Effect of Relative Humidity on the OH-Initiated Heterogeneous

Subscriber access provided by BOSTON UNIV

Article

Effect of Relative Humidity on the OH-initiated Heterogeneous Oxidation of Monosaccharide Nanoparticles Hanyu Fan, Mark R. Tinsley, and Fabien Goulay J. Phys. Chem. A, Just Accepted Manuscript • DOI: 10.1021/acs.jpca.5b06364 • Publication Date (Web): 16 Oct 2015 Downloaded from http://pubs.acs.org on October 19, 2015

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

The Journal of Physical Chemistry A is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Effect of Relative Humidity on the OH-Initiated Heterogeneous Oxidation of Monosaccharide Nanoparticles Hanyu Fan, Mark R. Tinsley, and Fabien Goulay* Department of Chemistry, West Virginia University, Morgantown, West Virginia 26506, USA *

Electronic mail: [email protected]

1 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 31

ABSTRACT The OH-initiated heterogeneous oxidation of solid methyl β-D-glucopyranoside nanoparticles (a cellulose oligomer surrogate) is studied in an atmospheric pressure gas flow reactor coupled to an aerosol mass spectrometer. The decay of the solid reactant relative concentration is measured as a function of OH exposure over a wide range of ambient relative humidity. The kinetic traces display an initial fast exponential decay followed by a slower decay. For long OH exposure, the fraction of a particle that reacts decreases from 90% at RH=30%, 60% at RH=20%, and 40% at RH=10%. A computational model based on the diffusion and reaction of the radical, monosaccharide, and water is developed in order to further examine the experimental data. The model parameters and validity are discussed based on previous literature data. The experimental data are consistent with a diffusion-controlled heterogeneous oxidation. These findings are discussed toward a better understanding of mass transport in semi-solid organic material and their effect on chemical change, in particular during the thermal transformation of cellulosic materials to useful chemicals.


Page 3 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1. INTRODUCTION Biomass (i.e. cellulose) is a very promising alternative and renewable source of energy due to the availability of the initial material and the fact that the overall energy production is carbon neutral. A transition toward the use of cellulosic fuels as a clean and renewable energy source still requires being able to identify and understand the rate limiting elementary chemical and physical steps governing the biomass-to-fuel transformation. Accepted mechanisms for cellulose thermal transformation suggest that the organic polymer first decomposes into smaller polysaccharide chains forming viscous droplets.1 Under gasification conditions (in presence of a limited amount of oxygen), complex heterogeneous chemical schemes, including free radical reactions (i.e. OH), in the bulk and at the surface of the particles will lead to the formation of smaller molecular weight products such as alcohol, furan and furfuran derivatives.1-3 Models trying to reproduce and predict the conversion of biomass to fuels are still overly simplistic, mostly due to the lack of knowledge about the heterogeneous chemistry involved during the multiphase conversion of polysaccharides to gaseous and liquid products.4-6 In particular, investigating how changes in solid reactants’ diffusion coefficients can affect radical heterogeneous chemistry is central toward a better understanding of semi-solid oxidation processes. The reaction kinetics of small oxidant species (OH, O3, NO3) with semi-solid viscous organic compounds commonly displays an initial exponential decay of reactant followed by arrested kinetics during which removal of reactant is severely slowed or quenched.7-12 In the case of well-mixed particles, the decay is exponential over the entire OH exposure range. The reaction quenching is commonly associated with the absence of a liquid phase in the aerosol. For example, studies of the oxidation of wet vs. dry particles of succinic acid by OH show a dramatic increase in the fraction of the reactant consumed with water content, though arrested kinetics do occur in both cases.10 Other studies of the oxidation of liquid vs. solid phase alkane aerosols indicate arrested kinetics in the solid phase but not in the liquid phase.13 Oxidation by ozone at low relative humidity (RH) is much more pronounced using oleic acid, 3 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 31

which is a liquid at standard conditions of pressure and temperature, compared to using either maleic or arachidonic acid. These latter two species or their oxidation products form a solid phase resulting in a lowering of the uptake of ozone.14 The most commonly cited mechanism for the occurrence of arrested kinetics is that the outer region of an aerosol undergoes a phase transition to a semi-solid or glassy state. This increases its viscosity and leads to reduced diffusion of reactant towards the surface of the particle.13,15,16 Alternatively, the formation of oxidation products changes the composition of aqueous aerosols enough to induce a phase change such as efflorescence.10 Other authors suggest that this behavior is due to saturation of surface adsorption sites or that aerosols contain, or develop, phase heterogeneities that restrict the reactant movement.7,17,18 There are a significant number of studies investigating variations in diffusion coefficients in solutions, or in aerosols, composed of organic species and water.19-25 Diffusion coefficients in these materials are known to vary exponentially over a range of water content,23,26,27 or organic content.20,22,28 These variations are considered to be induced by the formation of a hydrogen bond matrix amongst the organic molecules that eventually leads to a glass transition. Surface glassy crusts have been observed to form following rapid drying of sucrose particles.19 This phenomenon restricts the diffusion of water into and out of the particle, ultimately leading to an extremely slow equilibration of the particle with the surrounding atmosphere. Changes in composition of a particle, in terms of either the molar mass or O:C ratio of the reactants, can also lead to significant variations in the glass temperature and therefore mass transfer processes.29 Because of the sensitive dependence of the viscosity on chemical and physical structure, small changes in solid phase composition can change the reactant and product diffusion coefficients by orders of magnitudes, leading to complex kinetic behaviors. While studies have explicitly


Page 5 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


simulated changes in diffusivity during wetting and drying of particles,26,27 relatively few of them have attempted to simulate the impact of changes in diffusivity on oxidative aging.18,30-32 The present experimental study investigates the effect of relative humidity on the decay of βmethylglucopyranoside (MGP, C7H14O6), a cellulose oligomer surrogate, by measuring the relative reactant number density as a function of gaseous OH radical concentration. Arrested kinetics is observed for relative humidity ranging from 10% to 30%. Motivated by previous studies on diffusion gradients in semi-solid particles,19 we have developed a simple empirical model to investigate the effect of variations in the reactant diffusion coefficient on the chemical behavior. The model includes the reaction and diffusion of OH, H2O and monosaccharide molecules in the semi-solid phase with different initial conditions. The model parameters are extracted from literature data when available. Variation in the functional relationship between the water content and its diffusion coefficient is used to fit the observed kinetic trends. The experimental findings, supported by the modeling results, are discussed toward a better understanding of the role of molecular diffusion on the chemical reactivity of semi-solid materials and their implications for cellulose thermal transformations. 2. EXPERIMENTAL The experiments are performed using an atmospheric pressure slow flow reactor coupled to an aerosol mass spectrometer at the Chemical Dynamics Beamline at the Advanced Light Source synchrotron. The apparatus has been described in detail elsewhere16,32-34 and only a brief overview is presented here. Methyl β-D-glucopyranoside (MGP) organic aerosol is formed by nebulizing a 1 mg/ml MGP-aqueous solution using a constant-output atomizer (TSI, Model 3076). The droplets then pass through a room-temperature diffusion dryer to remove the water vapor and dry the particles. The residence time in the dryer is about 15 s. The aerosol is injected into a type-219 quartz flow tube reactor with an inner diameter of 2.5 cm and a length of 130 cm. OH radicals are generated along the flow tube by photolysis of ozone in the presence of 5 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 31

water by four mercury lamps (λ=254 nm, UVP. LLC) located outside the flow tube reactor. Ozone is generated by a mercury pen-ray lamp (UVP, LLC) or a commercial corona discharge ozone generator (OzoneLab Instruments). The amount of OH radical formed is controlled by the flow rate of molecular oxygen through the ozone generator or by the intensity of the corona discharge. The RH in the flow tube is set by flowing a known amount of nitrogen gas through a water bubbler. The water-saturated gas is then mixed with dry nitrogen, ozone, molecular oxygen and the particle flow. A small amount of hexane gas (150 ppb) is added to the flow in order to measure the OH exposure (the integral of the OH number density over the total reaction time).33,34 The total gas flow rate entering the flow tube reactor is set to 1 slm (standard liter per minute) corresponding to a total reaction time of 37 s. Upon exiting the reaction flow tube, the aerosol passes through an ozone denuder to remove the remaining ozone. A fraction of the flow is sent to a Scanning Mobility Particle Sizer (SMPS, TSI, Model 3936) for particle characterization. The typical surface-weighted mean diameter of the particles is 228±14 nm. Typical particle diameter distributions are displayed in Figure S1 of the supplementary information. A portion of the aerosol flow is sampled through an aerodynamic lens into a pulsed time-offlight aerosol mass spectrometer coupled to synchrotron Vacuum-Ultraviolet (VUV). The resulting aerosol beam is vaporized by contact with a copper tip heated at 423 K at a pressure of 3×10-7 Torr. The resulting plume is then ionized by the VUV light of the synchrotron and analyzed by time-of-flight spectrometry. A fraction of the remaining flow is sent to a gas chromatograph (GC) equipped with a flame ionization detector (FID) (SRI model 8610C) for monitoring the loss of hexane. Ozone and aerosol are removed prior to entering the GC by a potassium iodide (KI) trap and particle filter. The OH exposure is calculated from the decay of hexane concentration analyzed with the GC.33,34


Page 7 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


3. EXPERIMENTAL RESULTS Figure 1 displays the mass spectrum obtained upon vaporization of unreacted MGP nanoparticles at 10.5 eV using the aerosol mass spectrometer. The photoionization mass spectrum is similar to that obtained upon electron impact ionization at 70 eV although higher masses are detectable in the case of soft ionization. The photon energy was selected in order to maximize the ion signal while minimizing the dissociative ionization of the parent molecule. The highest mass observed in Figure 1 is at m/z=176 corresponding to dissociative ionization by loss of a water molecule. The other fragments indicated in Figure 1 have been used as tracers of the relative monosaccharide abundance for the kinetic measurements. The signal at m/z=163 is likely to be due to the loss of neutral CH3O upon ionization. Lower mass cations require a more complex fragmentation process, such as successive water, OH or methanol losses.

Figure 1 Mass spectrum of unreacted methyl-β-D-glucopyranoside nanoparticles obtained at 10.5 eV. Figure 2 displays the relative abundance of the MGP reactant as a function of OH exposure for RH=30% by (a) plotting the relative signal of 5 different tracer masses and (b) averaging the relative signal of these 5 masses for two independent data sets. The kinetic decay is found to be independent of the selected mass fragment mass. This confirms that the observed behavior at high 7 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 31

OH exposure is not due to the contribution from the increase of a product signal at the same mass. The thick red line in Figure 2 (b) is an exponential fit to the experimental data up to an OH exposure of 0.5×1012 cm-3 s and extrapolated to higher OH exposure. The exponential fit leads to a pseudo first order coefficient of 2.0(±0.2)×1012 cm3 s-1. The reactant decay rate coefficient is found to be independent of the gas phase molecular oxygen concentration. The exponential trace is within the 2σ error bars up to an OH exposure of about 1.25×1012 cm-3 s and deviates significantly from the data at higher OH exposure.

Figure 2 Relative signal of unreacted methyl-β β -D-glucopyronoside as a function of OH exposure obtained (a) for mass fragments at m/z=60 (blue triangles), 73 (black open circles), 121 (black filled circles), 144 (green diamonds), and 163 (red squares) and (b) by averaging the ion signal at five different masses (see text). The error bars in (b) are 2σ σ about the mean value. The red line is


Page 9 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


obtained by fitting the experimental data up to 0.5× ×1012 cm-3 s with an exponential function and extrapolated to higher OH exposure. Figure 3 displays (a) the relative surface-weighted diameter, (b) the remaining mass fraction, and (c) the relative abundance of the reactant as a function of OH exposure for 10% (blue squares), 20% (red triangles) and 30% (black circles) relative humidity. In Figure 3(a), the surface-weighted diameter changes by less than 1% for OH exposure lower than 0.3× ×1012 cm-3 s with no observable effect of the relative humidity. This corresponds to a decay of the mass fraction of the aerosol by 7% (Figure 3). At higher OH exposure, the mass loss slows down and is dependent on the relative humidity. In Figure 3(c), all three traces display a similar behavior as to that seen in Figure 2 with an exponential decay at low OH exposure and a slower decay at high OH exposure. Neglecting any change in the particles’ surface-weighted diameter at low OH exposure (see Figure 3), the reactive uptake coefficient γ can be calculated by fitting the initial part of the experimental data (OH exposure < 0.75× ×1012 cm-3 s) to an exponential function and normalizing by the surface-weighted particle diameter.33 The reactive uptake coefficients at low OH exposure are found to be 0.92(±0.3) at RH=10%, 1.2(±0.3) at RH=20% and 1.9(±0.3) at RH=30%. These values are not corrected for gas phase diffusion of the OH radicals. It is important to note that these uptake coefficients are calculated assuming a particle is well-mixed over the timescale of the reaction.35 In the case of viscous organic aerosols, the reactive uptake coefficient is a function of the molecular reactant diffusion coefficient and reacto-diffusion length.35 Because of the large uncertainty in the monosaccharide diffusion coefficient and its dependence on water concentration, the calculated uptake coefficients are only indicative of the relative particle reactivity at low OH exposure. Nonetheless, the observed dependence on gas RH suggests that water plays a significant role during the heterogeneous reaction, either chemically or by changing the mass transport properties of the 9 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 31

particle. In Figure 3, in addition to displaying slower decays at low OH exposure, traces recorded at low RH exhibit significant arrested kinetics with non-exponential decays at high OH exposure. At an OH exposure of 1.8×1012 cm-3 s, the fraction of reacted monosaccharide molecules decreases from 90% at RH=30% to 60% at RH=20% and 40% at RH=10%.

Figure 3 Relative (a) surface-weighted diameter, (b) remaining mass fraction and (c) relative ion signal of methyl-β β -D-glucopyronoside as a function of OH exposure for RH=10 % (blue squares), RH=20% (red triangles), and RH=30% (black circles). In panels (a) and (b), the error bars are 2σ σ 10 ACS Paragon Plus Environment

Page 11 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


about the mean obtained by averaging three SMPS measurements. In panel (c), the error bars are 2σ σ about the mean obtained by averaging the signal of five different fragments. Figure 4 displays the mass spectrum of the (a) unreacted and (b) reacted samples over the 130 to 180 m/z range. In panel (b), the reacted particle spectrum is obtained for an OH exposure of 0.4×1012 cm-3 s and RH=30%. Small peaks are detected in addition to the fragment peaks of the reactants. The additional signals are likely to be due to fragmentation of the products upon soft ionization, as observed for highly oxidized molecules.16,32,36 No signal is detected at masses higher than that of the parent reactant. This suggests that either the products fragment upon ionization or that they all have smaller molecular weights than the unreacted molecule.

Figure 4 Mass spectrum of methyl-β β -D-glucopyranoside nanoparticles obtained at 10.5 eV for (a) unreacted and (b) reacted samples over the 130 to 180 m/z range. The reacted sample is obtained for an OH exposure of 0.4× ×1012 cm-3 s and RH=30%.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 31

Figure 5 displays the normalized signal of product ions at (a) m/z=77, (b) m/z=102, and (c) m/z=142 as a function of OH exposure for RH=10 % (blue squares), RH=20% (red triangles), and RH=30% (black circles). All the signals reach a maximum at values of OH exposure that are less than 2.5× ×1012 cm-3 s. The rise time is much shorter than the timescale associated with the reactant decay shown in Figure 3. After reaching a maximum, the product signals slowly decay with a rate that is independent of the relative humidity. There is no evidence of secondary product formation in the solid phase.


Page 13 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 5 Relative signal of the products at (a) m/z=77, (b) m/z=102, and (c) m/z=142 as a function of OH exposure for RH=10 % (blue squares), RH=20% (red triangles), and RH=30% (black circles). 4. MODELING OF THE REACTANT DIFFUSION AND REACTION While the main focus of this report is the communication of our experimental findings, we also describe a simple empirical PDE model capable of reproducing the observed trends in the experimental data. The model investigates the impact of diffusion on the observed arrested kinetics and the effect of relative humidity on the diffusive flux within an aerosol. Variation in the functional relationship between water content and its diffusion coefficient is used to fit the observed trends in the experimental behavior. The model assumes that the chemical reaction is limited by diffusion of the reactant to the surface, which is typical in systems where the diffusive timescale of the oxidant is slow compared to the timescale of its reaction.37,38 It uses a pair of coupled reaction-diffusion equations to represent the concentration profiles of the reactant monosaccharide, [MGP], and the oxidant, [OH], within an aerosol particle. Each particle is assumed to have spherical symmetry and a radius of 100 nm. The reaction-diffusion equations are written using spherical polar coordinates:38

∂ [OH ] 1 ∂  2 ∂ [OH ]  = 2  r DOH  − k [OH ] [ MGP ] r ∂r  ∂t ∂r 

(E1)

∂ [ MGP ] 1 ∂  2 ∂ [ MGP ]  = 2  r DMGP  − k [OH ] [ MGP ] ∂t r ∂r  ∂r 

(E2)

No-flux boundary conditions are used for both [MGP] and [OH] at the particle center, r = 0, and at the surface, r = R, for [MGP]. Henry’s Law is used to determine the surface [OH] with an Henry’s Law coefficient of 30 M/atm, based upon literature estimates using measurements of OH uptake onto water surfaces.39,40 The reaction rate constant, k=5× ×10-12 cm3 s-1, for the reaction of the OH radical with the 13 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 31

carbohydrate molecule, is estimated based on the structure-activity relationship for the oxidation of MGP in an aqueous solution.41 The model is formulated using equations E1 and E2 with an additional diffusion equation expressing the local water content. Under equilibrium conditions, the water activity, aH 2O , at the surface is calculated using Raoult’s Law in conjunction with the partial pressure of the water in the gas phase. The activity of the water as a function of the MGP-reactant mass fraction is determined using AIOMFAC-web calculations for the considered molecule (see supplementary information).42-44 DMGP and DH 2O are both assumed to be functions of aH 2O ; DH 2O is empirically determined based upon the functional dependence given by Zobrist et al. (see supplementary information).45 The reaction-diffusion equations are solved using the MATLAB routine PDEFUN with appropriate initial and boundary conditions. Our experimental findings indicate only a small, 105) as a species approaches its glass temperature. The second point is that the ratio DMGP DH 2O is set to be 0.1 over the entire wH 2O range. Therefore, a smaller DH 2O value is necessary to see an impact on the value of DMGP. Literature reports indicate that the ratio between carbohydrate and water diffusion coefficients is about 0.1 for values of wH 2O higher than 0.3.47 At lower water weight fractions and close to the glass transition temperature, decoupling between the two diffusion coefficients could lead to much smaller ratios.47 At low water weight fraction, a higher water diffusion coefficient, similar to that measured for maltose, could be compensated for in the model by a much lower ratio between the water and monosaccharide diffusion coefficients.

Figure 9 Diffusion coefficients for water in maltose (circles)47 and diffusion coefficient for water (thick red line) and MGP (dashed green line) used in the present model. 5. DISCUSSION AND IMPLICATIONS FOR HETEROGENEOUS CHEMISTRY The OH + β-MGP reaction has been studied in the liquid phase by identifying the final products as well as the reaction intermediates.53-56 Electron paramagnetic resonance experiments demonstrated that the initial reaction step is abstraction of a hydrogen atom from the molecule to form an organic radical and a water molecule.53 In the presence of oxygen, the most likely product is the α-hydroxy-peroxyl radical.56 19 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 31

These peroxy radicals can then decompose to give a carbonyl compound and a HO2 radical. Alternatively, the ROO organic radical may react through intra and/or inter molecular reactions. Reaction products are likely to be mainly glucose and organic acids through the loss of small alcohol and ketone molecules as observed in aqueous solution.55 The RH-independent fast rise of the product signal observed in Figure 5 is consistent with carbohydrate oxidation at the particle surface. The observed product decay is likely to be due to further reaction with the OH radicals. Propagation of the chemistry to the particle bulk would likely lead to higher product concentrations as well as formation and detection of secondary products. The absence of secondary products is therefore consistent with a chemical reaction scheme mostly located at the particle surface with negligible bulk reactivity. At the high OH concentrations used in the present work, it is likely that the primary reaction products rapidly react with OH and break down to small molecular mass species and desorb to the gas phase. In order to test the diffusion-controlled reaction hypothesis, the model includes surface-only reaction through a fast reaction of the OH radicals. The assumption that every OH radical reaching the particle surface leads to reaction has been shown to be valid for OH reacting with most saturated molecules for which the initial chemical step is the fast abstraction of an hydrogen atom.35 The reaction with the OH radicals is modeled to occur only within the first 2 nm of the particle surface with no impact on the bulk chemical composition. This reactive thickness agrees with the reactodiffusion length calculated using the formula of Houle et al.35 assuming DMGP =1 10-12 cm2 s-1. The model also neglects the vaporization of the products by assuming a constant diameter and constant total mass. Desorption of the products from the surface would expose a new monosaccharide layer with no effect on the particle bulk. Although mass-loss processes are not negligible, the qualitative agreement between the data in Figure 3 and the model in Figure 7 agrees with the diffusion of the monosaccharide from the bulk to the surface being a rate-limiting step. 20 ACS Paragon Plus Environment

Page 21 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


The computational model developed here strongly links the occurrence of arrested kinetics to water content. This remains true even if an initially uniform water weight fraction is assumed. In this case, the diffusion of water across the particle is fast, but, provided DMGP is sufficiently decoupled from DH 2O , the amount of water present still dictates the value of DMGP throughout the particle. All behavior is extremely sensitive to the functional form of the diffusion coefficient vs. weight fraction. The model shows that small changes in the functional forms can significantly enhance the formation of a surface skin and its role in slowing down the kinetics. Under such conditions, the monosaccharide will be relatively mobile on the inside but will be prevented from reaching the surface due to the high near surface viscosity. At low water weight fraction (

Effect of Relative Humidity on the OH-Initiated Heterogeneous

Recommend Documents