Electron transfer and Multi-Atom Abstraction Reactions between Atomic Metal Anions and NO, NO2 and SO2

J. M. Butson, S. Curtis and P.M. Mayer

Department of Chemistry and Biomolecular Sciences, University of Ottawa, Ottawa Canada K1N 6N5

Keywords: metal anion, electron transfer, NO, NO2, SO2

Abstract

The atomic metal anions Fe¯, Cs¯, Cu¯ and Ag¯ were reacted with NO, NO2 and SO2 to form

intact NO¯, NO2¯ and SO2¯ with no fragmentation. Yields for the molecular anions ranged from

4 to 97% and were found to correlate to the exothermicity of the electron transfer process.

Sequential oxygen atom extraction was found to take place between the metal anions and NO

and NO2. Reactions between NO2 and Fe¯ resulted in FeO¯, FeO2¯ and FeO3¯ while reactions of

Cu¯ with NO2 resulted in CuO¯ and CuO2¯. Reactions of Cu¯ and Ag¯ with NO resulted in

CuO¯ and AgO¯ respectively.

1

1

23

4

56

7

8

10

11

12

13

14

15

16

17

18

12

Introduction

Electron attachment processes are most significant at low electron energies,

approximately 15 eV and below.[1] The process of attaching electrons to neutral molecules can

be described in terms of the resonance model, where a neutral molecule (AB) captures an

incoming electron to form an excited state (AB*¯) (Supplemental Information, Figure S1, A) or a

neutral molecule in an excited state captures an incoming electron to form a stable product anion

(Supplemental Information, Figure S1,B).[2-5] This attachment can result in two different

‘trapping mechanisms’ with which the electron is attached, and are referred to as shape or

Feshbach resonances.[4] Once an electron is attached to a molecule, elastic/inelastic scattering as

well as dissociative/non-dissociative processes can occur.[4] Dissociative attachment occurs

when there exists a resonant state of the correct asymptotic form that prevents electron emission

and allows dissociation to occur (Supplemental Information, Figure S1, C).

Numerous electron attachment studies of NO, NO2 and SO2 (with electron affinities of

0.024, 2.2730 and 1.1070 eV, respectively)[5-7] have demonstrated that these neutral molecules

dissociate into various fragments under electron impact at low electron kinetic energies in the gas

phase via the scheme in ‘Supplemental Information, Figure S1, C’.[8-15]

Dissociative Electron Attachment to NO, NO2 and SO2

Fragments associated with electron attachment to NO were first detected by Tate et al. in

1932.[8] They inferred that electron attachment had occurred by measuring the atomic anion

currents associated with O¯ formation from NO.[8] This assessment was confirmed by a number

of authors as mentioned by Sambe et al.[9] O¯ is observed around electron resonance energies

2

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

12

of 8-9 eV, as well as a Feshbach resonance associated with N¯.[9] Of note is the study performed

by Chantry et al. where the kinetic-energy distribution of the O¯ ions demonstrated that the

dissociative electron attachment process in NO led exclusively to the production of O¯ and a N

atom in an excited state.[10]

Electron attachment to NO2 yields O¯ at resonance energies of 1.4, 3.1 and 8.3 eV.[11]

The authors also observe NO¯ and O2¯ anions, similar to the results reported by Abouf et al.[11,

12] Electron attachment to SO2 yields O¯, S¯ and SO¯ as observed in an electron attachment

study reported by Spyrou et al.[13] In the study they report onset energies of 4.55 and 7.3 eV for

the formation of O¯, 4.2 and 7.24 eV for the formation of S¯ and 4.85 eV for the formation of

SO¯, within experimental error of two similar studies reporting such values.[14, 15]

While both electron attachment and electron transfer result in the transfer of electrons,

many key differences exist. These differences, as well as their implications, will be discussed in

the next section.

Atomic Electron Transfer

Electron transfer reactions are similar to electron attachment reactions in that the transfer

of an electron from one molecule/atom to the next occurs according to the Franck-Condon

condition, i.e. at constant atomic coordinates,[16] according to the following relation:

Scheme 1

where M¯ is the molecular/atomic anion, AB is the initial neutral molecule, M¯[AB] and

M[AB]¯ represent the encounter complex before and after the electron transfer (Supplemental

3

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

12

Information, Figure S1, D) and E‡ represents an energy barrier to electron transfer.[16, 17] The

rate constants kc and kb represent the initial formation and dissociation of the encounter complex

while kp represents the kinetic bottleneck that can occur when an internal energy barrier for

electron transfer is present due to the geometries of M¯[AB] differing from M[AB]¯.[16-18] For

exothermic reactions, when the chemical barrier to electron transfer (E‡) is low, kp >> kb and

formation of the encounter complex is rate limiting and proceeds at near collision rates.[19,20]

Since formation of the encounter complex is nearly independent of temperature, these reactions

will proceed with little to no temperature dependence.[19,20] In situations where an exothermic

reaction possesses a large internal energy barrier, kp << kc or kb, passage over this barrier

becomes rate limiting and a positive temperature dependence is associated with the reaction rate.

[21-24] Prior experiments and calculations performed by Kebarle et. al. found that rate constant

calculations regarding exothermic electron transfer reactions between simple molecules and

NO2/SO2 were in good agreement with the theoretical predictions.[16]

As noted from electron attachment experiments to NO, NO2 and SO2, even small

structural differences between the neutral molecule and product anion can prevent the formation

of the intact product anion (Supplemental Information, Table S1). Dissociative electron

attachment typically stems from the nature by which electrons approach the neutral molecule

(reaction kinetics) and changes to the molecule’s internal energy.[18-20] Electronic transitions

are rapid on the nuclear time scale, in the Franck-Condon region.[20, 21] Due to this, the nuclear

geometry associated with the ground state of the neutral molecule will not be immediately

altered from an electronic transition creating an anion.[20, 21] The resultant anion is thus formed

in an energy level above the ground state of the intact molecular anion (Supplemental

Information, Figure S1, A).[20, 21] As noted by Marcus other factors beyond Franck-Condon,

4

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

12

such as internal energy also affect charge transfer.[19] To account for this, he added a term

within his theory describing the vibrational modes of the reaction coordinate (as well as the

solution surrounding it, which is not applicable in the gas phase) to account for changes in

internal energy of the transition state.[18] Therefore, if the electronic transition of the reaction

coordinate corresponds to a high enough vibrational energy level, the anion can then follow a

dissociative pathway (Supplemental Information, Figure S1,C).[20, 21] This effect is displayed

within the aforementioned cases of electron attachment to NO, NO2 and SO2, where the

difference in bond length, bond angles (Supplemental Information, Table S1) and internal energy

between the neutral and anionic molecules are such that a dissociative pathway (Supplemental

Information, Figure S1, C) becomes the saddle point of the transition state and the molecules are

unable to form intact product anions. One method to avoid this transition would be to excite the

vibrational states of the molecule prior to electron attachment,[5] or transfer an electron via

another atom/molecule (Supplemental Information, Figure S1, D).[16, 18] Via scheme 1 (see

also Supplemental Information, Figure S1, D) the use of atomic anions to transfer electrons to

neutral molecules alters the time scale significantly, allowing the correct geometries and internal

energies to be achieved through the formation of an initial encounter complex, allowing an intact

product anions to be formed.[18-20, 22-24] Characteristics of both the incident anion and neutral

target molecule/atom affect electron transfer, one of which is electron affinity.[20]

Atomic metals have some of the lowest electron affinities of the periodic table.[5]

Therefore, they have the potential to transfer electrons to neutral reactants with high efficiency,

enabling the formation of intact product anions due to the dynamics of electron transfer.[18]

Recently, a method developed by Curtis et. al. and Attygale et al. have allowed these atoms to be

easily studied in the gas phase.[25, 26]

5

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

12

Production of Atomic Metal Anions

Pioneering work to produce a significant flux of AMAs was performed by Sallans et. al. in 1983.

[27] They reported that atomic metal anions can be produced in good yields in a Fourier

transform mass spectrometer (FTMS) utilizing collision-induced dissociation (CID) of anionic

metal carbonyl complexes and argon.[27] Their initial experiments utilized Cr(CO)5¯ formed by

a 12.9 eV, 150-ms electron beam pulse on Cr(CO)6 at approximate pressures of 5 x 10-8 torr.

After another 100 ms, an 18.5 eV CID pulse was applied for 0.100 ms, and a 100-ms CID

interaction time was used against Ar at 6 x 10-6 torr.[27] Following this, all other ions were

ejected from the FTMS except Cr¯, allowing subsequent reactions of Cr¯ with neutral substrates

via leak valves.[27] The sequential loss of CO ligands from a metal anion complex produces the

atomic metal anion as per the following general scheme:

M(CO)x + e¯ M(CO)x¯ M¯ + x(CO)

where M is the metal center and ‘x’ is the number of CO ligands. In this manner, Sallans et al.

were able to generate the atomic transition-metal anions V¯, Cr¯, Fe¯, Co¯, Mo¯, and W¯ via

CID of the corresponding metal carbonyl negative ions in a FTMS.[28]

The electron affinity of CO is likely negative, based on G3MP2B3 level calculation [5,

6], meaning that the production of atomic metal anions via CID of metal carbonyl anions is

exothermic. CO2 is another small molecule with negative EA. The utility of CO2 as a leaving

group was noted by Curtis et al. after observing the decomposition of metal-oxalate complexes in

the gas phase.[29] Curtis et al. performed collision induced dissociation of K and Ag oxalate

anion complexes in the gas phase and were able to produce atomic metal anions via the loss of

6

12

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

12

neutral CO2 using an electrospray ionization source in a triple quadrupole mass spectrometer.

[29] The process can be generalized as follows:

M(CO2)x¯ M¯ + x(CO2)

where (CO2)x represents the oxalate dianion and M is the metal center. It is theorized by Curtis et

al. that the high production of AMAs is enabled largely via the negative electron affinity of CO2

(-0.6 eV),[30] but is also affected by the electron affinity of the metal, the size of the metal, and

the strength of the bond between the carbon and metal within the complex.[29, 31]

In addition to metal-oxalate complexes, other dicarboxylic acid salts such as maleate,

fumarate and succinate were observed to produce AMAs such as Na¯, K, Cs¯, Ag¯ via CID at

ambient temperatures using an ESI source by Attygale et. al.[25]

Previous studies have characterized the reactions of various AMAs with metal

carbonyls[28] and acids,[28, 32] simple thiols, sulfides and disulfides,[33] methyl halides

containing a single atom of either F, Cl, Br or I,[34] halogenated and nitro containing alkanes

[26] and alcohols and hydrocarbons.[35] Both electron transfer and halogen abstraction has been

previously observed with large fluorinated compounds, such as pentafluorophenol and

pentafluoroaniline. [36] In the current study we examine the utility of atomic metal anions for

forming polyatomic anions that cannot be generated by electron transfer, notably NO, NO2 and

SO2.

Materials and Methods

Experimental Method

7

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

12

Electrospray ionization (ESI) mass spectrometry experiments were accomplished using a

Micromass Quattro-LC triple-quadrupole mass spectrometer equipped with a Z-spray source,

while running the MassLynx 3.5 operating system. Metal oxalate solutions were prepared by

combining 2 moles of oxalic acid with 1 mole of a metal salt at concentrations on the order of 10 -

1 mol/L in methanol. Solutions were subsequently placed on a Daiger Vortex-Genie 2 shaker and

allowed to shake for approximately 30 minutes before being diluted to a final working

concentration of 10-4 mol/L or lower. Solutions containing iron typically required a 3:1 mole ratio

for best results. This is due to the binding of two oxalic acid molecules to the 3+ oxidation state

of the iron salt used, rather than one molecule of oxalic acid required for 1+ oxidation state

metals, to produce a net negative charge. A flow rate of 50 ul/min was used to pump solutions

through the capillary tube of the ESI source. Capillary, cone and extractor voltages of the ESI

source were set to 2.96 kV, 42 V and 7 V, respectively. The ESI source temperature was set to

80⁰C in order to limit clustering.[37]

The atomic metal anion’s m/z ratio was selected with the first quadrupole (Supplemental

Information, Figure S2, Metal Anion Selector). Entrance and exit voltages of the hexapole

collision cell (Supplemental Information, Figure S2, Collision Cell) were set to 50 eV while the

collision energy remained variable. The purity of the metal anion beam was then tested via

collisions with argon at 8 x 10-4 bar, using lab frame collision energies up to 75 eV. Polyatomic

impurities at the set m/z will dissociate at these collision energies while the pure atomic metal

anions will not. Only if no fragment ions were detected was the selected m/z then allowed to

undergo reactions with neutral gas-phase reactants in the hexapole collision cell, in the absence

of argon. Neutral gaseous molecules were bled into the hexapole collision cell via a gas line

attached to the cell. Lab frame collision energies, up to 50 eV, were tested during each

8

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

12

experiment. The pressure of the neutral reactants were kept constant throughout an experiment.

Reaction products were then detected via the third quadrupole (Figure S2, Product Anion

Selector). The median intensity of the signals present in a mass spectrum is first computed to

estimate the background present. Only m/z signals possessing an intensity greater than twice the

median spectrum intensity are used for further analysis. Relative intensities of individual m/z

signals were calculated by dividing the specific ion’s intensity by the total ion intensity present

in the spectrum (excluding background intensities). All reactants tested were at least UHP grade.

NO, NO2, and SO2 geometries were optimized using the UHF/3-21G* level of theory.

using the Gaussian 09 suite of programs.[38]

Results and Discussion

Electron Transfer to NO, NO2 and SO2

Reactions of AMAs with NO, NO2 and SO2 yielded anionic/neutral products resulting

from electron transfer, adduct formation and atom abstraction. All reactions of Ag¯ and Cu¯

were confirmed for both isotopes.

Figure 1 displays typical product anions of reactions between gaseous AMAs and neutral

NO. Reactions with NO and atomic Cs¯, Fe¯ and Cu¯ yielded electron transfer to NO. Electron

transfer from Cs¯ and Cu¯ to NO required threshold energies of 0.92 eV and 0.60 eV centre-of-

mass collision energy (Ecom) respectively, while Fe¯ required no threshold energy and Ag¯ did

not react. Fe¯ transferring an electron to NO is an endothermic reaction; therefore a threshold

9

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

12

energy should be observable. The lack of such an observation suggests that the energy

distribution of the AMA is so broad that the threshold is not observable under the current

reaction conditions. The maximum relative abundances (based on percentage yield of all ions in

the spectrum) of electron transfer from Fe¯, Cs¯, Cu¯ and Ag¯ to NO were 56.4% at 9.5 eV Ecom

(Fe¯), 18.4% at 9.2 eV Ecom (Cs¯), 3.3% at 15 eV Ecom (Cu¯) and 0% (Ag¯).

Figure 2 shows selected mass spectra of typical product anions resulting from reactions

between gaseous AMAs and neutral SO2. Threshold energies for all electron transfer reactions

were 0 eV Ecom. Maximum relative abundances for electron transfer to SO2 producing SO2¯ from

Fe¯, Cs¯, Cu¯ and Ag¯ were 76% at 16 eV Ecom (Fe¯), 61% at 4.8 eV Ecom (Cs¯), 72% at 14.8 eV

Ecom (Cu¯ ) and 24% at 22 eV Ecom (Ag¯).

Figure 3 exhibits typical product anions resulting from reactions between gaseous AMAs

and neutral NO2. Similar to reactions with SO2, onset energies for all electron transfer reactions

were 0 eV. Maximum relative abundances for electron transfer from Fe¯, Cs¯, Cu¯ and Ag¯ to

produce NO2¯ were 100% at 13.5 eV Ecom (Fe¯), 97% at 0 eV Ecom (Cs¯), 84% at 0 eV Ecom (Cu¯)

and 100% at 14.84 eV Ecom (Ag¯).

Formation of [FeO]¯, [FeO2]¯ & [FeO3]¯

The formation of [FeO]¯, [FeO2]¯ and [FeO3]¯ from reactions with neutral NO2 implies a

multiple collision reaction mechanism of Fe¯ with NO2 or the presence of a triply oxygenated

neutral species, such as N2O4. The equilibrium between NO2 and N2O4 has been well

documented,[39-42] the gas phase equilibrium constant for the reaction being 1.7 x 102 mol/L.

[43] At such low relative pressures of N2O4, a sequential mechanism is most likely. However, it

is known that reactions can take place between NO2 and air and water in the inlet system,[44]

10

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

1718

19

20

21

22

23

24

12

and so reactions with impurities cannot be ruled out. A study of [FeO]¯, [FeO2]¯ and [FeO3]¯

using photoelectron spectroscopy revealed that each sequential addition of oxygen increases the

electron affinity of the molecule.[45] The electron affinities of the ground states of FeO, FeO2,

and FeO3 were found to be 1.50, 2.36 and 3.26 eV.[45] Semi-empirical calculations suggest that

[FeO2]¯ is of the form [O-Fe-O]¯ rather than [Fe(O2)]¯.[46] The authors of the previous photo-

detachment study suggest that the relatively high experimentally derived electron affinity of

FeO2 is inconsistent with the low electron affinities associated with Fe and O2 and therefore of a

side bonded Fe(O2) (I) or a linear/bent Fe-O-O (II) type of structure.[47] They go on to suggest

that the removal of a non-bonding d electron from a linear [FeO2]¯ molecule of the form of O2--

Fe3+-O2- would result in the high electron affinity that was measured experimentally; due to the

stability of the d5 configuration associated with a Fe3+ oxidation state.[47] Furthermore, based on

laser ablation studies performed by Andrews et al., the most stable form of [FeO2]¯ was

determined to be [O-Fe-O]¯ with a bond angle of about 141 degrees.[48] The photoelectron

spectrum of [FeO3]¯ obtained by Wu et al. suggests a highly symmetric trigonal planar structure

of [FeO3]¯, similar to that of D3h FeO3.[45] Interestingly, the electron affinity of FeO4 does not

increase in the same manner as FeO, FeO2 and FeO3.[45] Wu et al. suggest that FeO3 contains the

maximum known oxidation state of the Fe atom (+6).[45] As such, any additional oxygen atoms

will be unable to further oxidize the Fe centre, clearly illustrated by the levelling off of the

electron affinity of FeO4.[45] This lack of a further increase in electron affinity explains why the

[FeO4]¯ is never observed within this experiment and the highest Fe adduct observed is [FeO3]¯.

[45] Additionally, increasing Ecom favours the formation of [FeO]¯ over [FeO2]¯, while both are

favoured over the formation of [FeO3]¯ (Figure 4). This may be the result of an entropically

driven reaction mechanism, or fewer collisions as a result of the higher collision energies. It

11

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

12

should be noted that the formation of [FeO]¯, [FeO2]¯ and [FeO3]¯ are clearly less favourable

than electron transfer to NO2, occupying at maximum ~15% of the observed relative abundance

(Figure 4).

Formation of [AgO]¯, [CuO]¯ & [CuO2]¯

Collisions of Cu¯ with NO2 produced [CuO]¯ and [CuO2]¯ (Figure 3C). The neutral reactant in a

single collision reaction could be NO2 or N2O4 molecule (or other impurity) present in the supply

of NO2. Similar to the reaction of Fe¯, addition of O via sequential collisions is most likely. The

molecular connectivity of the [CuO2]¯ anion could be either [O-Cu-O]¯ or [Cu(O2)]¯. A previous

study using LPES to probe electron affinities found that of the two isomers, Cu(O2) had an

electron affinity of 1.50 eV, while O-Cu-O had an electron affinity of 3.46 eV, suggesting that it

is the thermodynamically favoured product isomer.[49] CuO possesses an electron affinity of

1.77 eV as measured via LPES.[50] NO possesses an electron affinity of 0.024 as measured via

LPES.[5] Therefore, both reaction products, [CuO]¯ and [CuO2]¯, should retain a negative charge

after a reaction with NO2, as noted in Figure 3C. This trend within the electron affinity is

comparable to Fe¯ becoming increasingly stable with the sequential addition of O atoms,

resulting in a thermodynamically driven reaction scheme. Increasing the Ecom of these collisions

favours the formation of [CuO]¯ over [CuO2]¯ (Figure 4), suggesting its formation is

entropically favoured or fewer collisions occur at higher collision energies. Both [CuO]¯ and

[CuO2]¯ formation are less favourable than electron transfer, occupying at maximum ~2% of the

observed relative abundance (Figure 4). Ag¯ and Cu¯ also react with NO to form [AgO]¯ (in

high yield) and [CuO]¯ (trace) (Figure 1C & D).

12

1

2

3

4

56

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

12

Trends Resulting from Electron Transfer to NO, NO2 and SO2

When the relative abundance of product anions [NO]¯, [NO2]¯ and [SO2]¯ produced via

electron transfer from AMAs (Cs¯, Fe¯, Cu¯ and Ag¯) are compared to the electron affinities of

the AMAs, an obvious trend emerges (Table 1). Within the results, increasing electron affinity of

the neutral molecule increases the relative intensity of [NO]¯, [NO2]¯ or [SO2]¯ produced from

collisions with each respective AMA. Relative reaction enthalpies support this conclusion as

well, and have been included in the Supplemental Information, Tables S2 and S3. The

experimental results, relative electron affinities and enthalpies suggest that for a specific neutral,

R, increasing the EA of the metal does not necessarily decrease the electron transfer to form R¯.

Electron transfer appears to be most favorable for the reactions that are closest to thermoneutral,

i.e., involving the metal with the EA closest to that of the molecule. The exception is the high

yield of [NO2]¯ from reactions with Cs¯.

Further information can be gleaned from the trends resulting from analysis of the

temperature dependence of the electron transfer reactions. The reactions of Cs¯/Cu¯ with NO

producing NO¯ and Cs/Cu show a positive dependence on the Ecom, while the reaction of Fe¯

with NO (producing NO¯ and Fe) shows little to no change in NO¯ relative abundance with

increasing Ecom (Figure 5). This is direct experimental evidence of the endothermic nature of

electron transfer between the AMAs and NO. All metal anions show a positive dependence on

the centre-of-mass collision energy when transferring electrons to SO2, even when,

thermodynamically, the reactions are exothermic (Supplemental Information, Table S3) (Figure

6). This indicates the existence of an energy barrier in the electron transfer step (scheme 1).[16]

Interestingly, all metal anions show a negative temperature dependence or no temperature

dependence (for the case of Fe¯) when reacting with NO2 (Figure 7). This provides experimental

13

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

12

evidence of the exothermic nature of these reactions (Supplemental Information, Table S3) while

suggesting no internal energy barriers exist for the reactions.[16]

Conclusions

The reactions of AMAs with small neutral molecules (NO, SO2 and NO2) possessing

electron affinities approximately below (NO)/similar to (SO2)/well above (NO2) the metal anions

demonstrates that AMAs are able to transfer electrons to neutral molecules with high efficiency.

Moreover, reactions of Fe¯/Cu¯ with NO2 were found to produce multiply substituted metal

centred product anions, though the participation of impurities caused by reactions of NO2 with

the inlet system cannot be ruled out. Further investigations into the reactivity of AMAs will no

doubt yield findings of applicability and interest to the field of anion/electro-chemistry.

Acknowledgements

PMM thanks the Natural Sciences and Engineering Research Council of Canada for continuing

financial support. Acknowledgment is made to the Donors of the American Chemical Society

Petroleum Research Fund for partial support of this research.

Supporting Information

Supplementary data associated with this article can be found, in the online version, at

http://dx.doi.org/XXXXX, consisting of Figure S1 (potential electron attachment mechanisms),

14

1

2

3

45

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

12

Table S1 (product ion geometry changes), Table S2 (product/reactant electron affinities) and

Table S3 (reaction enthalpies of electron transfer reactions).

15

1

2

3

12

References

1. Christophorou, L.G., Electron Attachment and Detachment Processes in Electronegative Gases. Beiträge aus der Plasmaphysik, 1987. 27(4): p. 237-281.

2. Massey, H.S.W., Negative Ions, in Advances in Atomic and Molecular Physics, R.B. David and B. Benjamin, Editors. 1979, Academic Press. p. 1-36.

3. H. S. W. Massey, F.R.S., Negative Ions. 2 ed. 1950, London: The Syndics of the Cambridge University Press. 133.

4. Bardsley, J.N. and F. Mandl, Resonant scattering of electrons by molecules. Reports on Progress in Physics, 1968. 31(2): p. 471.

5. Smirnov, B.M., Negative Ions. 1982: McGraw-Hill Inc. .6. Webbook, N.C., NIST Standard Reference Database Number 69. 2005: National Institute of

Standards and Technology: Gaithersburg, MD.7. Nimlos, M.R., Photoelectron spectroscopy of SO2

-, S3-, and S2O-, G.B. Ellison, Editor. 1986: J. Phys. Chem.

8. Tate, J.T. and P.T. Smith, The Efficiencies of Ionization and Ionization Potentials of Various Gases Under Electron Impact. Physical Review, 1932. 39(2): p. 270-277.

9. Sambe, H. and D.E. Ramaker, Dissociative electron attachment in NO. The Journal of Chemical Physics, 1991. 94(4): p. 2548-2556.

10. Chantry, P.J., Dissociative Attachment in CO and NO. Physical Review, 1968. 172(1): p. 125-136.11. Rangwala, S.A., E. Krishnakumar, and S.V.K. Kumar, Dissociative-electron-attachment cross

sections: A comparative study of NO_{2} and O_{3}. Physical Review A, 2003. 68(5): p. 052710.12. Abouaf, R., R. Paineau, and F. Fiquet-Fayard, Dissociative attachment in NO 2 and CO 2. Journal

of Physics B: Atomic and Molecular Physics, 1976. 9(2): p. 303.13. Spyrou, S.M., I. Sauers, and L.G. Christophorou, Dissociative electron attachment to SO[sub 2].

The Journal of Chemical Physics, 1986. 84(1): p. 239-243.14. Cadez, I.M., V.M. Pejcev, and M.V. Kurepa, Electron-sulphur dioxide total ionisation and electron

attachment cross-sections. Journal of Physics D: Applied Physics, 1983. 16(3): p. 305.15. Orient, O.J. and S.K. Srivastava, Production of negative ions by dissociative electron attachment

to SO[sub 2]. The Journal of Chemical Physics, 1983. 78(6): p. 2949-2952.16. Kebarle, P. and S. Chowdhury, Electron affinities and electron-transfer reactions. Chemical

Reviews, 1987. 87(3): p. 513-534.17. Sharma, D.K.S. and P. Kebarle, Chloronium ions as alkylating agents in the gas-phase ion-

molecule reactions with negative temperature dependence. Journal of the American Chemical Society, 1982. 104(1): p. 19-24.

18. Marcus, R.A., On the Theory of Oxidation-Reduction Reactions Involving Electron Transfer. I. The Journal of Chemical Physics, 1956. 24(5): p. 966-978.

19. Marcus, R.A., Electron Transfer Reactions in Chemistry: Theory and Experiment (Nobel Lecture). Angewandte Chemie International Edition in English, 1993. 32(8): p. 1111-1121.

20. Condon, E.U., Nuclear Motions Associated with Electron Transitions in Diatomic Molecules. Physical Review, 1928. 32(6): p. 858-872.

21. Franck, J. and E.G. Dymond, Elementary processes of photochemical reactions. Transactions of the Faraday Society, 1926. 21(February): p. 536-542.

22. Hughes, B.M., C. Lifshitz, and T.O. Tiernan, Electron affinities from endothermic negative-ion charge-transfer reactions. III. NO, NO[sub 2], SO[sub 2], CS[sub 2], Cl[sub 2], Br[sub 2], I[sub 2], and C[sub 2]H. The Journal of Chemical Physics, 1973. 59(6): p. 3162-3181.

16

1

2

3456789

10111213141516171819202122232425262728293031323334353637383940414243444546

12

23. Berkowitz, J., W.A. Chupka, and D. Gutman, Electron Affinities of O[sub 2], O[sub 3], NO, NO[sub 2], NO[sub 3] by Endothermic Charge Transfer. The Journal of Chemical Physics, 1971. 55(6): p. 2733-2745.

24. Armentrout, P.B., The power of accurate energetics (or thermochemistry: what is it good for?). J Am Soc Mass Spectrom, 2013. 24(2): p. 173-85.

25. Attygalle, A.B., F.U. Axe, and C.S. Weisbecker, Mild route to generate gaseous metal anions. Rapid Communications in Mass Spectrometry, 2011. 25(6): p. 681-688.

26. Curtis, S., et al., Reactions of Atomic Metal Anions in the Gas phase: Competition between Electron Transfer, Proton Abstraction and Bond Activation. The Journal of Physical Chemistry A, 2011. 115(48): p. 14006-14012.

27. Sallans, L., et al., Preparation and reactions of chromium(1-) ion. The chromium-hydrogen bond strength. Journal of the American Chemical Society, 1983. 105(20): p. 6352-6354.

28. Sallans, L., et al., Generation and reaction of atomic metal anions in the gas phase. Determination of the heterolytic and homolytic bond energies for the vanadium, chromium, iron, cobalt, and molybdenum hydrides VH, CrH, FeH, CoH, and MoH. Journal of the American Chemical Society, 1985. 107(15): p. 4379-4385.

29. Curtis, S., et al., Old acid, new chemistry. Negative metal anions generated from alkali metal oxalates and others. Journal of the American Society for Mass Spectrometry, 2010. 21(11): p. 1944-1946.

30. Knapp, M., et al., Formation of long-lived CO−2, N2O−, and their dimer anions, by electron attachment to van der waals clusters. Chemical Physics Letters, 1986. 126(3-4).

31. Kahwa, I.A. and A.M. Mulokozi, The thermal decomposition temperatures of ionic metal oxalates. Journal of thermal analysis, 1981. 22(1): p. 61-65.

32. Stevens Miller, A.E., et al., Reactions of Fe− with acids: gas-phase acidity and bond energy of FeH. International Journal of Mass Spectrometry and Ion Processes, 1993. 123(3): p. 205-216.

33. Sallans, L., K.R. Lane, and B.S. Freiser, Gas-phase reactions of iron(1-) and cobalt(1-) with simple thiols, sulfides, and disulfides by Fourier-transform mass spectrometry. Journal of the American Chemical Society, 1989. 111(3): p. 865-873.

34. Stevens Miller, A.E., et al., Methyl bonding to Fe and Fe(CO)n:: reactions of Fe− and Fe(CO)n− with methyl halides. International Journal of Mass Spectrometry, 2000. 195–196(0): p. 341-349.

35. Halvachizadeh, J., A. Mungham, and P.M. Mayer, The Dehydrogenation of Alcohols and Hydrocarbons by Atomic Metal Anions. Eur. J. Mass spectrom., 2015. 21: p. 487-495.

36. Butson, J.M. and P.M. Mayer, Electron attachment to pentafluorophenol and pentafluoroaniline via reaction with atomic metal anions. Chemical Physics Letters, 2014. 614(0): p. 186-191.

37. Cech, N.B. and C.G. Enke, Practical implications of some recent studies in electrospray ionization fundamentals. Mass Spectrometry Reviews, 2001. 20(6): p. 362-387.

38. Frisch, M.J., et al., Gaussian 09, Revision B.01. 2009: Wallingford CT.39. Geers-Müller, R. and F. Stuhl, On the kinetics of the reactions of oxygen atoms with NO2, N2O4,

and N2O3 at low temperatures. Chemical Physics Letters, 1987. 135(3): p. 263-268.40. Vander Auwera, J. and M. Herman, Spectroscopic investigation of the response of the NO2-N2O4

chemical system to optical laser pumping. Journal of Photochemistry, 1987. 38(0): p. 15-33.41. Kołodziejczyk, W., D. Wójcik-Pastuszka, and R. Berkowski, Experimental and theoretical study of

the reaction system of 2NO2&#xa0;↔&#xa0;N2O4 and some fluorinated derivatives of tert-butanol in the gas phase. Journal of Molecular Structure, 2014. 1071(0): p. 71-78.

42. Wójcik-Pastuszka, D., W. Kołodziejczyk, and R. Berkowski, Experimental and theoretical study of the reaction system of 2NO2 ↔ N2O4 and some fluorinated derivatives of butanols in the gas phase. Journal of Molecular Structure, 2013. 1052(0): p. 1-7.

17

123456789

1011121314151617181920212223242526272829303132333435363738394041424344454647

12

43. Chao, J., R.C. Wilhoit, and B.J. Zwolinski, Gas phase chemical equilibrium in dinitrogen trioxide and dinitrogen tetroxide. Thermochimica Acta, 1974. 10(4): p. 359-371.

44. Melko, J.J., et al., Exploring the Reactions of Fe+ and FeO+ with NO and NO2. The Journal of Physical Chemistry A, 2012. 116(47): p. 11500-11508.

45. Wu, H., S.R. Desai, and L.-S. Wang, Observation and Photoelectron Spectroscopic Study of Novel Mono- and Diiron Oxide Molecules:  FeOy- (y = 1−4) and Fe2Oy- (y = 1−5). Journal of the American Chemical Society, 1996. 118(22): p. 5296-5301.

46. Blyholder, G., J. Head, and F. Ruette, Semiempirical calculation of iron-oxygen interactions. Inorganic Chemistry, 1982. 21(4): p. 1539-1545.

47. Fan, J. and L.-S. Wang, Photoelectron spectroscopy of FeO[sup - ] and FeO[sup - ][sub 2]: Observation of low-spin excited states of FeO and determination of the electron affinity of FeO[sub 2]. The Journal of Chemical Physics, 1995. 102(22): p. 8714-8717.

48. Andrews, L., et al., Reactions of Laser-Ablated Iron Atoms with Oxygen Molecules:  Matrix Infrared Spectra and Density Functional Calculations of OFeO, FeOO, and Fe(O2). Journal of the American Chemical Society, 1996. 118(2): p. 467-470.

49. Wu, H., S.R. Desai, and L.-S. Wang, Two isomers of CuO[sub 2]: The Cu(O[sub 2]) complex and the copper dioxide. The Journal of Chemical Physics, 1995. 103(10): p. 4363-4366.

50. Polak, M.L., et al., Photoelectron spectroscopy of copper oxide (CuO-). The Journal of Physical Chemistry, 1991. 95(9): p. 3460-3463.

18

123456789

10111213141516171819

20

12

Table 1. Product distribution at the collision energy corresponding to the maximum relative abundances of NO¯, SO2¯ and NO2¯. The electron affinity (EA) of each reactant and metal are also listed.

Reactant(R)

M¯ EA (M)(eV)

Relative Abundance (%)

R- M- MOn-

NO

EA 0.024 eV

Fe- 0.15 38 62

Cs- 0.47 20 80

Cu- 1.22 4 96

Ag- 1.3 0 69 31

SO2

EA 1.1070 eV

Fe- 0.15 65 35

Cs- 0.47 69 31

Cu- 1.22 73 27

Ag- 1.3 4 96

NO2

EA 2.2730 eV

Fe- 0.15 82 13 5

Cs- 0.47 97 3

Cu- 1.22 84 14 2

Ag- 1.3 96 4

19

123

4

5

12

Figure Captions

Figure 1. Characteristic mass spectra of (A) Cs¯ (B) Fe¯, (C) 65Cu¯ and (D)107Ag¯ reactions with neutral NO gas. Note that the appearance of NO2¯ in the reaction with Cu¯ was determined to be an artifact. The Y axis represents total ion abundance (absolute intensity) while the X axis represents mass to charge ratio (m/z). Pressures of NO gas range from; (A) 2.2 x 10-4, (B) 1.5 x 10-4, (C) 1.6 x 10-4 and (D) 3.5 x 10-4 torr. Centre of mass collision energy was set to (A) 0.92 eV, (B) 0 eV, (C) 0.95 eV and (D) 0 eV.

Figure 2. Characteristic mass spectra of (A) Cs¯ (B) Fe¯, (C) 65Cu¯ and (D)107Ag¯ reactions with neutral SO2 gas. The Y axis represents total ion abundance (absolute intensity) while the X axis represents mass to charge ratio (m/z). Pressures of SO2 gas range from; (A) 1.2 x 10-4, (B) 1.4 x 10-4, (C) 1.3 x 10-4 and (D) 1.8 x 10-4 torr. Centre of mass collision energy was set to (A) 1.92 eV, (B) 0 eV, (C) 0 eV and (D) 3.75 eV.

Figure 3. Characteristic mass spectra of (A) Cs¯ (B) Fe¯, (C) 65Cu¯ and (D)109Ag¯ reactions with neutral NO2 gas. The Y axis represents total ion abundance (absolute intensity) while the X axis represents mass to charge ratio (m/z).Pressures of NO2 gas range from; (A) 2.8 x 10-4, (B) 2.0 x 10-4, (C) 1.1 x 10-4 and (D) 1.4 x 10-4 torr. All centre of mass collision energies were set to 0 eV.

Figure 4. Relative intensities of (A) Fe¯ and (B) 65Cu¯ adducts from reactions with NO2¯. The Y axis represents relative intensity (%) while the X axis represents centre of mass collision energy (eV). Pressures of NO2 gas range from (A) 1.4 x 10-4 and (B) 1.1 x 10-4 torr.

Figure 5. Relative intensities of (A) 65Cu¯, (B) Cs¯ and (C) Fe¯ with NO. The Y axis represents relative intensity (%) while the X axis represents centre of mass collision energy (eV). Pressures of NO gas range from; (A) 1.6 x 10-4, (B) 2.2 x 10-4 and (C) 1.5 x 10-4 torr.

Figure 6. Relative intensities of (A) 107Ag¯, (B) 65Cu¯, (C) Cs- and (D) Fe¯ with SO2. The Y axis represents relative intensity (%) while the X axis represents centre of mass collision energy (eV). Pressures of SO2 gas range from; (A) 1.8 x 10-4, (B) 1.3 x 10-4, (C) 1.4 x 10-4 and (D) 1.2 x 10-4

torr.

Figure 7. Relative intensities of (A) 109Ag¯, (B) 65Cu¯, (C) Cs¯ and (D) Fe¯ with NO2¯. The Y axis represents relative intensity (%) while the X axis represents centre of mass collision energy (eV). Pressures of NO2 gas range from (A) 3.5 x 10-4, (B) 1.1 x 10-4, (C) 2.0 x 10-4 and (D) 1.4 x 10-4 torr.

20

1

234567

89

101112

13141516

171819

202122

23242526

27282930

31

32

12

Figure 1. Characteristic mass spectra of (A) Cs¯ (B) Fe¯, (C) 65Cu¯ and (D)107Ag¯ reactions with neutral NO gas. Note that the appearance of NO2¯ (noted in the reaction with Cu¯ and Cs¯) was determined to be an artifact. The Y axis represents total ion abundance (absolute intensity) while the X axis represents mass to charge ratio (m/z). Pressures of NO gas range from; (A) 2.2 x 10 -4, (B) 1.5 x 10-4, (C) 1.6 x 10-4 and (D) 3.5 x 10-4 torr. Centre of mass collision energy was set to (A) 0.92 eV, (B) 0 eV, (C) 0.95 eV and (D) 0 eV.

21

10000

8000

6000

4000

2000

010020

10

8

6

4

2

0

x103

8040

4000

3000

2000

1000

080604020

3000

2500

2000

1500

1000

500

010020

A B

C D

65Cu

-

Fe-

NO2-

Cs-

107Ag

-

NO-

NO-

NO-

CuO-

AgO-

Abs

olut

eIn

tens

ity

m/z1

234567

8

9

12

Figure 2. Characteristic mass spectra of (A) Cs¯ (B) Fe¯, (C) 65Cu¯ and (D)107Ag¯ reactions with neutral SO2 gas. The Y axis represents total ion abundance (absolute intensity) while the X axis represents mass to charge ratio (m/z). Pressures of SO2 gas range from; (A) 1.2 x 10-4, (B) 1.4 x 10-4, (C) 1.3 x 10-4 and (D) 1.8 x 10-4 torr. Centre of mass collision energy was set to (A) 1.92 eV, (B) 0 eV, (C) 0 eV and (D) 3.75 eV.

22

50

40

30

20

10

0

x103

10040

10

8

6

4

2

0x1

03

807060504030

12

10

8

6

4

2

0

x103

7570656055

8000

6000

4000

2000

012080

A B

C D

B

65Cu

-

SO2-

Fe-

SO2-

SO2-

SO2-

Cs-

107Ag

-

m/z

Abs

olut

eIn

tens

ity1

2

3

45678

9

12

Figure 3. Characteristic mass spectra of (A) Cs¯ (B) Fe¯, (C) 65Cu¯ and (D)109Ag¯ reactions with neutral NO2 gas. The Y axis represents total ion abundance (absolute intensity) while the X axis represents mass to charge ratio (m/z).Pressures of NO2 gas range from; (A) 2.8 x 10-4, (B) 2.0 x 10-4, (C) 1.1 x 10-4 and (D) 1.4 x 10-4 torr. All centre of mass collision energies were set to 0 eV.

23

1000

01009080

1600

1200

800

400

01008060

40

30

20

10

0

x103

140120100806040

4000

0504540

80

60

40

20

0

x103

6560555045

25

20

15

10

5

0

x103

120100806040

2000

1000

0135130

1200

800

400

0110100

A B

C D

65Cu

-

Fe-

Cs-

109Ag

-

NO2-

FeO2-

FeO3-

FeO-

NO2-

CuO-

CuO2-

NO2-

NO2-

Abs

olut

eIn

tens

ity

m/z1

2345

6

7

8

9

12

Figure 4. Relative intensities of (A) Fe¯ and (B) 65Cu¯ adducts from reactions with NO2¯. The Y axis represents relative intensity (%) while the X axis represents centre of mass collision energy (eV). Pressures of NO2 gas range from (A) 1.4 x 10-4 and (B) 1.1 x 10-4 torr.

24

4

2

0

6543210

1

0

6543210

FeO-

FeO2-

FeO3-

CuO-

CuO2-

Rel

ativ

e In

tens

ity (%

)

Centre of Mass Collision Energy (eV)

1

345

12

10

8

6

4

2

0

151050

1009080

151050

1009080

840

40

30

20

10

0

840

80

60

40

20

0

151050

NO-

Cu-

Cs-

Fe-

A

B

CRel

ativ

e In

tens

ity (%

)

Centre of Mass Collision Energy (eV)

Figure 5. Relative intensities of (A) 65Cu¯, (B) Cs¯ and (C) Fe¯ with NO. The Y axis represents relative intensity (%) while the X axis represents centre of mass collision energy (eV). Pressures of NO gas range from; (A) 1.6 x 10-4, (B) 2.2 x 10-4 and (C) 1.5 x 10-4 torr.

25

1

234

5

6

12

80604020

0

20100

80604020

20100

80

60

40

20

20100

100806040200

20100

A

B

C

D

SO2-

Ag-

Cu-

Cs-

Fe-

Rel

ativ

e In

tens

ity (%

)

Centre of Mass Collision Energy (eV)

Figure 6. Relative intensities of (A) 107Ag¯, (B) 65Cu¯, (C) Cs- and (D) Fe¯ with SO2. The Y axis represents relative intensity (%) while the X axis represents centre of mass collision energy (eV). Pressures of SO2 gas range from; (A) 1.8 x 10-4, (B) 1.3 x 10-4, (C) 1.4 x 10-4 and (D) 1.2 x 10-4

torr.

26

1

2345

12

10080604020

0

1086420

10080604020

0

121086420

10080604020

0

1086420

10080604020

0

14121086420

A

B

C

D

NO2-

Ag-

Cu-

Cs-

Fe-

Rel

ativ

e In

tens

ity (%

)

Centre of Mass Collision Energy (eV)

Figure 7. Relative intensities of (A) 109Ag¯, (B) 65Cu¯, (C) Cs¯ and (D) Fe¯ with NO2¯. The Y axis represents relative intensity (%) while the X axis represents centre of mass collision energy (eV). Pressures of NO2 gas range from (A) 3.5 x 10-4, (B) 1.1 x 10-4, (C) 2.0 x 10-4 and (D) 1.4 x 10-4 torr.

27

1

2

3456

12