J . Am. Chem. SOC.1991, 113, 8765-8770
8765
Gas-Phase Acidities of (CO)5MnH, (C0)4FeH2,and ( CO)~COH Amy E.Stevens Miller**'and J. L. Beauchamp Contribution No. 6563 Jrom t h e A r t h u r Amos Noyes Laboratory of Chemical Physics, Cali$ornia Institute of Technology, P a s a d e n a , California 91 125. Received March 25, 1991
Abstract: The gas-phase acidities of (CO)5MnH, (CO),FeH2, and (CO),CoH are determined by ion cyclotron resonance studies of their deprotonation by reference anions. These compounds a r e all extremely strong acids in the gas phase, with AHacid[(CO)5MnH] = 318 f 4 kcal/mol, AHaCid[(CO),FeH2] = 319 f 5 kcal/mol, and AHaCid[(CO),CoH] 5 314 f 1 kcal/mol. When combined with the known metal-hydrogen homolytic bond energies, the acidities determine the electron affinities, EA[(CO)5Mn] = 2.4 & 0.3 eV, EA[(CO),FeH] I2.8 f 0.8 eV, and EA[(CO),Co] 2 2.3 eV. Negative-ion mass spectra and negative-ion-molecule reactions of these complexes are discussed. The neutral hydrides dissociatively attach thermalized (298 K) electrons, with each of the hydrides showing loss of CO on the attachment. Loss of H2to produce (CO),Fe- is competitive with CO loss on electron attachment to (C0),FeH2. In a process consistent with its strong gas-phase acidity, loss of an H atom to produce (CO),Co- is competitive with detachment of CO in electron attachment to (CO),CoH.
Introduction The determination of gas-phase acidities, as defined by the free energy change of reaction 1, has led to a n understanding of intrinsic substituent effects and the influence of solvation on this important molecular p r ~ p e r t y . ~ - A I ~classic example is from the AH A- + H+ AGacid = AHacid - TAsa,id (1) gas-phase acidities of the alcohols, which follow the order neopentyl > tert-butyl > isopropyl > ethyl > methyl, exactly the opposite of their solution behavior." This is interpreted as resulting from poor solvation of t h e anion with increasing bulkiness, which overrides the effect of alkyl group polarizability stabilizing the anionic ~enter.~.I~J~ Gas-phase acidities are also important in t h a t t h e enthalpy change for the acidity, AHaCid[AH], is related to the electron affinity of A , EA[A], and the homolytic bond energy, D[A-HI, as shown in Scheme 1 and eq 2; determining any two of these quantities allows additional thermochemistry to be derived. AHacid[A-H] = D[A-HI - EA[A] + 313.6 kcal/mol ( 2 ) We report here the gas-phase acidities of three organometallic carbonyl hydrides, ( C O ) 5 M n H , (CO),FeH,, and (CO),CoH,
Scheme I AH
L A -
+
H+
t
+
( I ) Present address for correspondence: PL/LID, Hanscom AFB, MA 0173 1-5000. Permanent address: Department of Chemistry and Biochem-
istry, University of Oklahoma, Norman, OK 73019-0370. (2) Bartmess, J. E.; Mclver, R. T., Jr. The Gas-Phase Acidity Scale. In Gas Phase Ion Chemistry; Bowers, M. T., Ed.; Academic Press: New York, 1979; Vol. 2. (3) Bartmess, J. E.; Scott, J. A.; McIver, R. T., Jr. J . Am. Cfiem. SOC. 1979, 101, 6056. (4) Moylan, C. R.; Brauman, J. 1. Annu. Reo. Pfiys. Cfiem. 1983, 34, 187. (5) Caldwell, G.; McMahon, T. B.; Kebarle, P.; Bartmess, J. E.; Kiplinger, J. P. J . Am. Cfiem. Soc. 1985, 107, 80. (6) Henchman, M.; Viggiano, A. A.; Paulson, J. F.; Freedman, A,; Wormhoudt, J. J . Am. Cfiem. SOC.1985, 107, 1453. (7) Sallans, L.; Lane, K . R.; Squires, R. R.; Freiser, B. S. J . Am. Chem. SOC.1985, 107, 4379. (8) Paulson, J. F.; Viggiano, A. A.; Henchman, M.; Dale, F. In Symposium on Atomic and Surface Physics, Obertraun, Austria, February 1986; Howorka, F., Lindinger, W., MBrk, T. D., Eds.; Studia Studienf~rderungsgesellschaftm. b. H.: Innsbruck, Austria, 1986. Henchman, M.; Viggiano, A. A.; Paulson, J. F.; Dale, F.; Deakyne, C. Faraday Discuss. Cfiem. Soc. 1988,85,87 ff. Viggiano, A. A,; Henchman, M. J.; Dale, F.; Deakyne, C. A.; Paulson, J. F. J . Am. Cfiem.Soc., submitted for publication. (9) Lias, S.G.; Bartmess, J . E.; Liebman, J. F.; Holmes, J. L.; Levin, R. D.; Mallard, W. G . Gas-Phase Ion and Neutral Thermochemistry. J . Pfiys. Cfiem. Ref Data 1988, 17, Suppl. No. 1. (IO) Wetzel, D. M.; Salomon, K. E.; Berger, S.;Brauman, J. 1. J . Am. Cfiem. SOC.1989, 111. 3835. ( 1 1 ) Han,C.-C.; Brauman. J. I . J . Am. Cfiem. SOC.1989, I l l , 6491 and references therein. (12) Ervin, K. M.; Gronert, S.;Barlow, S. E.; Gilles, M. K.; Harrison, A. G.; Bierbaum, V. M.; DePuy. C. H.; Lineberger, W. C.; Ellison, G. B. J . Am. Cfiem. SOC.1990, 112, 5750. (13) Brauman, J. 1.; Blair, L. K. J . Am. Cfiem. SOC.1970, 92, 5986. (14) Boand, G.;Houriet, R.; GBumann, T. J . Am. Cfiem. SOC.1983,105, 2203.
0002-7863/91/ 1513-8765$02.50/0
A + H
, A + H + + ~ IP[H] = 313.6 kcaUmol
determined by using t h e techniques of ion cyclotron resonance
spectro~copy.'~Acidities of these complexes in aqueousl6I8 and a ~ e t o n i t r i l e solutions '~ have been reported; both (C0)5MnH and (CO),FeH2 are weak acids, and (CO),CoH is a strong acid, being fully dissociated in aqueous solution.20 Poor solvation of the bulky carbonyl anions is expected t o be a major factor in determining the aqueous acidities of these complexes; it is of interest to quantify this effect. These data and the electron affinities could provide a relatively direct means of corroborating the metal-hydrogen bond dissociation energies which have been reported for (CO)5MnH and (C0),CoH2' and a means of determining the bond energy in ( C 0 ) 4 F e H 2 .
Experimental Section The techniques of ion cyclotron resonance (ICR) spectroscopy have been d e ~ c r i b e d . 'The ~ spectrometer used in this study was built in our laboratories and utilizes a 15411. electromagnet capable of 23.4-kG maximum field. Standard marginal oscillator detection was employed. Pressures were measured with a Schulz-Phelps-type ion gauge calibrated against an MKS Baratron Model 90Hl-E capacitance manometer; the estimated uncertainty in absolute pressure is f20%. All electron energies are uncalibrated. All reactions were identified by double-resonance technique^;'^ neutrals were not detected. These experiments were performed at ambient temperature (298 K). Negative- and positive-ion mass spectra were recorded in the drift mode. Negative-ion mass spectra with thermalized electrons were taken in the trapping mode, and spectra were recorded following a 3-ms pulse of 70-eV electrons. The electrons were thermalized by using high pressures of C 0 2 , as described in the literature.22 Pressures of the metal (15) Beauchamp, J . L. Annu. Rev. Phys. Chem. 1971, 22, 527. (16) Hieber, W.; Wagner, G. 2. Naturforsch. 1958, 138, 339. (17) Hieber, W.; Hiibel, W. 2. Elektrocfiem. 1953, 57, 235. (18) Galembeck, F.; Krumholz, P. J . Am. Chem. SOC.1971, 93, 1909. (19) Moore, E. J.; Sullivan, J. M.; Norton, J. R. J . Am. Cfiem. SOC.1986, 108, 2257. (20) A pK,, of 6.8 for (C0)4FeH, has been determined in methanol: Walker, H. W.; Kresge, C. T.; Ford, P. C.; Pearson, R. G. J . Am. Cfiem. Soc. 1979, 101, 7428. (21) Martinho Simks, J. A,; Beauchamp, J. L. Cfiem. Reo. 1990,90,629 and references therein. (22) Woodin, R. L.; Foster, M. S.;Beauchamp, J. L. J . Cfiem. Pfiys. 1980, 72, 4223.
0 1991 American Chemical Society
Stevens Miller and Beauchamp
8766 J . Am. Chem. SOC.,Vol. 113, No. 23, 1991 complexes were kept low (( 1-4) X IO-* Torr) to ensure electron thermalization before attachment and to prevent rapid reaction of the anions by cluster-forming reactions. All ion-molecule reactions reported here were studied by using the trapping mode and typically observed for 500-1000 ms following a 3-10-ms electron pulse. Reagent anions were made by electron impact on an appropriate neutral: F from NF3, CI' from CCI4, Br- from CH3Br, I- from CH31, and CFICO< from CFIC0 2 H . Pressures of the reagent neutrals were in the range of 2-10 times the pressure of the metal carbonyl hydrides, and the electron-impact energy (typically near 5 eV) was optimized for the highest yield of each desired reagent anion. Under these conditions, the initial electron pulse produced predominantly reagent anions rather than metal-containing anions. Scattered free electrons can also be trapped in the ICR spectrometer and undergo dissociative attachment reactions with the metal carbonyl hydrides. We found that optimization of all reaction conditions (pressures of reagent and metal complexes, electron pulse length and energy, trapping voltages) proved more effective than electron ejection techniques2' at minimizing complications from the electron attachment reactions. Furthermore, the electron attachment reactions are extremely rapid, and the free electrons are scavenged within 100 ms of the initial electron pulse, while all reactions reported here were confirmed by double resonances at long reaction times. The kinetics of the reactions reported here could not be determined, since gaseous decomposition products precluded measurements of the neutral hydride pressure. Even in the event that decomposition products were not present, cluster-forming reactions between the metal anions and the parent neutral hydride and ligand displacement, ligand exchange, and nucleophilic displacement reactions involving the other neutral reagents made the reaction schemes too complex to determine accurate reaction rates. All of the protontransfer reactions reported here are qualitatively fast (at or near the collision rate), which is indicative of exothermic proton transfer. The cluster-forming reactions appeared qualitatively slow. Pentacarbonylhydridomanganesewas prepared by reaction of Na[ (CO)5Mn] with phosphoric acid.24 Final purification was effected by pumping on the dry ice/acetone-cooled sample until no impurities were observed in the mass spectrum. Tetracarbonyldihydridoiron was prepared by acidification of (CO),FeH- as described in the literature.25 The sample was purified by cooling with dry ice/acetone and pumping on the sample to remove volatile impurities. The sample of (C0)4FeH2was stored on dry ice. A fresh sample was prepared each day by warming the stock sample to -38 OC and vacuum-transferring the (C0)4FeH2. This sample was held at -38 "C during the experiments. Despite all these precautions, substantial decomposition occurred, as evidenced by CO and Fe(CO)5 in the mass spectra; this was attributed in large part to decomposition in the inlet system and near the hot filaments of the ICR cell. Tetracarbonylhydridocobaltwas prepared from [Co(pyridine),] [Co(CO)4]2 as described in the literature.26 (CO),CoH was purified by several vacuum distillations from the sample held at -45 OC and stored in dry ice. A sample bulb was prepared daily from this stock by warming it to 0 "C and vacuum-transferring the (C0)4CoH. This bulb was held at 0 "C during the course of the experiments; colder baths gave insufficient vapor pressure for easy handling. The sample was degassed several times during use. Again, decomposition was extensive, with CO+ and C 0 2 + being the most abundant ions in the positive-ion mass spectrum. Decacarbonyldimanganese was obtained from Alfa Inorganics and used as supplied. All other reagents were used without purification, except for degassing with several freeze-pumpthaw cycles.
-
Results Mass Spectra and Negative-Ion-Molecule Chemistry of (CO),MnH. T h e positve-ion mass spectra of ( C O ) , M n H recorded on o u r instrument agree well with those reported in t h e literaT h e 70-eV negative-ion mass spectrum of ( C O ) , M n H , taken in the drift mode a t a pressure of 3 X lo-' Torr, shows the ions (CO),Mn- (6%), ( C 0 ) 4 M n H - (91%), and (CO)3MnH- (3%). T h e spectrum is unchanged from 7 0 to 0.5 eV. T h e negative-ion spectrum with thermalized electrons (1:90 to 1:400 mixture of ( C O ) S M n H with C 0 2 )shows 100% (CO),MnH-. (23) Beauchamp. J. L.;Armstrong, J. T. Rev. Sci. Insfrum. 1969,40, 123. (24) Davison, A.; Faller, J. W. Inorg. Chem. 1967, 6 , 845. (25) Farmery, K.; Kilner, M. J . Chem. Soc. A 1970,634. (26) Sternberg, H. W.; Wender, 1.; Friedel, R. A.; Orchin, M. J . Am. Chem. SOC.1953, 75. 2717. (27) Saalfeld. F. E.: McDowell, M. V.; DeCorpo, J . J.; Berry, A. D.; MacDiarmid, A. G. Inorg. Chem. 1973, 12, 48. (28) Mays, M. J.; Simpson, R. N . F. J . Chem. SOC.A 1967, 1936.
Table I. Gas-Phase Acidities of Reference Swcies'
HX
AHacid[HXl 371.5 f 0.2 333.4 f 0.2 323.5 f 0.1
HF HCI HBr
HX CFjCO2H HI
AHacid[HXI 322.9 f 4.1 314.4 f 0.1
"All data are in kcal/mol at 298 K and are taken from ref 9. Enthalpies for HF, HCI, HBr, and HI are calculated from bond energies and electron affinities as given by Scheme 1 and eq 2. AGaCidfor CF,C 0 2 H (316.3 f 1.9 kcal/mol) is determined by ion-molecule equilibrium reactions and used to determine AHaddby eq 1, using a calculated entropy (which incurs the additional error in the enthalpy). The equilibrium method applied to HBr suggests that CF,C02H may be less acidic than HBr by -0.5 k ~ a l / m o l ,but ~ this is so slight that it does not change the acidity of (CO)5MnH reported here, for which CF3C02His used as a reference. Table 11. Selected Values for Gas-Phase Acidities of Hydrides" complex AHac,d[HX] complex (C0)5MnH 318 f 4 (C0)4CoH (C0)4FeH2 319 f 5 (C0)4FeH "All data are in kcal/mol at 298 K.
Metal Carbonyl AHacid[HX] 1314 f 1 2319 f 5
A study of the negative-ion-molecule chemistry of ( C 0 ) 5 M n H at 4 X T o r r shows a t long reaction times t h e formation of very small amounts of M n 2 ( C 0 ) 9 - and Mn2(CO)*- arising from ( C 0 ) 4 M n H - and of M n 2 ( C 0 ) , - arising from ( C O ) 3 M n H - (eqs 3a, 3b, a n d 4). W e could not unequivocally attribute these (CO),MnK
+
(C0)3MnH-
(CO),MnH
c
+ (CO)sMnH
Mn~(C0)9-+ H2 (major channel) (3a) Mn2(COh- + CO + H2 (minorchannel) (3b)
-
Mn2(C0)f
+ CO + H2
(4)
products to reactions with the hydride; t h e reactions are slow, and addition of M n 2 ( C O ) , o increases the abundance of these ions dramatically. These facts suggest t h a t these ions could be due to electron transfer with M n 2 ( C O ) l o , possibly present in trace a m o u n t s from decomposition of ( C O ) s M n H . Acidity of (CO),MnH. T h e negative ions X- = F, CI-, a n d CF3CO; all react with (CO)SMnH to give only the proton-transfer product (CO),Mn-, according to eq 5. In a 1O:l mixture of (CO),MnH
+ X-
-
(CO)SMn-
+ XH
(5) Torr,
CF3C02Hwith (CO),MnH a t a total pressure of 4.4X t h e ion intensity ratio of (CO),Mn- to CF3C0; was 1.3:l after a 500-ms trapping time; ion loss from reactions and trapping
inefficiency is too great to allow observation of the ions a t longer times. T h e double-resonance experiments indicate that virtually all (CO)5Mn- observed a t this time is a result of proton abstraction by CF3C02-(eq 3,b u t no CF3C0c is produced by t h e reverse of reaction 5. This means that t h e reaction is still proceeding in the forward direction a n d an equilibrium situation has not been achieved. A limit on the free energy change for this particular proton-transfer reaction is therefore given by AGO I-RT In { [(CO),Mn-] [ C F 3 C 0 2 H ]/ [ C F 3 C 0 2 - ][ ( C O ) S M n H ]), with t h e concentration ratios given by the ion intensities and the measured neutral pressures. A slight entropy change (-0.8 eu) is calculated from the rotational symmetry numbers of reactants a n d products a n d used in conjunction with t h e limit of AGO, resulting in AH 5-1.8 kcal/mol, that is, AHaa[(CO)sMnH] 5 A H a d [ C F 3 C 0 2 H ] - 1.8 kcal/mol. For X- = I-, neither reaction 5 nor any other reaction occurs. T h e reverse reaction, in which (CO)SMn- would abstract a proton from HI to give I-, cannot be observed on the ICR time scale ( k Ilo-" cm3.molecule-~.s-~),nor is any reaction of (CO)sMn- with H I observed. T h e acidity of ( C O ) s M n H is bracketed between that of HI and 1.8 kcal/mol less than that of CF3C02H;AHacidvalues for the A value of AHacidreference acids a r e given in Table I.
Gas-Phase Acidities of Carbonyl Hydrides Table Ill. 70-V Positive-Ion Mass Spectral % ion mle ion current m/e 169 (C0)4FeH+ 1 85 168 (CO),Fe+ 3 84 140 (CO),Fe+ 2 68 11 3 (C0)2FeHt 14 56 112 (CO)>Fe+ 15
J . Am. Chem. SOC.. Vol. 113, No. 23, 1991 8767
-
according to eq IO. For X- = I-, neither reaction IO nor any (C0),FeH2 + X(CO),FeHHX (10)
Data for (CO),FeH2 5% ion ion current (CO)FeH+ 12 (CO)Fe+ 27 FeC+ 3 Fe+ 23
[(CO),MnH] = 318 f 4 kcal/mol is chosen and is presented in Table 11. The absence of additional reference species with an acidity within this range precluded further refinement of this val ue.29 Particularly in the absence of either an equilibrium situation or observation of the back-reactions, there is concern that (CO),Mn- arises by reaction with any Mn2(CO),o from decomposition of the hydride. Two pieces of evidence counter this possibility. First, although the kinetics of proton abstraction could not be determined quantitatively, the observed reactions of F, C1-, and CF3C02-,as given by eq 5, are qualitatively much too fast to be attributed to reaction with any trace Mnz(CO)loimpurity. Second, we examined the reactions of the halides, including I-, with Mn2(CO)lo. Reactions of the halides do occur, producing (CO),Mn-, as shown for 1- in eq 6. The fact that (CO),MnMn2(CO)lo+ I-
-
(CO)5Mn- + (CO),MnI
(6)
was not observed from the reaction of I- with the sample of (CO),MnH gives some confidence that Mn2(CO)lois not a contributor to the reactions we observed with the other X- ions. Mass Spectra and Negative-Ion-Molecule Chemistry of (C0),FeH2. The 70-eV positive-ion mass spectrum of (C0),FeH2 has not been reported; the relative ion abundances from the spectrum obtained by using ICR spectroscopy are reported in Table 111. Ion abundances are somewhat uncertain, as these were determined by subtracting ion intensities due to (CO),Fe from the observed mass spectrum; typically, -50% of the total ion current of iron-containing ions was attributed to (CO),Fe. The spectra also show CO+ as 20-60% of the total ion current. The (CO),,FeH2+ions are not observed (abundance
