Nitric Oxide Reduction to Ammonia by TiO2 Electrons in Colloid

Nitric Oxide Reduction to Ammonia by TiO2 Electrons in Colloid Solution via .... with TiO 2. V. Nahuel Montesinos , Natalia Quici , Hugo Destaillats ,...
0 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/JPCA

Nitric Oxide Reduction to Ammonia by TiO2 Electrons in Colloid Solution via Consecutive One-Electron Transfer Steps Sara Goldstein,† David Behar,† Tijana Rajh,‡ and Joseph Rabani*,† †

Institute of Chemistry and the Accelerator Laboratory, The Hebrew University of Jerusalem, Jerusalem 91904, Israel Center for Nanoscale Materials, Argonne National Laboratory, Argonne, Illinois 60439, United States

‡

ABSTRACT: The reaction mechanism of nitric oxide (NO) reduction by excess electrons on TiO2 nanoparticles (eTiO2−) has been studied under anaerobic conditions. TiO2 was loaded with 10−130 electrons per particle using γ-irradiation of acidic TiO2 colloid solutions containing 2-propanol. The study is based on time-resolved kinetics and reactants and products analysis. The reduction of NO by eTiO2− is interpreted in terms of competition between a reaction path leading to formation of NH3 and a path leading to N2O and N2. The proposed mechanism involves consecutive one-electron transfers of NO, and its reduction intermediates HNO, NH2O•, and NH2OH. The results show that eTiO2− does not reduce N2O and N2. Second-order rate constants of eTiO2− reactions with NO (740 ± 30 M−1 s−1) and NH2OH (270 ± 30 M−1 s−1) have been determined employing the rapid-mixing stopped-flow technique and that with HNO (>1.3 × 106 M−1 s−1) was derived from fitting the kinetic traces to the suggested reaction mechanism, which is discussed in detail.

■

INTRODUCTION TiO2 is a transition metal oxide with band gap in the UV region of 3.0, 3.2, and 3.3−3.4 eV for Rutile, Anatase, and Brookite, respectively.1−3 The fundamental mechanism of TiO2 nanocrystals mediated light absorption reactions is quite wellunderstood.1 When TiO2 absorbs UV light, conduction band electrons and valence band hole-pairs are first produced and become quickly localized at the nanocrystal surface as less mobile states. The spectroscopic features of the electrons and holes in TiO2 have been intensively studied.4−17 Both electrons and holes have broad optical spectrum expanding from the UV to the IR enabling in many cases to follow their reaction kinetics. The holes and electrons can oxidize and reduce many molecules, respectively, so that TiO2 mediates the conversion of photon energy into chemical reaction. The application of TiO2 nanoparticles for decontamination of wastewater from organic and inorganic hazardous metal ions has become a mature research field.14,18−20 Most photocatalytic works focus on detoxification via oxidation of the pollutants in aerated media. Much less is known about the reductive processes, although CO2 conversion to organic materials for energy storage21 and nitrate reduction22−30 have attracted many researchers. Nitrate concentration in surface and ground waters has been found to increase regularly in many agricultural lands, © 2015 American Chemical Society

mainly because of an intensive use of fertilizers. Reduction of nitrate to N2 and to the less desired ammonia involves the formation of several nitrogen intermediates with known physical and chemical properties, including NO32−, •NO2, NO22−, NO2−, NO, HNO, NH2O•, NH2OH, and N2O (Scheme 1). Relatively little research has been carried out on the timeresolved kinetics of TiO2 electrons (eTiO2−) reactions with nitrogen compounds, although most, if not all, the reduction steps of NO3− by eTiO2− are thermodynamically downhill processes.31,32 Gao et al. applied pulse radiolysis to produce eTiO2− and to study its reaction with nitrate and nitrite.22 Mohamed et al. applied photolysis to produce eTiO2− and to study its reaction with nitrate, suggesting that the reduction proceeds via multielectron transfer process.30 Reduction of nitrogen species by consecutive one-electron transfer reactions or by multielectron process may lead to the same end-products but with different reaction kinetics. For example, according to Scheme 1, reduction of nitrate to nitrite can proceed via Received: October 12, 2014 Revised: February 28, 2015 Published: March 2, 2015 2760

DOI: 10.1021/jp5102863 J. Phys. Chem. A 2015, 119, 2760−2769

Article

The Journal of Physical Chemistry A

monitored after 5 min incubation at 95 °C and subsequent cooling. Calibration curve was prepared with a standard solution of hydroxylamine hydrochloride. Ammonia was assayed using a modification of the Berthelot reaction based on indophenol formation via the reaction of phenol with alkaline hypochlorite.35 Briefly, phenol (6 g), and sodium nitroprusside (20 mg) were dissolved in 100 mL phosphatecitrate buffer at pH 12 (3 g Na3PO4·12H2O, 3 g Na3C6H6O7· 2H2O, 0.3 g EDTA), and 0.3% hypochlorite was prepared in 0.57 M NaOH. A two mL sample was mixed with 0.8 mL phenol-nitroprusside and 1.2 mL alkaline hypochlorite. The solutions were incubated for 15 min at 40 °C, and those containing TiO2 were centrifuged to separate the precipitated TiO2. The absorption was monitored at 635 nm, applying ε635 = 2.3 × 104 M−1 cm−1.35 Control experiments showed that the presence of TiO2 does not interfere with the analysis. NO-saturated solutions were prepared in gastight syringes, which were further diluted with Ar-saturated solutions using the syringe technique. NO-saturated solution contained less than 0.3% nitrite determined using the Griess reagent after removing NO by purging the solution with Ar. The concentration of NO was determined by employing a spectroscopic assay using aerated solutions of 1 mM ABTS 2− as a reductant (ε660(ABTS•−) = 12000 M−1 cm−1, 60% yield)36 or using oxygenated Griess reagent for determination of nitrite formed via autoxidation of NO. The reduction of Cu(neoc)2+ to Cu(neoc)2+ under anoxic conditions was employed for the determination of [eTiO2−] using ε445 = 7500 M−1 cm−1 obtained by reducing Cu(neoc)22+ with ascorbic acid in acidic solutions. Mixing of reagents under strict deaerated conditions was carried out using the syringe technique (i.e., transferring deaerated solution from one syringe to another through a Teflon stopcock and glass hose with water lubricated taper joints. This method introduced less than 1 μM O2 calculated by mixing NO-saturated solution with deaerated water and measuring nitrite concentration after removing the NO in a large number of control tests. In most cases, such a contamination of oxygen could be ignored. Another mixing method involved a flow system (SP230iw syringe pump, World Precision Instruments) with a substantial part of a HDPE hose. These two mixing techniques and the rapid-mixing stopped flow method gave similar results. Methods. Radiolysis. γ-Radiolysis experiments were carried out at room temperature using a 137Cs source (6.4 Gy min−1). Unless otherwise stated, TiO2 electrons were produced upon irradiation of Ar-saturated TiO2 colloidal suspension in 0.01 M HCl containing 1 M 2-propanol in 10 mL syringes with no gas phase. Kinetics. Stopped-flow kinetic measurements were carried out with the Bio SX-17MV Sequential Stopped-Flow from Applied Photophysics with a 1 cm optical path at 25 °C. All experiments were carried out while flushing the syringe cups with Ar. The syringes were flushed with Ar-saturated water before each experiment. Each value given is an average of at least three measurements, where eTiO2− in 0.01 M HCl was mixed with various aqueous substrates at a 1:1 volume ratio. The decay of eTiO2− was followed at 660 nm. Modeling of the experimental results was carried out using a noncommercial program developed at Brookhaven National Laboratories by Dr. H. A. Schwarz.

Scheme 1. Electron Balance in Nitrate Reduction to Ammonia

reactions 1−3 (first-order decay of eTiO2−) or via a two-electron transfer reaction (second-order decay of eTiO2−). NO is an important intermediate produced from nitrite and nitrate by one and three electron reduction steps, respectively. The present study concerns the reduction mechanism of NO by eTiO2−, demonstrating unequivocally that its reduction to NH3 proceeds via five consecutive one-electron transfer reactions.

■

EXPERIMENTAL SECTION Materials. All chemicals were of analytical grade and used as received. Water for preparation of the solutions was deionized and purified using a Milli-Q purification system. All chemicals were of analytical grade and were used as received: titanium tetrachloride, neocuproine (neoc), ascorbic acid, phenol, sodium nitroprusside, 8-hydroxyquinoline, hydroxylamine hydrochloride, and 2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonate) (ABTS2−) from Sigma-Aldrich, Ar 99.999%, N 2 99.999%, and N2O (99.9%) from Maxima, Israel. NO (Matheson Gas Products) was purified by passing it through a series of deaerated scrubbing bubblers containing two bubblers of 0.5 M NaOH, one of 0.1 M HCl and one of purified water in this order. Cu(neoc)22+ was freshly prepared by mixing equivalent amounts of neoc and CuCl2 in aqueous solutions. Colloidal stock TiO2 was prepared by slowly introducing TiCl4 under a stream of argon to vigorously stirred 0.1 M HCl, followed by dialysis of the resulting optically transparent colloidal solutions at pH 2.5. Subsequently, the solution was heated for 96 h at 55 °C to get the desired nanoparticle size.33 Analytical Methods. Generally, centrifugation was employed when precipitation of TiO2 occurs during analysis. Nitrite concentration was assayed by mixing equal volumes of the sample and the Griess reagent prepared in HCl. The absorption at 540 nm was read 15 min after mixing the sample with the reagent. Calibration curves were prepared using known concentrations of nitrite. Hydroxylamine was assayed using 8hydroxyquinoline method.34 Briefly, 0.5 mL of 1% 8hydroxyquinoline in ethanol and 0.5 mL of 1 M Na2CO3 were added to 1 mL sample. The absorption at 708 nm was 2761

DOI: 10.1021/jp5102863 J. Phys. Chem. A 2015, 119, 2760−2769

Article

The Journal of Physical Chemistry A

Figure 1. (A) XRD pattern of prepared TiO2 colloid sample (blue line), XRD of 64.39% anastase (An) and 35.61% brookite (Br) (red line), difference between 64.39% An and 35.61% Br and sample (gray line). Miller indexes of An and Br are specified for the most intensive peaks only. (B) TEM micrograph of as-prepared TiO2. (C) Size distribution of as-prepared TiO2 obtained from TEM micrograph shown in (B).

■

RESULTS

structure with polygonal geometry. The average short diameter was found to be 4.09 nm and the average longer diameter to be 5.57 nm. This is in close agreement with the weighted average diameter of 5.52 nm obtained using XRD. As XRD measurements average a much larger number of particles, 5.52 nm was used for the calculation of average particle surface area (95.7 nm2), particle volume (88.0 nm3), and average particle weight (3.4 × 10−19 g), assuming ellipsoid shape and density of anatase TiO2 of 3.9 g cm−3. Under our experimental conditions where we used 0.8−2 g/L TiO2, the number of TiO2 particles will consequently be (2.4−5.9) × 1018 particles/L.

Characterization of TiO2 Nanoparticles. The crystalline character of the TiO2 nanoparticles was investigated by XRD (Figure 1A). The sample exhibits diffraction patterns characteristic of 64.4% crystal phase of pure anatase (5.7 nm particle size) and 35.6% crystal phase of pure brookite (5.2 nm particle size). The weighted average diameter has been determined to be 5.52 nm. The size and size distribution of the TiO2 particles was also determined using transmission electron microscopy (TEM) measurements (Figure 1, panels B and C). The TEM measurements show that the particles have elongated faceted 2762

DOI: 10.1021/jp5102863 J. Phys. Chem. A 2015, 119, 2760−2769

Article

The Journal of Physical Chemistry A Optical Absorption of eTiO2− Produced by Radiolysis. In irradiated aqueous solutions, a mixture of reducing and oxidizing radicals is formed initially according to eq 17. The values in parentheses are the G-values, which represent the yields of the species in 10−7 M Gy1−.

50 electrons per particle where most kinetic tests have been carried out. An average value of ε660 = 550 M−1 cm−1 was used for determination of [eTiO2−] using the stopped-flow method. Reaction of eTiO2− with NO. The extent of eTiO2− oxidation by NO expressed as the concentration ratio of reacted electrons and NO, Δ[e TiO2 − ]/Δ[NO], depends on their initial concentration ratio, [eTiO2−]o/[NO]o. It increases upon increasing [eTiO2−]o/[NO]o, approaching a value of 5 at [eTiO2−]o/[NO]o > 12 (Figure 3A). Figure 3A represents experiments carried out using various mixing methods

H 2Ô → e−aq (2.6), •OH(2.7), H•(0.6), H3O + (2.6), H 2O2 (0.72)

(17)

When acid is present, reaction 18 converts the hydrated electrons into H atoms. In the presence of 1 M 2-propanol, reactions 19 and 20 produce (CH3)2C•OH radicals. TiO2 nanocrystals react with eaq− (reaction 21) and H• (reaction 22). The (CH3)2C•OH reacts with TiO2 according to reaction 23. Thus, the radiation induced reactive radical species end up as eTiO2− either by direct reaction with TiO2 or via the reducing intermediates. eaq − + H+ → H• •

(18)

OH + (CH3)2 CHOH → H 2O+(CH3)2 C•OH

(19)

H• + (CH3)2 CHOH → H 2 + (CH3)2 C•OH

(20)

eaq − + TiO2 → eTiO2−

(21)

H• + TiO2 → e TiO2− + H+

(22)

(CH3)2 C•OH + TiO2 → eTiO2− + (CH3)2 CO + H+ (23)

The overall result of the radiolysis of TiO2 in the presence of 1 M 2-propanol and absence of O2 is the accumulation of relatively stable excess of electrons on colloid TiO2 nanoparticles, which allows the study of both slow and fast reactions of these electrons with solutes.6 The absorption spectrum of eTiO2− has been previously reported using both photolysis and radiolysis generation methods.6,22,30,33,37 We found that the extinction coefficient decreases upon increasing the number of electrons per particle (Figure 2, inset), its value varies by ∼10% in the range of 10−

Figure 3. (A) Effect of [eTiO2−]o/[NO]o on Δ[eTiO2−]/Δ[NO] obtained using different mixing methods and [TiO2]. (○) Syringe technique, 0.2−0.67 g/L TiO2; (●) syringe technique 0.8−2.0 g/L TiO2; (□) flow mixing, 0.42 g/L TiO2; (▲) rapid-mixing stopped-flow 0.42 g/L TiO2; (white square within dark square) rapid-mixing stopped-flow 0.8 g/L TiO2; (★) rapid-mixing stopped-flow 1.25 g/L TiO2. The solid line represents the calculated ratio based on the suggested mechanism. (B) Dependence of Δ[eTiO2−]/Δ[NO] on [eTiO2−]o/[NO]o at various electron densities. Number of electrons per particle are (●) 8−9; (○) 10−13; (△) 30−36; (■) 40−48; (white square within dark square) 59−63; (★) 95−106; ◇ 113−128. The data was obtained using the rapid-mixing stopped-flow method. The solid line represents the calculated ratio based on the suggested mechanism.

Figure 2. Average extinction coefficient of eTiO2− obtained for 2.5 g/L TiO2 and [eTiO2−] = 0.1−1 mM determined via the reduction of Cu(neoc)22+. The variation is due to the dependence of the extinction coefficient on the average number of electrons per TiO2 particle observed by varying the γ radiation dose. This is shown in the inset for ε660 using 2.5 g/L TiO2 (●) and 0.84 g/L TiO2 (Δ). The number of TiO2 particles was derived using the average diameter 5.52 nm. 2763

DOI: 10.1021/jp5102863 J. Phys. Chem. A 2015, 119, 2760−2769

Article

The Journal of Physical Chemistry A including the syringe technique, the flow mixing, and the rapidmixing stopped flow. Although there is significant scatter of the results, Δ[eTiO2−]/Δ[NO] follows a similar systematic dependency on [eTiO2−]o/[NO]o at all TiO2 concentrations. When applying the syringe and the flow mixing techniques, the radiation produced eTiO2− in a narrow concentration range of (3.4−3.7) × 10−4 M. Therefore, the initial number of electrons per TiO2 particle is inversely proportional to [TiO2], implying that there is only little if any effect of the number of electrons per particle on Δ[eTiO2−]/Δ[NO]. On the other hand, the measurements of [eTiO2−] from the kinetic traces was carried out at various [eTiO2−]o (Figure 3B). Figure 3 (panels A and B) demonstrates that the profile of Δ[eTiO2−]/Δ[NO] versus [eTiO2−]o/[NO]o does not depend on [TiO2] as well as on the number of excess electrons per TiO2 particle. No detectable amounts of NH2OH could be observed at [eTiO2−]o/[NO]o > 2 as under such conditions a residual concentration of eTiO2− remains after completion of the NO reduction process. However, a small concentration of NH2OH was observed at [eTiO2−]o/[NO]o = 0.3 and 1.25, amounting to 1.7% and 3.4% of [eTiO2−]o, respectively. Since reduction of NO to NH2OH requires 3 electrons, 5% and 10% of eTiO2− reduced NO to NH2OH under these conditions. Accumulation of ammonia is observed under all experimental conditions. The fraction of NO reduced to NH3 increases upon increasing [eTiO2−]o/[NO]o, approaching a value of one at high [eTiO2−]o/[NO]o ratio (Figure 4), indicating that NH3 becomes the sole reduction product at a sufficiently high excess of electrons.

Figure 5. Typical kinetic traces for the reaction of 1.8 × 10−4 M eTiO2− with 12.5−132 μM NO in 5 mM HCl. The upper trace shows the stability of the electrons using the stopped-flow apparatus in the given time range.

produced under Ar showed 98% of the eTiO2− absorption spectrum 70 min after replacing Ar with N2. Similarly, irradiation of the above TiO2 under N2O instead of Ar produced a slightly lower [eTiO2−] within the same irradiation time. The small difference is apparently due to the O2 contaminant in the commercial N2O. Evidently, eTiO2− is inert toward both N2 and N2O on the timescale of 70 min. Figure 5 also shows typical kinetic traces observed when eTiO2− reacts with various concentrations of NO. The residual absorption in the presence of excess of eTiO2− shows that Δ[eTiO2−]/Δ[NO] increases upon increasing the ratio [eTiO2−]o/[NO]o in agreement with the dependency observed using other methods of mixing (Figure 3A). In the presence of sufficient excess of NO, the eTiO2− absorption decayed according to a first-order rate law, and kobs depends linearly on [NO]o, yielding a slope of (1.22 ± 0.03) × 103 M−1 s−1 (Figure 6). The partial decay of eTiO2− at [eTiO2−]o/[NO]o > 10 has been analyzed in terms of a pseudo-first-order reaction. Due to the multielectron reduction of NO, [eTiO2−]o cannot be considered constant, nevertheless the pseudo-first-order rate constant (kobs) has been evaluated for comparative purpose. Figure 7

Figure 4. Fraction of NO reduced to NH3 as a function of [eTiO2−]o/ [NO]o. Results were obtained using the flow mixing technique and 0.42 g/L TiO2 and 1.8 × 10−4 M eTiO2−in 5 mM HCl. The solid line represents the calculated ratio based on the suggested mechanism.

The commercial NO cylinder is contaminated with N2O, which is not removed during its purification. This prevents quantitative determination of N2O, originating from NO reaction with eTiO2−. Nevertheless, when [eTiO2−]o/[NO]o < 2, ca. 20% of NO yielded N2O, although the initial N2O level is almost 90% of the total N2O measured. Stopped-Flow Kinetic Study. The stability of eTiO2− using the stopped-flow apparatus is demonstrated in Figure 5 where a constant absorption is monitored at 660 nm during 200 s after mixing the colloidal blue suspension of eTiO2− with Ar-saturated water. The stability of the stored electrons was unaffected when water was saturated with either N2O or N2. This is in agreement with the experiments where the absorption of eTiO2−

Figure 6. Observed first order rate constant of the decay of the absorption at 660 nm when 1.8 ± 0.1 × 10−4 M eTiO2− reacted with NO in 5 mM HCl. 2764

DOI: 10.1021/jp5102863 J. Phys. Chem. A 2015, 119, 2760−2769

Article

The Journal of Physical Chemistry A shows that kobs increases linearly with [eTiO2−]o, suggesting that the reduction of NO is first-order in [eTiO2−].

Figure 9. Typical kinetic traces for the reaction of 1.3 × 10−4 M eTiO2− with various NH2OH in 5 mM HCl. Figure 7. Pseudo-first-order rate constant as a function of [eTiO2−]o obtained by the reaction of 6−30 μM NO with 130−260 μM eTiO2−.

constant depends linearly on [NH 2OH]o, yielding the bimolecular rate constant of 550 ± 10 M−1 s−1 (Figure 10).

Figure 8 shows that kobs/[ eTiO2−]o and kobs/[NO]o under limiting concentrations of NO and eTiO2−, respectively, do not

Figure 10. Observed first-order rate constant of the absorption decay at 660 nm in the reaction of eTiO2− with NH2OH in 5 mM HCl.

Figure 8. Effect of electron density on kobs/[eTiO2−]o and kobs/[NO]o demonstrated under limiting concentrations of NO (○) and limiting concentrations of eTiO2− (●), respectively.

Under limiting concentrations of NH2OH, the partial decay of eTiO2− obeyed first-order kinetics, resulting in a second-order rate constant of 270 ± 30 M−1 s−1. The 2-fold ratio between the second-order constants calculated under excess of NH2OH and excess of eTiO2−, respectively, results from the two-electron transfer stoichiometry, where the rate-determining step is one electron transfer to NH2OH.

depend on the excess electron density in agreement with Figure 3 (panels A and B). A plot of kobs versus [eTiO2−]o2 yields a curve, which does not pass through the origin when forced to be linear. This, together with the first-order-decay profiles observed under limiting concentrations of NO (Figure 5) or eTiO2− (not shown) as well as the lack of electron density effect on kobs/[eTiO2−]o (Figure 8) rule out second-order reaction in eTiO2−. TiO2 electrons are also capable of reducing NH2OH. Figure 9 shows typical kinetic traces observed when eTiO2− reacted with various concentrations of NH2OH. When 2.5 × 10−5 M NH2OH reacted with 1.3 × 10−4 M eTiO2−, the change in the absorbance at 660 nm was 0.025 (Figure 9). As NH2OH is reduced by two eTiO2−, one calculates that ε660 = 500 M−1cm−1 in good agreement with the value determined using the reduction of Cu(neoc)22+. In the presence of excess of NH2OH, the absorption of eTiO2− decayed via a first-order reaction, implying that the rate-determining step involves one electron. The observed pseudo first-order rate

■

DISCUSSION In the present work, the total concentration of TiO2 electrons, [eTiO2−]T, is much higher than the TiO2 nanocrystals’ concentration. If the reduction sequence of NO is confined to the surface of one given nanoparticle, the kinetics and, consequently, the nature of the end-products must depend only on the electron density at the surface of the given particle, namely on [eTiO2−]T/[TiO2]. The situation is different if NO and its reduction intermediates encounter a number of TiO2 particles before reacting with eTiO2−. In this case, the reaction rates are proportional to the frequency of encounters with TiO2 nanoparticles (i.e., [TiO2]). In adition, the probability for eTiO2− reaction is proportional to the electron density at the TiO2 surface (i.e., to the ratio [eTiO2−]T/[TiO2]). The overall reaction rate is thus proportional to [eTiO2−]T. Our results show 2765

DOI: 10.1021/jp5102863 J. Phys. Chem. A 2015, 119, 2760−2769

Article

The Journal of Physical Chemistry A that the profiles of Δ[eTiO2−]/Δ[NO] and [NH3]/[NO]o versus [eTiO2−]o/[NO]o (Figure 3 (panels A and B) and Figure 4) as well as the decay rates of eTiO2− (Figure 5) depend on [eTiO2−]T rather than on the number of electrons per nanoparticle. This shows that the reactions are not confined to a given nanocrystal but rather involve a number of TiO2 nanoparticles. Therefore, the kinetic analysis is in terms of [eTiO2−]T and not [TiO2] or [eTiO2−]T/[TiO2]. Scheme 1 summarizes the reaction paths by which NO is converted to NH3. NO is reduced predominantly to NH3 in the presence of sufficient excess of [eTiO2−]. Since both N2O and N2 are inert toward eTiO2−, the only path leading to NH3 is reactions 6, 11, 13 and 14 (see Scheme 1), namely, the reduction of NO to NH3 involves the formation of HNO, NH2O•, NH2OH, and •NH2OH− as intermediates where only NH2OH is sufficiently stable to enable the direct study of its reaction with eTiO2−. One-electron reduction of NO produces NO− (reaction 6), which becomes readily protonated (pKa = 11.438). HNO is further reduced by one electron (reaction 11) producing after protonation NH2O• (pKa = 12.639). Reaction 11 competes with reactions 7 and 8 (see Scheme 1), which eventually lead to the formation of N2O and NO2−.40 As nitrite reacts with eTiO2− much faster than NO (k4 ∼ 106 M−1 s−122), the reduction of nitrite regenerates one NO molecule so that the net reaction involves two electrons and two NO molecules producing one N2O molecule. NH2O• produced by reaction 11 reacts with an additional electron (reaction 13), yielding after protonation NH2OH in competition with its recombination forming N2 (reaction 12). The competing reactions 7 and 8 as well as reaction 12 contribute to the decrease in NH3 evolution at the lower [eTiO2−]o/[NO]o (see Scheme 1). The reduction of NH2OH to NH3 requires two electrons. The NH2OH system was studied separately, demonstrating that eTiO2− decays via first-order reaction under limiting concentrations of eTiO2−, where kobs increases linearly with [NH2OH]o as well as under limiting concentrations of [NH2OH]o, where kobs increases linearly with [eTiO2−]o. The derived bimolecular rate constant is twice higher in the presence of excess [NH2OH]. This is expected if the rate-determining step involves reaction of one NH2OH molecule with one eTiO2−, producing NH2OH− as a transient intermediate (reaction14a), which readily reacts with an additional eTiO2− (reaction 14b) or dismutates (reaction 14c). e TiO2− + NH 2OH → NH 2OH−

(14a)

e TiO2− + NH 2OH− + 2H+ → NH3 + H 2O

(14b)

2NH 2OH− + 2H+ → NH3 + NH 2OH + H 2O

(14c)

Figure 11. Decay of 150 μM eTiO2− in the presence of 10.5 μM NO. Simulation using k4 = 106 M−1 s−1, k6 = 740 M−1 s−1, k8 = 5.8 × 106 M−1 s−1, k9 = 5.4 × 109 M−1 s−1, k12 = 1.4 × 108 M−1 s−1, k14a = 270 M−1 s−1, k11/k8 = 0.22, and k13/k12 > 7 × 10−3.

Figure 12. Decay of 180 μM eTiO2− in the presence of 640 μM NO. Simulation using k4 = 106 M−1 s−1, k6 = 740 M−1 s−1, k8 = 5.8 × 106 M−1 s−1, k9 = 5.4 × 109 M−1 s−1, k12 = 1.4 × 108 M−1 s−1, k14a = 270 M−1 s−1, k11/k8 = 0.22, and k13/k12 > 7 × 10−3.

determination of k11/k8 = 0.22 and k13/k12 > 7 × 10−3 does not depend on the state of the related intermediates whether in the bulk or under equilibrium between adsorbed and bulk species. In the latter case, the apparent rate constants include the adsorption constants. The same kinetic parameters show excellent fits with the reactants and products analyses (Figure 3 (panels A and B) and Figure 4). Although the mechanism is complex, the degrees of freedom in choosing the rate constants are limited considering that the same rate constants must fit to a wide range of conditions. The competition between reactions 8 and 11 (see Scheme 1) is most significant for both kinetics and reactants and products analyses (Figure 3 (panels A and B) and Figure 4). Ignoring reaction 8 makes it impossible to fit the experimental data with any computed curve in the full range of conditions. The requirement k13/k12 > 7 × 10−3 implies that the path leading to N2 is minor. Because of experimental difficulties in determining the yields of N2O and N2, the exact ratio of k13/ k12 could not be derived. The values k7 = 8 × 106, k8 = 5.8 × 106, and k9 = 5.4 × 109 M−1 s−1 have been measured in TiO2 free homogeneous solutions.40 Since k7 and k8 are comparable, the reaction scheme can be somewhat simplified by ignoring reaction 7. The slope of the line in Figure 6 (i.e., 1.22 ± 0.03 × 103 M−1 −1 s ) is higher than the value of k6, since under the given experimental conditions NO is reduced by more than one

If reaction 14b and/or 14c are considerably faster than 14a, two electrons are consumed and the pseudo-first-order rate constant is kobs = 2k14a[NH2OH]o. On the other hand, in the presence of excess eTiO2−, the observed rate constant is determined by the decay rate of NH2OH and kobs = k14a[eTiO2−]o. Reactions 14b and 14c are kinetically undistinguishable as only eTiO2− time-dependent absorbance was followed. Simulations of the complex mechanism presented in Scheme 1 (Figures 11 and 12) demonstrates excellent fits to kinetic data for all tested [eTiO2−]o/[NO]o ratios using k4 = 106 M−1 s−1,22 k6 = 740 ± 30 M−1 s−1, k11/k8 = 0.22 ± 0.02, k13/k12 > 7 × 10−3, k11/k13 > 10, and k14a = 270 ± 30 M−1 s−1. Note that the 2766

DOI: 10.1021/jp5102863 J. Phys. Chem. A 2015, 119, 2760−2769

Article

The Journal of Physical Chemistry A

decay time profile of [eTiO2−] (Figures 11 and 12) obeys a firstorder rate law. Simultaneous multielectron transfer would have resulted in a higher reaction order and the decay in the presence of excess NO would remarkably deviate from a firstorder rate law. Therefore, the one-electron reduction product of NO (i.e., HNO) must be produced as an intermediate. HNO reduction by eTiO2− (reaction 11) competes with reactions 7 and 8 (see Scheme 1), which do not involve electron transfer. As simultaneous multielectron transfer must involve only one TiO2 nanoparticle, the number of electrons per particle rather than the average [eTiO2−] in the solution is expected to determine the reaction rate. Therefore, in the case of simultaneous multielectron transfer, the competition between reactions 7 and 8 and reaction 11 must strongly depend on the number of electrons per nanoparticle (see Scheme 1). Figure 3 (panels A and B) shows no systematic dependecy of Δ[eTiO2−]/ Δ[NO] versus [eTiO2−]0/[NO]0 on [TiO2] when nearly constant concentration of eTiO2− is mixed with various [NO] or when the electron density at the nanoparticles’ surface is varied, ruling out the possibility that the reduction of HNO involves more than one electron. For example, if HNO undergoes a two-electron reaction, increasing the electron density by a factor of 10 would have increased the reaction rate by a factor of 100. Therefore, one must conclude that NH2O• is actually produced as an intermediate. Similar argumentation rules out the transfer of more than one electron to NH2O•. The observed accumulation of NH2OH under certain conditions is expected only if NH2O• reacts with only one electron. The study of NH2OH reduction to NH3 clearly showes that it involves consecutive two electrons as discussed above. Thus, simultaneous multielectron reactions with NO and any of its reduction intermediates are ruled out under the conditions of this work. The reaction rate constants of NO and its reduction intermediates, which are involved in competing reactions, are far below the diffusion controlled rate. These reactions involve electrons from different TiO 2 nanoparticles and obey homogeneous reaction kinetics. Mohamed et al.30 studied the multielectron reduction of nitrate by eTiO2− using anatase TiO2 particles of 1−3 nm average diameter loaded with an average of 6−9 electrons and suggested that NH3 is produced according to reactions 1−6 followed by 7, 10, 15, and 16 (see Scheme 1). However, under the conditions of the present work, N2 and N2O are inert toward eTiO2−. The difference, if real, may be attributed to the different types of TiO2, particularly the number of electrons per particle and the particle size. This is also expressed in the spectral differences. Trapped electrons show a peak in the red region (750 nm), while the absorption of mobile electrons increases steadily in the visible and infrared regions up to at least 2500 nm.13 The absorption of the trapped electrons may vary depending on the nature of the TiO2 crystal defects. The spectrum observed by Mohammed et al.30 shows a broad absorption with a peak changing from about 550 to 650 nm at low and high [eTiO2−], respectively, as the deeper traps are being filled first. In the present work, the absorption steadily increases toward the IR region, approaching a plateau at 900 nm (Figure 2), suggesting superposition of the absorptions of trapped and mobile electrons. In the earlier work,30 freshly prepared TiO2 was used with a relatively high number of defects, while in the present work, the TiO2 suspension was heated for 96 h at 55 °C and further aged for many months in order to minimize the defects. Another difference is the number of electrons per particle of up to 130 in the present work

electron. It is also significantly higher than the slope of the line in Figure 7 (i.e., 350 ± 20 M−1 s−1), which is due to the relatively small excess of [eTiO2−] over [NO] given that NO is reduced by almost five electrons (i.e., kobs/[eTiO2−]o is an underestimation). Both experiments and simulations show that despite the complex nature of the mechanism, the decay of eTiO2− in the presence of NO obeys quite closely a first-order rate law due to reaction 6 being the initial and rate-determining step. Except for NH2OH, which reacts with eTiO2− somewhat slower than NO (2k14a = 550 M−1 s−1 compared to k6 = 730 M−1 s−1), other eTiO2− reactions are considerably faster than reaction 6 (see Scheme 1). The difference between 2k14a and k6 is not sufficiently large to cause significant deviation from the firstorder rate law. The above mechanism is in good agreement with the experimental observations presented in Figure 3 (panels A and B) and Figure 4 (i.e., the solid lines representing the numerical solutions of the relevant kinetic differential equations fit quite well to the experimental data), demonstrating that at low ratios of [eTiO2−]o/[NO]o the path leading to N2O prevails and Δ[eTiO2−]/Δ[NO] approaches the value one. When the ratio [eTiO2−]o/[NO]o increases, the formation of NH3 is enhanced, approaching the initial concentration of NO at [eTiO2−]o/ [NO]o > 12 (Figure 4). We observed low yields of NH2OH only under the conditions where the main path of reaction sequence does not lead to NH3. This is not surprising in view of the comparable reactivity of NO and NH2OH toward eTiO2−. The small yields of NH2OH are in agreement with the computed values (i.e., [NH2OH]/[eTiO2−]o at 0.3 and 1.25 [eTiO2−]o/[NO]o are 0.017 and 0.059, respectively), and the respective experimental values are 0.017 and 0.034. In the presence of a large excess eTiO2−, both NO and NH2OH are completely reduced. Concecutive One-Electron Versus Multielectron Transfer Paths. The proposed mechanism assumes that all electron transfer reactions in this system are first order in [eTiO2−]. In view of the large number of electrons per nanoparticle, simultaneous multielectron transfer to NO and its reduction intermediates has to be considered. Reaction of more than one electron at a time must involve only one TiO2 nanoparticle, and the kinetics are determined by the numbers of electrons and reactant molecules at the nanoparticle domain, which affect both the slower (rate-determining) reactions as well as the eTiO2− reactions with intermediates. If the intermediates are involved in competing reactions, the local concentrations may affect the competition and hence the different end products. Figure 3 shows that the Δ[eTiO2−]/ Δ[NO] versus [eTiO2−]0/[NO]0 profile at different [TiO2] and nearly constant number of electrons per particle does not depend on [TiO2]. If NO or its reduction products are adsorbed to the TiO2, increasing [TiO2] at constant [eTiO2−]o is expected to slow the reaction rate due to the dilution of the reactants in the nanovolume. This is correct for both oneelectron transfer and simultaneous multielectron transfer processes. The lack of a systematic effect of [TiO2] suggests that adsorption is not important in this system. Similarly, Figure 3B shows no systematic effect of the number of electrons per particle on the above profile, ruling out simultaneous multielectron transfer. Specifically, NO must react with one electron, otherwise reaction 6 would not have been first order in [eTiO2−], which is in contradiction with the straight line in Figure 7 and the horizontal lines in Figure 8. In addition, the 2767

DOI: 10.1021/jp5102863 J. Phys. Chem. A 2015, 119, 2760−2769

Article

The Journal of Physical Chemistry A compared to 6−9 in the earlier work,30 which are more suitable for simultaneous transfer of more than one electron, although this is not supported by the experiments. In conclusion, the reduction of NO to NH3 by excess electrons on TiO2 proceeds via consecutive one- electron transfer reactions.

■

Reactive Species in Photoexcited Nanocrystalline TiO2 Films by WideWavelength-Range (400−2500 Nm) Transient Absorption Spectroscopy. J. Phys. Chem. B 2004, 108, 3817−3823. (14) Arbour, C.; Sharma, D. K.; Langford, C. H. Picosecond Flash Spectroscopy of Titania Colloids with Adsorbed Dyes. J. Phys. Chem. 1990, 94, 331−335. (15) Skinner, D. E.; Colombo, D. P., Jr.; Cavaleri, J. J.; Bowman, R. M. Femtosecond Investigation of Electron Trapping in Semiconductor Nanoclusters. J. Phys. Chem. 1995, 99, 7853−7856. (16) Colombo, J.; Philip, D.; Bowman, R. M. Does Interfacial Charge Transfer Compete with Charge Carrier Recombination? A Femtosecond Diffuse Reflectance Investigation of TiO2 Nanoparticles. J. Phys. Chem. 1996, 100, 18445−18449. (17) Tamaki, Y.; Furube, A.; Murai, M.; Hara, K.; Katoh, R.; Tachiya, M. Dynamics of Efficient Electron−Hole Separation in TiO 2 Nanoparticles Revealed by Femtosecond Transient Absorption Spectroscopy under the Weak-Excitation Condition. Phys. Chem. Chem. Phys. 2007, 9, 1453−1460. (18) Lang, X.; Ma, W.; Chen, C. Selective Aerobic Oxidation Mediated by TiO2 Photocatalysis. Acc. Chem. Res. 2014, 47, 355−363. (19) Hoffmann, M. R.; Martin, S. T.; Choi, W.; Bahnemann, D. W. Environmental Applications of Semiconductor Photocatalysis. Chem. Rev. 1995, 95, 69−96. (20) Mohamed, H. H.; Bahnemann, D. W. The Role of Electron Transfer in Photocatalysis: Fact and Fictions. Appl. Catal., B 2012, 128, 91−104. (21) Roy, S. C.; Varghese, O. K.; Paulose, M.; Grimes, C. A. Toward Solar Fuels: Photocatalytic Conversion of Carbon Dioxide to Hydrocarbons. ACS Nano 2010, 4, 1259−1278. (22) Gao, R. M.; Safrany, A.; Rabani, J. Reactions of TiO2 Excess Electron in Nanocrystallite Aqueous Solutions Studied in Pulse and Gamma-Radiolytic Systems. Radiat. Phys. Chem. 2003, 67, 25−39. (23) Kudo, A.; Domen, K.; Maruya, K.-I.; Onishi, T. Photocatalytic Reduction of NO3− to Form NH3 over Pt-TiO2. Chem. Lett. 1987, 1019−1022. (24) Gao, W.; Jin, R.; Chen, J.; Guan, X.; Zung, H.; Zhang, F.; Guan, N. Titania-Supported Bimetalic Catalysts for Photocatalytic Reduction of Nitrate. Catal. Today 2004, 90, 331−336. (25) Rengaraj, S.; Li, X. Z. Enhanced Photocatalytic Reduction Reaction over Bi3+−TiO2 Nanoparticles in Presence of Formic Acid as a Hole Scavenger. Chemosphere 2007, 66, 930−938. (26) Ergas, S. J.; Rheinheimer, D. E. Drinking Water Denitrification Using a Membrane Bioreactor. Water Res. 2004, 38, 3225−3232. (27) Zhang, F.; Pi, Y.; Cui, J.; Yang, Y.; Zhang, X.; Guan, N. Unexpected Selective Photocatalytic Reduction of Nitrite to Nitrogen on Silver-Doped Titanium Dioxide. J. Phys. Chem. C 2007, 111, 3756− 3761. (28) Kominami, H.; Nakaseko, T.; Shimada, Y.; Furusho, A.; Inoue, H.; Murakami, S.-y.; Kera, Y.; Ohtani, B. Selective Photocatalytic Reduction of Nitrate to Nitrogen Molecules in an Aqueous Suspension of Metal-Loaded Titanium(Iv) Oxide Particles. Chem. Commun. 2005, 2933−2935. (29) Kominami, H.; Furusho, A.; Murakami, S.-y.; Inoue, H.; Kera, Y.; Ohtani, B. Effective Photocatalytic Reduction of Nitrate to Ammonia in an Aqueous Suspension of Metal-Loaded Titanium(IV) Oxide Particles in the Presence of Oxalic Acid. Catal. Lett. 2001, 76, 31−34. (30) Mohamed, H. H.; Mendive, C. B.; Dillert, R.; Bahnemann, D. W. Kinetic and Mechanistic Investigations of Multielectron Transfer Reactions Induced by Stored Electrons in TiO2 Nanoparticles: A Stopped Flow Study. J. Phys. Chem. A 2011, 115, 2139−2147. (31) Dimitrijevic, N. M.; Savic, D.; Micic, O. I.; Nozik, A. J. Interfacial Electron-Transfer Equilibria and Flatband Potentials of aFerric Oxide and Titanium Dioxide Colloids Studied by Pulse Radiolysis. J. Phys. Chem. 1984, 88, 4278−83. (32) Lide, D. R. Handbook of Chemistry and Physics, 73rd ed.; CRC Press: London, 1992−1993.

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Tel: 972-2-6585292. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was generously supported by the US-Israel BSF under Contract no. 2012158. We acknowledge with thanks the invaluable assistance by Dr. Inna Popov from the Center for Nanoscience and Nanotechnology, at the Hebrew University of Jerusalem.

■

REFERENCES

(1) Chang, L.; Mohamed, H. H.; Dillert, R.; Bahnemann, D. Kinetics and Mechanisms of Charge Transfer Processes in Photocatalytic Systems: A Review. J. Photochem. Photobiol., C 2012, 13, 263−276. (2) Kim, D.-H.; Kim, W.-S.; Kim, S.; Hong, S.-H. Brookite TiO2 Thin Film Epitaxially Grown on (110) YSZ Substrate by Atomic Layer Deposition. ACS Appl. Mater. Interfaces 2014, 6, 11817−11822. (3) Magne, C.; Cassaignon, S.; Lancel, G.; Pauport, T. Brookite TiO2 Nanoparticle Films for Dye-Sensitized Solar Cells. ChemPhysChem 2011, 12, 2461−2467. (4) Rothenberger, G.; Moser, J.; Gratzel, M.; Serpone, N.; Sharma, D. K. Charge Carrier Trapping and Recombination Dynamics in Small Semiconductor Particles. J. Am. Chem. Soc. 1985, 107, 8059−8054. (5) Serpone, N.; Lawless, D.; Khairutdinov, R.; Pelizzetti, E. Subnanosecond Relaxation Dynamics in TiO2 Colloidal Sols (Particle Sizes Rp = 1.0−13.4 nm). Relevance to Heterogeneous Photocatalysis. J. Phys. Chem. 1995, 99, 16655−16661. (6) Safrany, A.; Gao, R. M.; Rabani, J. Optical Properties and Reactions of Radiation Induced TiO2 Electrons in Aqueous Colloid Solutions. J. Phys. Chem. B 2000, 104, 5848−5853. (7) Bahnemann, D.; Henglein, A.; Lilie, J.; Spanhel, L. Flash Photolysis Observation of the Absorption Spectra of Trapped Positive Holes and Electrons in Colloidal Titanium Dioxide. J. Phys. Chem. 1984, 88, 709−711. (8) Kolle, U.; Moser, J.; Graetzel, M. Dynamics of Interfacial ChargeTransfer Reactions in Semiconductor Dispersions. Reduction of Cobaltoceniumdicarboxylate in Colloidal TiO2. Inorg. Chem. 1985, 24, 2253−2258. (9) Kamat, P. V.; Bedja, I.; Hotchandani, S. Photoinduced Charge Transfer between Carbon and Semiconductor Clusters. One-Electron Reduction of Ca in Colloidal TiO2 Semiconductor Suspensions. J. Phys. Chem. B 1994, 98, 9137−9142. (10) O’Regan, B.; Graetzel, M.; Fitzmaurice, D. Optical Electrochemistry. 2. Real-Time Spectroscopy of Conduction Band Electrons in a Metal Oxide Semiconductor Electrode. J. Phys. Chem. 1991, 95, 10525−10528. (11) O’Regan, B.; Graetzel, M.; Fitzmaurice, D. Optical Electrochemistry I. Steady-State Spectroscopy of Conduction-Band Electrons in a Metal Oxide Semiconductor Electrode. Chem. Phys. Lett. 1991, 183, 89−93. (12) Tamaki, Y.; Hara, K.; Katoh, R.; Tachiya, M.; Furube, A. Femtosecond Visible-to-Ir Spectroscopy of TiO2 Nanocrystalline Films: Elucidation of the Electron Mobility before Deep Trapping. J. Phys. Chem. C 2009, 113, 11741−11746. (13) Yoshihara, T.; Katoh, R.; Furube, A.; Tamaki, Y.; Murai, M.; Hara, K.; Murata, S.; Arakawa, H.; Tachiya, M. Identification of 2768

DOI: 10.1021/jp5102863 J. Phys. Chem. A 2015, 119, 2760−2769

Article

The Journal of Physical Chemistry A (33) Gao, R. M.; Safrany, A.; Rabani, J. Fundamental Reactions in TiO2 Nanocrystallite Aqueous Solutions Studied by Pulse Radiolysis. Radiat. Phys. Chem. 2002, 65, 599−609. (34) Palling, J. D.; Hollocher, T. C. Basis and Optimization of the Indooxine Method to Determine Oximes. Mikrochim. Acta 1985, II, 137−144. (35) D, S. Determination of Ammonia and Kjeldahl Nitrogen by Indophenol Method. Water Res. 1976, 10, 31−36. (36) Goldstein, S.; Czapski, G. Kinetics of Nitric-Oxide Autoxidation in Aqueous-Solution in the Absence and Presence of Various Reductants: The Nature of the Oxidizing Intermediates. J. Am. Chem. Soc. 1995, 117, 12078−12084. (37) Mohamed, H. H.; Dillert, R.; Bahnemann, D. W. Reaction Dynamics of the Transfer of Stored Electrons on TiO2 Nanoparticles: A Stopped Flow Study. J. Photochem. Photobiol., A 2011, 217, 271− 274. (38) Shafirovich, V.; Lymar, S. V. Nitroxyl and Its Anion in Aqueous Solutions: Spin States, Protic Equilibria, and Reactivities toward Oxygen and Nitric Oxide. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 7340− 7345. (39) Lind, J.; Merenyi, G. Kinetic and Thermodynamic Properties of the Aminoxyl (NH2O•) Radical. J. Phys. Chem. A 2006, 110, 192−197. (40) Lymar, S. V.; Shafirovich, V.; Poskrebyshev, G. A. One-Electron Reduction of Aqueous Nitric Oxide: A Mechanistic Revision. Inorg. Chem. 2005, 44, 5212−5221.

2769

DOI: 10.1021/jp5102863 J. Phys. Chem. A 2015, 119, 2760−2769