Gas Phase Reactions of Ions Derived from Anionic Uranyl Formate and Uranyl Acetate Complexes

Abstract

The speciation and reactivity of uranium are topics of sustained interest because of their importance to the development of nuclear fuel processing methods, and a more complete understanding of the factors that govern the mobility and fate of the element in the environment. Tandem mass spectrometry can be used to examine the intrinsic reactivity (i.e., free from influence of solvent and other condensed phase effects) of a wide range of metal ion complexes in a species-specific fashion. Here, electrospray ionization, collision-induced dissociation, and gas-phase ion-molecule reactions were used to create and characterize ions derived from precursors composed of uranyl cation (U^VIO₂ ²⁺) coordinated by formate or acetate ligands. Anionic complexes containing U^VIO₂ ²⁺ and formate ligands fragment by decarboxylation and elimination of CH₂=O, ultimately to produce an oxo-hydride species [U^VIO₂(O)(H)]^-. Cationic species ultimately dissociate to make [U^VIO₂(OH)]⁺. Anionic complexes containing acetate ligands exhibit an initial loss of acetyloxyl radical, CH₃CO₂•, with associated reduction of uranyl to U^VO₂ ⁺. Subsequent CID steps cause elimination of CO₂ and CH₄, ultimately to produce [U^VO₂(O)]^–. Loss of CH₄ occurs by an intra-complex H⁺ transfer process that leaves U^VO₂ ⁺ coordinated by acetate and acetate enolate ligands. A subsequent dissociation step causes elimination of CH₂=C=O to leave [U^VO₂(O)]^–. Elimination of CH₄ is also observed as a result of hydrolysis caused by ion-molecule reaction with H₂O. The reactions of other anionic species with gas-phase H₂O create hydroxyl products, presumably through the elimination of H₂.

Introduction

The speciation and reactivity of uranium are topics of sustained interest [1] because of their importance to the development of nuclear fuel processing methods [2] and a more complete understanding of the factors the govern the mobility and fate of the element in the environment [3, 4]. Electrospray ionization (ESI) provides easy access to a wide range of gas-phase complexes containing uranium in high oxidation states (+5 and +6) for studies of intrinsic structure and reactivity (i.e., outside of the influence of solvent or other condensed phase effects) in a species specific fashion. For example, the transfer of uranium as mono-positive, pentavalent U^VO₂ ⁺ from solution to the gas phase using ESI was first reported in 1992 [5] and since then, advances in the fundamental understanding of uranyl coordination chemistry have been made using the ionization method [6–21]. Most importantly, ESI has been used to generate gas-phase, doubly charged complexes containing UO₂ ²⁺, such as [UO₂(L)_n]²⁺, with L = acetonitrile (acn) or acetone (aco) and n = 4 or 5, for collision-induced dissociation (CID) and ion-molecule reaction experiments using tandem mass spectrometry [15].

ESI and tandem mass spectrometry have also been used to explore the intrinsic dissociation and ion-molecule reactions of transuranics [22–25]. For example, ESI was used to generate acn and aco complexes of PuO₂ ²⁺ and UO₂ ²⁺ solutions using similar experimental conditions [24]. In accord with relative stabilities of the actinyl ions, UO₂ ²⁺ > PuO₂ ²⁺ > NpO₂ ²⁺, the yields of plutonyl complexes with acn or aco were significantly lower than those of uranyl, and dipositive neptunyl complexes were not observed. CID of the di-positive coordination complexes in a 3-D quadrupole ion trap produced doubly and singly charged fragment ions; the fragmentation products revealed differences in underlying chemistries of PuO₂ ²⁺ and UO₂ ²⁺, including the lower stability of Pu(VI) compared with U(VI).

Recently, gas-phase organoactinyl complexes were synthesized in an ion trap by collisional activation of anionic An^VIO₂ complexes where An = U, Np, Pu [26]. In these precursors, the formally An^VIO₂ ²⁺ actinyl core was coordinated by three carboxylate ligands, with R = CH₃ (methyl), C≡C-CH₃ (butynl), C₆H₅ (phenyl), or C₆F₅ (pentafluorophenyl). One CID pathway observed for these precursor ions was decarboxylation to generate [(R)AnO₂(O₂C-R)₂]^– products that presumably feature discrete An-C bonds. Isolation of the [(R)AnO₂(O₂C-R)₂]^– products and subsequent reaction with gas-phase H₂O causes hydrolysis to generate species with composition [AnO₂OH(O₂C-R)₂]^–. The hydrolysis reactions provided support for the hypothesis that true organoactinyl species were generated.

Here we report the fragmentation of anionic uranyl complexes with general formula [UO₂(O₂C-R)₃]^–, R = H (formate) and CH₃ (acetate), in a linear ion trap (LIT) mass spectrometer. We have recently reported that the use of the 2-D ion trap provides access to fragmentation pathways and reactions not observed in earlier studies with 3-D ion traps. For example, our past studies of the dissociation behavior of gas-phase actinyl complexes using a 3-D ion trap were complicated by a preponderance of product ions obviously generated by interactions with background H₂O. Collisions with H₂O in the ion trap create hydrated product ions or led to charge reduction reactions that form [UO₂OH]⁺ and [UO₂]⁺ and complexes containing these cations [15, 16, 24]. In a prior CID study of acetone-coordinated UO₂ ²⁺, ligand addition reaction rates were so fast that complexes with 2 or 3 coordinating aco ligands hydrated to generate heterogeneous, tetra-, or pentacoordinate complexes [15]. As a consequence, gas-phase complex ions containing UO₂ ²⁺ and two or fewer ligands could not be generated by CID in a “top-down” approach.

However, in a recent investigation of the fragmentation behavior of UO₂ ²⁺ complexes containing acn [27], we showed that the amount of adventitious H₂O in the LIT is significantly lower than in the 3-D ion trap employed in our previous studies. CID was used to drive complete acn ligand elimination to generate UO₂ ²⁺, and this was the first reported generation of the bare di-cation by multiple-stage CID. More importantly, CID of [UO₂(acn)]²⁺ generated [UO₂NC]⁺, which subsequently fragmented to produce NUO⁺. While the bare NUO⁺ ion, which is isoelectronic with UO₂ ²⁺, had previously been prepared in a “bottom-up” approach by insertion of U⁺ into NO in an ion-molecule reaction [28], our experiments demonstrated that the species can also be created in a CID reaction. Formation of the nitrido species by transfer of N from cyanide was confirmed using precursors labeled with ¹⁵N, and a mechanism involving a cyanate intermediate was proposed based on density functional theory (DFT) calculations. Similar experiments have now been conducted with aco-coordinated UO₂ ²⁺, and novel products such as [UO₂-HCO]⁺ and [UO₂-CH₂CH₃]⁺ have been identified using a combination of isotope labeling and high-resolution/high accuracy mass measurements [29].

Our goal in this study was to map out more thoroughly the dissociation behavior of formate- and acetate-coordinated UO₂ ²⁺, and the reactions of dissociation products with background H₂O in the LIT. To the best of our knowledge, the dissociation behavior of [UO₂(O₂C-H)₃]^– has not yet been reported. CID of [UO₂(O₂C-CH₃)₃]^– in a 3-D quadrupole ion trap has been reported [26, 30], and the structure of the species has been determined using infrared multiple-photon photodissociation (IRMPD) spectroscopy [31]. However, the experiments reported here demonstrate that the chemistry of product ions from both systems can be explored more accurately in the 2-D LIT. As discussed below, our results clear up some ambiguity in the intrinsic dissociation pathways for the uranyl–acetate complex [30], and reveal some apparent product ions that are better described as products of reactions with background H₂O.

Experimental

Samples of uranyl formate, d₁-formate, acetate, and d₃-acetate were prepared by combining ca. 2–3 mg of UO₃ (Strem Chemicals, Newburyport, MA, USA), corresponding to approximately 7 × 10^–6 to 1 × 10^-5 moles, with a 2-fold mole excess of the respective acids (purchased from Sigma Aldrich, St. Louis, MO, USA) and 400 μL of deionized/distilled H₂O contained in a glass scintillation vial. The solutions were allowed to incubate on a hot plate at 70 °C for 12 h. We chose to synthesize the compounds because uranyl formate (native or d-labeled) and uranyl d₃-acetate are not commercially available. Caution: uranium oxide is radioactive (α- and γ-emitter), and proper shielding, waste disposal, and personal protective gear should be used when handling the material.

When cooled, 20 μL of the resulting solution was diluted with 800 μL of 50:50 (by volume) H₂O/EtOH and used without further work-up as the spray solution for ESI-MS. Ethanol was used for these experiments to avoid any ambiguity when looking for potential molecular O₂ adducts to product ions [32–35]. With the limited mass measurement accuracy of the LIT, O₂ and CH₃OH adducts cannot be distinguished.

ESI and CID experiments were performed on a ThermoScientific (San Jose, CA, USA) LTQ-XL LIT mass spectrometer. The uranyl-carboxylate spray solutions were infused into the ESI-MS instrument using the incorporated syringe pump at a flow rate of 5 μL/min. In the negative ion mode, the atmospheric pressure ionization stack settings for the LTQ (lens voltages, quadrupole, and octopole voltage offsets, etc.) were optimized for maximum transmission of the singly-charged ions [UO₂(O₂C-R)₃]^–, R = H, D, CH₃, or CD₃, to the ion trap mass analyzer by using the auto-tune routine within the LTQ Tune program. In the positive ion mode, the transmission was optimized using [UO₂(O₂C-R)(CH₃CH₂OH)₂]⁺. Helium was used as the bath/buffer gas to improve trapping efficiency and as the collision gas for CID experiments.

For CID, precursor ions were isolated using an isolation width of 1.0 to 1.5 mass to charge (m/z) units. The exact value was determined empirically to provide maximum ion intensity while ensuring isolation of a single isotopic peak. To probe CID behavior in general, the (mass) normalized collision energy (NCE, as defined by ThermoScientific) was set between 5% and 18%, which corresponds to 0.075–0.27 V applied for CID with the current instrument calibration. The activation Q, which defines the frequency of the applied radio frequency potential, was set at 0.30 and the activation time employed was 30 ms.

To probe gas-phase reactions of selected precursor ions with background H₂O, ions were isolated using widths of 2–4 m/z units. The specific width used was chosen empirically to ensure maximum ion isolation efficiency. The ions were then stored in the LIT for periods ranging from 1 ms to 10 s. Despite the lower H₂O levels in the 2-D ion trap, there is still a sufficient partial pressure of the neutral to permit an investigation of ion-molecule reactions, particularly when using long isolation times. For both CID and ion-molecule reaction experiments, the mass spectra displayed represent the accumulation and averaging of at least 30 isolation, dissociation, and ejection/detection steps.

Results and Discussion

Tandem MS of [UO₂(O₂C-H)₃]^–

The ESI mass spectrum generated from uranyl formate in H₂O/CH₃CH₂OH is shown in Figure 1a. The dominant negative ion created by ESI was [U^VIO₂(O₂C-H)₃]^- at m/z 405, and larger clusters containing UO₂ ²⁺ and formate were also observed. CID of the larger clusters generated [U^VIO₂(O₂C-H)₃]^– (data not shown) as the dominant product and was not examined in detail.

Figure 1

Negative ESI and CID spectra derived from uranyl formate in 50/50 H₂O/CH₃CH₂OH: (a) Full ESI spectrum, (b) CID of [U^VIO₂(O₂C-H)₃]^- at m/z 405, (c) CID (MS³ stage) of [U^VIO₂(H)(O₂C-H)₂]^- at m/z 361, and (d) CID (MS⁴ stage) of [U^VIO₂(O)(O₂C-H)]^– at m/z 331. In the spectra, the circles and arrows illustrate the MSⁿ pathway. In each spectrum, the bold peak label indicates the precursor selected for CID whereas labels in italics represent the products from dissociation or ion-molecule reactions as indicated in the text

The multiple-stage (MSⁿ) CID spectra generated by initially isolating [U^VIO₂(O₂C-H)₃]^– at m/z 405 are shown in Figure 1b–d. Data generated using deuterium-labeled formic acid (HO₂C-D) are provided in Supplementary Figure S1a–c of the Supporting Information. The two product ions generated by CID of [U^VIO₂(O₂C-H)₃]^– (MS/MS stage, Figure 1b) appeared at m/z 361 and 331. Owing to the low-mass cut-off of the of the LIT, we were not able to determine the extent to which [U^VIO₂(O₂C-H)₃]^– dissociates to create formate at m/z 45.

As discussed below, the peak at m/z 331 is a product generated by dissociation of the ion at m/z 361. The ion at m/z 361 peak is likely formed by elimination of CO₂ and hydride transfer to U^VIO₂ ²⁺ to generate a species with formula [U^VIO₂(H)(O₂C-H)₂]^- via Reaction 1.

$$ {\left[{\mathrm{U}}^{\mathrm{VI}}{\mathrm{O}}_2{\left({\mathrm{O}}_2\mathrm{C}-\mathrm{H}\right)}_3\right]}^{\hbox{--}}\to {\left[{\mathrm{U}}^{\mathrm{VI}}{\mathrm{O}}_2\left(\mathrm{H}\right){\left({\mathrm{O}}_2\mathrm{C}-\mathrm{H}\right)}_2\right]}^{\hbox{--} }+{\mathrm{CO}}_2 $$

(1)

The loss of 44 mass units was also observed for CID of [U^VIO₂(O₂C-D)₃]^– (MS/MS stage, Supplementary Figure S1a), with transfer of D^– to the U^VIO₂ ²⁺ core. Unlike previous reports [26, 30] of the dissociation behavior of uranyl-acetate anion ([U^VIO₂(O₂C-CH₃)₃]^–), elimination of formyloxyl radical, •O₂C-H, with associated reduction of uranyl ion to the U(V) dioxo cation, U^VO₂ ⁺, was not observed. The electron affinities (EA) of both formyloxyl radical [36, 37] and acetyloxyl radical [38] have been measured using negative ion photoelectron/photodetachment spectroscopy. The apparent lack of reduction/radical elimination for the uranyl-formate anion precursor is consistent with the higher EA of formyloxyl radical relative to acetyloxyl radical (3.498 ± 0.0015 eV and 3.250 ± 0.010 eV, respectively).

For CID of the [U^VIO₂(H)(O₂C-H)₂]^– ion at m/z 361 (MS³ stage, Figure 1c), the only product ion detected appeared at m/z 331. The neutral loss at this stage is 30 mass units, likely CH₂=O, and the product ion composition is assigned as [U^VIO₂(O)(O₂C-H)]^–. The dissociation reaction would, therefore, involve heterolytic cleavage of a formate C–O bond, retention of O^2- and hydride-transfer to create the neutral O=CH₂ neutral species as shown in Reaction 2. The neutral loss shifted from 30 to 32 units following CID of [U^VIO₂(D)(O₂C-D)₂]^– (MS³ stage, Supplementary Figure S1b), consistent with the pathway depicted in Reaction 2.

$$ {\left[{\mathrm{U}}^{\mathrm{VI}}{\mathrm{O}}_2\left(\mathrm{H}\right){\left({\mathrm{O}}_2\mathrm{C}-\mathrm{H}\right)}_2\right]}^{\hbox{--}}\to {\left[{\mathrm{U}}^{\mathrm{VI}}{\mathrm{O}}_2\left(\mathrm{O}\right)\left({\mathrm{O}}_2\mathrm{C}-\mathrm{H}\right)\right]}^{\hbox{--} }+\mathrm{O}={\mathrm{CH}}_2 $$

(2)

Subsequent CID of the ion at m/z 331 (MS⁴ stage, Figure 1d) caused elimination of 44 mass units (CO₂) to create a product ion at m/z 287 [U^VIO₂(O)(H)]^–. To maintain an overall –1 charge on the product ion, we propose that the anion at m/z 287 is [U^VIO₂(O)(H)]^–, that is, U^VIO₂ ²⁺ coordinated by oxide and hydride ligands. The loss of 44 units (CO₂), and a shift of the product ion m/z value to 288, was observed following CID of the deuterium labeled precursor [U^VIO₂(O)(O₂C-D)]^– (MS³ stage, Supplementary Figure S2c).

To probe the general reactions with H₂O in the LIT, we isolated the respective product ions generated by multiple stage CID of [U^VIO₂(O₂C-H)₃]^– for periods ranging from 1 ms to 10 s. The product ion spectra generated by isolation and storage of [U^VIO₂(O)(H)]^– at m/z 287 (the final product ion observed in Figure 1d) are shown in Figure 2. No adduct species were observed when [U^VIO₂(O)(H)]^– was allowed to react with H₂O for 1 or 10 ms (Figure 2a and b). At 100 ms isolation time (Figure 2c), a product ion at m/z 303 appeared, and the same ion was the base peak at 1 s isolation time (Figure 2d). At 10 s isolation time, a product ion at m/z 321 dominated the spectrum (Figure 2e).

Figure 2

Product ion spectra (MS⁵ stage) generated by isolation and storage of [U^VIO₂(O)(H)]^–, m/z 287, without imposed collisional activation, for reaction with background H₂O. Spectra were generated using isolation/reaction times of: (a) 1 ms, (b) 10 ms, (c) 100 ms, (d) 1 s, and (e) 10 s. In the spectrum, the circles and vertical arrows illustrate MSⁿ dissociation pathway. Horizontal arrows indicate ion-molecule reaction steps. The bold peak labels indicate the precursor selected for CID whereas labels in italics represent the products from dissociation or ion-molecule reactions as indicated in the text

The m/z 303 product in Figure 2c and d is 16 mass units greater than the m/z 287 precursor, and we attribute this peak to the formation of [U^VIO₂(O)(OH]^– by a process that involves H₂O addition and H₂ elimination, as shown in Reaction 3.

$$ {\left[{\mathrm{U}}^{\mathrm{VI}}{\mathrm{O}}_2\left(\mathrm{O}\right)\left(\mathrm{H}\right)\right]}^{\hbox{--} }+{\mathrm{H}}_2\mathrm{O}\to {\left[{\mathrm{U}}^{\mathrm{VI}}{\mathrm{O}}_2\left(\mathrm{O}\right)\left(\mathrm{O}\mathrm{H}\right)\right]}^{\hbox{--} }+{\mathrm{H}}_2 $$

(3)

The product ion at m/z 321, which is 18 units higher in m/z than the m/z 303 peak, may either be an H₂O adduct (formally [U^VIO₂(O)(OH)(H₂O)]^–) that appears at long isolation times, or a tris-hydroxy species [U^VIO₂(OH)₃]^–. Determination of the true composition and structure for this species would benefit from an IRMPD investigation. Subsequent CID of the ion at m/z 321 caused elimination of H₂O to re-form the species at m/z 303 (spectrum not shown), a pathway that could be possible for either species.

Although the reaction of [U^VIO₂(O)(H)]^– with H₂O has not yet been reported, the products generated by exposure of the similar species UO₃ ^– to CH₃OH in the gas phase has been investigated [39]. The results of most reactions between “uranate” species and CH₃OH involved elimination of CH₂=O and H₂O. However, for some product ions, particularly those that potentially contained hydride ligands, H₂ elimination reactions reminiscent of those observed here for [U^VIO₂(O)(H)]^– were observed.

It is important to note that an O₂ adduct to the [U^VIO₂(O)(H)]^- ion, which would appear at m/z 319, was not observed even at isolation/reaction times as long at 10 s. Earlier studies have shown that cationic U^VO₂ ⁺ complexes will reversibly bind molecular O₂, whereas U^VIO₂ ²⁺ species do not [32–34]. Similar behavior with respect to O₂ addition has since been demonstrated for negatively charged species that contain U^VO₂ ⁺ [35]. The tendency to add O₂ in an ion-molecule reaction can, therefore, be used to reveal those species that formally contain a U^VO₂ ⁺ core. The fact that the anion at m/z 287 in the present study does not add O₂ strongly suggests that it contains a U^VIO₂ ²⁺ core, and supports our composition assignment.

Isolation of the species at m/z 331 (attributed to formation of [U^VIO₂(O)(O₂C-H)]^– in Figure 1b and c) to react with H₂O was also investigated. The only product observed, however, was an apparent H₂O adduct using isolation times ranging from 1 ms to 10 s (product ion spectra using isolation times of 1 ms to 1 s are provided in Supplementary Figures S2a–d). As for the product generated by reaction of [U^VIO₂(O)(OH)]^– with H₂O, the composition and structure of the peak at m/z 349 is difficult to assign unambiguously. The species may be a simple H₂O adduct (i.e., [U^VIO₂(O)(O₂C-H(H₂O)]^–), or a di-hydroxy species such as [U^VIO₂(OH)₂(O₂C-H)]^–. Subsequent CID of the ion at m/z 349 caused simple H₂O elimination (Supplementary Figure S2e). As for the m/z 287 species, no O₂ adduct was observed for the m/z 331 ion, suggesting that the anion also contains a U^VIO₂ ²⁺ core coordinated by formate and oxide ligands.

The ion-molecule reactivity of the product ion at m/z 361, generated by elimination of CO₂ from [U^VIO₂(O₂C-H)₃]^–, was more interesting. Although no products attributable to an ion-molecule reaction were observed at 1 ms isolation/reaction time, as shown in Figure 3, isolation of the species for 10 ms through 10 s (Figure 3a–d, respectively) and reaction with H₂O led to formation of an ion at m/z 377. This species is assigned as a U^VIO₂ ²⁺ complex with hydroxide and formate, with corresponding elimination of H₂, generated via Reaction 4.

Figure 3

Product ion spectra generated by isolation and storage of [U^VIO₂(H)(O₂C-H)₂]^–, m/z 361, without imposed collisional activation, for reaction with background H₂O. Spectra in (a)–(d) are for isolation times of 1 ms, 10 ms, 1 s, and 10 s, respectively. The spectrum in (e) shows the product ions generated by CID of the m/z 377 ion. In the spectrum, the circles and vertical arrows illustrate MSⁿ dissociation pathway. Horizontal arrows indicate ion-molecule reaction steps. The bold peak labels indicate the precursor selected for CID whereas labels in italics represent the products from dissociation or ion-molecule reactions as indicated in the text

$$ {\left[{\mathrm{U}}^{\mathrm{VI}}{\mathrm{O}}_2\left(\mathrm{H}\right){\left({\mathrm{O}}_2\mathrm{C}-\mathrm{H}\right)}_2\right]}^{\hbox{--} }+{\mathrm{H}}_2\mathrm{O}\to {\left[{\mathrm{U}}^{\mathrm{VI}}{\mathrm{O}}_2\left(\mathrm{O}\mathrm{H}\right){\left({\mathrm{O}}_2\mathrm{C}-\mathrm{H}\right)}_2\right]}^{\hbox{--} }+{\mathrm{H}}_2 $$

(4)

Product ions at m/z 333 and 331 were observed after subsequent CID of the ion at m/z 377 (MS⁴ stage, Figure 3e). The product ion at m/z 331 exhibited the same ion-molecule reactivity as discussed above (and shown in Supplementary Figure S2a–e), leading to the conclusion that it is [U^VIO₂(O)(O₂C-H)]^– generated by intra-complex proton transfer and elimination of (neutral) formic acid by Reaction 5.

$$ {\left[{\mathrm{U}}^{\mathrm{VI}}{\mathrm{O}}_2\left(\mathrm{O}\mathrm{H}\right){\left({\mathrm{O}}_2\mathrm{C}-\mathrm{H}\right)}_2\right]}^{\hbox{--}}\to {\left[{\mathrm{U}}^{\mathrm{VI}}{\mathrm{O}}_2\left(\mathrm{O}\right)\left({\mathrm{O}}_2\mathrm{C}-\mathrm{H}\right)\right]}^{\hbox{--} }+{\mathrm{HO}}_2\mathrm{C}-\mathrm{H} $$

(5)

The product ion at m/z 333 (generated by CID of [U^VIO₂(OH)(O₂C-H)₂]^– at m/z 377) is formed by elimination of 44 mass units (CO₂), and likely represents the creation of [U^VIO₂(OH)(H)(O₂CH)]^– by Reaction 6a. Subsequent CID of the ion at m/z 333 generated the species at m/z 331 (spectrum not shown) by elimination of H₂. Isolation of the product ion at m/z 333 for 1 s, without imposed collision activation (Supplementary Figure S3a), led to the generation of a peak at m/z 349 (16 mass units higher), and likely represents formation of a dihydroxy species through reaction with H₂O as shown in Reaction 6b. Subsequent CID of the m/z 349 ion (Supplementary Figure S3b) caused elimination of 18 mass units (presumably H₂O through Reaction 6c) to create the m/z 331 ion, for which the ion-molecule reaction behavior was identical to that described above.

$$ {\left[{\mathrm{U}}^{\mathrm{VI}}{\mathrm{O}}_2\left(\mathrm{O}\mathrm{H}\right){\left({\mathrm{O}}_2\mathrm{C}-\mathrm{H}\right)}_2\right]}^{\hbox{--}}\to {\left[{\mathrm{U}}^{\mathrm{VI}}{\mathrm{O}}_2\left(\mathrm{O}\mathrm{H}\right)\left(\mathrm{H}\right)\left({\mathrm{O}}_2\mathrm{C}-\mathrm{H}\right)\right]}^{\hbox{--} }+{\mathrm{CO}}_2 $$

(6a)

$$ {\left[{\mathrm{U}}^{\mathrm{VI}}{\mathrm{O}}_2\left(\mathrm{O}\mathrm{H}\right)\left(\mathrm{H}\right)\left({\mathrm{O}}_2\mathrm{C}-\mathrm{H}\right)\right]}^{\hbox{--} }+{\mathrm{H}}_2\mathrm{O}\to {\left[{\mathrm{U}}^{\mathrm{VI}}{\mathrm{O}}_2{\left(\mathrm{O}\mathrm{H}\right)}_2\left({\mathrm{O}}_2\mathrm{C}-\mathrm{H}\right)\right]}^{\hbox{--} }+{\mathrm{H}}_2 $$

(6b)

$$ {\left[{\mathrm{U}}^{\mathrm{VI}}{\mathrm{O}}_2{\left(\mathrm{O}\mathrm{H}\right)}_2\left({\mathrm{O}}_2\mathrm{C}-\mathrm{H}\right)\right]}^{\hbox{--}}\to {\left[{\mathrm{U}}^{\mathrm{VI}}{\mathrm{O}}_2\left(\mathrm{O}\right)\left({\mathrm{O}}_2\mathrm{C}-\mathrm{H}\right)\right]}^{\hbox{--} }+{\mathrm{H}}_2\mathrm{O} $$

(6c)

No O₂ adducts to the ions at m/z 333, 349. or 377 were observed with isolation/reaction times up to 10 s, suggesting that each contains a U^VIO₂ ²⁺ (i.e., U^VI) core.

Tandem MS of [UO₂(O₂C-H)(CH₃CH₂OH)₂]⁺

The positive ion mass spectrum derived from uranyl-formate in H₂O/CH₃CH₂OH is shown in Figure 4a. The most abundant positively charged species generated by ESI appeared at m/z 407 and was initially assigned as [U^VIO₂(O₂C-H)(CH₃CH₂OH)₂]⁺. However, because formate and ethoxide have the same nominal mass, the species could also be assigned the composition [U^VIO₂(OCH₂CH₃)(CH₃CH₂OH)₂]⁺. The dominant pathways for the multiple-stage CID of the m/z 407 precursor (MS/MS and MS³ stages, Figure 4b and c) involved elimination of single CH₃CH₂OH ligands, leading ultimately to a product ion at m/z 315 (Figure 4c). Minor peaks at m/z 379 and 389 were observed in the CID spectrum of [U^VIO₂(O₂C-H)(CH₃CH₂OH)₂]⁺ (Figure 4b) that appear to involve the decomposition of CH₃CH₂OH into CH₂=CH₂ and H₂O were observed. Subsequent CID of these species generated the m/z 361 product ion that is the base peak in Figure 4b.

Figure 4

Positive ESI and CID spectra derived from uranyl formate in 50/50 H₂O/CH₃CH₂OH: (a) Full ESI spectrum, (b) CID of [U^VIO₂(O₂C-H)(CH₃CH₂OH)₂]^– at m/z 407, (c) CID (MS³ stage) of CID of [U^VIO₂(O₂C-H)(CH₃CH₂OH)]^– at m/z 361, and (d) CID (MS⁴ stage) of CID of [U^VIO₂(O₂C-H)]^– at m/z 315. In the spectra, the circles and arrows illustrate the MSⁿ pathway. In each spectrum, the bold peak label indicates the precursor selected for CID whereas labels in italics represent the products from dissociation or ion-molecule reactions as indicated in the text

With the mass accuracy of the LIT used for these experiments, it is not possible to distinguish formate from ethoxide. However, assignment can be made with greater confidence by inspection of the CID pattern generated from the m/z 315 ion (Figure 4d). To begin with, we recently investigated the fragmentation of uranyl-methoxide and uranyl-ethoxide cations, [U^VIO₂(OCH₃)]⁺ and [U^VIO₂(OCH₂CH₃)]⁺ [40]. Our experiments demonstrated that [U^VIO₂(OCH₃)]⁺ dissociates by reductive elimination of methoxy radical (•OCH₃) to make U^VO₂ ⁺, or by transfer of H^– from the methyl group to make U^VIO₂H⁺. The [U^VIO₂(OCH₂CH₃)]⁺ ion instead dissociates by pathways that involve the elimination or retention of CH₃ and CH₂O. The retention of CH₃ leads to a product ion at m/z 285, whereas retention of CH₂O creates a product ion at m/z 300: these specific product ions can, therefore, be considered diagnostic for a [U^VIO₂(OCH₂CH₃)]⁺ precursor ion. As is clear from inspection of Figure 4d, product ions at m/z 285 and 300 appear at less than 5% relative intensity, which strongly suggests that the population of ions contributing to the peak at m/z 315 in Figure 4c does not include a significant fraction of [U^VIO₂(OCH₂CH₃)]⁺. The m/z 315 product is, therefore, mostly uranyl-formate cation, [UO₂(O₂C-H)]⁺.

The dominant product ion generated by CID of [U^VIO₂(O₂C-H)]⁺ appears at m/z 287, which as a cation can be assigned the composition [U^VIO₂(OH)]⁺. Formation of [U^VIO₂(OH)]⁺ presumably involves intramolecular H atom transfer and elimination of CO via Reaction 7. We note that the m/z value of the product ion shifted by one mass unit to 288 for CID of the [U^VIO₂(O₂C-D)]⁺ precursor ion, consistent with the proposed fragmentation reaction.

$$ {\left[{\mathrm{U}}^{\mathrm{VI}}{\mathrm{O}}_2\left({\mathrm{O}}_2\mathrm{C}-\mathrm{H}\right)\right]}^{+}\to {\left[{\mathrm{U}}^{\mathrm{VI}}{\mathrm{O}}_2\left(\mathrm{O}\mathrm{H}\right)\right]}^{+}+\mathrm{C}\mathrm{O} $$

(7)

No O₂ adduct was observed when the m/z 287 (or 288 derived from the d-labeled precursor) was isolated without imposed collision activation. This again suggests that the species contains a U^VIO₂ ²⁺ core. Formation of [U^VIO₂(OH)]⁺ by intramolecular H atom transfer was also observed as a minor pathway in our earlier investigation of the dissociation of [U^VIO₂OCH₂CH₃]⁺ [40].

Tandem MS of [UO₂(O₂C-CH₃)₃]^–

The ESI-MS spectrum generated from the uranyl acetate solution is shown in Figure 5a. As for formate, ESI generated primarily [U^VIO₂(O₂C-CH₃)₃]^– at m/z 447 and larger clusters containing UO₂ ²⁺ and acetate. The multiple-stage CID spectra generated by initially isolating [U^VIO₂(O₂C-CH₃)₃]^– at m/z 447 are shown in Figure 5b–e. Data generated using deuterium-labeled acetic acid (HO₂-CD₃) are provided in Supplementary Figure S4a–e.

Figure 5

Negative ESI and CID spectra derived from uranyl acetate in 50/50 H₂O/CH₃CH₂OH: (a) Full ESI spectrum, (b) CID of [U^VIO₂(O₂C-CH₃)₃]^– at m/z 447, (c) CID (MS³ stage) of [U^VO₂(O₂C-CH₃)₂]^– at m/z 388, (d) CID (MS⁴ stage) of [U^VO₂(CH₃)(O₂C-CH₃)]^– at m/z 344, and (e) CID (MS⁵ stage) of [U^VO₂(O₂C=CH₂)]^– at m/z 328. In the spectra, the circles and arrows illustrate the MSⁿ pathway. In each spectrum, the bold peak label indicates the precursor selected for CID whereas labels in italics represent the products from dissociation or ion-molecule reactions as indicated in the text

The dominant product ion generated by CID of [U^VIO₂(O₂C-CH₃)₃]^– at m/z 447 appeared at m/z 388, through a neutral loss of 59 mass units. The m/z 388 product is, therefore, generated by a dissociation reaction that involves elimination of acetyloxyl radical, CH₃CO₂•, and reduction of U^VIO₂ ²⁺ to U^VO₂ ⁺. A minor peak (less than 5% relative intensity) was observed at m/z 403, which corresponds to the elimination of CO₂ to create [U^VIO₂(CH₃)(O₂C-CH₃)₂]^–. Again, because of the low mass cutoff of the LIT in these experiments, the degree to which [U^VIO₂(O₂C-CH₃)₃]^– dissociates to produce the acetate anion could not be determined.

The elimination of CH₃CO₂• and CO₂ from [U^VIO₂(O₂C-CH₃)₃]^– at the MS/MS stage are consistent with earlier CID studies [26, 30], including an investigation of decarboxylation and synthesis of organometallic actinide (An) species performed using a 3-D quadrupole ion trap [26]. Of three possible CID pathways for [An^VIO₂(O₂C-CH₃)₃]^–, An = U, Np, and Pu, considered in the earlier study (decarboxylation, anion ligand loss, and neutral ligand loss), the dominant CID channel was neutral ligand loss (elimination of acetyloxyl radical) to produce [An^VO₂(O₂C-CH₃)₂]^–. The CID reaction also leads to reduction of An^VIO₂ ²⁺ to An^VO₂ ⁺, and the favored dissociation pathway was consistent with DFT calculations of reaction energies.

CID of [U^VO₂(O₂C-CH₃)₂]^– at m/z 388 (MS³ stage, Figure 5b) caused the loss of 44 mass units (CO₂) to generate a product ion at m/z 344. The neutral loss was also 44 mass units following CID of the deuterium-labeled analogue, [U^VO₂(O₂C-CD₃)₂]^– (Supplementary Figure S4c). The composition of the ion at m/z 344 is, therefore, assigned as [U^VO₂(CH₃)(O₂C-CH₃)]^– ([U^VO₂(CD₃)(O₂C-CD₃)]^– for the deuterium-labeled analogue). A minor peak at m/z 346 was also observed, the identity of which is discussed below.

Subsequent CID of the product ion at m/z 344 caused the elimination of 16 mass units, which we attribute to loss of CH₄, to leave an ion at m/z 328. The neutral loss at this CID stage shifted to 20 mass units for the deuterium labeled version of the precursor ion (Supplementary Figure S4d), demonstrating that the loss of methane involves intramolecular proton/deuteron transfer from an acetate ligand to the CH₃/CD₃ ligand. The product ion at m/z 328 is, therefore, assigned the composition [U^VO₂(O₂C=CH₂]^–: U^VO₂ ⁺ coordinated by an acetate enolate ligand. Subsequent CID of the m/z 328 ion generates a product ion at m/z 286, for which a composition assignment of [U^VO₂(O)]^–, created by elimination of ketene (O=C=CH₂), is reasonable. We note that there is precedent for products ions that are metal complexes with acetate enolate. For example, Meyer and coworkers reported the generation of lithium acetate enolate from malonic acid precursors using CID [41].

At this point, we can compare our CID results using the LIT to those of Luo et al. [30], who investigated the detection of uranyl species in natural water samples using extractive electrospray ionization and an instrument similar to the one employed in our study. In a discussion of the intrinsic fragmentation pathways for [U^VIO₂(O₂C-CH₃)₃]^–, the elimination of (neutral) CH₃CO₂• at the MS/MS stage to create [U^VO₂(O₂C-CH₃)₂]^– at m/z 388 was noted, as was the relatively low abundance of the decarboxylation product at m/z 403. CID of [U^VO₂(O₂C-CH₃)₂]^– was then reported to generate product ions at m/z 344 and 346 by elimination of 44 and 42 mass units, respectively, to leave [U^VO₂(CH₃)(O₂C-CH₃)]^– and [U^VO₂(OH)(O₂C-CH₃)]^–. As we describe below, it is more likely that the latter product ion, [U^VO₂(OH)(O₂C-CH₃)]^–, is generated by a reaction of [U^VO₂(CH₃)(O₂C-CH₃)]^– with H₂O after its generation by decarboxylation of [U^VO₂(O₂C-CH₃)₂]^–. In fact, there appear to be significant differences in the relative intensities of the m/z 344 and 346 product ions in our experiments compared with the earlier study by Luo and coworkers [30]. The m/z 346 ion is the base peak in the previous experiments, whereas the same ion appears at only 15% relative intensity in our experiments. The differences can probably be attributed to lower levels of adventitious H₂O in our experiment.

In the earlier study, CID of both the m/z 344 and 346 ions was reported to result in generation of product ions at m/z 302 and 304, which then ultimately dissociate to leave a terminal product ion at m/z 286 [30]. The m/z 286 product, assigned as [U^VO₂(O)]^–, is consistent with our experiments. However, the data shown here in Figure 5d indicate that under the conditions used in our study, the dominant dissociation pathway for the m/z 344 ion is the generation of the ion at m/z 328 through elimination of CH₄. The ions at m/z 302 and 304 are not observed as product ions from the dissociation of [U^VO₂(CH₃)(O₂C-CH₃)]^–. However, as we discuss below, both are observed as CID products of the m/z 346 product ion after it is generated by reaction of [U^VO₂(CH₃)(O₂C-CH₃)]^– with H₂O.

As noted above, in our experiments the peak at m/z 346 was observed at ca. 15% relative intensity in the spectrum shown in Figure 5c. This peak is two mass units higher (the difference between CH₃ and OH) than the species at m/z 344, suggesting that [U^VO₂(CH₃)(O₂C-CH₃)]^– (m/z 344) is hydrolyzed by reaction with background H₂O to generate [U^VO₂(OH)(O₂C-CH₃)]^– (m/z 346). Hydrolysis of [U^VO₂(CH₃)(O₂C-CH₃)]^– to generate [U^VO₂(OH)(O₂C-CH₃)]^– is consistent with the earlier study of actinyl-organometallic ions in the gas phase [26]. To test this hypothesis further, [U^VO₂(CH₃)(O₂C-CH₃)]^– at m/z 344 was isolated and stored in the LIT for periods ranging from 1 ms to 10 s. The resulting product ion spectra are shown in Figure 6, and it is clear that with increasing isolation and reaction time there is a decrease in the relative intensity of the m/z 344 ion, and increase in the intensity of the ion at m/z 346. This strongly suggests that [U^VO₂(OH)(O₂C-CH₃)]⁺ is generated from [U^VO₂(CH₃)(O₂C-CH₃)]^– by Reaction 8.

Figure 6

Product ion spectra (MS⁵ stage) generated by isolation and storage of [U^VO₂(CH₃)(O₂C-CH₃)]^–, m/z 344, without imposed collisional activation, for reaction with background H₂O. Spectra were generated using isolation/reaction times of: (a) 1 ms, (b) 10 ms, (c) 100 ms, (d) 1 s, and (e) 10 s. In the spectrum, the circles and vertical arrows illustrate MSⁿ dissociation pathway. Horizontal arrows indicate ion-molecule reaction steps. The bold peak labels indicate the precursor selected for CID whereas labels in italics represent the products from dissociation or ion-molecule reactions as indicated in the text

$$ {\left[{\mathrm{U}}^{\mathrm{V}}{\mathrm{O}}_2\left({\mathrm{CH}}_3\right)\left({\mathrm{O}}_2\mathrm{C}-{\mathrm{CH}}_3\right)\right]}^{\hbox{--} }+{\mathrm{H}}_2\mathrm{O}\to {\left[{\mathrm{U}}^{\mathrm{V}}{\mathrm{O}}_2\left(\mathrm{O}\mathrm{H}\right)\left({\mathrm{O}}_2\mathrm{C}-{\mathrm{CH}}_3\right)\right]}^{+}+{\mathrm{CH}}_4 $$

(8)

For isolation of [U^VO₂(CD₃)(O₂C-CD₃)]^– (m/z 350, data not shown), reaction with H₂O generated [U^VO₂(OH)(O₂C-CD₃)]⁺ at m/z 349, consistent with addition of OH and elimination of CD₃H.

Additional peaks at m/z 361 and 378 in Figure 6d and e represent the formation of the O₂ adducts [U^VO₂(O₂C-CH₃)(O₂)]^– and [U^VO₂(OH)(O₂C-CH₃)(O₂)]^–. As noted in the discussion of product ions from the uranyl-formate system, formation of O₂ adducts is known to occur for complex ions that contain the reduced U^VO₂ ⁺ core [32–35]. The appearance of the O₂ adducts in spectra shown in Figure 6d and e, therefore, demonstrate that both the m/z 344 and 346 ions contain U^VO₂ ⁺ cores, as expected following the initial elimination of acetyloxyl radical from [U^VIO₂(O₂C-CH₃)₃]^- (Figure 5a).

Subsequent isolation and CID of the presumed hydrolysis product, [U^VO₂(OH)O₂C-CH₃)]⁺, generated several product ions (Supplementary Figure S5b). One appeared at m/z 328, representing the elimination of H₂O, presumably to generate [U^VO₂(O₂C=CH₂)]^– as described above. Additional product ions from the CID of [U^VO₂(OH)O₂C-CH₃)]⁺ included the species at m/z 304 and 302 (both species were noted as apparent dissociation products from the m/z 344 and 346 precursors by Luo et al. [30]). The former is assigned as [U^VO₂(OH)₂]^–, which presumably is generated via intramolecular proton transfer and elimination of O=C=CH₂. The latter ion is assigned as [U^VO₂(O₂)]^–. CID of the ion at m/z 302 (spectrum not shown) caused elimination of 16 mass units to generate [U^VO₂(O)]^– at m/z 286.

Taken as a whole, the CID results reported here in Figure 5, and the ion-molecule reaction observations discussed above, provide a more accurate description of the intrinsic fragmentation pathways for the uranyl-acetate anion, [U^VIO₂(O₂C-CH₃)₃]^–. The primary fragmentation pathway for the intact precursor is elimination of acetyloxyl radical, CH₃CO₂•, with reduction of U^VIO₂ ²⁺ to U^VO₂ ⁺. Subsequent CID of [U^VO₂(O₂C-CH₃)₂]^– causes decarboxlation to generate [U^VO₂(CH₃)(O₂C-CH₃)]^–. Though some generation of [U^VO₂(OH)(O₂C-CH₃)]^– (m/z 346) by elimination of 42 mass units cannot be ruled out as a CID pathway, our experiments suggest that the species is more likely formed by a reaction of [U^VO₂(CH₃)(O₂C-CH₃)]^– with H₂O. CID of [U^VO₂(CH₃)(O₂C-CH₃)]^– causes loss of CH₄ to create [U^VO₂(O₂C=CH₂)]^-, which subsequently fragments to generate [U^VO₂(O)]^-.

Tandem MS of [UO₂(O₂C-CH₃)(CH₃CH₂OH)]⁺

The most abundant positively charged species generated by ESI of uranyl acetate dissolved in 50/50 H₂O/CH₃CH₂OH (data not shown) appeared at m/z 421 and was assigned as [U^VIO₂(O₂C-CH₃)(CH₃CH₂OH)₂]⁺. CID of the m/z 421 precursor (MS/MS and MS³ stages, Supplementary Figure S5a and b) caused elimination of single CH₃CH₂OH ligands to generate a product ion at m/z 329. Unlike the case for the formate analog, there was no ambiguity as to the assignment of the ion at m/z 329 as uranyl-acetate cation, [U^VIO₂(O₂C-CH₃)]⁺. Subsequent CID of [U^VIO₂(O₂C-CH₃)]⁺ generated a product ion at m/z 287, again assigned as [U^VIO₂(OH)]⁺. The product ion shifted to m/z 288 when CID of [U^VIO₂(O₂C-CD₃)]⁺ (Supplementary Figure S5e) was examined, again suggesting a formation pathway that involves intramolecular H⁺ transfer and elimination of O=C=CH₂ (O=C=CD₂ for the deuterium labeled analogue). Isolation of [U^VIO₂(OH)]⁺ or [U^VIO₂(OD)]⁺ for periods up to 10 s caused formation of only H₂O adducts (data not shown). The lack of O₂ adducts is indicative of the U^VI oxidation state.

Conclusions

To summarize, ESI, CID, and gas-phase ion-molecule reactions were used to make and characterize species derived from precursors composed of uranyl cation (UO₂ ²⁺) coordinated by formate or acetate ligands. The low levels of adventitious H₂O in our linear ion trap allowed (1) the intrinsic dissociation pathways for both systems to be determined, and (2) the identification of apparent dissociation products that are better described as species generated by ion-molecule reactions.

Anionic complexes containing UO₂ ²⁺ and formate ligands fragment by decarboxylation and by elimination of CH₂=O, ultimately to produce an oxo-hydride species [UO₂(O)(H)]^–. Elimination of formyloxyl radical, with associated reduction of U^VIO₂ ²⁺ to U^VO₂ ⁺, was not observed. Product ions generated by CID, which include [U^VIO₂(O)(H)]^– and [U^VIO₂(OH)(O₂C-H)]^–, undergo subsequent reactions that involve H₂O addition and elimination of H₂.

Our experiments allow for a revision of the intrinsic dissociation pathways for [U^VIO₂(O₂-CH₃)₃]^–. The primary fragmentation pathway for the intact precursor is elimination of acetyloxyl radical, CH₃CO₂•, with reduction of U^VIO₂ ²⁺ to U^VO₂ ⁺. Subsequent CID of [U^VO₂(O₂C-CH₃)₂]^– causes decarboxlation to generate [U^VO₂(CH₃)(O₂C-CH₃)]^–. Though some formation of [U^VO₂(OH)(O₂C-CH₃)]^– (m/z 346) by direct elimination of 42 mass units during CID is possible, our experiments suggest that the species is instead formed by a reaction of [U^VO₂(CH₃)(O₂C-CH₃)]^– with H₂O. CID of [U^VO₂(CH₃)(O₂C-CH₃)]^– causes loss of CH₄ to create [U^VO₂(O₂C=CH₂)]^–, which subsequently fragments to generate [U^VO₂(O)]^-. Loss of CH₄ occurs by an intra-complex H⁺ transfer process that leaves UO₂ ⁺ coordinated by acetate and acetate enolate anions. Subsequent dissociation step causes elimination of CH₂=C=O to furnish [UO₂(O)]^–. Elimination of CH₄ is also observed as a result of hydrolysis caused by ion-molecule reactions with H₂O.

For both systems, CID of the positively charged precursor [U^VIO₂(O₂-C-R)(CH₃CH₂OH)₂]⁺, R = H, D, CH₃, or CD₃, causes elimination of CH₃CH₂OH ligands in successive dissociation stages. Using a comparison to the known fragmentation pattern of [U^VIO₂(OCH₂CH₃)]⁺, CID of the m/z 315 product ion derived from [U^VIO₂(O₂-C-H)(CH₃CH₂OH)₂]⁺ can be used to assign the composition as the uranyl-formate cation, [UO₂(O₂C-H)]⁺ (rather than uranyl-ethoxide cation). CID of both [UO₂(O₂C-H)]⁺ and [UO₂(O₂C-CH₃)]⁺ generate [U^VIO₂(OH)]⁺, presumably through intramolecular H transfer and elimination of (neutral) CO or CH₂=C=O for the formate and acetate precursors.