of 8

Revealing the Nature of Trapping Sites in Nanocrystalline Titanium Dioxide by Selective Surface Modification †

25 views
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Share
Description
Revealing the Nature of Trapping Sites in Nanocrystalline Titanium Dioxide by Selective Surface Modification †
Transcript
  Revealing the Nature of Trapping Sites in Nanocrystalline Titanium Dioxide by SelectiveSurface Modification † Nada M. Dimitrijevic,* Zoran V. Saponjic, David M. Bartels, Marion C. Thurnauer, David M. Tiede, and Tijana Rajh* Chemistry Di V  ision, Argonne National Laboratory, Argonne, Illinois 60439 Recei V  ed: January 10, 2003; In Final Form: April 3, 2003 Excess electrons in nanocrystalline TiO 2  were studied in bare and dopamine-capped TiO 2  nanoparticles byelectron-beam pulse radiolysis. Reaction of hydrated electrons with dopamine-capped TiO 2  nanoparticles wasfound to be at the diffusion-controlled limit,  k   )  1  ×  10 11 M - 1 s - 1 , while the reaction with 1-hydroxy-1-methylethyl radicals, (CH 3 ) 2 C ˙ OH, was 2 orders of magnitude slower,  k   )  4  ×  10 8 M - 1 s - 1 . The reactionsresult in injection of electrons into the conduction band of TiO 2  nanoparticles. Optical absorption spectra of injected excess electrons in dopamine-capped nanoparticles display monotonic featureless wavelengthdependence up to 1800 nm. In contrast, bare particles have shown two preferential optical transitions withenergies in the visible region (  λ max ) 670 nm and  λ max ) 900 nm). Flat band potential of dopamine-cappedTiO 2  nanoparticles was shifted by 100 mV to more negative values. The strong coupling of dopamine tosurface Ti atoms was also found to improve the separation of photogenerated charges. This was demonstratedby the enhanced efficiency of photogenerated electrons in reducing silver cations to metallic silver in systemslinked via a dopamine bridge, compared to the same systems linked through carboxyl groups. Introduction Control of surface chemistry, particle shape, and particleorganization of nanocrystalline materials has gained considerableattention primarily because of their unique chemistry (enhancedchemical reactivity, specific syntheses, enhanced catalyticactivity, hardening of materials, specific bindings, etc.). Titaniumdioxide is the photocatalytic material that has been studied mostextensively over the past 10 years because it is an inexpensive,nontoxic, and photostable material. 1 Nanoparticulate TiO 2  hasbeen studied as a light-harvesting material for potential use inphotocatalytic removal of hazardous industrial byproducts 2 - 4 or in nanocrystalline solar cells. 5 - 9 The mechanism of semi-conductor-assisted photocatalysis is based on the principle thatparticulate semiconductors behave as miniature photoelectro-chemical cells. Although TiO 2  is very effective from an energeticpoint of view, it is a relatively inefficient photocatalyst. Themain energy loss is due to the recombination of chargesgenerated upon excitation of TiO 2 , which is manifested as therelatively low efficiency of long-lived charge separation. Inaddition, because of its large band gap (  E  g  )  3.2 eV), TiO 2 absorbs less than 5% of the available solar light photons. Hence,the main research interests concerning application of nanoc-rystalline TiO 2  are focused on improvement of the separationof charges, as well as the efficiency for visible light absorption.Two approaches, based on surface modification of TiO 2  withdye molecules or multifunctional ligands, have been pursued.Dye sensitization of wide band gap semiconductors via photo-induced interfacial electron transfer from excited dyes 5 - 9 hasbeen a topic of continuing interest because of the extension of the semiconductor response into the solar spectrum due to thedye absorption. The second approach involves adjusting theelectronic properties of nanocrystalline particles by adsorptionof electron-donating bidentate enediol ligands. 10 - 13 This resultsin instantaneous long-range charge separation between organicligands covalently linked to the surface atoms and the conductionband of semiconductor nanoparticles. This approach also extendssemiconductor response into the solar spectrum due to theelectronic coupling of   π   states of ligands with the conductionband states of semiconductor nanoparticles.Enhanced coupling of multifunctional ligands results also inpassivation of the surface states. 13 X-ray absorption spectro-scopic (XAS) studies have shown that the surface states in TiO 2 nanocrystallites are actually coordinately unsaturated Ti sitesthat are formed upon surface reconstruction of nanoparticles. 14 Electron paramagnetic resonance (EPR) studies have shown thatthe majority of photogenerated electrons are trapped on theseundercoordinated surface sites. 13,15 However, when the surfaceTi atoms in nanoparticles are coupled with enediol ligands, XASstudies indicate restructuring of the undercoordinated surfacesites into their optimal octahedral structure. Concomitantly, EPRspectra indicate removal of the surface undercoordinated sites.Illumination of nanoparticles surface-modified with enediolligands resulted in the appearance of two very strong and narrowaxially symmetric EPR signals in the  g  <  2 region [Ti(III)character] well resolved in the parallel orientation characteristicof electrons in an anatase lattice environment.In this paper we describe the study of the effect of enediolmodification of surface Ti atoms on the electronic propertiesof excess electrons injected by electron-beam radiolysis. Wehave used an electron pulse radiolysis technique to selectivelyinject electrons in dopamine-capped TiO 2  and study their opticalproperties. This technique enables examination of excesselectrons without interference of positive holes present inphotoexcitation experiments. Comparison of the optical absorp-tion spectra in bare and capped TiO 2  leads to understanding of the nature of surface trapping sites and reports on the ability of enediol ligands to repair the coordination of surface Ti atomsto a bulklike structure isoenergetic with the conduction band † Part of the special issue “Arnim Henglein Festschrift”. 7368  J. Phys. Chem. B  2003,  107,  7368 - 737510.1021/jp034064i CCC: $25.00 © 2003 American Chemical SocietyPublished on Web 06/04/2003  of TiO 2  nanoparticles. Additionally, ligation of the surface Tiatoms with the enediol ligands can change reduction propertiesof charge carriers in small-particle TiO 2 . We have determinedthe flat band potential of the dopamine-capped TiO 2  in orderto address the reducing power of these particles. The effect of strong coupling of dopamine modifier on the efficiency of charge separation was monitored by the deposition of metallicsilver. Experimental Section All chemicals used were of highest commercial purity andwere used without further purification. ASTM type I water orD 2 O (for transient absorption in near-IR) was used. Thepreparation and characterization of dopamine-modified TiO 2  isdescribed elsewhere 13 and only outlined here. The seeds of colloidal TiO 2  were prepared by dropwise addition of titanium-(IV) chloride to cooled water with an apparatus developed forcontrolling temperature and rate of mixing reaction compo-nents. 11,15 Slow growth of the particles was achieved by dialysisat 4  ° C against water until the pH of the solution reached 3.5,when the growth of crystalline particle TiO 2  was complete.Capping of TiO 2  particles by surface modification that resultedin the formation of a charge-transfer complex was achieved byaddition of surface-active ligands at concentrations required tocover all surface sites. The dopamine-capped TiO 2  colloids(TiO 2  /DA), onset of absorption  λ ≈ 760 nm, were then dialyzedagainst H 2 O or D 2 O, pH 3.5. The pH was adjusted with HClO 4 .The titania particles were 42.8 ( 3.5 Å in diameter. Addition-ally, TiO 2  particles with partial dopamine coverage of surfaceTi centers were prepared by the addition of appropriate amountsof dopamine into TiO 2  colloidal suspensions prepared in D 2 O.The modification of TiO 2  particles with carboxyethyl-   -cyclodextrin, CE-   -CD, was obtained by mixing solutions of the two components in desired proportions. The conjugation of CE-   -CD onto TiO 2  /DA particles via carboxyl groups oncyclodextrin and amino groups on dopamine was performed bya condensation reaction 16 via the intermediate  N  -hydroxysuc-cinimide ester modified and optimized for linking to TiO 2 nanoparticles. Typically, 20 mg of   O -(  N  -succinimidyl)-  N  ,  N  ,  N  ,  N  -tetramethyluroniumtetrafluoroborate (TSU) and 30  µ L of   N  ,  N  -diiodopropylethylamine were consecutively added into 2 mLof 0.1 mM CE-   -CD. The solution was dialyzed against water(molecular weight cutoff 500) in order to remove unreactedTSU. After dialysis, 0.4 mg of dopamine was added, and thesolution was incubated for 24 h in the dark. After incubation,0.2 mL of solution was added to 1 mL of 10 mM TiO 2  (molarconcentration). The solution of TiO 2  /DA/CE-   -CD was dialyzedagainst water, pH 3.5 (3500 MW cutoff), in order to removenonbound CE-   -CD. All dialysis steps were performed undernitrogen atmosphere. The coverage of the surface by dopaminewas 5.7%, and each TiO 2  particle was linked to approximatelytwo molecules of CE-   -CD. Apparatus.  Electron pulse radiolysis experiments wereperformed at the Argonne 20 MeV linear accelerator. Twodifferent pulse radiolysis setups were used, one for the short-time resolution (up to 100  µ s) and another for longer times (from1 ms to 10 s). In both cases transient species were detectedoptically.For short-time experiments, photomultiplier (HamamatsuR-1913), interference filters (bandwidth 50 nm), and silicon orInGaAs photodiodes (EG&G FND100 and germanium powerdevice GAP520, respectively) were used as optical detectors.The silicon photodiode was used for measurements of the rateconstant of hydrated electrons with TiO 2  /DA particles in H 2 Osolution at 650 nm. The InGaAs photodiode was used formeasurements of the spectra of excess electrons in nanoparticlesin D 2 O solutions in the wavelength region up to 1800 nm. Theelectron pulses were collinear with the analyzing light beambut in the opposite direction.A commercial Olis RSM 1000 rapid-scanning monochromatorsystem was used for long-time experiments. It consists of asubtractive double monochromator ( 1  /  4  m, f/4.4) with a rotatingsector slit at the intermediate focal plane to scan the wavelength,a beam splitter, a pair of photomultipliers, and 14-bit 1 MHzdigitizers to acquire  I  t   and  I  0  signals. A 250-nm-wide spectrumis acquired every millisecond. For these experiments, electronsentered the cell perpendicular to the analyzing light. Details aregiven elsewhere. 17 The dose in the cell was measured by use of aerated solutionsof thiocyanate and hexacyanoferrate(II). 18,19 For dosimetry inlong-time experiments the “super Fricke” solution was used. 18 The absorbed dose in a sample was varied by changing theduration of electron pulses and ranged from 21 to 288 Gy/pulse.All measurements were done at room temperature.For the experiments in heavy water, deuterated  tert  -butylalcohol, (CD 3 ) 3 COD, was used as a scavenger of OD and Dradicals.FTIR spectra were measured on a Nicolet 510 Fouriertransform infrared spectrophotometer equipped with a SpectraTech diffuse reflectance accessory. Samples were collected asdry material and measured as 8 wt % in KBr matrix. Typically100 scans were performed for each spectrum. The resolutionwas 4 cm - 1 . Results are presented as Kubelka - Munk plots.Samples were illuminated with a Xe 300-W lamp (ILC).Specimens for observation in the transmission electronmicroscope (TEM) were prepared by placing 1 or 2 drops of colloidal solution onto holey carbon films supported on coppergrids. These specimens were allowed to air-dry for at least 12h. Specimens were imaged in a Jeol 100CXII TEM operatingat 100 kV. High-resolution TEM measurements were carriedout on a Jeol 4000EXII TEM operating at 400 kV to recordlattice images of individual particles. Results and Discussion Excitation of nanocrystalline TiO 2  with photon energies largerthan its band gap results in formation of conduction bandelectrons and valence band holes. Localization of conductionband electrons into lower energy electronic sites occurs in lessthan 30 ps. 20 - 23 Due to the large number of electron surfacetrapping states in nanocrystalline particles, electrons localizepreferentially at the TiO 2  surface. 13,24,25 Surface-trapped elec-trons exhibit broad absorption in the visible region of thespectrum with a maximum around 620 nm in acidic solution. 26 - 29 The position of absorption maximum varies some 50 nm withthe size of the particles and the number of excess electrons. 29 On the other hand, the absorption of electrons in larger particlesincreases steadily from 400 nm toward longer wavelengths,having a plateau near 900 nm. This absorption feature wasattributed to delocalized electrons. 25,28,29 To probe the nature of surface trapping sites and theirselective removal upon binding of surface-active species, westudied and compared optical absorption spectra of excesselectrons in bare TiO 2  and dopamine-capped TiO 2  (TiO 2  /DA)nanoparticles using an electron pulse radiolysis technique. Ourapproach is to investigate the electronic structure of trappingsites by controlled injection of excess electrons into emptyelectron states of nanoparticles while selectively blocking thesurface trapping sites by surface modification with bidentateTrapping Sites in Nanocrystalline Titanium Dioxide  J. Phys. Chem. B, Vol. 107, No. 30, 2003  7369  ligands. The electron pulse radiolysis of aqueous solutionsenables selective formation of reducing radicals and, as a time-resolved technique, allows observation of the dynamics of charge-transfer reactions between reducing radicals and nano-particles. Reducing radicals formed during radiolysis have, ingeneral, high reducing potentials and transfer electrons to thenanoparticles, thereby raising the equilibrium chemical potentialof the whole system. As a result, excess electrons are stableand do not decay unless electron acceptors (such as molecularoxygen) are released into the system. Consequently, electronpulse radiolysis enables determination of electronic structureand redox properties of dispersed nanoparticles.The principle behind electron-beam radiolysis of aqueoussolutions is that highly energetic electrons lose their excessenergy in collision with water molecules, forming highlyenergetic radicals (e aq - , H, OH), and molecular (H 2 O 2 , H 2 )species homogeneously distributed throughout the irradiatedsample. The reducing radical species, hydrated electrons andH atoms, have a large negative potential for electron donation,  E  ° (e aq - )  ) - 2.77 V (vs NHE) 30 and  E  ° (H)  ) - 2.31 V (vsNHE), 31 respectively, and are capable of injecting electrons intothe conduction band of TiO 2  /DA particles. Oxidizing OHradicals can be efficiently scavenged by alcohols such as2-propanol,  tert  -butyl alcohol, etc. Excess Electrons in Dopamine-Capped TiO 2  Particles. TiO 2  is a semiconductor that has particularly reactive photo-generated positive holes (  E  ° VB ) 2.93 V vs NHE at pH 3.5) andmoderately reactive photogenerated electrons (  E  ° CB )- 0.27 Vvs NHE at pH 3.5). 32 When the metal oxide particles are in thenanocrystalline regime, a large fraction of the atoms thatconstitute the nanoparticle are located at the surface withsignificantly altered electrochemical properties. Due to thetruncation of the crystal units at the surface and their weakercovalent bonding with solvent species compared to the bondingwithin the lattice, the energy level of the surface species arefound in the mid-gap region, thereby decreasing their reducing/ oxidizing abilities. 33 In this paper we investigate optical proper-ties of electrons injected into these mid-gap empty electronicstates in TiO 2  nanoparticles from radiolytically generatedreducing radical species e aq - and H atoms. Also, electrons wereinjected into TiO 2  nanoparticles capped with enediol ligandssuch as dopamine that were shown to remove surface trappingsites in nanocrystalline TiO 2 . We have investigated capped TiO 2  / DA particles that were 4.3 nm in diameter, with dopaminecovering 90% of the existing 405 titanium surface sites/particle(maximum coverage). 13 Pulse radiolysis experiments were carried out in nitrogen-saturated aqueous solution of colloidal particles at pH 3.5 andin the presence of a high concentration of   tert  -BuOH (2 M), inwhich the following reactions occur:The oxidizing radical species OH reacts with  tert  -BuOH, whichis an efficient scavenger of OH ( k  1 ) 4 × 10 8 M - 1 s - 1 ) and toa lesser extent of H radicals ( k  1 ) 10 5 M - 1 s - 1 ); the product of reaction 1 is inert toward TiO 2  and/or dopamine. 34,35 Theconcentration of TiO 2  /DA in solution was kept high (0.16 mMparticle concentration) in order to overcome possible competitionbetween the reactions of nanoparticles and H + ions withhydrated electrons ( k  3 ) 2.0 × 10 10 M - 1 s - 1 ). In this case, theconcentration of free dopamine in solution was  e 0.1 mM,determined from the association constant of the dopamine withtitania particles,  K  ) 7.9 × 10 3 M - 1 . 13 Under these conditions,approximately half of the generated H atoms will not bescavenged with  tert  -BuOH. Remaining hydrogen atoms willparticipate directly in reducing TiO 2  /DA particles, based on therate constant for the reaction of H atoms with bare TiO 2 particles,  k   )  2  ×  10 9 M - 1 s - 1 , 34 or they can react with freedopamine (DA):Reduced dopamine radicals, DA - , have a sufficiently largenegative redox potential of   - 1.9 V vs NHE 35 to be able totransfer electrons into TiO 2  /DA particles:Dopamine radical anions exhibit absorption below 400 nm and,thus, do not interfere with our measurements at wavelengthslonger than 600 nm.Whatever the fate of H atoms that are not scavenged with tert  -BuOH,  the net effect is generation of excess electrons intothe particles  (reactions 2 - 6). However, because the yield of Hatoms is at least 5 times lower than that of e aq - , the majority of electrons injected into TiO 2  /DA particles come from reaction2. Free dopamine does not compete with either H + ions ornanoparticles for hydrated electrons because of the very lowrate of any possible reaction,  e  2  ×  10 4 s - 1 . 35 The absorption of excess electrons in 43 Å anatase nanopar-ticles was measured in the range 600 - 1800 nm, after the decayof hydrated electron absorption. A typical transient absorptionsignal of hydrated electrons at 650 nm is presented in Figure1a. In the absence of nanoparticles, the decay of absorption isdue to the reaction of hydrated electrons with hydronium ions,reaction 3. When TiO 2  or TiO 2  /DA particles are present in thesolution, the fast decay of transient absorption corresponds tothe reaction of e aq - with particles. The excess electrons inparticles exhibit long-lived transient absorption (see Figure 1b)throughout the whole range of examined wavelengths.Transient absorption spectra measured at 0.8  µ s after the pulsein N 2 -saturated heavy water solution of 0.16 mM nanoparticlesand 2 M  tert  -BuOH- d  10  are presented in Figure 2. The dosewas kept low, 152.2 Gy/pulse, to produce less than 1 electron/ particle. The spectrum of excess electrons in bare TiO 2  showsa broad band with a maximum around 670 nm and flatabsorption in the range 1200 - 1800 nm. The appearance of abroad band in the absorption spectrum correlates to the spectraof trapped electrons in TiO 2  observed previously in the range400 - 1200 nm. 26 - 29 It was found that these surface trappingsites in nanosize TiO 2  particles experience an adjustment in thecoordination geometry of the Ti atoms near the particle surfacefrom octahedral to square-pyramidal in order to accommodatefor large surface curvature. 13 The coordination sphere of thesurface titanium atoms is incomplete and thus exhibits highaffinity for oxygen-containing ligands to form chelating struc-tures. X-ray absorption near edge structure (XANES) revealsthat surface modification with enediol ligands such as dopaminerestores the pre-edge features of octahedrally coordinated Ti inthe anatase crystal environment and removes surface trappingsites. OH/H + (CH 3 ) 3 COH f  H 2 O/H 2 + C 4 H 8 OH (1)e aq - + TiO 2  /DA f  (e - )TiO 2  /DA (2)e aq - + H + f  H (3)H + TiO 2  /DA f  (e - )TiO 2  /DA + H + (4)H + DA f  DA - + H + ,  k  ) 4.2 × 10 9 M - 1 s - 1 35 (5)DA - + TiO 2  /DA f  (e - )TiO 2  /DA + DA (6) 7370  J. Phys. Chem. B, Vol. 107, No. 30, 2003  Dimitrijevic et al.  When the surface of the nanoparticles is partially coveredwith dopamine (surface coverage of only 70%), the maximumabsorption at 670 nm observed for excess electrons in bare TiO 2 disappears and a new maximum corresponding to excesselectrons at 900 nm appears, while the flat region of thespectrum in the range 1200 - 1800 nm remains unchanged(Figure 2,  b ). The lower yield of absorption that correspondsto trapped electrons, as well as the shift of absorption maximumtoward lower energies, indicates that the deepest surface sitesare repaired primarily and that the electrons localize at theremaining surface centers. This spectrum is identical to thespectrum of excess electrons in large particles or aged anatasecolloids, reported previously and assigned to the conductionband electrons. 25,28,29 When the surface of particles is completely covered with amonolayer of dopamine (TiO 2  /DA particles) that convertsundercoordinated Ti sites into anatase-like octahedral Ti surfaceatoms, the spectrum of excess electrons in titania becomesfeatureless, appearing as a monotonic absorption in the wholerange from 600 to 1800 nm. This spectrum is characteristic forelectrons that are not localized on surface trapping sites whoseenergy of localization is  e 0.7 eV below the conduction band.The dependence of the absorption coefficient on the wavelengthdoes not show a characteristic exponential Drude behavior (  λ 2 ) 36 expected for free carrier absorption. This result suggests thatthe overall spectrum is a consequence of either (i) superpositionof the multiple absorption features created from broad distribu-tion of energies of shallow trapping states in disordered particlesor (ii) superposition of free carriers and intersubband absorptionsin indirect semiconductors. In the first case the particle disorderwould reflect on the EPR spectra, causing broadening of thesignals of excess electrons. Our EPR measurements do notsupport this mechanism because vary narrow signals ( ∆  H  pp ) 2.5 G) 10,13 of photogenerated electrons were observed. However,contribution of long-range disorder to overall absorption spectracannot be excluded. 37 In the second case, contribution of intersubband transitions to overall absorption spectrum inindirect semiconductors 36 results in the flat monotonic absorptionin the region 0.6 - 2.0 eV, assuming energy difference 1.0 eV <  ∆  E   <  1.5 eV between the two lowest subbands of anataseconduction band. XANES measurements in nanoparticulateTiO 2 , 10,12 as well as multiple scattering calculations of XAS inbulk anatase TiO 2 , 38 report the energy difference of subbandsas ∼ 1.4 eV. In this case the dominant contribution of free carrierabsorption is expected at wavelengths further into the IR (  λ g 2.5  µ m). Our future research in mid- and far-IR regions willaddress these issues.The disappearance of the maximum absorption confirms thatsurface trapping sites have been repaired by the bidentate DAligands. The ability of DA to replace surface five-coordinatedTi (defect sites) into octahedral surface Ti states promotesformation of the bulklike surface, making nanocrystallites intobulklike material. From our results we can conclude thatmodification of TiO 2  with dopamine suppresses/repairs defectsites that are deeper than ∼ 0.6 eV below the conduction band.The absorption spectra reported here are also in agreement withEPR measurements of the photogenerated electrons in TiO 2 particles modified with enediol ligands. 13 The EPR spectra of electrons in dopamine-modified TiO 2  have shown delocalizedsites, in addition to localized lattice-Ti 3 + centers. As mentionedabove, measurements in mid- and far-IR are needed for fullunderstanding of the ability of enediol ligands to suppresstrapping sites reported earlier to be 0.42 - 0.5 eV below theconduction band. 25,39 In Figure 3a, optical signals at 1000 and 1700 nm for asolution containing TiO 2  /DA particles are presented. The 1000nm signal corresponds to the decay of hydrated electrons inD 2 O, while the signal at 1700 nm corresponds to the formationof lattice electrons in well-crystallized TiO 2  particles coveredwith dopamine (Figure 3b). The rates of decay and formationagree within 20%. The signals were not corrected for the delayedsecondary response of the InGaAs photodiode. 40 For theseparticular two wavelengths, the distortion of optical signals isalmost identical and small, amounting to 10% of the true optical Figure 1.  (a) Transient absorption signals at 650 nm measured in N 2 -saturated solution of 2 M  tert  -BuOH, pH 3.5, in the absence of particles(water) and in the presence of 43 Å TiO 2  or TiO 2  /DA particles. Particleconcentration is 160  µ M. Dose was 138.0 Gy/pulse. (b) Transientabsorption signal at 650 nm of 80  µ M TiO 2  /DA and 2 M  tert  -BuOHrecorded up to 1 s. Dose was 123.8 Gy/pulse. Figure 2.  Transient absorption spectra observed 0.8  µ s after pulse inN 2 -saturated D 2 O solution of 2 M  tert  -BuOH- d  10  and 160  µ M ( 2 ) bareTiO 2 , and ( [ ) capped TiO 2  /DA particles, pH 3.5. The spectrumpresented with circles ( b ) is obtained for TiO 2  particles with only 70%of surface sites covered with dopamine. Dose was 152.2 Gy/pulse. Thefigure contains absorption spectra of dopamine-modified TiO 2  particleswith 70% (green line) and full coverage (blue line) of the surface states. Trapping Sites in Nanocrystalline Titanium Dioxide  J. Phys. Chem. B, Vol. 107, No. 30, 2003  7371  signal. 40 The agreement between rates of decay and formationconfirms that injection of excess electrons into capped TiO 2  / DA particles proceeds predominantly through reaction withhydrated electrons. Reaction Rates.  The rate constant for the reaction of hydratedelectrons with TiO 2  /DA, reaction 2, was determined by changingthe concentration of particles and measuring the decay of hydrated electron absorption at 650 nm. The particle concentra-tion was determined from the average particle diameter 42.8 ( 3.5 Å and total concentration of TiO 2 . Figure 4 presents transientabsorption signals detected after a pulse of electrons in a pH3.5 aqueous solution containing 2 M  tert  -BuOH and variousconcentrations of TiO 2  /DA. The bimolecular rate constant,  k  1 ) (1.0 ( 0.3) × 10 11 M - 1 s - 1 , was determined from the slopeof the linear plot of pseudo-first-order rates (Figure 4, inset).We have found that the reaction of hydrated electrons with TiO 2  / DA particles is 2 times slower than the corresponding reactionwith bare TiO 2 . The value of 2.1 × 10 11 M - 1 s - 1 was estimatedfrom the decay of e aq - absorption in the presence of 0.16 mMTiO 2  particles. At pH 3.5, dopamine is a charged molecule (p K  )  8.9); 35 therefore, it can alter the surface charge of modifiedTiO 2  particles and alter the reaction rate of particles withnegatively charged solvated electrons. This value is approxi-mately 2 times smaller than the one reported previously 34 forhighly charged 4.7 nm bare TiO 2  particles at pH 2.65, derivedfrom competition kinetics. On the other hand, these values of the rate constant for bare TiO 2  in acidic solution are 2 ordersof magnitude larger than the one at pH 10, when the surface of TiO 2  is negatively charged. 41 The rate constant for the reaction of 1-hydroxy-1-methylethylradicals, (CH 3 ) 2 C ˙ OH, with TiO 2  /DA was estimated from thegrowth of the absorption after the pulse in N 2 -saturated solutioncontaining 2 M acetone, 1 M 2-propanol, and 0.1 mM TiO 2  / DA particles. In this system the following reactions occur:The (CH 3 ) 2 C ˙ OH radicals,  E  ° [(CH 3 ) 2 C ˙ OH]  ) - 1.8 V (vsNHE), 42 formed in reactions 7 and 8 inject electrons into cappedTiO 2  /DA particles. The bimolecular rate constant for the reactionof (CH 3 ) 2 C ˙ OH radicals with TiO 2  /DA particles,  k  10 ) 4 × 10 8 M - 1 s - 1 , was determined from the growth of the transientabsorption at 650 nm that corresponds to the injected excesselectrons in particles (Figure 5). The value agrees with the datareported for 4.7 nm bare TiO 2  particles. 34 The electron-transfer reaction on the surface of semiconductorparticles is a heterogeneous process that involves the movement Figure 3.  (a) Transient absorption signals detected at 1000 and 1700nm after 152.2 Gy electron pulse in N 2 -saturated D 2 O solution of 2 M tert  -BuOH- d  10  and 160  µ M TiO 2  /DA particles. (b) High-resolution TEMimages of TiO 2  particles. Figure 4.  Absorption - time profiles (650 nm) of hydrated electronsat different particle concentrations of capped TiO 2  /DA. Dose was 138.0Gy/pulse. Inset shows the dependence of the observed decay rates asa function of particle concentration. Figure 5.  Growth of absorption at 650 nm after the 251 and 110 Gyelectron pulses in N 2 -saturated solution of 2 M acetone, 1 M 2-PrOH,and 0.1 mM TiO 2  /DA particles, pH 3.5. e aq - + (CH 3 ) 2 CO + H 2 O f  (CH 3 ) 2 C ˙ OH + OH - (7)OH/H + (CH 3 ) 2 CHOH f  H 2 O/H 2 + (CH 3 ) 2 C ˙ OH (8)2(CH 3 ) 2 C ˙ OH f  (CH 3 ) 2 CHOH + (CH 3 ) 2 CO (9)(CH 3 ) 2 C ˙ OH + TiO 2  /DA f  (e - )TiO 2  /DA + (CH 3 ) 2 CO + H + (10) 7372  J. Phys. Chem. B, Vol. 107, No. 30, 2003  Dimitrijevic et al.
Related Search
Advertisements
Related Docs
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks