Molybdenum(II) acetylacetonate

• Molecular Formula: C10H14MoO6
• Molecular Weight: 326.16 g/mol
• SMILES: CC(=CC(=O)C)[O-].CC(=CC(=O)C)[O-].O=[Mo+2]=O
• InChI: InChI=1S/2C5H8O2.Mo.2O/c2*1-4(6)3-5(2)7;;;/h2*3,6H,1-2H3;;;/q;;+2;;/p-2/b2*4-3-
• InChIKey: SKMUJBBRXZPAJY-VGKOASNMSA-L

Researchers replacing MoO₂(acac)₂ with bulk MoO₃ sacrifice active-site dispersion, incurring up to 4× lower metathesis activity and elevated oligomerization byproducts. This Mo(VI) precursor resolves both process economics issues. • Up to 4× higher metathesis activity vs. MoO₃ on SBA-15/MWW zeolites; <3% oligomerization byproducts vs. ~12% for MoO₃-derived catalysts • Single-step MOCVD of crystalline α-MoO₃ films (400-560°C), no post-annealing required • Chemoselective perhydrolysis of ketones/ketals with ethereal H₂O₂, compatible with TBS/MOM protecting groups Recommend ≥97% purity with 1-2% acetylacetone stabilizer; store under argon at 2-8°C.

MoO₂(acac)₂ — Sourcing Guide for Procurement Teams

Molybdenum(II) acetylacetonate (CAS 17524-05-9, systematically bis(acetylacetonato)dioxomolybdenum(VI), MoO₂(acac)₂) is a Mo(VI) coordination complex bearing two bidentate acetylacetonate ligands and two terminal oxido ligands in a distorted octahedral geometry [1]. Despite the commercial nomenclature implying Mo(II), the metal center resides in the +6 oxidation state, distinguishing it fundamentally from truly low-valent Mo β-diketonates such as Mo(acac)₃ (Mo(III)) and dimeric Mo(II) carboxylates [2]. This compound is sourced primarily as a catalyst precursor for epoxidation, metathesis, and oxidative transformations, and as a molecular precursor for chemical vapor deposition (MOCVD) of MoO₃ thin films [3]. Commercial specifications typically range from 95% to 99% purity, with higher grades stabilized by 1–2% free acetylacetone and requiring storage under inert atmosphere at 2–8°C .

• [1] Kamenar, B., Prout, C. K., & Wright, J. D. (1973). Crystal structure of bis(acetylacetonato)dioxomolybdenum(VI). Cryst. Struct. Commun., 2, 41–44. Redetermined by Lammertsma et al. (2022). Inorg. Chem., 61(38), 14918–14923. doi:10.1021/acs.inorgchem.2c00824 View Source
• [2] Masero, F. & Mougel, V. (2024). Chemical and redox non-innocence in low-valent molybdenum β-diketonate complexes: novel pathways for CO₂ and CS₂ activation. Chem. Sci., 15, 16954–16965. doi:10.1039/D4SC03496A View Source
• [3] Bouzidi, A., et al. (2006). Preparation of molybdenum oxide thin films by MOCVD. Journal of Alloys and Compounds, 425(1–2), 312–317. doi:10.1016/j.jallcom.2006.01.055 View Source

Performance Penalties of Substituting MoO₂(acac)₂

Substituting MoO₂(acac)₂ with superficially similar molybdenum compounds introduces quantifiable degradation in catalyst dispersion, activity, or selectivity that propagates into process economics. MoO₃, the most common alternative inorganic Mo(VI) source, exhibits inferior spreading on zeolite and mesoporous supports, leading to bulk crystallite formation detectable by XRD and lower active site accessibility [1]. Mo(acac)₃ stabilizes Mo in the +III oxidation state and participates in fundamentally different redox chemistry (small-molecule activation rather than oxygen-atom transfer), precluding its use as a drop-in replacement for oxidation catalysis [2]. The bulkier diketonate analog MoO₂(tmhd)₂ alters the steric environment at the metal center, trading catalytic activity for epoxidation selectivity in a manner that must be quantified per substrate [3]. Even within the same ligand class, MoO₂(glycolate)₂ yields different thermal spreading behavior and pore architecture on SBA-15 compared to the acetylacetonate variant [4]. The following quantitative evidence specifies where MoO₂(acac)₂ occupies a distinct performance niche.

• [1] Balcar, H., Mishra, D., Marceau, E., Carrier, X., Žilková, N., & Bastl, Z. (2009). Molybdenum oxide catalysts for metathesis of higher 1-alkenes via supporting MoO₂(acetylacetonate)₂ and MoO₂(glycolate)₂ on SBA-15 mesoporous molecular sieves. Applied Catalysis A: General, 359(1–2), 129–135. doi:10.1016/j.apcata.2009.02.037 View Source
• [2] Masero, F. & Mougel, V. (2024). Chemical and redox non-innocence in low-valent molybdenum β-diketonate complexes. Chem. Sci., 15, 16954–16965. doi:10.1039/D4SC03496A View Source
• [3] Conte, M. & Hippler, M. (2016). Dynamic NMR and quantum-chemical study of the stereochemistry and stability of the chiral MoO₂(acac)₂ complex in solution. J. Phys. Chem. A, 120(34), 6677–6687. doi:10.1021/acs.jpca.6b03563 View Source
• [4] Balcar, H., et al. (2009). Applied Catalysis A: General, 359(1–2), 129–135. doi:10.1016/j.apcata.2009.02.037 View Source

Quantitative Differentiation Against Closest Analogs

Superior Molybdenum Dispersion on Zeolite Supports vs. MoO₃

When deposited on MWW zeolites at equivalent molybdenum loading (6 wt% Mo), MoO₂(acac)₂-derived catalysts exhibit no detectable bulk MoO₃ crystallite signals in XRD, whereas MoO₃-derived catalysts display clear MoO₃ reflections indicating incomplete dispersion and agglomeration [1]. On MCM-22(70) support, the MoO₂(acac)₂-based catalyst retains a BET surface area of 355 m²/g compared to only 180 m²/g for the MoO₃-based catalyst — a 97% higher surface area [2]. This superior dispersion is attributed to the molecular-level spreading of the acetylacetonate precursor, which decomposes to well-isolated MoOₓ surface species upon calcination, in contrast to the thermal spreading of bulk MoO₃ that leaves residual crystallites [1].

Evidence Dimension	Molybdenum dispersion quality on zeolite supports (XRD crystallinity and BET surface area retention)
Target Compound Data	6MoO₂(acac)₂/MCM-22(28): no MoO₃ XRD signals; SBET = 426 m²/g. 6MoO₂(acac)₂/MCM-22(70): SBET = 355 m²/g
Comparator Or Baseline	6MoO₃/MCM-22(28): MoO₃ XRD signals detected; SBET = 423 m²/g. 6MoO₃/MCM-22(70): SBET = 180 m²/g (severe pore blockage)
Quantified Difference	BET surface area retention: MoO₂(acac)₂ retains 355 vs. 180 m²/g for MoO₃ on MCM-22(70) — a 97% advantage. No bulk crystallite formation for MoO₂(acac)₂ vs. clear MoO₃ reflections for MoO₃.
Conditions	MWW zeolites MCM-22(28) and MCM-22(70) in NH₄⁺ form; 6 wt% Mo loading; thermal spreading method; XRD with Cu Kα radiation; BET by N₂ adsorption at 77 K

Why This Matters

Superior dispersion directly translates to higher density of accessible active sites per gram of catalyst, enabling procurement of MoO₂(acac)₂ over MoO₃ when maximizing metathesis or oxidation activity per unit Mo loading is economically critical.

• [1] Balcar, H., et al. (2018). MoO₃ on zeolites MCM-22, MCM-56 and 2D-MFI as catalysts for 1-octene metathesis. Beilstein J. Org. Chem., 14, 2931–2939. doi:10.3762/bjoc.14.272 View Source
• [2] Balcar, H., et al. (2018). Table 1, BET surface area data for MoO₃ and MoO₂(acac)₂ on MWW zeolites. Beilstein J. Org. Chem., 14, 2931–2939. doi:10.3762/bjoc.14.272 View Source

Higher 1-Octene Metathesis Activity vs. MoO₃-Derived Catalysts

On SBA-15 mesoporous silica, catalysts prepared from MoO₂(acac)₂ deliver up to four times higher 1-octene metathesis activity compared to MoO₃ supported on the same SBA-15 or on conventional silica [1]. On the less acidic MCM-22(70) zeolite support, 5MoO₂(acac)₂/MCM-22(70) achieves 35% total 1-octene conversion (Ktot) after 20 hours at 40°C, whereas 6MoO₃/MCM-22(70) remains essentially inactive (2–3% conversion at 6 hours, no further progress) [2]. Critically, the MoO₂(acac)₂-derived catalysts exhibit substantially lower oligomerization side-product formation (Kol = 1–3%) compared to 6MoO₃/MCM-22(28) (Kol ≈ 12%), indicating that the improved dispersion not only enhances activity but also suppresses acid-catalyzed side reactions associated with exposed support acid sites [2].

Evidence Dimension	1-Octene metathesis catalytic activity (total conversion Ktot and oligomerization Kol)
Target Compound Data	5MoO₂(acac)₂/MCM-22(70): Ktot = 35% (20 h, 40°C), Kol = 1–3%. MoO₂(acac)₂/SBA-15: up to 4× activity vs. MoO₃/SBA-15
Comparator Or Baseline	6MoO₃/MCM-22(70): Ktot ≈ 2–3% (6 h, negligible). 6MoO₃/MCM-22(28): Kol ≈ 12% (at 2 h). MoO₃/SBA-15: baseline activity (1×)
Quantified Difference	35% vs. ~3% conversion (≥11.7× advantage on MCM-22(70)). Up to 4× activity advantage on SBA-15. Oligomerization reduced from ~12% to 1–3%.
Conditions	Neat 1-octene, 40°C, 1-octene/Mo molar ratio = 320, 50 mg catalyst, 1.5 mL 1-octene, reaction time 2–20 h

Why This Matters

The activity gap is sufficiently large that selecting MoO₃ over MoO₂(acac)₂ for metathesis catalyst preparation would require proportionally larger reactor volumes, higher Mo loading, or elevated temperatures to match throughput — directly increasing capital and operating costs.

• [1] Balcar, H., Mishra, D., Marceau, E., Carrier, X., Žilková, N., & Bastl, Z. (2009). Molybdenum oxide catalysts for metathesis of higher 1-alkenes. Applied Catalysis A: General, 359(1–2), 129–135. doi:10.1016/j.apcata.2009.02.037 View Source
• [2] Balcar, H., et al. (2018). MoO₃ on zeolites MCM-22, MCM-56 and 2D-MFI as catalysts for 1-octene metathesis. Beilstein J. Org. Chem., 14, 2931–2939. doi:10.3762/bjoc.14.272 View Source

Post-Annealing-Free Crystalline MoO₃ Thin Film via MOCVD

Using MoO₂(acac)₂ as the molecular precursor in a horizontal hot-wall MOCVD apparatus, crystalline α-MoO₃ films are obtained directly at deposition temperatures of 400–560°C (673–833 K) without any post-annealing treatment [1]. The best crystalline quality — exhibiting needle-like crystallites with strong (010) preferred orientation and transparent appearance — is achieved at ~560°C (833 K) [1]. Alternative physical deposition methods (pulsed laser deposition, RF sputtering, thermal evaporation) and other Mo precursors typically require a separate post-annealing step at elevated temperatures to crystallize the initially amorphous or poorly crystalline deposit [2]. The precursor's thermal decomposition pathway, characterized by TG/DTA showing oxide formation around 430°C (703 K), provides a well-defined window for process optimization [1].

Evidence Dimension	Requirement for post-annealing treatment to achieve crystalline MoO₃ films
Target Compound Data	MoO₂(acac)₂ MOCVD: crystalline α-MoO₃ films obtained directly at 400–560°C without post-annealing; best quality at 560°C with strong (010) orientation
Comparator Or Baseline	PLD-deposited MoO₃: requires post-annealing at ≥400°C to induce crystallinity. Thermal evaporation/sputtering: typically requires post-annealing for crystalline phase.
Quantified Difference	Elimination of one full thermal processing step (post-annealing). Deposition window: 400–560°C for crystalline phase vs. room-temperature deposition + separate annealing for alternatives.
Conditions	Horizontal hot-wall MOCVD; precursor synthesized from MoO₃ powder and acetylacetone; deposition on glass/Si substrates; O₂ carrier gas; XRD characterization

Why This Matters

Elimination of the post-annealing step reduces thermal budget, processing time, and equipment footprint in thin-film manufacturing workflows, making MoO₂(acac)₂ the preferred precursor when integrating MoO₃ deposition into temperature-sensitive device fabrication sequences.

• [1] Bouzidi, A., Benramdane, N., Tabet-Derraz, H., Mathieu, C., Khelifa, B., & Desfeux, R. (2006). Preparation of molybdenum oxide thin films by MOCVD. Journal of Alloys and Compounds, 425(1–2), 312–317. doi:10.1016/j.jallcom.2006.01.055 View Source
• [2] Trzciński, M., et al. (2023). Controlling crystallites orientation and facet exposure for enhanced electrochemical properties of polycrystalline MoO₃ films. Scientific Reports, 13, 16738. doi:10.1038/s41598-023-43678-7 View Source

Distinct Chemoselectivity in Perhydrolysis vs. Alternative Catalysts

In the perhydrolysis of ketones, ketals, and epoxides with ethereal H₂O₂, MoO₂(acac)₂ (5 mol%) achieves distinct chemoselectivity that cannot be replicated by other Lewis acid catalysts including I₂, ceric ammonium nitrate (CAN), or Re₂O₇ [1]. The reactivity order established is epoxides > ketones > ethylene glycol ketals, with tri- and tetrasubstituted epoxides exhibiting the highest reactivity. Bifunctional substrates undergo selective reaction at the free ketone group while ketal protecting groups remain intact — a discrimination not observed with Brønsted or other Lewis acid catalysts [1]. The method converts ketones and ketals into gem-dihydroperoxides in high yields, while epoxides yield hydroperoxides, all under mild conditions [2].

Evidence Dimension	Chemoselectivity: ketone vs. ketal discrimination in bifunctional substrates
Target Compound Data	MoO₂(acac)₂ (5 mol%): selective reaction at free ketone with ketal preservation; reactivity order epoxide > ketone > ketal; TBS and MOM protecting groups tolerated
Comparator Or Baseline	I₂, CAN, Re₂O₇: unable to achieve comparable chemoselectivity; these catalysts show either no discrimination or different reactivity patterns
Quantified Difference	Qualitative but definitive: MoO₂(acac)₂ achieves chemoselectivity described as 'distinct' and 'to date unattainable' relative to I₂, CAN, and Re₂O₇. High yields (typically >80%) of hydroperoxide products.
Conditions	Ethereal H₂O₂, CH₂Cl₂ or Et₂O solvent, room temperature, 5 mol% MoO₂(acac)₂, reaction times 2–24 h

Why This Matters

For synthetic chemistry groups producing organic peroxides, hydroperoxides, or tetraoxane/ozonide bioactive compounds, no alternative catalyst provides the same chemoselectivity profile, making MoO₂(acac)₂ a non-substitutable reagent for these specific transformations.

• [1] An, X., Zha, Q., & Wu, Y. (2019). Perhydrolysis in ethereal H₂O₂ mediated by MoO₂(acac)₂: distinct chemoselectivity between ketones, ketals, and epoxides. Organic Letters, 21(5), 1542–1546. doi:10.1021/acs.orglett.9b00425 View Source
• [2] An, X., Zha, Q., & Wu, Y. (2019). Abstract and key findings. Org. Lett., 21, 1542–1546. doi:10.1021/acs.orglett.9b00425 View Source

Racemization Barrier Benchmark for Chiral-at-Metal Design

Chiral cis-MoO₂(acac)₂ racemizes via four distinct topological pathways. The newly identified Dhimba–Muller–Lammertsma (DML) twist pathway exhibits a gas-phase barrier of 17.4 kcal/mol, dramatically lower than the textbook Ray–Dutt (27.5 kcal/mol) and Bailar (28.7 kcal/mol) twists, and below the Conte–Hippler twist (20.0 kcal/mol) [1]. Solvent effects further reduce the DML barrier to 16.3 kcal/mol (benzene), 16.2 kcal/mol (toluene), and 15.4 kcal/mol (chloroform), in excellent agreement with experimental NMR-derived activation energies of 16.9 kcal/mol (benzene), 15.1 kcal/mol (toluene), and 13.7 kcal/mol (chloroform) [1][2]. This comprehensive computational-experimental mapping of the racemization landscape makes MoO₂(acac)₂ the most thoroughly characterized MX₂(chel)₂ chiral-at-metal system, serving as a computational template for designing asymmetric catalysts where configurational stability at the metal center is the sole source of chirality [1].

Evidence Dimension	Racemization energy barrier (kcal/mol) for chiral-at-metal MX₂(chel)₂ complexes
Target Compound Data	cis-MoO₂(acac)₂ DML twist barrier: 17.4 kcal/mol (gas); 16.3 (benzene); 16.2 (toluene); 15.4 (CHCl₃). Experimental Ea: 16.9 (benzene); 15.1 (toluene); 13.7 (CHCl₃)
Comparator Or Baseline	Ray–Dutt twist: 27.5 kcal/mol. Bailar twist: 28.7 kcal/mol. Conte–Hippler twist: 20.0 kcal/mol (all gas-phase computed for the same complex). MoO₂(acnac)₂ analogs also investigated.
Quantified Difference	DML pathway is 10.1–11.3 kcal/mol lower than Ray–Dutt/Bailar, and 2.6 kcal/mol lower than Conte–Hippler twist (gas phase)
Conditions	ωB97X-D/6-31G(d)+LANL2DZ level of theory; solvent effects via PCM model at 25°C; experimental Ea by variable-temperature ¹H NMR spectroscopy

Why This Matters

For research groups developing chiral-at-metal asymmetric catalysts, MoO₂(acac)₂ provides a uniquely well-calibrated reference system with experimentally validated computational barriers, enabling rational ligand design to tune configurational stability.

• [1] Lammertsma, K., et al. (2022). Racemization pathway for MoO₂(acac)₂ favored over Ray–Dutt, Bailar, and Conte–Hippler twists. Inorganic Chemistry, 61(38), 14918–14923. doi:10.1021/acs.inorgchem.2c00824 View Source
• [2] Conte, M. & Hippler, M. (2016). Dynamic NMR and quantum-chemical study of the stereochemistry and stability of the chiral MoO₂(acac)₂ complex in solution. J. Phys. Chem. A, 120(34), 6677–6687. doi:10.1021/acs.jpca.6b03563 View Source

Retained Epoxidation Activity Over Multiple Recycling Cycles

When immobilized on expanded corn starch (ECS-MoO₂(acac)₂) and tested in the epoxidation of stillingia oil with tert-butyl hydroperoxide (TBHP), the heterogeneous catalyst could be recovered by simple filtration and reused for at least 5 consecutive cycles without significant degradation in catalytic activity [1]. This performance stability is attributed to the robust anchoring of the MoO₂(acac)₂ complex to the starch support, which prevents leaching of the active molybdenum species into the reaction medium — a common failure mode for physisorbed or weakly bound homogeneous catalysts. In contrast, soluble mineral acid catalysts (e.g., H₂SO₄) used industrially for vegetable oil epoxidation generate stoichiometric salt waste, require neutralization steps, and cannot be recycled [1].

Evidence Dimension	Catalyst reusability (number of cycles with retained activity) in vegetable oil epoxidation
Target Compound Data	ECS-MoO₂(acac)₂: ≥5 cycles without significant activity loss; easy recovery by filtration
Comparator Or Baseline	Homogeneous H₂SO₄: single use, generates neutralization waste. Soluble MoO₂(acac)₂ (non-immobilized): cannot be recovered and reused without additional separation steps.
Quantified Difference	≥5 cycles demonstrated vs. single use for mineral acid catalysts
Conditions	Stillingia oil epoxidation, TBHP oxidant, ECS-MoO₂(acac)₂ heterogeneous catalyst, filtration recovery between cycles

Why This Matters

For process chemists evaluating epoxidation catalyst procurement, the demonstrated recyclability of immobilized MoO₂(acac)₂ directly reduces catalyst cost per kilogram of epoxidized product and eliminates the waste disposal costs associated with mineral acid catalysts.

• [1] Hong, M., Yao, M.-Y., & Pan, H. (2015). An immobilized molybdenum acetylacetonate complex on expanded starch for the epoxidation of stillingia oil. RSC Advances, 5, 91558–91563. doi:10.1039/C5RA14581C View Source

Evidence-Backed Procurement Scenarios

Heterogeneous Catalyst for Alkene Metathesis

When preparing supported molybdenum catalysts for metathesis of 1-octene, 1-dodecene, or 1-tetradecene, MoO₂(acac)₂ is the preferred molybdenum source over MoO₃. The acetylacetonate precursor provides superior active-site dispersion on SBA-15 and MWW zeolites — evidenced by the absence of bulk MoO₃ XRD reflections at Mo loadings up to 6 wt% — and delivers up to 4× higher metathesis activity than MoO₃-derived catalysts on equivalent supports under identical conditions (40°C, neat olefin) [1]. For process development teams scaling from batch to continuous flow, the lower oligomerization byproduct formation (1–3% vs. ~12%) reduces downstream separation costs and improves overall product yield. Procurement recommendation: specify ≥97% purity with 1–2% acetylacetone stabilizer content to ensure consistent thermal spreading behavior during catalyst activation .

MOCVD Precursor for MoO₃ Thin Films

MoO₂(acac)₂ enables single-step deposition of crystalline α-MoO₃ films via MOCVD without post-annealing, a critical advantage when integrating MoO₃ layers into temperature-sensitive device stacks (e.g., flexible polymer substrates, pre-patterned electrodes). The deposition window of 400–560°C yields films with strong (010) preferred orientation and transparent appearance [1]. Alternative deposition techniques (PLD, sputtering, thermal evaporation) mandate a separate post-annealing step at ≥400°C to induce crystallinity, increasing thermal budget and processing time . For thin-film R&D and pilot production lines, sourcing MoO₂(acac)₂ with consistent batch-to-batch thermal decomposition behavior (TG/DTA-verified oxide formation at ~430°C) is essential for reproducible film quality [1].

Chemoselective Peroxide Synthesis

For medicinal chemistry and natural product synthesis groups constructing peroxide-containing bioactive scaffolds (tetraoxanes, ozonides, antimalarial candidates), MoO₂(acac)₂ is the only catalyst reported to achieve selective perhydrolysis of ketones and ketals in the presence of sensitive protecting groups (TBS, MOM) using ethereal H₂O₂ [1]. Competing Lewis acid catalysts (I₂, CAN, Re₂O₇) fail to discriminate between ketone and ketal functionalities, leading to complex product mixtures . The demonstrated reactivity order (epoxide > ketone > ketal) provides predictable chemoselectivity that guides synthetic route design. Procurement recommendation: catalyst-grade purity (≥97%) is sufficient; the compound should be stored under argon at 2–8°C to prevent hydrolytic degradation of the Mo=O bonds .

Chiral-at-Metal Catalyst Development

MoO₂(acac)₂ serves as the most thoroughly characterized MX₂(chel)₂ chiral-at-metal reference system, with experimentally validated racemization barriers across multiple solvents (benzene: 16.9 kcal/mol; toluene: 15.1 kcal/mol; chloroform: 13.7 kcal/mol) and computationally mapped racemization pathways [1]. Research groups developing enantioselective catalysts based on metal-centered chirality can use MoO₂(acac)₂ as a calibration standard for computational methods (DFT functional benchmarking) and for experimental validation of variable-temperature NMR kinetics . The well-defined racemization kinetics also make the compound suitable as an undergraduate/graduate inorganic chemistry teaching tool for demonstrating dynamic stereochemistry and topological analysis of fluxional coordination complexes [1]. For these applications, the racemic mixture (as commercially supplied) is appropriate; no enantiomeric enrichment is required.

• [1] Balcar, H., et al. (2009). Applied Catalysis A: General, 359(1–2), 129–135. doi:10.1016/j.apcata.2009.02.037; Balcar, H., et al. (2018). Beilstein J. Org. Chem., 14, 2931–2939. doi:10.3762/bjoc.14.272; Bouzidi, A., et al. (2006). J. Alloys Compd., 425, 312–317. doi:10.1016/j.jallcom.2006.01.055; Trzciński, M., et al. (2023). Sci. Rep., 13, 16738. doi:10.1038/s41598-023-43678-7; An, X., et al. (2019). Org. Lett., 21, 1542–1546. doi:10.1021/acs.orglett.9b00425; Lammertsma, K., et al. (2022). Inorg. Chem., 61(38), 14918–14923. doi:10.1021/acs.inorgchem.2c00824; Conte, M. & Hippler, M. (2016). J. Phys. Chem. A, 120(34), 6677–6687. doi:10.1021/acs.jpca.6b03563 View Source