Article pubs.acs.org/JPCA
Laboratory Studies of CHF2CF2CH2OH and CF3CF2CH2OH: UV and IR Absorption Cross Sections and OH Rate Coefficients between 263 and 358 K María Antiñolo, Sergio González, Bernabé Ballesteros, José Albaladejo, and Elena Jiménez* Departamento de Química Física, Facultad de Ciencias y Tecnologías Químicas, Universidad de Castilla-La Mancha, Avda. Camilo José Cela, s/n. 13071 Ciudad Real, Spain S Supporting Information *
ABSTRACT: Fluorinated alcohols, such as 2,2,3,3-tetrafluoropropanol (TFPO, CHF2CF2CH2OH) and 2,2,3,3,3-pentafluoropropanol (PFPO, CF3CF2CH2OH), can be potential replacements of hydrofluorocarbons with large global warming potentials, GWPs. IR absorption cross sections for TFPO and PFPO were determined between 4000 and 500 cm−1 at 298 K. Integrated absorption cross sections (Sint, base e) in the 4000−600 cm−1 range are (1.92 ± 0.34) × 10−16 cm2 molecule−1 cm−1 and (2.05 ± 0.50) × 10−16 cm2 molecule−1 cm−1 for TFPO and PFPO, respectively. Uncertainties are at a 95% confidence level. Ultraviolet absorption spectra were also recorded between 195 and 360 nm at 298 K. In the actinic region (λ > 290 nm), an upper limit of 10−23 cm2 molecule−1 for the absorption cross sections (σλ) was reported. Photolysis in the troposphere is therefore expected to be a negligible loss for these fluoropropanols. In addition, absolute rate coefficients for the reaction of OH radicals with CHF2CF2CH2OH (k1) and CF3CF2CH2OH (k2) were determined as a function of temperature (T = 263−358 K) by the pulsed laser photolysis/laser induced fluorescence (PLPLIF) technique. At room temperature, the average values obtained were k1 = (1.85 ± 0.07) × 10−13 cm3 molecule−1 s−1 and k2 = (1.19 ± 0.03) × 10−13 cm3 molecule−1 s−1. The observed temperature dependence of k1(T) and k2(T) is described by the following expressions: (1.35 ± 0.23) × 10−12 exp{−(605 ± 54)/T} and (1.36 ± 0.19) × 10−12 exp{−(730 ± 43)/T} cm3 molecule−1 s−1, respectively. Since photolysis of TFPO and PFPO in the actinic region is negligible, the tropospheric lifetime (τ) of these species can be approximated by the lifetime due to the homogeneous reaction with OH radicals. Global values of τOH were estimated to be of 3 and 4 months for TFPO and PFPO, respectively. GWPs relative to CO2 at a time horizon of 500 years were calculated to be 8 and 12 for TFPO and PFPO, respectively. Despite the higher GWP relative to CO2, these species are not expected to significantly contribute to the greenhouse effect in the next decades since they are short-lived species and will not accumulate in the troposphere even as their emissions grow up.
1. INTRODUCTION Hydrofluorocarbons (HFCs) have been used as short-term replacements for the ozone depleting chlorofluorocarbons (CFCs). Even though HFCs are not considered ozone-depleting substances,1 they have some environmental hazards and health risks, as reviewed by Tsai.2 One of these environmental hazards is the potential to enhance the Earth’s average temperature, contributing to global warming. Currently, fluoropropanols (FPOs), such as 2,2,3,3-tetrafluoropropanol (TFPO, CHF2CF2CH2OH), and 2,2,3,3,3-pentafluoropropanol (PFPO, CF3CF2CH2OH), are potential replacements of HFCs in several industrial and electronic applications. 3 The atmospheric fate of these species is expected to be essentially dominated by the gas-phase reaction with hydroxyl (OH) radicals: OH + CHF2CF2CH2OH → products k1 (1) OH + CF3CF2CH2OH → products
k2
© 2012 American Chemical Society
To our knowledge, no kinetic data on k1 have been reported yet. In contrast, the temperature dependence of k2 has been studied both experimentally and theoretically. As far as we know, two different groups reported the temperature dependence of k2 by relative4 and absolute5 methods between 250 and 430 K, which are in reasonable agreement. Additionally, Hurley et al.6 reported a relative room temperature k2 of (1.02 ± 0.10) × 10−13 cm3 molecule−1 s−1, in good agreement with previous data.4,5 The theoretical temperature dependence of k2 has been reported by Wang et al.7 using a dual-level dynamics method. In this work, the first measurement of k1 is reported as a function of temperature (T = 263−358 K) by the pulsed laser photolysis/laser induced fluorescence (PLP-LIF) technique. Special Issue: A. R. Ravishankara Festschrift Received: November 20, 2011 Revised: January 20, 2012 Published: January 23, 2012
(2) 6041
dx.doi.org/10.1021/jp2111633 | J. Phys. Chem. A 2012, 116, 6041−6050
The Journal of Physical Chemistry A
Article
absorption cross sections were described in detail earlier.11−17 The UV apparatus consisted of a deuterium lamp (source of continuous UV radiation, 180−500 nm), a gas cell with an optical path length (S ) of 107 cm, and a 0.5-m focal length spectrograph coupled to a cooled coupled-charged device (CCD). The transmitted radiation (I) through the cell was dispersed in a 300 grooves/mm grating and detected in the cooled CCD at 253 K. The cooling of the CCD increases the signal-to-noise ratio by reducing the noise due to the dark current. UV absorption spectra were registered after 6000 accumulations under stationary conditions. The absorption cross sections at each wavelength (in base e), σλ, were obtained by applying the Beer−Lambert law:
Absolute absorption cross sections for TFPO and PFPO in the ultraviolet (UV) and infrared (IR) regions have also been determined in this work. For both FPOs, UV absorption cross sections (σλ) are determined between 195 and 360 nm. σλ values together with the calculated spectral actinic flux (Fλ) by the TUV model8 allow the estimation of the photolysis rate, J, in the actinic region (λ > 290 nm) if the photolysis quantum yields (Φλ) are known: J≅
∑ σλΦλFλΔλ
(3)
Lifetimes (τ) of TFPO and PFPO due to UV photolysis and their reaction with OH radicals can be then estimated from J and k1 and k2 (hereafter referred as ki), respectively. Additionally, the integrated IR absorption cross sections (Sint) at 10 cm−1 intervals were determined between 500 and 4000 cm−1 and used to calculate their radiative forcing (RF) using the narrow-band radiative forcing model reported by Pinnock et al.9 and Elrod.10 Hence, an assessment of the atmospheric implications of both the UV and IR absorption process by the title fluoropropanols (FPOs) and their reaction with OH radicals is reported here. Potential atmospheric degradation routes for the title FPOs are also discussed.
Aλ = −ln(I /I0)λ = σλS[FPO]
(4)
where Aλ is the absorbance and I0 and I are the transmitted light intensity measured at a given wavelength λ in the absence and in the presence of FPO, respectively. The pressure of FPO in the UV cell was varied between 1.8 and 9.6 Torr. A series of three to four independent experiments was carried out at each FPO concentration. Additionally, the most intense Hg line, 253.7 nm, of a pen-ray Hg/Ar lamp (Oriel) was also used to measure the single absorption cross section at that wavelength. The absorption cell described above was also employed in these measurements. A photodiode detector (Oriel, detection range 200−1100 nm) together a picoammeter (Keithley, model 6485) was used for detecting the transmitted light intensities I0 and I at 253.7 nm. An interference filter (Andover; centered at 254 nm with a fwhm of 10 nm) was mounted in front of the photodiode. The FPO concentration used in these measurements was limited by the maximum pressure by which the absorption cell could be filled (ca. 10 Torr, 3.2 × 1017 cm−3). The second experimental setup consists of a FTIR spectrometer with a Globar lamp, which was used as a source of IR radiation, a liquid nitrogen cooled mercury cadmium telluride (MCT) detector with Blackman-Harris 3-Term apodization, and a single pass stainless steel cell (FTIR.com, Mercury series) sealed with ZnSe windows. The diameter of the IR cell is 37.5 mm, and the optical path length (S ) is 10 cm. In some measurements, a deuterated triglycine sulfate (DTGS) detector was used to record the IR spectra of both FPOs. IR absorption spectra were recorded under static conditions by accumulation of 16 spectra. The absorption cross sections at each ν̃, σ(ν̃), were determined from the slope of the corresponding absorbance, A(ν̃),
2. EXPERIMENTAL SECTION For ease of presentation, this section has been divided into four subsections. The first subsection describes the fractional distillation of these FPOs and subsequent analysis by gas chromatography coupled to a mass spectrometry (GC-MS) technique. This procedure was carried out prior to the absorption and kinetic experiments in order to identify and quantify the impurities, if present, which may interfere in the determination of the absorption cross sections and the OH rate coefficients. The experimental systems used in the absorption and kinetic measurements are briefly described in the subsequent subsections. These setups have been described in detail elsewhere.11−19 2.1. Pretreatment and GC-MS Analysis of FPO Samples. Initial purities of the liquid samples were >99% for CHF 2 CF 2 CH 2 OH (Sigma-Aldrich) and >99% for CF3CF2CH2OH (Fluka). For the absorption measurements, high levels of a strong absorber may interfere in the determination of UV and IR absorption cross sections. On the other hand, the presence of a very reactive impurity, even at low concentrations, may have a large effect on the measured rate coefficients, because of the low reactivity of perfluoroalcohols, such as CF3CF2CH2OH, or low hydrogenated fluoroalcohols, such as CHF2CF2CH2OH. Therefore, the liquid samples of FPOs were subjected to a fractional distillation. The measured boiling points of CHF 2 CF 2 CH 2 OH and CF 3 CF 2 CH 2 OH at 709 mmHg were 73 and 92 °C, respectively. The distilled fraction was then separated by headspace gas chromatography (Thermo Electron Co., model Trace GC Ultra) and analyzed by a mass spectrometer (Thermo Electron Co., model DSQ II). The ratio of the area of the main peak of each FPO over the total area was taken as the purity of the gas sample. The purities of the TFPO and PFPO samples were found to be 99.91% and 99.8%, respectively. Only ethanol (0.09%) was identified as an impurity of TFPO. 2.2. Gas-Phase UV and IR Absorption Spectroscopy. The experimental setups employed to determine the UV and IR
A(ν̃) = −log T (ν̃) = σ(ν̃)S[FPO]
(5)
where T(ν̃) is the transmittance at that wavenumber. In the IR region, it is common to report integrated absorption cross section, Sint, at a certain wavenumber range. Sint values were obtained by applying the Beer−Lambert law from the integrated absorbance, Aint: A int =
ν̃
∫ν̃ 2 A ν̃ dν̃ = Sint S[FPO] 1
(6)
Sint values are determined for selected IR bands and every 10 cm−1 in order to use them in the GWP calculation for TFPO and PFPO. Figure S1 of the Supporting Information shows Beer−Lambert law plots (eq 6) for PFPO and TFPO at several 10 cm−1 ranges. The wavenumbers shown in Figure S1 are those at the center of the wavenumber range. 6042
dx.doi.org/10.1021/jp2111633 | J. Phys. Chem. A 2012, 116, 6041−6050
The Journal of Physical Chemistry A
Article
concentration of FPO for a given temperature and total pressure. Under these experimental conditions, ki′ is defined by the following expression:
Pure TFPO was used (1.0 and 4.3 Torr) to record the IR absorption spectra. However, diluted mixtures of PFPO in helium (He) were necessary to avoid a deviation of the Beer− Lambert law. The dilution factor f (f = pPFPO/{pHe + pPFPO}) for PFPO was 1.8 × 10−2. In the IR absorption cell the total pressure ranged from 15 to 91 Torr. Therefore, the concentration of FPOs is in the range from 3.2 × 1016 to 1.5 × 1017 molecules cm−3 for TFPO and from 9 × 1015 to 5.4 × 1016 molecules cm−3 for PFPO. 2.3. Gas-Phase Kinetics of the OH Reactions. The experimental system employed in the kinetic study of the OH reactions with the title FPOs has been previously described.12−16,18,19 Diluted FPOs, OH-precursor, and helium flowed through a jacketed reaction cell (V = 200 cm3). Through the external jacket, ethanol or water was flowed to get temperatures below and above 298 K, respectively. The temperature in the reaction cell was monitored by a chromelalumel thermocouple (±1 K). All gases were introduced in the reactor by means of calibrated mass flow controllers. Photolysis of H2O2 and HNO3 at 248 nm by a KrF excimer laser (Optex, Lambda Physik) was the source of the transient species, OH radicals: H2O2 + hν(λ = 248nm) → 2OH
(7)
HNO3 + hν(λ = 248nm) → OH + NO2
(8)
k i′ = k i[FPO] + k 0
where k0 (= k10[precursor] + k11) is the pseudofirst order rate coefficient due to reactions 10 and 11. k0 is determined from the decay of SOH recorded in the absence of FPO. At each temperature and pressure, the rate coefficients k i were determined from the slope of the plot of all measured ki′ − k0 versus [FPO]. k i′ − k 0 = k i[FPO]
(9)
−2
where N is the number of photons cm at 248 nm. Upper limits for the precursor were estimated to be [H2O2]0 < (0.9− 1.0) × 1014 molecules cm−3 and [HNO3]0 < (0.5−3.7) × 1015 molecules cm−3 as previously described.21 The H2O2 and HNO3 bubblers were kept at room temperature. In some experiments, the HNO3 bubbler was kept in a bath at 0 and 8 °C to reduce its vapor pressure and, consequently, its concentration in the reaction cell. Upper limits of [OH]0 ranged from 7 × 1010 to 4.9 × 1011 molecules cm−3. Generated OH radicals may react with both the photochemical precursor of OH radicals and the FPO and/or diffuse out of the detection zone: OH + FPO → products
ki
OH + precursor → products
k10
(10)
OH → loss
k11
(11)
(13)
Equation 13 was preferred since each ki′ was corrected with k0. FPO concentrations were determined by FTIR spectroscopy at the exit of the reactor and from measurements of the mass flow rates, total pressure, and temperature in the reaction cell. Total gas flow rates were measured using calibrated mass flow transducers and ranged from 110 to 540 sccm (standard cubic centimeter per minute). The linear gas flow velocity in the kinetic reaction cell ranged between 2 and 5.7 cm s−1, which was adequate to avoid accumulation of photoproducts between photolysis laser shots. Pressures were measured using 100-Torr capacitance manometers. The temperature of the gas in the reaction zone was measured with a thermocouple inserted directly in the gas flow (±1 K). Four IR bands were used to average the gas-phase concentration of FPOs, and the average was used to compare with those obtained from flow measurements. Both concentrations are in good agreement (within ±10%); therefore, the second method was preferred. The FPO concentration ranges were (0.1−1.9) × 1015 molecules cm−3 for TFPO and (0.1−2.4) × 1015 molecules cm−3 for PFPO. In a few experiments concentrated mixtures of FPOs were used to minimize the large contribution of k0. TFPO was placed in a bubbler, and a He flow (up to 12 sccm) was flown through it to get higher concentrations in the reaction cell. The TFPO concentrations measured by FTIR spectroscopy in these experiments ranged from 1.0 × 1016 molecules cm−3 to 2.7 × 1016 molecules cm−3. The use of concentrated mixtures of FPOs was limited by two key factors: (i) In the storage bulb, a limiting factor was the maximum partial pressure of FPO introduced in the 10-L bulb to prepare the gas mixture together with the optimum total pressure (pHe + pPFPO) required to make the mass flow controllers work efficiently. (ii) When concentrations of FPOs (on the order of 1016 cm−3) were introduced in the reaction cell by means of a glass bubbler, the quenching of the OH LIF was too fast to accurately monitor the OH temporal profile. For that reason, these data were disregarded. 2.4. Materials. He (99.999%) was used as supplied without further purification. Hydrogen peroxide was preconcentrated prepared by bubbling He through a H2O2 (Sharlab) sample initially at 50% (w/v) concentration for several days prior to use as described in ref 22. Aqueous solution of HNO3 (65%, Panreac) was not preconcentrated as H2O2. During the kinetic measurements, the solution of H2O2 and HNO3 was constantly bubbled in order to introduce the gas phase precursor into the reaction cell. Commercial fluoroalcohols used in this investigation were CHF2CF2CH2OH (>99%, Sigma-Aldrich) and CF3CF2CH2OH (>99%, Fluka). Distilled fluoroalcohols were degassed by repeated freeze−pump−thaw cycles.
The upper limit for the initial OH concentration, [OH]0, was estimated by using the measured laser fluence, E248nm (4.7− 9.4 mJ pulse−1 cm−2), the absorption cross section of the precursor at 248 nm (σ248nm), the quantum yield for OH production at the photolysis wavelength from H2O2 or HNO3 (ϕ248nm = 2 and 0.97, respectively),20 and the concentration of the precursor: [OH]0 = N σ248nmϕ248nm[precursor]0
(12)
(1−2)
Under pseudofirst order conditions with respect to the initial OH concentration ([FPO]0 and [precursor]0 ≫ [OH]0; [FPO]0/[OH]0 = 300−14 000), temporal profiles of the LIF signal from OH radicals, SOH, would follow a single exponential rate law, if no other side reactions take place. In Figures S2 and S3 of the Supporting Information, some examples of the OH LIF signal (SOH in log scale) versus reaction time are presented at several temperatures and FPO concentrations. The pseudofirst order rate coefficient (ki′) was obtained from the nonlinear least-squares analysis of such exponential curves at each 6043
dx.doi.org/10.1021/jp2111633 | J. Phys. Chem. A 2012, 116, 6041−6050
The Journal of Physical Chemistry A
Article
3. RESULTS AND DISCUSSION This section has been divided into several subsections to separate the results and discussion of the absorption cross sections of TFPO and PFPO (Subsection 3.1) from those of the temperature dependence of the OH-rate coefficients for reactions 1 and 2 presented in Subsection 3.2. A comparison with previous results on other fluorinated alcohols is also included in both subsections. 3.1. UV and IR Absorption Cross Sections. An example of the UV absorption spectra between 195 and 360 nm is shown for both FPOs in Figure 1a,b for TFPO and PFPO,
the photolysis rate J for these FPOs is expected to be negligible in the actinic region. Figure 2a,b shows the averaged IR absorption cross sections as a function of wavenumber for TFPO and PFPO, respectively. The integrated absorption cross sections for some IR selected bands in base e, Seint, are listed for both FPOs in Table 1. The absorption spectra were a bit noisy below 600 cm−1; thus, total Seint are reported to be (1.92 ± 0.34) × 10−16 cm2 molecule−1 cm−1 for TFPO and (2.09 ± 0.37) × 10−16 cm2 molecule−1 cm−1 for PFPO between 4000 and 600 cm−1. Uncertainties are at
Figure 2. Averaged IR absorption cross sections of CH2FCF2CH2OH (a) and CF3CF2CH2OH (b) overlapped with the calculated radiative forcing in the atmospheric window.
Table 1. Integrated Absorption Cross Sections, Seint (×10−17 cm2 molecule−1 cm−1), for Some Selected IR Bands for the Studied Fluoropropanols Figure 1. Example of the UV absorption spectrum for (a) CHF2CF2CH2OH, p = 6.5 Torr, and (b) CF3CF2CH2OH, p = 9.6 Torr. The solar actinic flux (dashed line) at the surface is overlapped to the spectra.
Δν̃, cm−1a CHF2CF2CH2OH 710−650 855−815 965−900 1500−1000 3040−2870 3700−3630 CF3CF2CH2OH 700−565 825−700 1500−850 3040−2840 3710−3620
respectively. The low absorption of TFPO and PFPO in this spectral region (
