SEARCH

SEARCH BY CITATION

Keywords:

  • mesosphere;
  • positive ions;
  • silicon;
  • oxcited oxygen

Abstract

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Experiment
  5. 3. Results and Discussion
  6. Acknowledgments
  7. References

[1] The rate constant for the reaction of Si+ with O2(a 1Δg) was measured in a selected ion flow tube (SIFT) between 200 and 500 K. The temperature dependence is T−(2.2±0.8). The rate constant at 500 K is below the detection limit (<1 × 10−11 cm3 molecule−1 s−1) while the rate constant at 200 K is 3.6 × 10−11 cm3 molecule−1 s−1. The product is SiO+, in accord with potential energy surfaces. These surfaces also show that the reaction of Si+ with O2(3Σg) only yields association products, SiOO+ or OSiO+ ions. An upper mesosphere/lower thermosphere model predicts that reaction with O2(a 1Δg) is the most important removal process for Si+ during the day between 87 and 107 km, which may explain the depletion of this ion relative to other ions of meteoric origin, such as Fe+ and Mg+. The reaction therefore has an important influence on the composition and lifetimes of sporadic E layers during daytime.

1. Introduction

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Experiment
  5. 3. Results and Discussion
  6. Acknowledgments
  7. References

[2] In the atmosphere, the first excited state of molecular oxygen, O2(a1Δg), is mainly produced in the mesosphere/lower thermosphere (MLT) by photolysis of O3 at wavelengths shorter than 320 nm [Ball et al., 1995]:

  • equation image

O2(a1Δg), which is excited by 0.97 eV, is long-lived in the MLT. The quenching life-time is more than 4 hours, much longer than the phosphorescence lifetime of 72 minutes for emission at 1.27 μm to the ground state [Newman et al., 2000]. Since the rate coefficient for O3 photolysis in the MLT is ∼8 × 10−3 s−1, during daytime the steady-state ratio [O2(a 1Δg)]/[O3] is ∼34 [Brasseur and Solomon, 2005]. After sunset, O2(a 1Δg) decays by an order of magnitude every 2.8 hours. While the conventional view is that O2(a 1Δg) does not play any significant role in the positive ion chemistry of the atmosphere, we will show that O2(a 1Δg) has an important impact on the ion-molecule chemistry of silicon.

[3] The major source of silicon in the earth's upper atmosphere is the ablation of the roughly 30 tonnes of interplanetary dust that enters each day from space [Plane, 2003]. Models predict silicon will enter the atmosphere as Si, SiO, and SiO2. Si+ ions form in the MLT through hyperthermal collisions of Si and SiO with air molecules, when these silicon species ablate from meteoroids with entry velocities above ∼35 km s−1 [Vondrak et al., 2008]. Other sources of Si+ and its oxides include charge transfer between SiO and O2+ in the lower E region (which is exothermic by 1.2 eV [Lias et al., 1988]) and photoionization of SiO at wavelengths shorter than 108 nm. In both cases, subsequent reaction of the SiO+ with O to form Si+ [Fritzenwallner and Kopp, 1998] occurs. Although silicon atoms can charge transfer with both NO+ and O2+, and photoionize below 153 nm, the concentration of neutral Si is very low because it reacts at essentially every collision with O2 to form SiO [Gómez Martin et al., 2009].

[4] Meteoric ablation also injects various metallic ions into the MLT, giving rise to sporadic E layers. These are thin layers of concentrated plasma, typically only 1–3 km wide between 95 and 140 km [Mathews, 1998]. The resulting sharp gradients in electron density can have a significant impact on radio communications, both by facilitating over-the-horizon high frequency communication and by obscuring space-to-ground communications. The electrons are balanced predominantly by atomic ions of meteoric origin −Fe+, Mg+, and Si+ [Grebowsky and Aikin, 2002; Kopp, 1997]. In fact, since there is ambiguity in mass spectrometric measurements at m/z = 28, which could be Si+ and/or N2+, the presence of this ion in a sporadic E layer confirms that it is Si+, since N2+ would undergo very rapid dissociative recombination [Kopp et al., 1995; Kopp, 1997].

[5] The concentration into layers occurs through a variety of mechanisms. At low- and mid-latitudes, sporadic E layers form through ion convergence at the null points of wind shears [Whitehead, 1989], often driven by atmospheric tides and gravity waves [Zhou and Mathews, 1995]. At high latitudes formation likely results from ion convergence in electric fields near auroral activity [Olesen et al., 1975], where charged particle precipitation produces excess ionization in the lower E region [MacDougall et al., 2000]. There is also some indication that sporadic E layers are associated with meteor showers [Grebowsky et al., 1998; Kopp, 1997].

[6] Atomic ions have lifetimes of at least several days in the thermosphere above 100 km, and can be transported to heights of over 400 km [Gardner et al., 1998; Grebowsky and Reese, 1989; MacDougall et al., 2000; McNeil et al., 1998]. The long lifetimes of atomic cations is a result of dielectric recombination (e.g., M+ + e [RIGHTWARDS ARROW] M + ) being very slow, with a rate coefficient ∼10−12 cm3 molecule−1 s−1 [Bates and Dalgarno, 1962]. Indeed, this inefficiency is what allows concentrated sporadic E layers to exist in the MLT.

[7] Below 100 km these atomic ions undergo reactions with species such as O3, O2 and H2O to form molecular ions which undergo dissociative recombination with electrons [Plane, 2003]. For Si+, several reactions are important. This paper addresses the reaction with O2(a 1Δg):

  • equation image

We have estimated the enthalpy of the reaction at 0 K using a SiO+ bond energy of 467 kJ mol−1, calculated at the Complete Basis Set (CBS-Q) level of electronic structure theory within the Gaussian 03 suite of programs [Frisch et al., 2003]. The same level of theory was used for all the calculated enthalpies reported in this paper. In contrast to reaction 2, Si+ can only associate with ground-state O2(X 3Σg) in the presence of a third body,

  • equation image

where M = N2 or O2 in the MLT. The kinetics of this reaction have been studied by Fahey et al. [1981], who obtained k3(300 K) = (1 ± 0.4) × 10−29 cm6 molecule−2 s−1, with M = He. An N2 buffer is expected to increase the rate by a factor of approximately 3 [Viggiano, 1984; Viggiano et al., 1985], which we use later in the paper. Si+ also reacts with O3,

  • equation image

with a rate coefficient k4(298 K) = (6.5 ± 2.1) × 10−10 cm3 molecule−1 s−1, recently measured at the University of Leeds (J. C. Gómez Martin and J. M. C. Plane, Kinetic studies of atmospherically relevant silicon chemistry. Part III: reactions of silicon ions, submitted to Physical Chemistry Chemical Physics, 2010). This reaction is fast, ∼50% of the collision rate [Langevin, 1905; Miller, 1995; Troe, 1987].

[8] Finally, Si+ also reacts relatively rapidly with H2O:

  • equation image

The rate coefficient is (2.3 ± 0.9) × 10−10 cm3 molecule−1 s−1 [Ferguson et al., 1981]. The rate of removal of Si+ by reactions 3–5 would very rapidly deplete Si+ around 100 km to undetectable levels, independent of whether reaction 2 occurs. Since Si+ is observed, it is clear that, as for the molecular ions of Fe+ [Woodcock et al., 2006] and Ca+ [Broadley et al., 2008], atomic O must play a critical role in reforming Si+. The critical reactions are,

  • equation image
  • equation image

The last important reaction is dissociative recombination,

  • equation image

Inspection of rocket mass spectrometric profiles of Si+, Fe+, and Mg+ shows that Si+ disappears completely below 95 km, in contrast to the metallic ions whose concentration ratios sometimes peak around 90 km [Kopp et al., 1995; Kopp, 1997]. One reason may be reaction 2; analogous reactions of Fe+ and Mg+ are endothermic and thus not of importance in the MLT. Here we report the first measurement of k2, and examine the atmospheric implications.

2. Experiment

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Experiment
  5. 3. Results and Discussion
  6. Acknowledgments
  7. References

[9] The measurements were made in a selected ion flow tube, which has been described in detail previously [Midey et al., 2008; Viggiano et al., 1990]. Briefly, ions are created in an external ion source, mass selected using a quadrupole, and injected into a flow tube through a Venturi inlet. The ions are thermalized and carried downstream by the helium. O2(a1Δg) is added 59 cm upstream from a sampling orifice. The primary and product ion count rates are monitored as a function of O2(a1Δg) concentration. The gas is pumped away by a roots blower and the ions are sampled into a quadrupole mass spectrometer and monitored by a discrete dynode particle multiplier.

[10] Si+ was generated from Si(CH3)4. Unfortunately, the ion could not be made directly in the moderate pressure ion source. Instead, SiCmHn+ ions from the source are injected (in a non-selective mode) at >100 eV, and stripped in collisions with the He. Small signals of several SiCmHn+ species persisted, but this method produced Si+ as >90% of the reactant ions. Rate constants are derived, under pseudo-first order conditions, from the slope of the semilogarithmic plot of the Si+ decrease as a function of the O2(a1Δg) concentration, assuring no interference from SiCmHn+ in rate constant measurements. Measurements at 200 K are made by pulsing liquid nitrogen through four sets of heat exchangers; measurements below 200 K were not attempted because previous experience indicates that O2(a1Δg) quenches at lower temperatures. Heating to 500 K is done by wrapping heating tape around the apparatus in several zones.

[11] O2(a1Δg) is produced as in equation (9),

  • equation image

This reaction is a well known source of O2(a 1Δg) [Khan and Kasha, 1963; McDermott et al., 1978; Midey et al., 2008; Seliger, 1960]. Briefly, a Cl2/He mixture is bubbled through a KOH/H2O2 solution kept at −20°C. A second trap (−70° C) removes H2O. 100% of the Cl2 is converted to a mixture of the two states of O2. After the traps, the He/O2 mixture flows into an emission cell [Midey et al., 2007] which has been calibrated to an absolute O2(a 1Δg) spectrometer. This setup has been described in detail elsewhere [Midey et al., 2007, 2008].

3. Results and Discussion

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Experiment
  5. 3. Results and Discussion
  6. Acknowledgments
  7. References

[12] The fraction of O2(a1Δg) in the O2 produced from reaction 9 varied from 6–27%. Addition of O2(a1Δg) produced mainly SiO+ (>60% of the products). However, imperfect trapping of H2O led to SiOH+ production from reaction 5. The measured values of k2 were corrected by the ratio of SiO+ to the sum of SiO+ and SiOH+. The correction may be influenced by the minor SiCmHn+ species. Therefore, we raise the typical error limits from ±35% to ±40% for absolute values and ±25% to ±35% for relative values [Midey et al., 2007, 2009]. No evidence of excited electronic states was found evidenced by linear kinetics.

[13] Figure 1 illustrates the temperature dependence of k2 between 200 and 500 K. The values of k2 are small; the rate constant at 300 K of 1.5 × 10−11 cm3 molecule−1 s−1 is 50 times smaller than the Langevin collision frequency of 7.6 × 10−10 cm3 molecule−1 s−1 [Langevin, 1905; Miller, 1995]. At 500 K only an upper limit to k2 is reported (<1 × 10−11 cm3 molecule−1 s−1). At 200 K, k2 is 3.6 × 10−11 cm3 molecule−1 s−1, approximately double the value at 300 K, giving a temperature dependence of T−(2.2±0.8).

image

Figure 1. A plot of the temperature-dependence of k2 between 200 and 500 K.

Download figure to PowerPoint

[14] Figure 2 illustrates the three reactive potential energy surfaces involved in the reactions of Si+ with O2(X 3Σg) and O2(a 1Δg). Reaction of Si+(2P) with O2(X 3Σg) occurs on both doublet and quartet surfaces. Initial formation of [Si-O2]+ where the O2 is side-on, is followed by a transition state where Si inserts into the O-O bond, leading to the most stable dioxide (linear [O-Si-O]+) on the 2A″ surface. This complex is 312 kJ mol−1 more stable than the Si+ + O2(X 3Σg) entrance channel, which explains why reaction 3 is unusually fast for an atom + diatomic recombination [Fahey et al., 1981]. The 4A″ surface is much higher in energy, with only partial insertion of the Si; the O-Si-O angle is 88°. The reaction of Si+ with O2(a 1Δg) occurs on a 2A′ surface. Initially, Si-O2+ forms with a O-Si-O bond angle of 64°. However, insertion of the Si+ between the O atoms involves an avoided crossing (70 kJ mol−1 below the reactants) onto another 2A′ surface leading to O-Si-O+ with a bond angle of 139°, which then connects with the SiO+ + O product.

image

Figure 2. Potential energy surfaces for the Si+ + O2(a1Δg, X3Σg) system, calculated at the CBS-Q level of theory. The position of the avoided crossing on the 2A′ surface was calculated at the B3LYP/6-311+G(2d,p) level.

Download figure to PowerPoint

[15] The absence of a barrier above the reactant entrance channel on the 2A′ surface is consistent with the negative temperature dependence of k2. The fact that reaction 2 is significantly slower than the Langevin frequency indicates that the majority of collisions between Si+ and O2(a 1Δg) are non-reactive. There are two possible reasons for this. First, the combination of Si+(2P) and O2(a 1Δg) generates six surfaces (3 2A′ and 3 2A″), only one of which appears to be reactive although it involves transition through an avoided crossing. Statistically the probability for a reaction to form SiO+ + O will be low. Successful transition through the avoided crossing will also be more likely at lower collision energies and hence temperatures. Second, spin-orbit constraints may also reduce the rate and enhance the negative T-dependence [Gómez Martin et al., 2009].

[16] The reaction of Si+ with O2(X 3Σg) produces SiO+ much more slowly than the reaction with O2(a 1Δg) because the former is endothermic by ΔrH0 = 27 kJ mol−1 while the latter is exothermic by 67 kJ mol−1 (Figure 2). A quasi-Arrhenius fit [Fahey et al., 1981] indicates that the O2(X 3Σg) rate constant at 300 K is ∼10−16 cm3 molecule−1 s−1, i.e., over 5 orders of magnitude lower than the O2(a 1Δg) rate constant. Instead, O2(X 3Σg) associates with Si+ (reaction 3) with a rate constant of ∼10−13 cm3 molecule−1 s−1 at the pressure of these measurements. These values guarantee that O2(X 3Σg) does not interfere in the experiment.

[17] Finally, we consider the atmospheric implications of the reaction between Si+ and O2(a 1Δg). Figure 3 illustrates the first-order removal rates of Si+ by reactions 2–5, as a function of altitude between 80 and 110 km. The vertical profiles of O2(a 1Δg), O2(X 3Σg), O3, and H2O, and the temperature and atmospheric density are taken from our MESOMOD model for midday (40°N, in April) [Murray and Plane, 2005]. Figure 3 shows that between 87 and 107 km reaction with O2(a 1Δg) is the most important removal process. Indeed, between 95 and 100 km where Si+ is relatively abundant, reaction 2 accounts for just under 60% of Si+ removal.

image

Figure 3. Vertical profiles in the mesosphere/lower thermosphere of the first-order removal rates of Si+. Conditions are 40°N, April at midday.

Download figure to PowerPoint

[18] Conversion of Si+ to SiO+ greatly increases the loss rate of ionized silicon species because atomic ions do not recombine efficiently with electrons, whereas molecular ions are lost rapidly through dissociative recombination, k ∼ 2 × 10−7 cm3 molecule−1 s−1 [Florescu-Mitchell and Mitchell, 2006].

[19] Si+ is depleted relative to other meteoric atomic ions below 95 km for several reasons. Si+ reacts with O2(a 1Δg) to form SiO+, with O3 to produce neutral SiO + O2+ with a ∼50% branching ratio, [J. C. Gómez Martin and J. M. C. Plane, University of Leeds, unpublished work] and following recombination of Si+ with O2, there is a limited charge transfer between OSiO+ and O2 to produce SiO2 + O2+ [Fahey et al., 1981]. For metallic ions such as Fe+ and Mg+, analogous reactions are endothermic. It is also difficult to reionize Si when neutralized because it reacts very rapidly with O2 [Gómez Martin et al., 2009], leaving little Si to undergo charge transfer with ambient E region ions such as NO+ and O2+ (in contrast to the layers of metallic atoms).

[20] Finally, it should be noted that O2(a1Δg) is also important in controlling the electron population in the lower E region because it reacts with O and O2 to detach electrons [Midey et al., 2008]. To our knowledge, these reactions and the reaction with Si+ are the only ion-molecule atmospheric processes where O2(a1Δg) has (so far) been identified to be important.

Acknowledgments

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Experiment
  5. 3. Results and Discussion
  6. Acknowledgments
  7. References

[21] JMCP was supported through award FA8655-10-1-3045 from the European Office of Aerospace Research and Development. RJB and AAV are grateful for the support of the Air Force Office of Scientific Research for this work. NE acknowledges funding from the Institute for Scientific Research of Boston College (FA8718-04-C-0055) and the Air Force Summer Faculty Fellowship Program.

References

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Experiment
  5. 3. Results and Discussion
  6. Acknowledgments
  7. References