9118
`
`J. Phys. Chem. A 1997, 101, 9118-9124
`
`Rate Constants for the Reactions of OH with HFC-245cb (CH3CF2CF3) and Some
`Fluoroalkenes (CH2CHCF3, CH2CFCF3, CF2CFCF3, and CF2CF2)
`
`Vladimir L. Orkin,*,† Robert E. Huie, and Michael J. Kurylo
`Physical and Chemical Properties DiVision, National Institute of Standards and Technology,
`Gaithersburg, Maryland 20899
`ReceiVed: June 18, 1997; In Final Form: September 15, 1997X
`
`The rate constant for the reaction of OH with HFC-245cb (CH3-CF2-CF3) was measured by the flash
`photolysis resonance fluorescence technique over the temperature range 287-370 K in order to ascertain its
`atmospheric lifetime. Given the potential for trace levels of olefinic impurities to introduce errors into results
`of the rate constant measurements for low reactivity HFCs, considerable emphasis was placed on HFC
`purification and on demonstrating the utility of vacuum UV spectroscopy as a sensitive tool for detecting
`olefinic impurities in HFC samples. Measurements were also made of the OH rate constants for CH2dCH-
`CF3, CH2dCF-CF3, and CF2dCF-CF3, over the temperature range 252-370 K, and for CF2dCF2 at 298
`K. Absorption spectra for the four fluoroalkenes as well as for ethene and propene were obtained from 160
`to 220 nm.
`
`Introduction
`The internationally legislated elimination of chlorofluorocar-
`bons (CFCs) from industrial applications, due to the established
`danger that they pose to Earth’s ozone layer, has stimulated
`considerable study of the atmospheric properties of possible
`chemical substitutes. Partially fluorinated hydrocarbons (hy-
`drofluorocarbons or HFCs) are among the leading environmen-
`tally acceptable CFC alternatives from the point of view of
`ozone depletion. Nevertheless, quantification of the possible
`role of HFCs as “greenhouse gases” requires accurate informa-
`tion on their atmospheric lifetimes, which are key parameters
`in determining the environmental consequences following their
`release into the atmosphere. HFCs typically have very low
`absorption cross sections at the wavelengths of solar radiation
`that penetrate the troposphere, and thus, their atmospheric
`lifetimes are dictated primarily by their reactivity with tropo-
`spheric hydroxyl radicals. This has provided the motivation
`for the present study of the OH reactivity toward HFC-245cb
`(CH3-CF2-CF3), which is currently a candidate for replacement
`of CFC-12 (CF2Cl2) in some refrigeration applications.
`Highly fluorinated alkanes are usually not very reactive
`toward OH, and consequently, the determination of the absolute
`rate constants can be easily compromised by the presence of
`reactive microimpurities in the samples studied. Fluorinated
`alkenes are the most likely reactive impurities in HFCs, being
`present as residual starting material from production or as
`decomposition products. Gas chromatography (GC) has been
`and currently is the principal technique for determining sample
`purity but, despite its high sensitivity, can have difficulties
`associated with incomplete separation of certain compounds.
`Thus, the possibility always exists that some olefinic impurities
`may not be chromatographically resolved from the main
`compound and,
`therefore, remain undetected. Unsaturated
`hydrocarbons, however, manifest very strong absorption bands
`in the far UV spectral region, thereby offering the possibility
`of using UV absorption as a sensitive diagnostic for detecting
`such impurities in nonabsorbing gases. We measured UV
`absorption spectra of several fluorinated propenes, perfluoro-
`
`† Also associated with the Institute of Energy Problems of Chemical
`Physics, Russian Academy of Sciences, Moscow, 117829.
`X Abstract published in AdVance ACS Abstracts, November 15, 1997.
`
`ethene, propene, and ethene to determine the effect of fluorina-
`tion on the UV absorption properties of alkenes and to ascertain
`the possible role of far UV absorption spectroscopy as an
`analytical supplement for determining the presence of unsatur-
`ated impurities in HFC samples.
`To quantify the uncertainty in the rate constant due to any
`remaining olefinic impurities, we also measured the OH
`reactivity of several fluoroalkenes. The gas-phase reactions of
`hydroxyl radicals with alkenes are of both fundamental and
`practical importance. OH-alkene reactions follow a complex
`mechanism involving adduct formation near room temperature,
`with direct hydrogen atom abstraction becoming increasingly
`important at higher temperatures.1 While there are many studies
`of OH reactions with a variety of unsaturated hydrocarbons,
`only limited information is available for fluorinated alkenes.
`For the ethenes, CH2dCHF2 and CH2dCF2
`3 have slightly lower
`OH reactivity than that of ethylene.4 However, in the case of
`hexafluoropropene (CF2dCF-CF3),5 the reactivity towards OH
`is more than an order of magnitude less than that of propene
`itself.4 There are no other studies of fluoroalkene reactivity.
`Hence, we present in this paper results of our measurements of
`rate constants for the OH reactions with three fluorinated
`propenes (CH2dCH-CF3, CH2dCF-CF3, and CF2dCF-CF3)
`and CF2dCF2.
`
`Experimental Section19
`
`Detailed descriptions of the apparatuses and the experimental
`methods used to measure the rate constants for the reactions of
`HFC-245cb and the various fluorinated olefins with OH and
`the olefin absorption spectra are given in previous papers.6-9
`Therefore, only brief descriptions are given here.
`OH Reaction Rate Constant Measurements. The principal
`component of the flash photolysis-resonance fluorescence
`(FPRF) apparatus is a Pyrex reactor (of approximately 50 cm3
`internal volume) thermostated with water or methanol circulated
`through its outer jacket. Reactions were studied in argon carrier
`gas (99.9995% purity) at a total pressure of 13.33 kPa (100.0
`Torr). Flows of dry argon, argon bubbled through water
`thermostated at 276 K, and HFC or fluoroalkene mixtures
`(containing 0.1-0.2% of the reactant diluted with argon or C2F6)
`were premixed and flowed through the reactor at a total flow
`
`S1089-5639(97)01994-4 This article not subject to U.S. Copyright. Published 1997 by the American Chemical Society
`
`Page 1 of 7
`
`Arkema Exhibit 1124
`
`

`
`Reactions of OH with HFC-245cb
`
`J. Phys. Chem. A, Vol. 101, No. 48, 1997 9119
`
`TABLE 1: Rate Constants Measured in the Present Work for the Reactions of OH with HFC-245cb and Several
`Fluoroalkenesa
`
`temp, K
`252
`277
`287
`298
`313
`330
`349
`370
`
`1.02 ( 0.05 (3)
`
`ki, 10-12 cm3 molecule-1 s-1
`CH2dCF-CF3
`CF2dCF-CF3
`CH2dCH-CF3
`1.10 ( 0.05 (1)
`2.83 ( 0.14 (2)
`1.72 ( 0.08 (1)
`1.10 ( 0.03 (3)
`2.50 ( 0.02 (3)
`1.59 ( 0.08 (2)
`1.23 ( 0.04 (1)
`1.54 ( 0.04 (3)
`1.12 ( 0.02 (6)
`2.17 ( 0.01 (5)
`1.54 ( 0.05 (4)
`2.00 ( 0.05 (1)
`2.54 ( 0.02 (3)
`1.15 ( 0.04 (2)
`1.88 ( 0.04 (3)
`1.43 ( 0.03 (1)
`3.44 ( 0.04 (2)
`4.67 ( 0.11 (3)
`1.18 ( 0.05 (4)
`1.73 ( 0.04 (4)
`1.37 ( 0.03 (2)
`a Error bars are levels of confidence of 95% and do not include estimated systematic errors. Numbers in parentheses indicate the number of
`experimental measurements.
`
`CF2dCF2
`
`k245cb, 10-15 cm3 molecule-1 s-1
`CH3-CF2-CF3
`
`rate between 0.5 and 1.2 cm3 s-1, STP. The concentrations of
`the gases in the reactor were determined by measuring the mass
`flow rates and the total pressure with a MKS Baratron
`manometer. Flow rates of both argon and the H2O/argon
`mixture were measured by calibrated Tylan mass flow meters,
`whereas those of the reactant/inert gas mixtures were determined
`by direct measurements of the rate of pressure change in a
`calibrated volume. Hydroxyl radicals were produced by the
`pulsed photolysis (1-4 Hz repetition rate) of H2O (introduced
`via the 276 K argon/H2O bubbler) by a xenon flash lamp focused
`into the reactor. The radicals were then monitored by their
`resonance fluorescence near 308 nm, excited by a microwave-
`discharge resonance lamp (280 Pa or 2.1 Torr of a ca. 2%
`mixture of H2O in UHP helium) focused into the reactor center.
`The resonance fluorescence signal was recorded on a computer-
`based multichannel scanner (channel width 100 (cid:237)s) as a
`summation of 1000-10 000 consecutive flashes. The radical
`decay signal at each reactant concentration was analyzed as
`described by Orkin et al.7 to obtain the first-order decay rate
`-1).
`due to the reaction under study ((cid:244)i
`UV Absorption Cross-Section Measurements. The absorp-
`tion spectra of the pure alkenes and the gas mixtures (i.e., HFC
`with alkene impurities) to be analyzed were measured over the
`wavelength range of 160-220 nm using a single-beam apparatus
`consisting of a 1 m vacuum monochromator equipped with a
`600 lines/mm grating. The radiation source was a Hamamatsu
`L1385 deuterium lamp, and the detector was a Hamamatsu R166
`photomultiplier. Spectra were recorded at increments of either
`0.5 or 0.1 nm at spectral slit widths of 0.5 and 0.1 nm. The
`pressure inside the 16.9 ( 0.05 cm absorption cell was measured
`by a MKS Baratron manometer at T ) 295 ( 1 K. Absorption
`spectra of the evacuated cell and of the cell filled with a gas
`sample were alternately recorded several times, and the absorp-
`tion cross sections were calculated from the differences. The
`complete spectra were constructed from data taken over several
`overlapping wavelength ranges. Data over each range were
`obtained at several pressures to verify adherence to the Beer-
`Lambert absorption law. The overall instrumental error associ-
`ated with uncertainties in the path length, pressure, temperature
`stability, and the measured absorbance was estimated to be less
`than 2% over most of the wavelength range, increasing to
`approximately 5-10 % at the long-wavelength ends of the
`spectra. Mixtures containing 2%, 10%, and 100% of the olefins
`under study were used at pressures in the cell ranging from 4
`Pa to 120 kPa (0.03-900.0 Torr).
`Sample Purity. All samples were first analyzed after several
`freeze/pump/thaw cycles by GC and GC/MS techniques.
`Hexafluoropropene, CF2dCF-CF3 (PCR Corp.), was 98.8%
`purity with the identified impurities being C3F8 (0.31%), CO2
`(0.28%), and CF2dCH-CF3 (ca. 0.21%). The sample of 3,3,3-
`trifluoropropene, CH2dCH-CF3 (Pfaltz & Bauer, Inc.) had no
`
`detectable impurities after outgassing. The original sample of
`tetrafluoroethene, CF2dCF2 (Union Carbide Corp.), was ca. 98%
`purity stabilized with ca. 1% of R-terpinene. Hexafluoro-
`cyclopropane, C3F6 (ca. 1%), and CO2 (0.2%) were found as
`the main impurities after purification of the original sample by
`passing through a -100 (cid:176)C cold trap to remove the inhibitor.
`Samples of propene for UV spectrum measurements were
`obtained from Matheson (>99.6% purity, with 0.38% of propane
`as a main impurity) and Philips Petroleum Co. (>99.9% purity,
`with propane as a main impurity). The sample of ethene was
`obtained from Philips Petroleum Co. (>99.99% purity, with
`methane and ethane as the main impurities).
`A sample of HFC-245cb (CH3-CF2-CF3), obtained from
`PCR Corp., had a stated purity of ca. 99%. Nevertheless, the
`sample was found to contain CH2dCF-CF3 (ca. 6.5%), CH3F
`(0.34%), CH3CF3 (0.2%), CO2 (0.03%), CHF2CHF2 (0.11%),
`(This original “mixture” was actually
`and CH2F2 (0.013%).
`used to measure the rate constant for the reaction between
`2,3,3,3-tetrafluoropropene and OH as well as to determine its
`UV cross sections.) A bromine purification/titration technique
`described in the Appendix was employed to prepare samples
`of CH3-CF2-CF3 of acceptable purity for kinetic measurements
`and also to quantify the concentration of CH2dCF-CF3. UV
`absorption measurements were used to analyze the residual
`olefin concentration following bromine purification. One puri-
`fied sample was sent out for independent analysis at Allied
`Signal Inc. using GC and GC/MS techniques.
`
`Results and Discussion
`
`Rate constants for the reactions of hydroxyl radicals with the
`four fluorinated alkenes at different temperatures are listed in
`Table 1 as well as those for the reaction with CH3-CF2-CF3.
`The data were fit to the Arrhenius equation, and the resulting
`parameters are given in Table 2. The data are plotted in
`Arrhenius form in Figure 1 (fluoropropenes) and Figure 2 (HFC-
`245cb). Also presented in Table 2 and Figure 1 are the only
`available literature data (for CF2dCF-CF3).
`Reactions of OH with Fluoroalkenes. The rate constants
`for the reaction between OH and CF2dCF-CF3 measured in
`the present work agree reasonably well with those from McIlroy
`and Tully,5 the only available study for comparison. Differences
`range from 5 to 15% over the common temperature range. Note
`that the results of these investigators over an extended temper-
`ature range5 show non-Arrhenius rate constant behavior, similar
`to that for other OH/alkene reactions.1 No other measurements
`of the reactions between OH and fluorinated propenes are
`available for comparison.
`All of the fluoropropenes studied here contain the -CF3
`group, but differ in the extent of fluorine substitution at the
`olefinic carbon atoms. From Table 3 it can be seen that the
`
`Page 2 of 7
`
`

`
`9120 J. Phys. Chem. A, Vol. 101, No. 48, 1997
`
`Orlrin et al.
`
`TABLE 2: Arrhenius Parameters for Reactions of OH with Fluoroalkenes and HFC-245ch'
`
`molecule
`
`A X 10“, cm’ molecule” s‘1
`
`E/R :1: AE/R, K
`
`km=c(298) x 1012 cm’ molecule” s‘1
`
`ref
`
`CHz=CH—CF;
`CHz=CF—CF3
`CI-‘2=CI-‘—CF3
`Cl-'2=CF2
`
`313:3};
`141:’: 3
`5_55jiE
`333:1 2b
`
`-183 :1: 26
`64 i 27
`-407 :t 85
`-281 :l: 49"
`
`1.52 :l: 0.02
`1.13 i 0.02
`2.21 :|: 0.08
`2.28 :l: 0.05"
`10.2 2|: 0.05
`
`this work
`this work
`this work
`Mcllroy and Tullys
`this work
`
`3
`
`o cr,=cr—cr,
`o cr2=cr-cr, (m.5)
`- cH2=cn-cr,
`o cH2=cr-cr,
`
`—I
`
`',
`I-I
`to
`:1O
`.9.0
`
`s? 1
`
`this work
`(1.52 :1: 0.04) x 10-3
`1690 1 60
`4_41:g;g
`Cl-I3—CF2—CF3
`‘Error bars are levels of confidence of 95% and do not include estimated systematic errors. ‘ An Arrhenius fit to the data over T = 293-378
`K only. The rate constant was measured by Mcllroy and Tullys over T = 293-831 K.
`TABLE 3: Rate Constants at 298 K for Radical Addition to
`Fluorinated Propenes and Ethenes (Relative to the Propylene
`and Ethylene, Respectively)‘
`addition of
`O atoms
`
`alkene
`
`addition of
`OH radicals
`
`addition of
`H atoms
`
`CHz=CH—CH3
`CHz=CF—CH3
`CF2=CH—CH3
`CHz=CH—CH2F
`CHz=CH—C1-'3
`CHz=CF—CF3
`CI-'2=CF—CI-'3
`CHz=CH2
`CHz=CI-IF
`CHz=CF2
`CI-IF=C1-IF
`CHF=Cl-'2
`CF2=CF2
`
`1.0
`0.5
`0.45
`0.21
`0.01
`
`0.007
`1.0
`0.5
`0.3
`0.4
`0.6
`1.4
`
`1.0
`
`0.051
`0.039
`0.072
`1.0
`0.61
`0.44
`
`1.13
`
`1.0
`0.95
`
`0.67
`0.3
`
`<0.02
`1.0
`0.44
`0.33
`
`0.06
`0.07
`
`' Evaluated on the basis of results of published studies listed in the
`NIST Kinetics Database“ as well as on results obtained in the present
`wmk.
`
`The fluoropropenes may also be considered as derived from
`ethene and fluoroethenes by substitution of CF; for H or F.
`Going from CH2=CH2 to CH2=CH—CF3 results in a reduction
`of the rate constant by a factor of 6; going from CHz=CI-IF to
`CH2=CF—CF3 lowers the rate constant by a factor of 5
`kc_.,n, g 0.3‘). Replacing the fluorine in CH2=CHF. CH2=CF2.
`and CFz=CF2 by CF; to make CHz=CH-CF3. CHz=CF-CF3.
`and CFz=CF-CF3 results in reducing the reactivity by factors
`of approximately 4.
`The effects of fluorine substitution in alkenes on their OH
`
`addition reactivity are very similar to those for O atom addition.
`This can be seen from Table 3. in which relative rate constants
`for the reactions between 0 and various fluoropropenes and
`fluoroethenes are also presented. Fluorination of any olefinic
`carbon results in small reactivity decreases (less than a factor
`of 2) in both the propenes and ethenes. However. fluorination
`of the -CH3 group in the propenes results in a more pronounced
`decrease in O atom reactivity (a factor of 5 in the case of -CH2F
`and 2 orders of magnitude in the case of -CF3). Thus. results
`of both the OH radical and O atom addition reactions studies
`
`suggest that fluorination of olefinic carbons does not afl'ect
`Jr-electron density in the double bond very much. in contrast
`with fluorination of the carbon adjacent to the olefinic group.
`Relative reactivity data for H atom addition to fluorinated
`ethenes and propenes are also presented in Table 3.
`In these
`cases, the reactions appear to be very sensitive to complete
`fluorination of the olefinic carbon atoms. with more moderate
`decreases in reactivity caused by fluorination of the -CH3 group
`in propylene.
`Reaction of OH with HFC-245eb (CH3—CF;—CF3). As
`discussed earlier. the purification procedure used in this work
`resulted in CH3-CF2-CF3 samples of suficient purity for
`accurate rate constant determinations. Even for the lowest
`
`N-1
`I
`
`EU
`
`O-1
`
`2.5
`
`3.0
`
`3.5
`
`4.0
`
`1000/T. K-1
`Figure I. Arrhenius plots of rate constants for the ractions of OH
`with fluoropropenes. Solid lines are the least-square lit: to our data;
`dashed lines are their 95% confidence intervals.
`
`£0
`
`8
`
`-I-l
`
`
`
`10-15crnarnolecule
`
`k245oh'
`
`2.5
`
`3.0
`
`8.5
`
`1000/T. K"
`Figure 2. Arrhenius plot of the measured k245d, values and the least-
`squares lit to our data (solid line) with its 95% confidence intervals
`(dashed hues).
`
`presence of the -CF; group results in more than an order of
`magnitude decrease in room—temperature reactivity toward OH
`relative to that for propylene. Fluorination of the olefinic carbon
`atoms results in only small additional changes in reactivity.
`similar to the behavior observed for the fluorinated ethenes.”
`
`Specifically. CHz=CF—CF3 is slightly less reactive than
`CH1=CH-CF3. while CF2=CF-CF; is slightly more reactive.
`
`residual reactive
`temperature rate constant measurements.
`impurities in the samples were at concentrations low enough to
`
`Page 3 of 7
`
`

`
`Reactions of OH with HFC-245cb
`
`J. Phys. Chem. A, Vol. 101, No. 48, 1997 9121
`
`avoid any significant overestimation of the measured rate
`constant. For example, the fluoropropene rate constants reported
`here indicate that a 0.002% residual olefinic impurity could
`contribute a maximum of 2% to our measured room-temperature
`rate constant for HFC-245cb. Recall that the 0.002% upper
`limit for residual olefinic impurities was verified for the samples
`by UV analyses. The purified sample analyzed at Allied Signal
`Inc. contained the highest level of residual unsaturated impuri-
`ties, a finding confirmed by our UV measurements. Neverthe-
`less, even the ca. 0.006% of residual tetrafluoropropene found
`in that sample could result in a maximum rate constant error of
`approximately 6% at room temperature. The order of magnitude
`lower concentration of another fully halogenated unsaturated
`compound (C3F4Br2) would have virtually no effect on the
`measured rate constant. The reactivity of CH2Br-CFBr-CF3
`toward OH should be similar to the reactivity of CH2Br-CF3
`10
`and therefore is sufficiently low that any possible error due to
`the trace remaining concentration of this primary product of
`bromination (0.012% as determined in the Allied Signal
`analysis) is negligible.
`There are no published data on the rate constant for the
`reaction of OH with HFC-245cb with which we can compare
`the present results. The A factor obtained from an Arrhenius
`fit to the data (4.4 (cid:2) 10-13 cm3 molecule-1 s-1) is smaller than
`obtained for other OH abstraction reactions involving a -CH3
`group in haloalkanes. However,
`there is no statistically
`significant curvature in the temperature dependence obtained
`which could result in the calculation of both low A and E/R
`factors. The obtained temperature dependence, E/R ) 1692 K,
`is similar to that reported10 for other F5 and higher fluorinated
`propanes with the exception of HFC-236fa (CF3-CH2-CF3).
`The E/R for HFC-236fa is approximately 2500 K, while E/R
`values for other HFC-236 isomers and several HFC-235 isomers
`range from 1330 to 1750 K.10
`We can compare the HFC-245cb rate constant with that for
`the reaction of OH with CH3-CF3 (HFC-143a), for which we
`would expect similar reactivity. However, it is first instructive
`to briefly examine the existing literature data for OH + HFC-
`143a. The Arrhenius parameters derived from a fit to our earlier
`HFC-143a data (A ) 0.95 (cid:2) 10-12 cm3 molecule-1 s-1; E/R )
`1980 K)7 differ slightly from the presently recommended
`parameters (A ) 1.8 (cid:2) 10-12 cm3 molecule-1 s-1; E/R ) 2170
`K) which are based on results of two absolute7,11 and one
`relative12 reaction studies. All the data points are in good
`agreement but result in somewhat different Arrhenius parameters
`derived from the individual data sets. The highest values of
`the A factor and E/R come from the absolute measurements by
`Talukdar et al.,11 who reported k143a ) 2.12 (cid:2) 10-12
`exp(-2200/T) cm3 molecule-1 s-1. These authors derived this
`value from a fitting of the data obtained in two independent
`sets of experiments performed by using different experimental
`techniques over somewhat different temperature ranges. How-
`ever, analysis of the data from the two techniques independently
`results in slightly smaller temperature dependencies and lower
`A factors. For example,
`the flash photolysis/laser-induced
`+0 39 (cid:2)
`fluorescence (FP/LIF) measurements result in A ) 1.01-0 28
`10-12 cm3 molecule-1 s-1 and E/R ) 2015 ( 93 K with the
`average temperature of 289 K, while the discharge flow/laser
`magnetic resonance (DF/LMR) measurements result in A )
`+0 64 (cid:2) 10-12 cm3 molecule-1 s-1 and E/R ) 2005 ( 132
`1.21-0 42
`K with the average temperature of 313 K. The higher values
`of both A and E/R factors reported by authors11 and used in
`data evaluations10 appear to be the result of combined fitting to
`the slightly higher DF/LIF data at higher temperatures together
`with the slightly lower RF/LIF data at lower temperatures.
`
`Figure 3. Ultraviolet absorption cross sections of propylene and
`fluoropropylenes at T ) 295 K.
`
`An Arrhenius fitting of the relative data by Hsu and DeMore12
`with methane as a reference compound using the latest evalu-
`ation10 of the rate constant for the reaction of OH with CH4
`+1 05 (cid:2) 10-12 cm3 molecule-1 s-1 and E/R
`results in A ) 1.04-0 52
`) 1998 ( 231 K, which are both slightly lower than those
`presented in the original paper.12 The relative rate data using
`CHF2-CF3 (HFC-125) as a reference compound12 yielded A
`+0 28 (cid:2) 10-12 cm3 molecule-1 s-1 and E/R ) 2070 ( 68
`) 1.21-0 22
`K. Thus, all of the individual absolute and relative studies of
`the reaction between OH and HFC-143a appear more consistent
`with an E/R of approximately 2000 K and an A factor of
`approximately 1.1 (cid:2) 10-12 cm3 molecule-1 s-1 than with the
`values currently recommended. An unweighted ln(k) vs 1/T
`fitting of all the available data points also results in k143a )
`+0 4 (cid:2) 10-12 exp{-(2030 ( 100)/T) cm3 molecule-1 s-1
`1.15-0 3
`with k143a(298) ) (1.27 ( 0.06) (cid:2) 10-15 cm3 molecule-1 s-1.
`(All uncertainties quoted above are the 95% confidence
`intervals).
`Thus the replacement of the -CF3 group (in HFC-143a) by
`-CF2-CF3 (to form HFC-245cb) results in a decrease in both
`A and E/R. The net result is a room-temperature rate constant
`for HFC-245cb that is actually higher by approximately 20%.
`The comparative effects of CF3 and C2F5 may also be derived
`from studies of the reactions of OH with CF3-CH2-CH2-
`CF3 (HFC-356mff)13,14 and CF3-CF2-CH2-CH2-CF2-CF3
`(HFC-55-10-mcff).13 The ratio of recommended rate constants10
`at room temperature (k55-10-mcff(298)/k356mff(298)) is almost
`identical to that for k245cb(298)/k143a(298). There are no tem-
`perature dependence data for k55-10-mcff available to compare
`changes in Arrhenius parameters. Additional measurements of
`such substituted analogs would be extremely useful to improve
`our understanding of the effects of fluorine substitution on
`molecular reactivity.
`An Atmospheric Lifetime of HFC-245cb (CH3-CF2-
`CF3). As mentioned earlier, reaction with tropospheric hydroxyl
`radicals is the primary removal process for HFCs in the
`atmosphere. Thus, the atmospheric lifetime of HFC-245cb
`((cid:244)245cb) can be estimated using a simple scaling procedure:
`OH ) kMC(277)
`(cid:244)245cb
`k245cb(277)
`OH ) 5.7 years are the atmospheric lifetimes
`where (cid:244)245cbOH and (cid:244)MC
`
`
`
`OH
`(cid:244)MC
`
`Page 4 of 7
`
`

`
`9122 J. Phys. Chem. A, Vol. 101, No. 48, 1997
`
`Orkin et al.
`
`2% cjHz=cH-c1=‘a
`
`E after I>!I.ri!i.°-tisl? Z
`
`165
`
`170
`
`l 75
`
`180
`
`I a
`
`bsorptioncross—aectlon.cmzmolecule
`
`
`
`
`
`wavelength. nm
`
`Figure 6. Absorption spectra of the “artificial” 2% mixture of 3,3,3-
`trifluompropene in CzFs before and alter “bromine” purification
`
`of CH3-CF2-CF3 and methyl chloroform (MC). respectively.
`due to reactions with hydroxyl radicals in the troposphere. and
`k245d,(277). kMc(277) = 6.69 x 10"” cm3 molecule"! 5” are
`the rate constants for the reactions of OH with these substances
`
`at T = 277 K. Such estimation results in raga, = 38.9 years.
`As we have discussed in previous paper.-’ the value of ‘raid,
`thus obtained should be a reasonably good estimate for the total
`atmospheric lifetime of such a well-mixed atmospheric com-
`pormd for which photodissociation by long-wavelength solar
`UV radiation is not appreciable.
`UV Spectra. The UV spectra measured in the present work
`for propene and three fluorinated propenes are presented in
`Figure 3. Propene exhibits an unresolved diffuse band at
`wavelengths below 190 nm. Variation in the instrumental
`spectral resolution by a factor of 5 resulted in no change in the
`spectrum obtained. The measured absorption cross sections
`coincide well (within 10% for I. < 200 nm) with those measured
`by Fahr and Nayak.” The fluorinatzed propenes similarly display
`continuous spectra with the highest absorption occurring at short
`wavelengths.
`Note that maximum absorption cross sections for the fluori-
`nated propenes between 160 and 180 nm exceed 2 x 10’"
`cm2 molecule”, typical for the majority of unsaturated hydro-
`carbons over the 170-180 nm range (see ref 16 for example).
`Nevertheless. fluorination of propene suppresses the short-
`wavelength absorption maximum somewhat and shifts the
`location of the maximrun to shorter wavelength. Perfluorination
`of ethene results in the largest change in maximum absorption
`cross section and a red shift in its location: 0.... g 0.75 x
`10"" cm2 molecule” at 2. 2 188.5 nm vs 0..., 2 6 x 10'"
`cm2 molecule"! at I g 170 nm for CHz=CH2 (see Figure 4).
`As indicated earlier,
`the strong UV absorption of the
`fluorinated alkenes provided a convenient diagnostic for check-
`ing purified samples for the presence of residual olefinic
`impurities. In particular. the high absorption cross section for
`2.3.3.3-tetrafluoropropene (a155...g 3.8 x 10_” crnz molecule_')
`permits its detection at the parts per million level in a 1 atm
`sample in our 17 cm absorption cell (every 1 ppm would result
`in a 1.5% absorption). Thus. detection of trace olefinic
`impurities at the ppm level is easily accomplished even with a
`single beam photometric technique using a relatively short
`absorption cell. Figure 6 shows an example of such application.
`This methodology may prove useful, particularly in industrial
`
`170
`
`180
`
`190
`
`200
`
`210
`
`220
`
`wavelength. nm
`Figure 4. Ultraviolet absorption cross sections of ethylene and
`tetrafluoroethylene at T = 295 K
`
`Irradiation time:
`
`0 min
`5 min
`I0 min
`15 min
`25 min
`60 min
`135 min
`
`350
`
`400
`
`450
`
`500
`
`550
`
`600
`
`wavelength, run
`
`-1-1
`
`a
`
`10'" ..
`
`.,
`
`10'
`
`7'‘ a
`
`
`
`
`..3absorptioncross-section.cmmolecule
`units
`
`absorption.abs.
`units
`
`absorptionat420nm.abs.
`
`O
`
`20
`
`40
`
`60
`
`80
`
`100
`
`120
`
`140
`
`irradiation time. min
`
`Figure 5. Absorption spectra ofa mixture of 13.73 kl’: (103 Torr) of
`an original HFC—245cb sample and 2.00 kPa (15.0 Torr) of Brz during
`irradiation Change of the absorption maximum in Figure la with the
`time. Dashed line is a result of fittingd = 0.56 + 0.43 exp{—t/6.2}.
`
`Page 5 of 7
`
`

`
`Reactions of OH with HFC-245cb
`
`J. Phys. Chem. A, Vol. 101, No. 48, 1997 9123
`
`applications, for analyzing the purity levels of fluorinated
`hydrocarbons being considered as possible substitutes for CFCs
`and halons.
`
`Acknowledgment. This work was supported in part by the
`U.S. Environmental Protection Agency and the Upper Atmo-
`sphere Research Program of the National Aeronautics and Space
`Administration. Although the manuscript has undergone an EPA
`policy review, it does not necessarily reflect the views of or
`infer an official endorsement by the Agency. We would like to
`thank Dr. Hillel Magid (Allied Signal, Inc.) for the analysis of
`one of the purified samples of HFC-245cb as well as for very
`fruitful discussions. We would also like to thank Dr. Thomas
`Buckley (NIST) for his helpful technical support in some of
`the experimental measurements.
`
`Appendix
`Bromination of the HFC-245cb Sample. A chemical
`titration of the olefinic group (CdC) with Br2 was used to
`determine the CH2dCF-CF3 concentration in mixtures used
`for both the OH reaction and UV spectrum measurements. For
`this determination, an HFC-245cb sample containing CH2dCF-
`CF3 was added to a 7.5 cm long, 1.9 cm i.d., absorption cell
`along with an excess of Br2 over the olefinic impurity. The
`molecular bromine concentration was determined from its
`absorption near a band maximum in the range 400-430 nm by
`using a Shimadzu UV-160 double-beam spectrophotometer. The
`pressure of Br2 in the cell as well as the total pressure after
`adding the mixture to be analyzed was measured using an MKS
`Baratron manometer. After verifying the absence of any dark
`reaction, the cell was exposed to irradiation from a tungsten
`lamp fitted with a glass filter with a cutoff at approximately
`350 nm. Molecular bromine photodissociation initiates a chain
`reaction by which the alkene double bonds become bromi-
`nated.
`
`(+h(cid:238))
`98
`
`¥Br + ¥Br
`
`Br2
`
`¥Br + CH2
`
`(+M)
`98
`dCF-CF3
`CH2Br-¥CF-CF3 (¥CH2
`
`-CFBr-CF3)
`
`
`
`CH2Br-¥CF-CF3 (¥CH2-CFBr-CF3) + Br2
`f
`CH2Br-CFBr-CF3
`(+M)
`¥Br + ¥Br 98
`
`Br2
`
`+ ¥Br
`
`Since the irradiation light wavelengths used were capable of
`photodissociating Br2 but not the bromoalkane product,10,17 this
`process should result in complete olefin removal from the
`mixture. Thus, the change in Br2 concentration, as measured
`via UV absorption after a sufficiently long period of irradiation,
`should be equal to the initial concentration of the olefin:
`dCF-CF3]t)0
`) [Br2]t)0
`- [Br2]tf¥
`
`[CH2
`
`The change in bromine UV absorption with irradiation time for
`a typical titration/purification experiment is shown in Figure
`5a,b. Experiments performed with various initial Br2 concentra-
`tions (i.e., different bromine/olefin initial concentration ratios)
`and irradiation intensities resulted in no significant difference
`in the olefin concentration determined.
`To prepare samples of CH3-CF2-CF3 acceptable for kinetic
`studies, the bromination process was performed in a 2 L glass
`
`bulb with an excess of molecular bromine over the olefinic
`impurity. Upon completion of the bromination process, the
`mixture contained only HFC-245cb, the saturated product of
`the bromination (the fluorobromoalkane), and molecular bro-
`mine. The product of bromination, CH2Br-CFBr-CF3, is a
`liquid at room temperature, and the condensed liquid product
`was removed from the bulb before further treatment. The
`sample was then washed with distilled water to remove the
`residual bromine. UV absorption measurements on both the
`gas sample and the outgoing water were used to determine the
`point at which all bromine was removed. The remaining gas-
`phase sample was purified via low-temperature fractional
`distillation to remove both water vapor and any fluorobromoal-
`kane that remained. The efficiency of this distillation process
`can be easily verified by GC analysis since, unlike the alkene
`impurity, the heavy corresponding bromoalkane can be fully
`separated on the chromatographic column.
`The efficiency of the whole purification process was tested
`on manometrically prepared mixtures of 0.2% and 2%
`CH2dCH-CF3 in C2F6 and 6.5% CF2dCF-CF3 in C2F6. A
`Shimadzu GC-9A gas chromatograph with a thermal conductiv-
`ity detector was used to analyze samples before and after
`purification. For these mixtures, the olefinic impurity peak was
`sufficiently resolved from the main C2F6 peak that residual
`impurities at the ppm level could be detected.
`In both cases,
`the residual concentration of fluoropropenes in purified mixtures
`was less than 0.001%. We also used UV absorption measure-
`ments to analyze for the presence of residual alkenes. Figure
`6 shows the result of such an analysis for the CH2CHCF3/C2F6
`mixture.
`Since the primary purpose of the bromine purification
`procedure was to prepare HFC samples of sufficient purity to
`be used for OH reaction studies, we measured OH decay rates
`using the bromine-treated test samples as a direct check on
`possible kinetic problems associated with residual bromoalkanes,
`alkenes, or molecular bromine. We first verified the efficiency
`of the molecular bromine wash-out by making kinetic measure-
`ments on a mixture of only C2F6 and Br2 (i.e., no alkene added).
`The rate constants for the reactions of OH with pure C2F6, C2F6
`treated with bromine, and C2F6 plus alkene treated with bromine
`were