O.25-in.-diemeter hols 7. one of the transmittu platu to allow various light sourcm em to the insideafthecell. Co’wasgenuatedbyfacusingthelrequency-doublod beam (530 nm) of a Quanta Ray NszAG laser onto a high-purity oobaltfailwhichwssouppatedbytheoppmitstrensmitterplete. Dotaih of the laser ionization esperiment have been described elsewhore.” Laser iaamuan can produce ions in excited states. and while star- on takentominimiutheirprmmce."thelormauonofnunuproductsfru excited Co‘ cannot be completely ruled out. The distributions of primary product ions listed in Tables lI-V are reproducible to within tlo‘li absolute far the ma)or products and £55 absolute for the minor products Product distributions of sum reactianoftheCo’uimeryreectionproductswcredeternunedby“ swept double resonance techniques" to isolate the ions of interest. The CID eaperinirms were performed by using the FTMS. Sen“ prmoereswuiaoetheorderoiIx10"torr.andtheergontargot's wasantheorderofl X IO”torr. Preesuresweremonitorcdwithe BayardoAIpert initiation gauge. Details of the CID experiments hen been described previously.“'“ The collision energy can be varied (N) (a) Cdy. R. 3.; Pr“. 3.5 1..). MessSpertrom. Icahn. use :1. I99. (b) Cody. R. 3.: Denise. R. C.; Presser. 3. S. Aasl.Chni. In. 4 96. (IS)Ceasady. C. 1.; Fraser. 3. S. vunpubliohedruuln. (I6)(aIComnerow.M..;3 Graesi.V :Purioad.G.MPhyertt.Im 57. 4|]. (bIMerahsILAG Canimrow. M. 3.; Par'uod.G 1 MP9)!- '2: :I. 4134. (c)Psr‘ntul.G.;Couaroe. M.3.Ah.flma5puvem. I .2I 102 Co’ and Cdllgmdl.’ Reaction: with Nitroelkener 1. Ant. Client. Soc. Vol. 106. No. 2!. I984 6l27 Tabb V. Listing of the Neutrals and Iota Resulting from the Primary Reactions of Co’ with Nitrobutane and Iutyl Nitrite Co’CJ‘LNO; -° CoR’ + neutral(s) Co' + C.H.NO, -. R" + neutral(s) rel % intermediate intermediate CoR’ or II." nautraIIs) (nitre) IoC.H.NO, 2-c.H,No, (€39,010, (nitrite) I-C.H.ONO C.H." CoNO, I 38 30 19 A 2 CoHNO,’ C.H. I 9 I3 2 o CoCJ'I.’ HNO, I I 3 I A I CoCJ-I,’ “1. NO, I III 6 l3 5 0 ciicnt.t It. HNO, I 3 4 o A 4 CoCH,’ c,H.N0, n o o 2 o CoC,H.N0,’ H, II i 2 2 o CoC,H.NO,° C,H, II 3 5 I o Coc,H,No,° cm. H I4 i i o H ' c,H,No, n 2 o o o CoCH,NO,‘ ,H, II 2 I o e 23 COCJ‘Iy’ canto, II 2 2 o o CoC,H,’ CH. NO, II 2 I 0 I 6 CoOH’ C.H.NO IV. v 7 1 3 o CoC.H.O‘ NO E I 3 I E 2 HNO E 2 4 I E 7 CoCJ-I,O‘ H3. NO E 3 5 0 E 2i CoCJflO’ H,. HNO E I 2 O E 6 CoCJ-Ip' 2H,. NO E 2 J 0 E 27 CoC,H.O’ CH,NO E I i 2 E t Sens-I mayalaobepoasiblel‘ortheproducts.0nthebasisofacon- . sideration of all of the data. however. including (i) subsequent C°°’C”s°”°"‘c“s°‘c*"° “005°” ion/molecule reactions of the primary products with the nitro- E II (ii) If . . I l l' . don I (iii) the . w' 00'" of Co‘ with other nitroalkanes. and (iv) the reactions of other CH.O‘Cf‘m ECoH'ocnpouo N CoOCH; mm ’9 a-ro-cd-uo—ociiocwo H. mo H typicallybetwesaOanleOeV. Thespradinionhinsticencgiasis tpmdantonthstotalaveragskineticenergyandisapprueimtelyJSS at I eV. I0$ at I0 eV. and 5% at 30 eV.I The studies of the reactions of cobalt-containing ions fanned by dactrenimpacton ' 4 ," ,' “10)weseperfa'madonanion cydotmemspectrornetarolooeventionaldeaignwhichhaspru- vieualybeandeacttbsdindetail.‘ Ion/molaalsreactiomandprearuua mideatifiedbyusingdoublemenanostechiu'qun Reportaddstaars results of product ions formed in a H mixture (by pressure) of Co(CO),NOtonitroaIkane.sttotaIpruauresbetweanSx IO‘ andl xlO”torr. Spectruwerealwaystahantomaasmgreatsrthan276umu. sincethe sum ot'themauol’CdCO),NO(l73 amu)andthelnrgut nitroalkane (C.H.NO,. IOJ emu) equals 276. Methyl nitrite and methyl-d, nitrite were prepared according tothe literature." Aflothachemialswerehigh-puritycommsreialsemplee whichmusedassuppliadeneaptformultiplafrusas-penp—thawm toremovenoneondensiblegaam. I‘m ud Disc-lea My Reaction. The primary for the reactions of Co’ with nitroalkanes are listed in Tables II-V. As an aid in ducidutingreactionnmhan'umthereactioruofCo’withmethyl nitrite and I-butyl nitrite were also studied. with their reaction products being given in Tables II and V. respectively. In the experiments with nitromethane and methyl nitrite. the empirical formulas of all the products were confirmed by using both deu- terated and undeuterated reagents. For several products from thelargerniuealkamcmpiricalformulasotherthanthoaelisted (I7) Hm W. T.; Mes-nan. M. M.; Enema. D. D. J. Chest. Phys. I971 54. 843. (ft) Hartuag. w. H.; cti-day. r. 'Orgamc‘ smarter; Wiley: New Vets. I943; Cornet Vol. II. pp tomes. metal ion. with nitroalkane." we believe the formulas listed are the most reasonable. Also listed in the tables is the probable struaure ofthe reaction intermediate that leads to the formation of each product. Structure I results from Co’ insertion into the C-NO, bond. Structures II and III (which are possible inter- o R-CHfCo‘N: I-CH.'C0.°W I A ncn,-<:o'-cn,~° ecu,-Co°-at,o~o n e 'o men-rib e-cu-ouo I (50' Co’ H i‘i m c ‘3 RCH,-N-Co°-O ncu,ou-Co'-o n o acme-came E nted'ntmntgptedbytheobast'vedmactioruofCo’withanuria') are products of metal insertion into the C-C and C-H bonds. respectively. Insertion of Co' into the NO bond. which has a bond strength" at 75 keel/moi. leads to intermediate IV. An analogous set of insertion isomers for Co’ with alkyl nitrites are also given (structures A-E). Throughout this paper. intermediate which follow from nitroalkanes will be labeled by l-V. and in- termediates from alkyl nitrites will be labeled A-E. As will be discussed below. some products from the reaction of nitroalkanes with Co‘ can be most eesily explained via inter- mediates A-E. suggesting that either a nitro-to-nitrite isomeri- (I9) Pop“. V. I; Matyuein. Y. N.: Lebubr. Y. A. Ian. Aknd. Need. Airs. 53R. 80. (Hot. Neath. I974. 8. lm. 103 6I28 J. Ant. Chant. Soc. Vol. I06. No. 2!. 1984 nation occurs or. in some other mechanistic step. a 'nitro intermediate“ is converted to a “nitrite intermediate". Nltrnmethae and Methyl Nitrite. From the data in Table II. two interesting points are evident about the reactions of Co’ with nitromethane. First. in comparison with the resume of Co’ with other monosubstituted methanes. nitromethane chemistry is outsiderably richer. having a large number of products. M . methyl alcohol.‘ and methylamine' have been found to be un- reactive with Co’. while methyl bromide.‘ methyl chloride.‘ and methyl ritercaptan‘2 each give only one to two products. in contrast to the reaction with nitromethane which yields three IIIIJOI products and seven minor products. Second. the products are not than that might be expected on the basis of C-N bond imertion. but appear to result primarily from insertion into the N-O bond. In order to explain the reaction products. It is useful to compare the results for nitromethane with those for methyl nitrite. Both compoundsareobservedtoyieldthesamemajorprimary woduct. CoOCH,*. as well as having four minor products in common. There '5 one distinguishing feature of methyl nitrite chemistry that affects both its solution and gas-phase chemistry. This is a very weak CH,O—NO bond. With a bond-dissociation energy of42InI/mol.”thisbondiscoraiderablyweakerthanthetypial C-C bond (ca. 88 bound"). the C-H bond (cs. 97 heal/moi"). the nitromethane C-N bond (6t kcal/moln). or the N-O bond (75 keel/mot"). Thus. one would expect that the reactions of Co’ with methyl nitrite would involve oxidative addition into the weak O-N bond. giving the insertion intermediate E. As shown inScheme I. the formationofEfolIowed byaB-H shiftontothe masltanexplainalloftheproductsofthemethylnitritereaction. Products resulting from the possible intermediates A and D are not observed. possibly due to the relatively strong bonds which would have to be broken. It is interesting that the NO group. which is a threeclectron donor” and has been found to bond strongly to Co’ in the gas phase.“ is lost preferentially to the OCH, group. Since the initial insertion into the O-N bond lads to an NO group which donates only one electron. the lifetime of the intermediate may not be sufficient to allow rearrangement in the linear three-electron donor complex geometry. Of relevance 'aastudyofWaltonetaI.” ' secondaryitmmspectromury (SIMS) on [Cr(NO)(CNCMe,),]PF, and several other similar complexes. It was found that while initial fragmentation occurs via loss of isocyanide ligands. after the third isocyanide ligand is lost. the fourth ligand lost is exclusively NO. No ions corre- sponding to Cr(NO)(CNCMe,)’ or Cr(NO)’ were observed. Their rationale for this observation was that while initial ligand In from large ions is determined by relative bond strengths. in smaller ions the ability of a ligand to delocalize charge bemmu animpirtantfactorinligandltn. Otherfaaasnichassyna'gistic ligand effects and the ability of a ligand to carry away excess energy may also play a role in ligand loss. In the reactions of Co‘ with nitromethane. products are ob- served due to intermediates predicted from both the nitro and nitrite functional groups. ,An intramolecular RNO, to RONO isomerization isobaervedintheetectronimpact(EI)mssspectra of many aromatic nitro compounds" and in a few aliphatic nitro compounds.” Such an isomerintioncanalsooccur thermlly" (20) Iatt. L; Robttlon. G. N. "I'hs Chem'ntry of Amino. Nitroeo. and Nitro Compouaa and their Derivativea': Patel. 3.; Ed.: Wiley: New York. m2; Part 2. p I075. (21) West. R. C.. Ed. 'Handboot of detry and Pbydu'. 56th ed; Charlu'cal Rubber Publishing Co.: CM I976. (22) Refer-tee 20. p l0”. ‘0!) Cotton. F. A.; Wilkinson. 0. “Advanced ".Cbeanstry 3rd lateracisnce: New York. I972; pp 7l3-7I9. ‘(20 Weddla.0 HcAlhrIJ; Ridgs.D. P..] “Gauche. I977. 9!. l0! (2S)Plertn.1 L. Wigley. D. EzWaImRAWdIIaIMI. I32 l26)(a)Iur-sey. M. M. McLaIIerty. F W. J.Ant. M.Sor. "66.88. 5023. (b) Meyeraon. S. mu Fielda.E. K Ibi'd.I966.88J914 (c) “MR Willis-an H..Vso.A. N H. Org. Mass. Spectrom I976. 3. I465. (d)Ieyoa.1. H. Iartraad.M.; CoohaR G. I. “Che—”See ”71.95. l7”. “7(27) Nibbertag. N. M. M.;daIear.Tb.1. 0g. MassSm. "76.1. Custody er 01. Scheme II creouo 6 c°6 l 2 0‘ -- CH,-N -CH, N -CH,ONO Co’ \ \ 0 Co’ 0 Ce' .0 .0 CH'°N\ 0 Co. -—ih—. Cfi.-\~\c CH‘O’CO"M c , 0° 0 MEMMTE E I! / ,0 Hog-C63 -—°CH,-C|'.o° '0 I 0 Scheme III ,0. l ? 9 ((00500, - [ca.~;---ce 1" emu-cone _. cmu-cr-o-n To Coon- cu.» C, J a l L'OCeCst, 0' o— i ol ”(0‘06 I cannot; most new or or, rW-Klortfil l Lamar“: orphotochemically 1’ However. EI-induced mutations arise due to charge/radical sitm in the molecule. and thermal and photochemical isomeriutions are believed to occur by an NO, dissociation from the alkyl group followed by a recombination. ItseemslikelyinthccaseofCo’ reacting withnilromethanethat the metal assists in such a nitro-to-nitrite isomerization. Possible mechanisms for the nitro-to-nitrite conversion are presented in Scheme II. There are two general pathways in which this con- version may occur. The first involves the coordination of Co’ to an oxygen while the molecule isomerizes (reacrion l in Scheme II). Once the nitrite isomer is formed. Co’ inserts intothe O—N bond to give intermediate E. The other possibilities involve participation of the metal (reactions 2 and 3 in Scheme II) by initial N-O or C—N insertion. Note that the actual nitrite isomc is not formed in these mechanisms. but only the metal-insertion intermediate (E) typical of nitrites is formed. Two major products. C00’ and CoOH’. and several mince products are seen in the nitromethane reaction which are not present in the methyl nitrite reactions. The most probable mechanism for the formation of these products. Scheme III. involves an intermediate with two oxygen atom bound to the nietaLintermediateV. Ouefactsuggmtingtheformatienofth'a i’ emu-cap H I intermediateasoppceedtoasmeswithonlyoneoxygen bound to the metal. intermediate IV. is that CoO’ formation from IV would involve loss of CH,NO. This process would be approxi- mately 30kcal/molendothermic.” TheformationofCoO’ from V with loss of HCN and H,O is I3 Itcal/mol exothermic. Also. the formation of the minor product Co(OH),‘ (or Co(O)(H,O)’) indicates that. at some point. two oxygen atoms must be to the metal. Collision-induced dissou'ation experiments were performed on the major primary reaction products of Co’ with nitromethane. Ascxpected.theproductsCoO’andCoOH’ losthndOH. (28) Reference 20. Part I. pp 434-4“ and referenc- therein. (29) Reference 20. Part I. pp 2I4—2l6 and referenc- therein. (JOIThermochatntealinIormatienutaheafremreIZOandRmsmtockst al.:Roaaustoeh.H..;.K.MDraxl ;Stsiaw...;IWH¢rea.1....TJPhyr Chews.Re/.Date.3tippl. I977..6I 104 C0‘ and Cofllgendl.’ Reaction with Nirroelftcnes Tfl VI. Ice and Neutrals Formed in Subsequent Reactions of the Prlnnry Products ofthe [Co’ + CH,NO;| Reaction (P - CH,NO,) CoR‘ + CH,NO, —- ctrit'r + neutral Coll’ CoR" neutral rel ‘5 C00‘ Co(OH);’ HCNO 40 Co(CH,NO,)' OH 30 CoP' 0 IS CoOP' I5 CoOH’ Co(OH)(l-Ip)’ HCNO Io Co(CH,NO,)’ Hp 40 Col” 0H 20 CoHP' 0 I0 Co(OH)P" 20 CoOCI-I' e CoH’ e CoNO’ NR‘ CcOCI-I.‘ CcP’ CHp I00 CoOCI-I,’ CoHP‘ Clip 100 CotOH){ CoOP‘ Hp 75 C°(0H)P’ OH 25 CoCHNO’ CoOP‘ HCN I00 CoCH1NO‘ Co(OH)P' HCN Im CoHNO,‘ Co(OH)P' NO If» 3rd Generation (Subsequent Reeetioe of Products Listed Above) Co(OH)(l-I;0)’ Co(OH)P' Hp Im CcCH,NO,‘ Co(NO)P' CHp I00 CcP' CoOP’ CH,NO l0 Co(OH)P’ CH,NO 70 Col-II” CH,NO, 20 Cell? Co(OH)P’ CH,NO It!) CoOP’ Co(OH)P' CH1NO, 70 (30004;)? HNO, Io Co(CH,N0,)P’ OH 5 CoP,‘ 0 IS 010le 01004;)!” "Not '0 Co(CH,NO¢)P‘ Hp 80 Co(OH)P,‘ 10 4th Generation (Subsequent Resetice of 3rd Gestation Reactions) Co(OCH,)P' a Co(OCH,)P' Co(OCHiXCH;)P' HNO, I00 Co(NO)P’ Co(CH,NO¢)P‘ HNO 30 CoP,’ N0 30 Co(NO)P,’ so Co(CH,NO;)P’ Co(CH9(CH,NO,)P' HNO, so Co(CH,N0,)P;’ so CcP,’ Co(CH,)P,’ HNO, so CoP,‘ $0 Co(OH)P{ Co(CH,NO,)P.‘ Hp ioo 5th Generation (Subeenamt Reactions duh Generation Rections) Co(OCH,)(CI-I;)P’ e e Co(NO)P,‘ Co(CH,)(Cl-I,NO.)P‘ e Co(CI~l,)P,’ e Co(CI-I,NO;)P‘ NR CoP,’ NR ‘Furtherreactioeofthisioncculdnotbedetatmined. ‘NR indi- etesthatthisiondidnctandsrgcasyfurtherreactionswithinthe timsseleUOslofthiseapei-imeit. respectively. to form Co‘ as the only CID product. CoOCH,‘ lost CHp to form Col-I’ exclusively with high-CID efficiency. Deuterium exclnnge experiments to probe for metal hydride character. i.e.. a structure such as u-wni’ I 5 H H movd inccndusivc. No deuterium ertchange was observed. Inch of deuterium exchange yields no information since FeH’. an ion ofobviousmetalhydridecharacter.doenotundergodeuterium exchange. CoH’. however. undergoedeuteium slowly. and NiH’ exchau. with deuterium readily." Although .I. Am. Chem. Soc. Vol. (06. Na. 2!. I984 6l29 Scheme IV C" o o O l l cu,cn,uo, —~cn,cn,~-¢.°-o —- cu,cmvc.s-o Com" c,~.~o 0 N I 04,01, N-cr-O l o Cocrflr'lo'w" I 11’ cmcu-cd-ou W"CN,CN°N'O i " .ECMM'CH,CM o CoC a icon 0 0,09th“, s s i \0’ l .. I anemone-no cu,cuc-c.‘-uo [Cri'c ‘c:°""'° .. mwfihmmo H 61 e (MW-v.0 Ceuo’ 'QHoO Cec.rs.o'~reo Scheme V Co’ 0 CH,CH,N0, -— C'H,CH,-Co’-NO. H I CH. ll weir-m, CoM‘QH. .. .-. l: CoCJt: mm. 1-... CH, H ‘cn-Cot—o GoC,H,’ . H, #1 Scheme VI Co’s cruel-two,» CH,-co° -cu,uc, I CoCl-l; exam; mace-crime. I + + iil cu,no, cr-i, Cocmiio; 4. C“. CoOCH,’ doe not react with D;. NiOCI-I,’. formed in the reaction of Ni’ with methyl nitrite. was found to undergo one deuterium exchange" This indiete that NiOCI-I,’ has a hydride structure and most probably CoOCH,’ has a similar structure based on the facile loss of Clip when activated. as well as the displacement of CH,O when CoOCI-I,’ reacts with CH,NO; (Table VI) and other Levis bases. Nitrnethane. For the reaction of Co’ with nitroethane (Table III). the majority (57%) of the primary products are analogous to those present in nitromethane (note that. as in the case of nitromethane. some products are best explained via the ‘nitrite- lilte' intermediate. E). Scheme IV suggests a mechanism for forming thee products starting with intermediate IV. Carbon-nitrogen bond insertion (intermediate I) is also an important process in the nitroethane reaction. In Scheme V. a mechanism is proposed that includes insertion of Co’ into the C-N bond followed by a d-H shift onto the metal. This mechanism could account for the formation of CoC,H.’ and CoHNoz’. which are l5% of the total primary product intensity. Carbon-carbon bond insertion (intermediate II) also appears to occur and requ in 14% of the products. including CoCH,’. CoCH2N02’. and CoCH,NO;’. The proposed C-C bond in- (3|) Carlin. T. 1.. Sallans. L; Camady. C. 1.; Incubate. D. I.: Free. I. S. J. Am. Client. See. I”). 105. 6320. 105 Cassndy er of. . .39 ow. . 28 2:9 .26 + 62.26 .88 52.... 2 + 8 + .62 .2688 .39 82 + 8 . 8.2.888 ...9 .2 . 82 + 8 + 3.2.888 .29 8 + .26 . 62.2688 .29 8 + .26 + 622.888 82.3 2 + 8. + 32.2.88 .89 8 + .388 .89 .26 + 8.. . 62.2.88 .29 8 . 82 + 8.2.88 82.... 2 . 8. . 62.2.88 .29 8. r .28 . 62.2.88 .29 622 + 8 + ..2688 8 + .388 28.8 .69 m8 . L8 .39 8 + L8 .39 .26 + . ozi 88.8 ......9 2 + 62.2.88 32.... 2 + 8 + . oz“: 88 .39 oz: + 8 in 3.268 .89 .26 + 62.2688 229 .28 + 8 + . oz :68 .....9 8 + ...8 82:: 2 + 8 + 62.2.88 .39 .2 + 82 + 8 + 3.2.88 . . . ....9 oz: . 8 + 3.268 .29 8 + 8 u 3.2.88 .39 2 + 3.28 + .2688 .89 2.8 + 8 . .6228 .29 .26 + .62.... 8 .29 8 + .26 + 32.2.88 32.88 + 2. + oz . .o 2.88 .29 28 + 8 . . oz 2 88 3.2.. 2 + 8 . .. 2.88 .2 .9 8 + .26 + 62.2.88 .39 .2 + oz + 8 + .3268 a. .9 .2 . oz . 8 + 3.268 .39 .28 + 8 + 62.2.88 .39 .2 + 62 + 8 + ..268 .39 82 + 8 . .o 2.88 .89 82 + 8 + 3.268 .29 02 + 8 r 3.2.88 ...9 8 + .26 + .6228 ....9 8 + .26 + .ch268 .39 8 + .26 + 62.2.88 .29 8 . .26 + .6228 .39 622 + ..2.888 .29 8 + .26 + 32 2.88 .88 o 08 />\ g — — {/2 ~389— 3583 e 8. + 3288 8. + 3288 8. + .3288 8. + .3288 8228.8 3.9 8 + 3288 .39 8 + .3288 8. + .328 8. + .328 ....9 8~ + .328 .39 8. + .328 8228.8 8 + .328 8 + .328 8 + .328 8 + .328 .8288 22 22 22 22 .828 .39 o8 + L8 22 22 .29 8 + .388 8 + .388 ...88 .29 8 + ...8 .29 8 . .28 + 62.2.88 .39 8 . L8 8.9 8 + oz . 3.2.88 .89 8 + 82 . 3.2.88 .29 .26 + 8 + .6228 .29 62.28 + .2 . .....888 .29 .26 . .62288 22.9 .26 + .62288 8 + .28 8 + a8 .88 e. A a... t 62.26 62.28 ...-.8: 6l30 J. Am. Chem. Soc. Vol. 106. No. 2!. I98! 2:03.: a 8:22 23¢ an- 825—1222 as! .248 3 8392.— 2382— .=> 03:. 106 C0‘ and Collignndi.’ Reactions with Nltmlkanes J. Am. Client Soc.. Vol. 106. No. 21. I954 613) senion mechanismisgivcninSchemeVl. Thaminorproduct. CoCH,N0, thzlnnnlCH. Methyleneisalsolormed inthereaetion leading toCoHNO, from nitromethaneThis isunusualsineeCHzisa high-energyndicslandarouldrequireeithersnn-Hshiltonto Co" oralternativelyontosnosygenvisacyclieintermediste: I/lz 0N0“ cu, ‘C\w re.T'h’e seenndmost abundant product (14%) ol‘ the nitroethane ction.,CoC,H .uloouldrea tl’romosidativ esdditionintoa C—Hhondfintumedistellhfollawdhya B—shiftsodlm oI'Hr Thislaaaol'H isthenfollowodhyacleavsgeoftheC—Nhond “Muslims shownin isn'narusunhla Co’iolau'vedtoii-Itreadilytinto -Hhonds ofamsfleka-nuzusndheeauae thedominan tproEessinthe witth’."whiehfavorsC-Hhond insertion."istheloasofH,sndN03. Analtcrnaiepsthway leadlngtoth'afamanon olCeC,H,’ via intumediste inSchemeV 0 II ‘ - JJ‘ - - . .2 A. thenisthree yetnitroethanoproducts. Again.not surprisingly. ClDonCoOH’ {ormedonly Co’sndnoCaH’. For CoHN01’. reactionsz Hioeumed. thormsthef tionofCo’raulting athigh-CIDui-‘gim NoCoNO‘waa («Minimum Como; T Cooii’ i no (2) c.‘ s mo. (3) liniito{D(Co‘-NO) D(Cc’-RNO,). Large Nltreelane. In contrast to the smaller nitroalkane. the larger nitroalkane react with Col...‘ not only by substitution but by bond cleavage procese. in this trend. butyl nitrite bond cleavage/ rearrangements occur for all CoL.’ ions. even Co- (CO),_,NO’. Such reactivity is typically observed for multi- functional organic molecule."”-" Alua‘horieoutaltrend' canheseau. Asthelengthof'the alkyl chain increase (C. -° C.). more reactions involving C-C inanions(intermadiate ll)areobeerved. Previousworksuggets tht the probability of reaction involving intermediate 1 should begretefesearstdaryandtertiarynitrnlhnethanforprinnry nitroalkane." ThisisappareutlythaceseasshowninTable V and VII. The rection demand in this work tend to involve interme- fiteinwhichtwoorbitalsareraqniru‘lonthematal(e.g..reein 2|) or reaction which require three orbitals on the metal (e.g., reectionslandZZ). ltisnosurpiaetlntsuchreectionprodues d'nappeerwhenligandsareaddedtothemetal. Reactions»- quiringthreemetalorbitalsoocnrforCo’. insomecasefor CoCO‘. andrarely fethaehermeal-ctnttain'mgspacieinTalna Vll. Rection requiring two orbitals on the metal (as opposd to three orbitals) seem to dominate (in addition to substitutitm rectioelwhenthareectantmetalcettenstwoemorelignrnh preent. as would he expected. Ninee'epene. There are many similaritie in the rection obserwdforCo’andCoCO‘withniu'opr'opaneandsomemaje diffeeeeaswell. TheproductsCol-lNO,’ OmofCJuthrough intermediate l) and CoC,l-l10’ are formed from both Co’ and CoCO’. Both Co’sndCoCO‘inertintoC—C bondszhowevar. CoCO‘ preferentially attach the C,l-l,-CH,NO, bond. Followir' this' nation. Co’ fore: the C, fragment while CoCO’ retains the C, fragment. This may imply active participation in the recumbythscarhonylgrmplethngnorroatypenimarmdata buttotheCoCO’inertioniutemediateshownin(38).' There ML ._.__..... Co -CNgNOa .«0 “ I ,xo‘ (so) cit ’/ ,- “cua areaunmbarofpoeihleneutrallese.(Cl-l,NO,+l-l)or (CH,NO,+H9. whichwouldganerstethebntadieee-lihamatal ctnnplex. MmefleadtheCOligandappearstobeanactive involvementof'theligandinthereection.‘ lnmostcase.COisnotinvolvadinthereaetiondirectlybut appentohaa‘specntor’mthsrneal." Eventpactatorligntfi caninfluence product distribution.incesewhere cleavageof thaM’-C0bondcompetewith|nocaeeinvolvingelimintion fromtheorganicspecie. Anexampleofthisisseeninthccese of 2-nitropropene. The Co+ ionreactstoform CoC,l-l,0’by eliminationofNOandl-i, ThisrectionalaooccursforCoCO"; howeve.theeapperttohetwopoeibleprocuefollowingN0 el'unination: loeofl-l,fremC,l-lyOorloeofCOfremthemeal. lntheeaeofCoCO‘withz-niu-opropanaJhaloeofNOfolloeed by theloss ofCO (to form CoC3H10’) predominnte; howeve. H, vs. CO elimination frequently appear to be competitive processe. Nineheane. Theoutstanding fetureinTahleVllregardin the nitrobntane is the Vertical trends'-i.e.. change in the chemistryofthematalenterasthenuntberofligandsonthe metalineeaaeJnnr-tielarfer 2-nitrobetaneand2-msthyl-2- nitropropane. eooo’ + comic. - Cassndy et al. For 2-nitrohutane. the predominant reaction intermediate change as the number of ligands increase. Co’ reacts through intermediates I. ll. IV. and E. With one or two carbonyls preertt on the metal. reacrions are observed via intermediate analogous to I. II. and E; CoCONO' only induces organic bond cleavage through an intermediate analogous to II. Note that only one C-C bond is attacked by CoCONO’. This may correpond to attack oftheweakeet bond (sinoein thecorresportding alliane. isopentane. the weakest C-C bond“ is the C,H;—C,H, bond). Presumably. therearenotasufl'rcient numberoforbitalsavailableonthemetal following C-C insertion of CoCONO’ to assist in a 6-H shift; thus. Call,- is lost. Similar studie with amine show analogous reults. The Z-methyl-Z-nitropropane rection also show a variety (1 interesting change as the number of ligands increase. Inter- mediate structure I. ll. [11. and E lead to the products observed for 2-methyl-2-nitropronna. Structure ll and Ill predominate when more ligands are present on the metal (e.g.. Co(CO){). Insertion into the C-NO, bond (structure 1) appears to be in- hibited due to steric effects of the ligands present on the metal and the bulky 2-methylo2-propyl group. Thus. the remaining options for the metal are inertion only into bonds such as C-C andC-H. TheloeolefrenC—H insertion)ismorepreminent when the number of ligands preent on the metal ion increase. The reaction of Co(CO),NO‘ produce CoCONO(NO,)* with alrnofOOandCJ-l, AtfusLonemaypredictaCo’inertioe into the C-N bond (structure I) as the intermediate. This in- termediate is not possible for several reasons. The cobalt ion already has three ligands present. which doe not leave enough amptymatalorhitalsfcrmetalinaetiontooccur. lftheinertion did occur. there are many Mr: available to shift and produce a strong HNO, ligand. No products. however. are observed reulting from a 6-H shift. A possible intermediate structure leading to the product ion CoCONO(NO,_)* is shown in the structure below “r“ i T° cucohuo’ + +no, - cu,—(l:—c—(':.’_co .. CH: N0: cucuuuoumi’ The mggeted inn-nudists cerepimds to irtasrtin of the CoCO' group into the C—NO, bond of the nitroalkane. In this inter- mediatetherearenol-latomswhichareonacarbon whichisd tothametalthatcouldshifttoprodncethe HNO, ligand. Also. there are no empty orbitals on the metal to assist in any rear- rangementofthemolecule. 1'hus.theorilyproductionobaer-ved is CoCONO(NO,)’ with the loss of C.H,CO as one ligand. In summary. several typical ligand effects have been observed inrectietsofnitroalknne. Productswhichraquirethreeormore enptyorbitalsonthemetalaresaentoberepresedordisappear as the number of ligands on the metal increase. Reactions oc- au'ringatbranchadsitearefavoredoveunbranchadsite. Steic effects were otnerved with 2-methyl-2-nitmpmpane when ligann were added to the metal ion. Intermediate in which the ligands must be actively involved in the insertion are seen. Also. when aligandispreentonthemetahseveraloptionsforreaction mechanisma(suchascompetition between lossofCOand "2) are available. 6 CgHfl° (39) WismadebyCJC and ESP. totheDivision of Chemical Science. Office of Basic Energy Science. United State Department of Energy (DE-AC02-805Rl0689). for sup- porting this reearch and to the National Science Foundation (CHE-8002685) for providing funds to purchase the FT MS. S.\V.M. and LA. acknowledge the National Science Foundation Min isopentanemayhselcelatedfrontdataobtn‘rnd inCoxandPileherénndMilleeandGnlden (a)Cea.J. D.; Pilcher. G. and W MOI-tenant": Acedenn‘e Hulda-Yakima (b)McMili-.D..F; Gold-Jinn“. ”was-um 110 6135 (CHE-8023704) and the Dow Cbernical Company Foundation 24-3: CgHrN03. 108-0L2; l-CJ'loNOn 621-054; l-C.H.ONO. $44- .. ...... m .... mpg,“ m m mv- éfi'éci'§$e§°t..3&:‘eii .zsna°e..39°;:;.:.: r F llowsh' s . v : - = ~ 1: ”mm” °' ° ‘p " Co(CO){’. aunt-o, CoNO’. aurora-z; CoCONO‘. 6l8l6-95-3; ecu-in Ne. CH,NO,. 73.52-5;cn,0No.62r.91.v. C;H,NO,. 19. cacoimor. atria-964; Co(CO),NO’. 52309434. APPENDIX C Appendix C ICR Studies of Trimethylaluminum The following manuscript is a study of the mass spectra of trimethylaluminum. This study resulted from an attempt to examine the Ziegler-Natta catalysts for olefin polymerization81 in which both trimethylaluminum (Al(CH3)3) and titanium chloride (TiCl4) were admitted into the ICR. The active bimetallic species formed by ion/molecule reactions could then be studied by admitting a series of olefins into the ICR. The Al(CH3)3/TiCl4 mixture did result in a large variety of bimetallic ions. No” double resonance responses, however, were- observed for these ions, suggesting that they must result from neutral-neutral reactions of trimethylaluminum with titanium chloride. A new inlet system was designed and constructed for the ICR in an attempt to keep the two gases separated until they were near the ICR cell. Double resonance responses were still not observed with the new inlet system. These same results for the Al(CH3)3/TiCl4 mixture were reported later by Staley et. al.82 The use of Fourier transform mass spectrometry and ion ejection techniques may facilitate the study of the Ziegler-Natta catalysts. The previous mass spectral studies of trimethylaluminum have suggested the presence of associated species in the gas phase (dimer, trimer, etc.). The ' ICR mass spectra and the ion/molecule reactions observed suggest that trimethylaluminum exists as a monomer only at low pressures as described in the following manuscript. Metal ion chemical ionization was also utilized in an attempt to determine the molecular weight of the species present. Triple quadrupole mass spectrometry (TQMS) was employed to obtain collision induced 111 112 dissociation (CID) spectra of several ions present in the mass spectrum of trimethylaluminum. 113 ABSTRACT The mass spectra of trimethylaluminum at various pressures are reported using ion cyclotron resonance (ICR) mass spectrometry. Ions indicative of the presence of the dimer are observed at higher pressures. Double resonance results, however, indicate that these ions are formed from ion/molecule reactions of fragment ions of the monomer with the neutral trimethylaluminum monomer. Evidence is also presented for the possible presence of the neutral dimer at higher pressures. The use of metal-containing ions as chemical ionization reagent ions is utilized in an attempt to determine the molecular weight of the species present in gas phase trimethylaluminum. Comparison of the results presented here with previous studies suggests a process in which trimethylaluminum is present as a dimer in solution, but vaporizes as a monomer with the possibility of dimerization occurring in the gas phase. INTRODUCTION Alkyl aluminum compounds have been the subject of a variety of studies because of the possibility of the existence of the molecule as a dimer due to the electron deficient nature of aluminum, and also because of their importance in the industrial Ziegler-Natta polymerization catalystsl. Early studies to determine the molecular weight of trimethylaluminum (A1(CH3)3) cryoscopically in benzene showed that the molecular weight corresponds to that of the dimer, A12(CH3)6293. The molecular weight of gaseous trimethylaluminum has been studied by Laubengayer et. al.4 using vapor pressure and density measurements. These results suggest that trimethylaluminum is a dimer in the gas phase at 70°C and a pressure of 135 mm. The apparent . molecular weight appeared to decrease, however, as the temperature was increased. This same effect has been observed more recently by Henrickson et. al.5 at a pressure of 30 mm. 114 Several temperature studies have been performed to determine the extent of association of trimethylaluminum. Almenningen et. al.6 reported that, at 215°C and 30 mm., trimethylaluminum exists as 9896 monomer. O'Brien et.al.7, using Raman spectroscopy, reported that trimethylaluminum exists primarily as the dimer at 70°C, but at 260°C and 1 atmosphere the dimer species is n_ot_ detectable. The bonding in the trimethylaluminum dimer has been found to be relatively weak («40.2 kcal/mole per Al—CHg-Al bridge, or 20.4 kcal/mole for the dimer)4 which may explain the dissociation to the monomer at high temperature. A more recent study by House8 estimates that trimethylaluminum is approximately 72% associated in the vapor phase. Thus, there appears to be an equilibrium between the monomer and dimer in the gas phase. The structure of the trimethylaluminum dimer has been studied by infrared9 and Raman7 spectroscopies, electron diffractions, X-ray diffractionlo‘lz, and 27A] nuclear quadrupole resonance13 in order to determine the structure and type of bridge-bonding in the trimethylaluminum dimer. Results from the early studies suggested the presence of H-bridging from the methyl groups, suggesting two structures with binding of the carbon through two hydrogens to the aluminum (structure 92 or through one hydrogen (structure 3)”. More recent results indicate that the three bonds to the hydrogens on the bridging carbon atom are tetrahedral and are symmetrical with respect to a vector directed to the center of the dimer”. Therefore, little or no bonding occurs ' between the bonding methyl hydrogen and the aluminum atom. This leads to the more accepted structure of A12(CH3)6 (structure Q) which consists of two methyl bridges. 115 H\ F H2 H H c , c — 3 3 \A.i,.4-I’ \AI/CHJ HJCQ H C\ ACHs H3C\ pcq CH3 Al Al / \ ,H , \ . / \ H30 , \H' CH3 H304 C-H ‘CHJ H30 ‘C( CH H2 H 3 3 .1. z. 3 Mass spectrometry has been utilized to study the ionization and subsequent dissociation of methyl-substituted metal compounds. The first mass spectrometric study of trimethylaluminum was performed by Winters and Kiser15 in 1967. This study was performed with the inlet at room temperature and the source temperature at 190°C with 70 eV electron impact (source pressure was not reported). The results are summarized in Table 1. The majority of the ions observed appear to result from the dissociation of trimethylaluminum monomer through simple bond cleavage processes (e.g., loss of CH3-). Several rearrangement ions are also observed ([Ale]+ and [AIHCH3]+). The dimer ion, [A12(CH3)6]+, was po_t observed, however, a small quantity of the ion [A12(CH3)5]+ (0.296 of the [Al(CH3)2]+ base peak) was observed suggesting the existence of dimeric species. The effect of electron energy on the mass spectrum of trimethylaluminum was also reported indicating an increase in the monomer molecular ion at low energy but no additional evidence for the existence of the dimer ion. Chambers et. al16 reported the mass spectrum of trimethylaluminum at source temperatures between 40°C and 200°C. Table 1 lists the reported spectra at 45°C and 195°C. Two ions indicative of the presence of a dimer were reported ([A12(CH3)3]"' and [A12(CH3)5]+) which were abundant at 40°C but diminished rapidly with an increase in the source temperature. This behavior 116 Table 1. Previously reported mass spectra of trimethylaluminum (% rel. int.) Tanakn & . Ion Possible Winters and Chambers and Comte»16 Smith" M Composition Risen-15 45-500c 195°C 60-10000 12 [c]+ 0.03 0.12 0.1 13 [CH]+ 0.15 0.17 0.5 14 [CH2]+ 0.48 0.48 1.3 15 [0113]+ 21.3 2.17 3.92 8.0 16 [CH4]*’ 1.74 3.98 9.7 26 [02112]+ 0.28 0.19 27 [A1]+,[C2H3]+ 36.5 12.96 , 18.99 10.5 28 [AIH]+,[CZH4]+ ' 1.04 0.91 1.2 29 [A1H2]+,[02H5]+ 2.3 1.85 1.67 1.5 30 [A1H3]+.[C2H5]+ 0.2 41 [1A.1c112]+.lc3115]+ 1.0 0.89 1.16 42 [14101130,[03116]+ 6.3 5.63 8.04 5.6 43 [AICH3H1+.[03H7]+ 3.4 4.63 5.40 2.7 56 [Alc2H5]+,[c4Hgl+ 2.68 1.76 57 [A1(CH3)2]+ 100.0 100.0 100.0 100.0 72 [Al(CI-I3)3]"' 4.9 9.13 7.1 8.0 99 [A12(CH3)3]+ 0.82 0.06 102 [A12(CH3)3H3]+ 0.2 115 [A12(CH3)4H1+ 0.2 129 [A12(CH3)5]+ 0.2 4.42 0.22 0.005 145 [A13(CH3)4H4]+ 0.2 203 [A14(CH3)5H5]+ 0.2 117 is analogous to that observed in the vapor phase studies described previously and is presumably due to thermal dissociation prior to ionization. Tanaka and Smith17 also studied the mass spectra of bridge-bonded aluminum compounds including trimethylaluminum. Their reported mass spectrum of trimethylaluminum at a source temperature of 60 to 100°C is included in Table 1. The ions with m/z > 72 (indicative of dimers) are of low relative intensity (< 0.296 rel. int.) and their sum is < 196 relative intensity. This is in contrast to other dimeric species (AlCl3, AlCH3C12, and A1(CH3)ZCI) whose mass spectra were also reported. In these dimeric species, ions indicative of the presence of the dimer are intense (> 5096 of the base peak). In the mass spectrum of trimethylaluminum, however, the [A12(CH3)5]+ ion is only 0.00596 of the base peak. Ions which could be fragment ions resulting from rearrangement of the dimer, [A12(CH3)3H3]+ and [A12(CH3)4H]+, are also observed in low abundance (0.296 rel. int.). Ions containing three and four aluminum atoms are also observed ([A13(CH3)4H4]+ and [Al4(CH3)6H5]+) but were believed to be a function of the source pressure suggesting the possibility that ion/molecule reactions account for these higher mass ions. Note that none of the species above m/z 72 could be assigned to simple fragmentation of the dimer, but probably contained Al-I-I bonds from rearrangement processes. The results from Tanaka and Smith17 indicate that the methyl bridge is either extremely weak or susceptible to electron impact, since when more than one aluminum is found it appears to be held together by H-bridges due ‘ to the presence of Al-H bonds rather than methyl bridges. This may suggest that while structure 3 is the probable structure for A12(CH3)6, structure A. or ’2' may be more likely for the dimer ion, [A12(CH3)6]+. Ion cyclotron resonance (ICR) mass spectrometry18 is a powerful technique for the study of bimolecular gas—phase ion/molecule reactions. Precursors 118 of ion/molecule reaction products may be determined using ion cyclotron double resonance techniquesls. ICR seems especially applicable to the study of the mass spectrum of trimethylaluminum since the entire ICR cell is operated at ambient temperature (i.e., the greatest amount of dimer should be present). This study utilizes ICR to study the ion/molecule reactions observed in trimethylaluminum which may lead to an understanding of the presence of the higher mass ions observed in the previous studies. A previous study of trimethylaluminum using ICR by Staley et. al.19 reports the presence of only one ion/molecule reaction product ([A12(CH3)5]+) from the reaction of [Al(CH3)2]+ and [Al(CH3)3]+ with (presumably) the monomer. No ions indicative of the presence of the dimer were reported. This study also utilizes metal-containing ions as chemical ionization (CI) reagent ions in ICRZO'22 in an attempt to obtain molecular weight information on the species present in gas phase trimethylaluminum (i.e., monomer or dimer). Collision induced dissociation (CID) was performed on several ions present in the mass spectrum of trimethylaluminum using triple quadrupole mass spectrometry (TQMS) in an attempt to obtain structural information on these ions. EXPERIMENTAL The ion/mOIecule reactions were studied in an ion cyclotron resonance (ICR) mass spectrometer of conventional design (used in the "drift mode" with a marginal oscillator detector) which was constructed at Michigan State University and is described elsewhere”. Pressures were measured with a ‘ Veeco RG 1000 ionization guage. Spectra were recorded from m/z 10 to 300. The trimethylaluminum and tricarbonylnitrosylcobalt(0) were purchased from Alfa Products and were used as supplied except for multiple freeze-pump-thaw cycles to remove noncondensible gases. The trimethylaluminum was sampled from the vapor above a reservoir of liquid 119 trimethylaluminum present on the inlet at room temperature. The vapor was admitted to the ICR through a Varian 951-5106 precision leak valve. The metal ion chemical ionization study was performed with a 3:1 mixture of Co(CO)3NO and Al(CH3)3 at a total pressure of 4 x 10'6 torr. Collison induced dissociation (CID) experiments were performed on an Extranuclear triple quadrupole mass spectrometer (ELQ 400-3). Electron impact fragment ions and possible ion/molecule reaction products were formed in the El ion source. Pressures were measured with an ionization guage located outside the source in the source housing. Parent ions of interest were mass selected by the first quadrupole and accelerated into the second quadrupole. (the CID chamber). The collision gas was argon (typical pressure was 3 x 10'3 torr). Collision energies ranged from 10 to 30 eV (lab). The products of collision induced processes were then mass analyzed by scanning the third mass filter. RESULTS AND DISCUSSION ICR Studies The normalized ICR mass spectra (70 eV) obtained at six different pressures ranging from 5.0 x 10‘7 torr to 1.2 x 10‘5 torr are listed in Table 2. The obvious difference between the ICR spectra and the previously reported mass spectra (Table 1) is the lack of alkyl fragments (m/z 12 to 16) in the ICR mass spectra. These alkyl fragments were observed initiially in the ICR studies, but following several flushings of the inlet with trimethylaluminum these fragments were no longer observed. This suggests that the alkyl fragments which are observed may be due to water which is present in the inlet, since trimethylaluminum reacts with water to produce CH4-. Also, the intensities of the alkyl fragments in the previous studies are quite similar to the ion intensities in the electron impact mass spectrum of 011424. It should be noted, however, that the sensitivity of ICR is lower at small m/z values which may account for the 120 Table 2. ICR mass spectra of trimethylaluminum at various pressures (% rel. int.). Pressure (torr) m/z Probable Ion 5.01110“7 1.8111045 4.0x10-6 6.0x10-6 7.5x10-6 1.2x10-5 27 [1411* 7.0 15.5 15.6 12.3 10.6 9.6 28 [21111]+ 2.8 2.3 1.1 29 [A1H2]+ 3.1 4.5 2.9 1.0 1.0 0.7 41 [1110112]+ 1.6 1.1 0.8 42 [AICH3]+ 4.5 3.6 3.3 2.0 1.9 1.1 43 [AlCH3H]+ 2.8 2.0 1.7 1.3 0.7 0.7 57 [A1(CH3)2]+ 100.0 100.0 100.0 100.0 100.0 100.0 72 [AI(CH3)3]+ 9.7 7.0 3.9 3.2 2.9 1.5 102 [A12(CH3)3H3]+ 0.6 0.8 113 [A12(CH3)3CH2]+ 0.3 0.6 0.9 0.9 115 [A12(CH3)4HI+ 1.1 1.3 1.0 1.3 0.7 129 [A12(CH3)5]+ 1.6 7.9 12.7 23.1 27.3 37.2 131 [Al3(CH3)3H5]+ 0.5 0.7 0.9 1.1 145 [A13(CH3)4H4]+ 0.3 0.4 149 [Al4(CH3)2H11]"' 0.4 0.6 1.2 171 [413(0113)5]+ 0.4 0.5 0.6 0.6 173 [A13(CH3)5H2]+ 0.3 0.4 0.6 0.8 187 [Al3(CH3)7H]+ 0.5 0.5 0.7 0.9 or [Al4(CH3)5H4]+ 203 [A14(CH3)6H5]+ 0.4 0.6 0.8 1.3 219 [AI4(CH3)7H6]+ 0.2 0.5 0.9 121 absence of these ions. At 5.0 x 10'7 torr, only one ion is observed above m/z 72 (molecular ion of the monomer). As the pressure is increased, however, up to twelve ions are observed over m/z 72. Several of these ions are identical to those observed- in previous studies. Double resonance experiments were performed to determine if these ions resulted from ion/molecule reactions or from electron impact on the associated species. These experiments indicate that all of the ion intensities observed larger than m/z 72 result from ion/molecule reactions, thus suggesting the absence of the associated species at these experimental conditions. The majority of the probable ion structures listed in Table 2 are suggested by the reaction of the precursor ions (which are determined in the double resonance experiments) with the monomeric species, A1(CH3)3. The majority of these structures follow the trend which was observed by Tanaka and Smith” in which the number of constituents on the aluminum atoms is 3n—l, where n is the number of aluminum atoms (e.g., [A12(CH;;)5]+ and [A13(CH3)6H2]+). The ion [Al(CH3)2]+ was found to react with (presumably) the monomer to yield the ion/molecule reaction products at m/z 102, 113, 115, and 129 as shown in reactions 1 through 4. [A1(CH3)21+ + A1(CH3)3 -... [A12(CH3)3H3]+(m/z 102) + 02113 (1) _y [A12(CH3)3CH2]+(m/z 113) + CH4 (2) _yIA12(CH3)4H]+(m/z 115) + CH2 (3) L, [A12(CH3)5]+ (m/z 129) (4) The product ion [A12(CH3)5]+ is also formed from the reaction of [Al(Cl*13)3]+ as reported previously19 in reaction 5. 122 [Al(CH3)3]+ + Al(CH3)3 ——-)> [A12(CH3)5]+ (m/z 129) + CH3' (5) Double resonance experiments indicate that [A1(CH3)2]+ is the only precursor for the formation of the product ions at m/z 131 and 145. The empirical formula for the m/z 131 product ion could be [A12(CH3)5H2]+, but this structure is improbable since there would be more than three ligands per aluminum atom. A more probable structure for m/z 131 is [A13(CH3)3H5]"’ (which follows the "3n—1" rule) which must be formed from the reaction of [AI(CH3)2]+ with the dimer, A12(CH3)6, as shown in reaction 6. [Al(CH3)2]+ + A12(CH3)6 —1-»[A13(CH3)3H5]+ (m/z 131) + 5 CH2 (6) Note that a more thermodynamically fovorable process for the loss of 5 CH2 radicals may be the loss of the neutrals ethylene and propene. The structure of the m/z 145 product ion MEG. have three aluminum atoms present. This suggests once again the reaction of [A1(CH3)2]+ with the dimer as shown in reaction 7. [A1(CH3)2]+ + A12(CH3)5 —-1-—[A13(CH3)4H41+(m/z145) + 4 CH2 (7) This product ion could also be formed by a secondary reaction of [A12(C113)5]+ (m/z 129) with the monomer as shown in reaction 8, but 73 double resonance ' response was observed at m/z 129 for the m/z 145 product ion. [A12(CH3)5]+ + Al(CH3)3 ——>[A13(CH3)4H4]+(m/z 145) + 4 CH2 (8) Several of the primary product ions were Observed to react further with 123 the trimethylaluminum. The product ion [A12(CI~13)4I-l]+ (m/z 115) reacts further with the monomer to yield the product ions at m/z 171, 173, and 187 as shown in reactions 9 to 11. [A12(CH3)4H]+ + A1(CH3)3 [A13(CH3)61+(m/z 171) + CH4 (9) [A13(CH3)6H21+(m/z 173) + 0112(10) [A13(CH3)7H]+ (m/z 187) (11) The m/z 131 product ion ([AI3(CH3)3H5]+) reacts further with the monomer to yield the product ions at m/z 187 and 203 as shown in reactions 12 and 13. [Al3(CHg)3Hs]+ + Al(CH3)3 t [Al4(CH3)5H4]+(m/z 187) + CH4 (12) [Al4(CH3)6H5]+ (m/z 203) (13) Note that the m/z 187 product ion apparently has two different structures, [A13(CH3)7H]+ and [A14(CH3)5H4]+, as seen in reactions 11 and 12, respectively. The most intense product ion formed, [A12(CH;;)5]+ (m/z 129), also reacts further to yield product ions at m/z 149 and 219. These apparently are due to the reaction with the dimer as shown in reactions 14 and 15. [.412(0113)5]+ + A12(CH3)6 -[:[A14((:113)2Hn]+ (m/z 149) + {7 CH2 + 2 CH} (14) [A14(CH3)7H6]+ (m/z 219) + {2 CH2 + 2 CH} (15) Note that all of the product ions which presumably result from reactions of the dimeric species, A12(CH3)6, are of relatively small intensity and are observed only at higher pressures. This may suggest a pressure dependence on the amount of dimer present in gas phase trimethylaluminum. The frequent loss of CH2 124 radicals and formation of Al—H bonds in the observed ion/molecule reactions suggests the presence of H-bridges (structure '1') in these ions. In summary, the ion/molecule reactions observed in this study indicate that the ions above m/z 72 which were observed in the previous mass spectral studies may have been due to ion/molecule reactions occurring in the source of the mass spectrometer. Several ion/molecule reactions are observed, however, which do indicate the possibility of the presence of the dimeric species, A12(CH3)6, at higher pressures. The structures predicted for the ion/molecule reaction products agree with the generalization made by Tanaka and Smith17 concerning the presence of H-bridges (structuresql and '2') when more than one aluminum atom is present. The differences in the mass spectra of trimethylaluminum reported in previous studies and here suggests that sample handling may be an important aspect in acquiring mass spectra of these type of compounds. For example, the amount of alkyl fragment ions varies widely among the reported spectra which may result from water present in the inlet. In this study, an evacuated bulb containing liquid trimethylaluminum was placed on the inlet. The first mass spectra obtained contained large amounts of alkyl ions (possibly due to the presence of water) and (presumably) dimeric ions above m/z 72. The intensity of these higher mass ions, however, quickly disappeared. This suggests a process in which trimethylaluminum exists as a dimer in solution, but vaporizes as a monomer with the possibility of dimerization in the gas phase. If the ‘ monomer molecules in the gas phase are present at high enough pressures and for a long enough period of time, it is possible for a monomer/dimer equilibration to occur. This type of process is consistent with both the vapor pressure and mass spectral studies of trimethylaluminum. The mass spectra of the monomer may therefore be obtained by pumping the monomer through the mass 125 spectrometer as they are formed by vaporization of the liquid and by maintaining a low pressure of trimethylaluminum in the mass spectrometer (attainable in ICR but possibly not in conventional MS sources). Metal Ion CI The use of metal and metal-containing ions formed by electron impact on metal carbonyls, M(C0)n, as chemical ionization (CI) reagent ions has been suggested previouslyzo‘zz. The ion/molecule reactions between the metal and metal-containing ions and the analyte may provide both molecular weight and structural information. Molecular weight information is obtained from the reaction of metal-containing ions, [M(C0)x]+, with the analyte molecule (P) as shown in reaction 16. [M(C0)x]"’ + P ———)>[M(C0)x.aP]"' + aCO (16) The neutral molecule P may displace carbonyl ligands which are present on the reactant metal ion. The determination of the product ions and their precursors (using double resonance techniques) can yield information on the molecular weight of the neutral reactant P. The reactions of metal-containing ions formed by 70 eV electron impact ionization on Co(C0)3NO with trimethylaluminum were studied in the ICR in an attempt to determine the molecular weight (and thus evidence for monomer or dimer) of the species present. One ion/molecule reaction product from 1 the reaction of the bare metal, [Co]+, with trimethylaluminum was observed at m/z 101 which corresponds to the loss of C2116 (reaction 17). [001* + Al(CH3)3 —-)-[CoAlCH3]+ + 02116 (17) 126 The predicted reaction of the monomer and dimer of trimethylaluminum with [Co(CO)2]+, for example, would yield products indicative of displacement of the monomer and dimer for one or two carbonyl ligands as shown in reactions 18 to 21. [Co(Co)2]+ + Al(CH3)3 —[:[Co(co)A1(CH3)3]+(m/z 159) + (:0 (18) [CoAl(CH3)3]+(m/z 131) + 2 co (19) [Co(CO)2]+ + A12(CH3)5 —[: [Co(CO)A12(CH3)5]+ (m/z 231) + CO (20) [001112301196]+ (m/z 203) + 2 co (21) A lower limit on the ratio of dimer to monomer could be calculated from the intensities of these product ions as shown in the equation below. dimer 1(ICOAIZ(CH3)6]+) + I([Co(CO)A12(CH3)5]+) monomer 1([CoAl(CH3)3]$) + I ([Co(CO)A1(CH3)3]+) The observed reactions of the cobalt-containing ions with trimethylaluminum are shown in reactions 22 to 27. [Cocor‘ + Al(CH3)3 -—)>[CoAl(CH3)3]+(m/z 131) + co (22) [Co(co)2]+ + Al(CH3)3 —’ [Co(CO)Al(CH3)3]+(m/z 159) + co (23) 1000017014" + Al(CH3)3 —a-—[C0(NO)A1(CH3)3]+(m‘/z 161) + (:0 (24) [Co(CO)2NO]+ + A1(CH3)3 [Co(NO)Al(CH3)3]+(m/z 161) + 2 co (25) ‘EICOCONOA1(CH3)3]+ (m/z 189) + co (26) [Co(co)3No]+ + Al(CH3)3 —>[CoCONOAl(CH3)3]+ (m/z 189) + 2 co (27) The monomer, AI(CH3)3, is seen to displace one or two carbonyl ligands from the metal—containing reactant ion. _lig product ions were observed which would 127 indicate the presence of the dimer, A12(CH3)6 (reaction 20 or 21). These results suggest that trimethylaluminum exists only as a monomer in the gas phase at these conditions. It is possible, however, that the dimer may dissociate upon complexation to the metal ion (due to the relatively low dissociation energy) which would prevent the formation of products indicative of the dimer during metal ion chemical ionization. Collision Induced Dissociation Studies The collision induced dissociation (CID) spectra for the major ions present in the mass spectrum of trimethylaluminum were obtained using a triple quadrupole mass spectrometer (TQMS). A typical 70 eV electron impact mass spectrum of trimethylaluminum (4.0 x 10'5 torr) obtained with the TQMS is listed in Table 3. The TQMS mass spectrum is in good agreement with the ICR spectra and the previously reported spectra. Daughter ion (CID) spectra were obtained for ions of m/z 42, 57, 72, and 129. Daughter scans of the other ions were not possible due to the low intensity of the parent ions. The CID of [AICH3]+ (m/z 42) yielded m/z 27 ([Al]+) corresponding to a loss of 15 (CH3-) as shown in reaction 28. [AICH3]+-—é-r-> [Al]"' + 0113- (28) Daughter ions at m/z 27 ([Al]+) and m/z 42 ([A1CH3]+) (loss of one or two methyl groups) were observed in the CID spectrum of [AI(CH3)2]+ (m/z 57) as shown I in reactions 29 and 30. [411(c113)2]+ £1: [AICH3]+ + CH3° (29) [1111+ + 2 0113- (30) The molecular ion of the monomer, [Al(CH3)3]+ (m/z 72), undergoes CID to 128 Table 3. Typical 70 eV El mass spectrum of trimethylaluminum obtained on the TQMS. m/z %rel. int. 27 17.2 28 0.6 29 0.2 42 4.3 43 1.1 57 100.0 72 2.8 99 1.0 113 0.6 129 21.5 145 0.6 149 0.4 203 0.3 129 yield [Al(CH3)2]+ (m/z 57) (reaction 31) but no [AlCH3]+ (m/z 42) (loss of two methyl groups) is observed. [41(c113)31+ A” [A1(CH3)2]+ + (3113- (31)- The last ion studied was m/z 129 which was proposed to be [A12(CH3)5]+ from the ICR studies. This ion undergoes CID to yield m/z 57 ([A1(CH3)2]+) at low collision energies (~10 eV). At higher collision energies («30 eV), a small amount of [AICH3]+ is also observed. These processes are shown in reactions 32 and 33. [A12(CH3)5]+ A” - )[A1(CH3)2]+‘+ Al(CH3)3 (32) L-D[AICH3]+ + A1(CH3)3 + CH3° (33) Reaction 32 is not unexpected since the previously reported value of D(Al—C-Al) = 10.2 kcal/mole4 is relatively small. ' All of the CID processes observed for the ions studied here appear to result from the simple cleavage of the Al-CH3 bonds and the relatively weak bridging Al—C-Al bonds in the [A12(CH3)5]+ dimer ion. 130 REFERENCES 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. J. Boor, Ziegler-Natta Catalysts and Polymerizations, Academic Press, New York (1979). K.S. Pitzer and H.S. Gutowsky, J. Am. Chem. Soc. 68, 2204(1946). G.E. Coates, Organometallic Compounds, Wiley, New York(1960). A.W. Laubengayer and W.F. Gilliam, J. Am. Chem. Soc. 63, 477(1941). C.H. Henrickson and D.P. Eyman, Inorg. Chem. 6, 1461(1967). A. Almenningen, S. Halvorsen, and A. Haaland, Chem. Commun., 644(1969). R.J. O'Brien and G.A. Ozin, J. Chem. Soc. A, 1136(1971). J.E. House, J. Organomet. Chem. 263, 267(1984). K.S. Pitzer and R.K. Sheline, J. Chem. Phys. 16, 552(1948). P.H. Lewis and R.E. Rundle, J. Chem. Phys. 21, 986(1953). R.C. Vranka and E.L. Amma, J. Am. Chem. Soc. 89, 3121(1967). J.C. Huffman and W.F.. Streib, Chem. Commun., 911(1971). M.J.S. Dewar and D.B. Patterson, Chem. Commun., 544(1970). S.K. Byram, J.K. Fawcett, S.L. Nyburg, and R.J. O'Brien, Chem. Commun., 16(1970). RE, Winters and R.W. Kiser, J. Organomet. Chem. 10, 7 (1967). D.B. Chambers, G.E. Coates, F. Glockling, and M. Weston, J. Chem. Soc. _A_, 1712(1969). J. Tanaka and SR. Smith, Inorg Chem. 8, 265(1969). T.A. Lehman and MM. Bursey, Ion Cglotron Resonance Spectrometry, Wiley Interscience, New York (1976). MM. Kappes, J.S. Uppal, and R.H. Staley, Organometallics 1, 1303(1982). R.C. Burnier, G.D. Byrd, and 8.8. Freiser, Anal. Chem. 52, 1641(1980). D.A. Peake and M.L. Gross, Anal. Chem. 57, 115(1985). M. Lombarski and J. Allison, Int. J. Mass Spectrom. Ion Proc. 65, 31(1985). 131 23. B. Radecki and J. Allison, J. Am. Chem. Soc. 106, 946(1984). 24. E. Stenhagen, S. Abrahamsson, and F.W. McLafferty Eds., Atlas of Mass Spectral Data, vol. 1, Wiley Interscience, New York (1961). APPENDIX D Appendix D Unsuccessful ICR Studies Nickel—Containing Anions In addition to the study of the ion/molecule reactions of iron, chromium, and cobalt—containing anions with polar organic molecules, an attempt was made to study the ion/molecule reactions of nickel-containing anions. The only anion present in the 70 eV electron impact mass spectrum of Ni(CO)4 is the l7-electron species Ni(CO)3‘. This ion was unreactive with the organic molecules studied in this dissertation. Low energy electron impact, which was found to be useful in the iron, chromium, and cobalt studies, does not form a sufficient amount of Ni(CO)2" to allow the study of its ion/molecule reactions. Another nickel-containing compound, Ni(PF3)4, was utilized in an attempt to form nickel-containing non-17 electrom species. The major anions present in the mass spectrum of Ni(PF3)4, however, are fluorine-containing fragments such as PF3‘, PF2' etc. (presumably due to the high electron affinity of fluorine). Nickel-containing anions formed from electron impact on Ni(PF3)4 are of relatively low abundance. Isomer Analysis Using Metal Ion Chemical Ionization A study was attempted in which metal and metal-containing positive ions were utilized as chemical ionization (CI) reagent ions to distinguish between two pairs of isomers: cis— and trans-1,2-dichloroethylene and cis- and trans-1,2-cyclohexanediol. The 70 eV electron impact fragment ions from Co(CO)3NO, Fe(CO)5, and Cr(CO)5 were used in this study. It was believed that the differences in the ion/molecule reactions observed between the metal species and the two isomers would provide structural information indicating 132 133 the type of isomer present. For example, Co+ has been observed to abstract two chlorine atoms from 1,2-dichloroethane to form CoC12+ 25. This product ion may also be expected in the reaction of cis-1,2-dichloroethylene but may not be expected for the trans—1,2-dichloroethylene isomer due to the spatial separation of the chlorine atoms. The ion/molecule reactions observed between the metal ions and the dichloroethylene and cyclohexanediol isomers were nearly identical. No differences were observed in the ion/molecule reactions of the isomers which might provide structural information. Studies of Ligand Effects in Metal Ion/Molecule Reactions The study of ligand effects on metal ion reactivity (metal species formed by electron impact on MLle species) has been limited to the presence of carbonyl ligands (from M(C0)n) and. nitrosyl ligands (from Co(CO)3NO). This study was initiated in an attempt to determine the effects of the presence of ligands other than CO and NO on metal ion reactivity. These metal-containing ions were reacted with several n-alkanes, n-alcohols, and primary amines to compare the ion/molecule reactions observed with those previously reported with M+ and MCO+. The ion/molecule reactions of nickel-containing ions from electron impact on Ni(PF3)4 are similar to those observed for Ni+ and Ni(CO)x+, i.e., the PF3 ligands were either "spectators" in the ion/molecule reaction or were lost in a ligand substitution process similar to CO loss with Ni(CO)x+. No product ions were observed in which the PF3 ligand changed the reactivity of the metal ion (i.e., no evidence for active participation of the PF3 ligand). Chromyl chloride, CrOZClg, was also utilized since a variety of metal species are formed during electron impact (e.g., CrO+, CrCl", CrOCl+ etc.). The major difficulty in the study of Chromyl chloride is the presence of neutral-neutral reaction products with the organic molecules due to the extreme 134 reactivity of chromyl chloride. The use of Fourier transform mass spectrometry and ion ejection techniques would make the study of the ion/molecule reactions of these ions possible. Tungsten carbonyl (W(CO)6) is unique in that a large amount of WC+ is formed during 70 eV electron impact. This allows the possible study of the effect of the C ligand. No ion/molecule reactions, however, were observed from WC+ with the organic molecules studied here. LIST OF FOOTNOTES AND REFERENC' 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. LIST OF FOOTNOTES AND REFERENCES Huang, S.K.; Allison, J. Organometallics 1983, _2_, 883. Jones, R.W.; Staley R.H. J. Am. Chem. Soc. 1980, 1_02, 3794. Jacobson, D.B.; Freiser, B.S. J. Am. Chem. Soc. 1983, 105, 7484. Wronka, J.; Ridge, D.P. J. Am. Chem. Soc. 1984, 106, 67. Armentrout, P.B.; Beauchamp, J.L. J. Am. Chem. Soc. 1981, M, 784. Halle, L.F.; Armentrout, P.B.; Beauchamp, J.L. nganometallics 1982, _1_, 963. Hanratty, M.A.; Beauchamp, J.L.; Illies, A.J.; Bowers, M.T. J. Am. Chem. Soc. 1985, l_07_, 1788. Aristov, N.; Armentrout, P.B. J. Am. Chem. Soc. 1984, 196, 4065. Allison, J.; Ridge, D.P. J. Am. Chem. Soc. 1979, 101, 4998. Heck, R.F. "Organotransition Metal Chemistry", Academic Press, New York, 1974; pp. 76-110. Allison, J.; Ridge, D.P. J. Am. Chem Soc. 1976, §_8_, 7445. Tsarbopoulos, A.; Allison, J. Organometallics 1984, 3, 86. Jacobson, D.B.; Freiser, B.S. J. Am. Chem. Soc. 1983, 1_0_§, 5197. Byrd, G.D.; Freiser, B.S. J. Am. Chem. Soc. 1982, m, 5944. Allison, J.; Freas, R.B.; Ridge, D.P. J. Am. Chem. Soc. 1979, 1_01_, 1332. Armentrout, P.B.; Halle, L.F.; Beauchamp, J.L. J. Am. Chem. Soc. 1981, _10_3, 6624. Allison, J.; Ridge, D.P. J. Am. Chem. Soc. 1977, §_9_, 35. Radecki, B.D.; Allison, J. J. Am. Chem. Soc. 1984, 106, 946. Burnier, R.C.; Byrd, G.D.; Freiser, B.S. Anal. Chem. 1980, 5_2, 1641. Burnier, R.C.; Byrd, G.D.; Freiser, B.S. J. Am. Chem. Soc. 1981, 103, 4360. 135 21. 22. 23. 24. 25. 26. 27 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 136 Carlin, T.J.; Wise, M.B.; Freiser, B.S. Inorg. Chem. 1981, 33, 2743. Cassady, C.J.; Freiser, B.S.; McElvany, S.W.; Allison, J. J. Am. Chem. Soc. 1984, 106, 6125. Hodges, R.V.; Beauchamp, J.L. Anal. Chem. 1976, Q, 825. Peake, D.A.; Gross, M.L. Anal. Chem. 1985, g, 115. Lombarski, M.; Allison, J. Int. J. Mass Spectrom. Ion Phys. 1983, g, 281. Lombarski, M.; Allison, J. Int. J. Mass Spectrom. Ion Proc. 1985, 33, 31. Cody, R.B.; Burnier, R.C.; Reents, W.D.; Carlin, T.J.; McCrery, D.A.; Lengel, R.H.; Freiser, B.S. Int. J. Mass Spectrom. Ion Phys. 1980, 33, 37. Beauchamp, J.L.; Foster, M.S. J. Am. Chem. Soc. 1971, 33, 4924. Allison, J.; Huang, S.K.; Lombarski, M.; McElvany, S.W.; Radecki, B.; Tsarbopoulos, A. American Society for Mass Spectrometry Meeting, Boston, 1983. Pignatoro, S.; Foffani, A.; Grasso, F.; Cantone, B. Z. Phys. Chem. 1965, :11, 106. Winters, R.E.; Kiser, R.W. J. Chem. Phys. 1966, 33, 1964. Compton, R.N.; Stockdale, J.A.D. Int. J. Mass Spectrom. Ion Phys. 1976, 33, 47. Dunbar, R.C.; Ennever, J.P.; Fackler, J.P. Inorg. Chem. 1973, 33, 2734. Foster, M.S.; Beauchamp, J.L. J. Am. Chem. Soc. 1975, 31, 4808. Weddle, G.H.; Ridge, D.P. American Society for Mass Spectrometry Meetigg, New York, 1980. Sallans, L.; Lane, K.; Squires, R.R.; Freiser, B.S. J. Am. Chem. Soc. 1983, 1_0_§_, 6352. Lane, R.R.; Lee, R.E.; Sallans, L.; Squires, R.R. J. Am. Chem. Soc. 1984, 30_6, 5767. McDonald, R.N.; Schell, P.L.; McGhee, W.D. Organometallics 1984, 3, 182. McDonald, R.N.; Chowdhury, A.K.; Schell, P.L. J. Am. Chem. Soc. 1984, 115, 6095. Stevens, A.E.; Feigerle, C.S.; Lineberger, W.C. J. Chem. Phys. 1983, _7_8_, 5420. Corderman, R.R.; Beauchamp, J .L. Inorg. Chem. 1977, _13, 3135. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 137 Lehman, T.A.; Bursey, M.M. "Ion Cyclotron Resonance Spectrometry", Wiley-Interscience, New York, 1976. Beauchamp, J.L. Annu. Rev. Phys. Chem. 1971, _23, 527. Wanczek, K.P. Dynamic Mass Spectrometry 1981, 3, 14. Dunbar, R.C. "Physical Methods in Modern Chemical Analysis", v.2, T. Kuwana Ed., Academic Press, New York, 1980; pp. 277-336. Beauchamp, J.L.; Anders, L.R.; Baldeschwieler, J.D. J. Am. Chem. Soc. 1967, 8_9_, 4569. Warnick, A.; Anders, L.R.; Sharp, T.E. Rev. Sci. Instr. 1974, 33, 929. Burrow, P.D.; Modelli, A.; Chiu, N.S.; Jordan, K.D. J. Chem. Phys. 1982, 31, 2699. Benson, S.W. "The Foundations of Chemical Kinetics", McGraw Hill Book Co., Inc., New York, 1960; p. 519. Dipole moment values were obtained from reference 51. West, R.C., Ed. "Handbook of .Chemistry and Physics", 56th ed.; CRC Press: Cleveland, 1976. Engelking, P.C.; Lineberger, W.C. J. Am. Chem. Soc. 1979, E1, 5569. Janousek, B.; Brauman, J.I. "Gas Phase Ion Chemistry", v.2, M.T. Bowers, Ed., Academic Press, N.Y., 1979. _ Collman, J.P. Acc. Chem. Res. 1975, 3, 342. Page, F.M. "Free Radicals in Inorganic Chemistry, Advances in Chemistry Series", R.F. Gould Ed., v.36, pp. 68-75, American Chemical Society, Washington D.C., 1962. ‘ Stull, D.R.; Prophet, H., Eds. "JANAF Thermochemical Tables" NSRDS—NBS 37, U.S. Gov't. Print. Off.: Washington D.C., 1971. Burrow, P.D.; Jordan, K.D. Chem. Phys. Lett. 1975, 33, 594. Thermodynamic values obtained from reference 59. Rosenstock, H.M.; Draxl, K.; Steiner, B.W.; Herron, J.T. J. Phys. Chem. Ref. Data Suppl. 1977, l, 6. Janousek, B.K.; Zimmerman, A.H.; Reed, R.J.; Brauman, J.I. J. Am. Chem. Soc. 1978, w, 6142. Ellison, G.B.; Engelking, P.C.; Lineberger, W.C. J. Phys. Chem. 1982, 33, 4873. Williams, J.M.; Hamill, W.H. J. Chem. Phys. 1968, Q, 4467. 63. 64. 65. 66. 67 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 138 Ferguson, E.E.; Dunkin, D.B.; Fehsenfeld, F.C. J. Chem. Phys. 1972, _5_7, 1459. Hotop, H.; Lineberger, W.C. J. Phys. Chem. Ref. Data. 1975, 3, 539. Compton, R.N.; Reinhardt, P.; Cooper, C.D. J. Chem. Phys. 1978, Q, 4360. Compton, R.N.; Reinhardt, P.; Cooper, C.D. J. Chem. Phys. 1975, 33, 3821. . Chipman, D.M. J. Phys. Chem. 1978, _8__2_, 1080. Siegal, M.W.-; Celotta, R.J.; Hall, F.L.; Levine, J.; Bennett, R.A. Phys. Rev. 1972, _A_3, 607. MacKay, G.I.; Lien, M.H.; Hopkinson, A.L.; Bohme, D.K. Can. J. Chem. 1978, _53, 131. Ellis, H.B.; Ellison, G.B. J. Chem. Phys. 1983, 13, 6541. Richardson, J.H.; Stephenson, L.M.; Brauman, J.I. Chem. Phys. Lett. 1974, g_5_, 321. Flesch, G.D.; White, R.M.; Svec, H.J. Int. J. Mass Spectrom. Ion Phys. 1969, 3, 339. Batt, L.; Robinson, G.N. "The Chemistry of Amino, Nitroso, and Nitro Compounds and their Derivatives"; Patai, S.; Ed.; Wiley: New York, 1982; Part 2, p.1075. Feigerle, C.S.; Corderman, R.R.; Bobashev, S.V.; Lineberger, W.C. 3._ Chem. Phys. 1981, 33, 1580. Corderman, R.R.; Engelking, P.C.; Lineberger, W.C. J. Chem. Phys. 1979, 33, 4474. - Refaey, K.M.A.; Franklin, J.L. Int. J. Mass Spectrom. Ion Phys. 1976, .22, 19. Tsuda, S.; Yokohata, A.; Michihiro, K. Bull. Chem. Soc. Jap. 1970, :13, 1649. Melton, G.E.; Rudolph, P.S. J. Chem. Phys. 1959, 31., 1485. Tsuda, S.; Yokohata, A.; Michihiro, K. Bull. Chem. Soc. Jap. 1969, :13, 607. R.R. Squires, personal communication. Boor, J. "Ziegler-Natta Catalysts and Polymerizations", Academic Press, New York, 1979. 139 82. Kappes, M.M.; Uppal, J.S.; Staley, R.H. Organometallics 1982, _l_, 1303.