`
`Journal of Fluorine Chemistry 82 (1997) 171-174
`
`
`
`ELSEVIER
`
`
`
`JOURNALOF
`ONE
`GREY
`
`
`Preparation of 2,3,3,3-tetrafluoropropene from trifluoroacetylacetone and
`sulphurtetrafluoride
`
`R. Eric Banks *. Michael G. Barlow, Mahmood Nickkho-Amiury
`Chemistry Department, University of Manchester fastitute of Science and Technalopy (UMIST). ManchesterM6G 10D, UK
`
`Received 26 July 1996; accepted 11 October 1996
`
`Abstract
`
`Practical details are presented for the laboratory preparation of 2,3,3.3}-tetrafluoropropene in good yield from trifluoroacetylacetone and
`sulphurtetrafluoride in the presence of hydrogen fluoride. The gas-phase IR spectrum of the tetrafluoropropene (b.p., ~ 28 °C) is reproduced
`and a detailed analysis of the clefin’s nuclear magnetic resonance (NMR) spectra ('H, °C, °F) is provided. © 1997 Elsevier Science S.A.
`“+
`2,3,3.3-Tetrafluoropropene, Triflucroacetylacetone (1.1, ] 4riffeoro-2,4-
`
`IR and NMR spectroscopy, Sulphur
`
`tetrafluoride:
`
`Keywords: Floorination;
`pentanedioane}
`
`1. Introduction
`
`unstated yields [1.2] by a sketchily described methodclogy
`culminating in preparative GLC, and its identity was con-
`firmed by nuclear magnetic resonance (NMR) CB, PRS
`Needing to establish quickly a procedure for making
`2.3,3,3-tetrafuoropropene (1), we chose io try the anusual
`spectroscopy.
`In our experience, adventitious hydrogen fuonde is diff-
`Li} fluormation-cum-cleavage reacuon CF,COCH,COCH,
`( 2) + SP, > CPCPaCH, (1) + SOR, [2] rather than seem-
`cult to avoid when working with sulphur tetrafluoride owing
`ta facile hydrolysis of this reagent. However, it should be
`ingly more attractive methods [3], since sot only were trt-
`noted that in the original work summiarsized above, no HE was
`
`fuoroacetylacetone sulphur—tetrafluoride(2) and
`
`
`deliberately
`added
`to CF,COCH.COR-~SF,
`reaction
`immediately available, but we were also well versed in the
`mixtures. By contrast,
`the same workers found that SF,
`use of thelatter reactant. Importantly, too, we have long been
`fluorination
`of
`the bistrifluoromethy!
`analogue of 2
`fully equipped at UMIST to pursue work with anhydrous
`only
`in
`anhydrous
`hydrogen
`fluoride
`proceeded
`hydrogen fluoride (AHF) which we suspected from the out-
`(21: CF,COCH,COCF, + SE,/HF - CF,CF--CHCF,CF,
`set might be needed as a catalyst/solvent. The results of
`(80% 3.
`successful pilot experiments carried out on three times (0.14
`Finding in our work that the Auorination of 2 with SF, in
`tamol of 2} the Herature scale [2] are reported here.
`the absence of added HFproceeded sluggishly, reactions were
`run on a2! g scale (see Table 1} at 60 °C in the presenceof
`HE, considered not to be absolutely anhydrous owing to the
`tmiethod used to load the Monel reaction vessel (see Sec-
`tion 4.2}. The best (gas chromatography (GC)}-estimated)
`yield of 2,3,3,3-tetraflucropropene (1) (82%, with no SF,
`detected chromatographically in the crude gaseous product)
`was achieved with a 2 to SFP,
`to HPF molar
`ratio of
`1.0:3.7:0.6, and after combining products from several
`similar experiments which had been largely or completely
`freed from HF, SF,, SOF, and CH,COF using a KF-H,0-
`NaQH-CaCl, scrubbing train, the Huoroalkene (b.p.. — 28
`°C) was isolaied in 75%yield and better than 99%GCpurity
`by precise fractional distillation. The identity of the product
`
`2, Results and discussion
`
`The Russian work [2] being followed established that
`Li l-qiffucrinated G-diketones CF,COCH,COR (R=CH,
`(2), CH. or (CH,),CH) react with about aitwo molar excess
`of SF, at 206 °C during 12 h to provide approximately 65 : 35
`mixtures (by gas-liquid chromatography (GLC) f¢uncali-
`brated) analysis of alkali washed products) of 2,3.3,3-tetra-
`flucropropene (1) and the corresponding alkanoy!] fluoride
`(RCOF). The tetraflucropropene (1) was
`isolated im
`
`* Corresponding author.
`0022-139 /97/ $17.00 ©
`21997 Elsevier Science S A AU rights reserved
`PIF S0022
`-1339(96)03556.-7
`
`Page 1 of 4
`
`ARKEMAEXHIBIT 1169
`
`ARKEMAv. HONEYWELL
`
`PGR2016-00011
`
`

`

`i72
`
`RE. Banks et al. / Journal of Fluorine Chemistry 82 (4997) F7)-1 74
`
`Table |
`Details of the conversion of CF2COCH.COCH, (2) to CF,CF=CH, (1) with SF,-HF mixtures in a Monel autoclave*
`
`
`
`
`
`
` 2 (g; mol) SF, (gi mo1) HF(g; mol} Temperature °C)” Time (4) Crude Product tg Yield of L(%)*
`
`21.0; 0.14
`39.0: 6.36
`20, 1.0
`60
`is
`28
`70
`21.0; 0.14
`30.0; 0.27
`20; 1.0
`60
`15
`30
`16
`24.0, 044
`26.9; 0.24
`16.0.8
`60
`18
`24
`82
`21.0, 0.14
`29.0: 0.265
`16,0.8
`6&0
`18
`3D
`76
`
`
`
`
`
`“After each run, the autoclave lid was removed and the black volatile acrid quid lying in the cup was weated with 4 M aqueous KOH (left in the vessel
`overnight}. The cup was then washed out thoroughly with water, cleaned with a mild abrasive and polished ready for use again.
`"Autoclave wall temperature; headspace value, 50°C.
`“Determined by weighing collection taps fromthe purification train.
`“According to GC peak areas (uncalibrated detector).
`
`
`pannearannette
`
`=.
`YT OUR
`|i
`i
`Pil
`
`{
`
`\|! {
`
`$700 om’!
`(C=C str.)
`
`3050 cm 7
`(C-H str)
`
`10
`
`08
`
`°Transmittance
`
`
`
`
`
`eenpnenenemyamsenwemyniamacinommtnnnaimanttt
`
`|i'
`
`vere brestenttnenetmnenabainnmathatwahnutrinaderenianneSinanindeian etyhemesabacantanaanaennmnvamninannerirsnan eacnatin’
`2000
`1O00
`4000
`SOG
`
`Wavenumber/ cr
`Fig. §. Gas phase FTIR specuum of 2.3.34-1etraflporopropene C1; CF,CP=Chi,}.
`
`4
`
`was confirmed speciroscopically (mass spectre
`TR (see Pig. |) and NMR (see Fig. 2
`
`The “H and '°F NMR spectra for asolution of Lin CDCI,
`
`metry (MS)
`
`comprise an ABPX, system in which one proton (H,, to
`fower field) shows no resolved coupling to the CF, group
`fluorines, whereas the other (Hy) does ()4/,¢) = 1.4 Hz:
`relative sign not clear cut). A complete set ofdatais provided
`in Fig. 2(a), the analysis bemg performed by Corio’s meth-
`ods [4], and the assignment of H, and Hy being based on
`ihe magnitades of the “/,,. coupling constants (Juang > fa
`[S}).
`
`3. Experimental details
`
`3.4. Starting materials
`
`3.44, Sources
`
`Trifluoroacetylacetone (1) 994), sulpher tetrafluoride
`(technical grade} and AHF (99.856 or better} were used as
`received fram Fluorochem Lid. (UK), Air Products and
`Chemicals Inc. (USA) and ICL (UK) respectively,
`
`Page 2 of 4
`
`342. Hazards
`Trifluoroacetylacetone (b.p.. J05-107 "C) is classed as a
`flammable irritant and its manipulation presents no special
`problems. By marked contrast, suiphur tetrafluoride (b.p.,
`~ 38 °C) is highly toxic, its inhalation toxicity being com-
`parable with that of phosgene 16}:
`this property, coupled
`with its susceptibility towards hydrolysis atroom temperature
`
`with the Hberation of hydrogen Aucride (SF, + HO - (last)
`2HF +. SOF,, SOF, +H,O-> (slow) 2HP+S0,), demands
`that it should be used only by appropriaiely well-trained per-
`sonnel in dedicated work areas of appropriate design (Ref.
`[6] contains useful information on the methodsofhandling).
`The hazardous nature of AHFis legendary [7] and anyone
`using this material must acquaint themselves thoroughly with
`acceptable safe practices, these must mclude making ade
`quate prior arrangements for medical treatment [7].
`Hudhick¥’s much-used manual [8] contains some useful
`information on howto transfer AHF (b.p., 19.5 °C) froma
`normal (no dip pipe) commercial metal cylinder to a poly-
`ethylene (PE) receiver, but gives inadequate detail about
`heating the cylinder gently to ensure a reliable discharge of
`the contenis. When necessary (which proved not to be the
`
`

`

`RE. Banks etal. / Journal of Fluorine Chemistry 82 (1997) 171-174
`
`$73
`
`man ice/salt bath, the valve on the storage cylinder was
`opened carefully (smooth working of the valve should be
`checked before making the connection to the PE bottle} and
`an appropriate amount of liquid AHF was collected (as
`judged visually with the aid of volume calibration marks
`made with a wax crayon; density of HF = 1.0015 gcm 7 at
`O°C (11]). The relatively small known amount of AHF used
`in each fluorination experiment (see Section 3.2.2) was
`obtained by transferring material from the cold PE collection
`vessel to a smaller chilled (ice/salt) pre-weighed stout PE
`boule equipped with a tight-fitting lid. Samples assembled in
`ihis manner can bestored, if necessary, in a refrigerator until
`required.
`The operations described above were carried out in a lab-
`oratory dedicated to work with AHF conjointly by two well-
`trained chemists, wearing appropriate personal protective
`equipment [9],
`
`3.2.1.2. Sulphur tetrafluoride
`Por each fluorination experiment (Section 3.2.2), afresh
`sample of SF, was transferred, using standard vacuum tech-
`niques, from a double-valved (important: for safety reasons,
`a fine-control stainless steel valve must be fitted to back up
`the main valve} commercial cylinder directly to a thoroughly
`dried, calibrated (volume scale; density of SF, = 1.9190 g
`cm”? at ~ 73°C [12]), pre-weighed Pyrex tube (150 cm*)}
`cooled in quid mtrogen (connections were made with the
`aid of copper, Pyrex and Viton® fluorubbertubing, as appro-
`priate}. Once charged, the cold Rotafio™™ mabe was quickly
`weighed to ascertain the approximate amount of SF, present
`(this determined the amount of wiflucroacetylacetone (2)
`placed im the fluorination vessel (see Section 3.2.2)), and
`then returned to its bath of liquid nitrogen to await use in a
`fluorination experiment. The whole operation,
`including
`weighing of the Rotaflo’™tube, was conductedin an efficient
`dedicated fume cupboard.
`
`3.2.2. Fluorination of trifluoroacetylacetone (2)
`In a typical experiment, cold (approximately —S °C)
`hydrogen fluoride (20 g, 1.0 mol) was poured from a PE
`storage bottle (Section 3.2.1.1) into a Monel autoclave (250
`cm") containing trifluoroacetylacetone (21.0 g, 0.14 mol).
`The vessel's lid carrying a closed valve, bursting disc assem-
`bly (nickel disc protected by a 0.1 mm thick PTFE
`membrane, 150 atm rating and a 0-250 atm gauge) was then
`quickly bolied on and standard vacuum transference tech-
`niques (autoclave cooled in liquid nitrogen) were used to
`charge « with sulphur tetrafluoride (39.0 g, 0.36 mol} from
`a Rotaflo tube container (Section 3.2.1.2). The autoclave’s
`vaive was then closed andthe vessel was allowed to warm to
`
`roomy temperature ina furne cupboard before it was heated
`vlectrically at 60 °C (outside wall temperature; vessel head-
`space temperature, 50 °C) for 15 h in a properly veotilated
`blast-proof cubicle. The autoclave was allowed to cool
`somewhat, and was then transferred to a fume cupboard and
`connected to a product recovery train, comprising an acid gas
`
`FC.
`io
`
`c=c
`
`Ha:
`*
`
`ay
`
`Hay
`
`EF
`fa) ‘Hand °F data (vge¥y = 21.3 He?
`oe
`<} waa
`ii
`Ho
`po" CF
`Y
`oe
`poe
`$
`o
`1
`sso PF eee OP ee
`(4
`i
`
` i L
`PS SD need
`ane
`
`(by UC daa?
`
`iro:
`i|:
`
`Cg
`Fig. 2. NMR chemical shifts (ppm,in parentheses) and coupling constants
`(Hz) for 2,3,3,3-tetrafluoropropene (1). Hat 300 MHz (Bruker AC-300
`instrument, ext. Me,Si cef.): ‘°F ar 198.8 MHz (Bruker AC-200: ext.
`CF,CO)Href., positive assignment downfield). Stepanovet al. [2] camed
`out a simplistic analysis of the spectra of f in CCl, on the basis of an AANX’-
`type system, quoting 5, (HMDS ref.) 5.44 ppm (/ye = 46 Hz) and &,
`(CPCI, ref.) - 72.2 Cd. CPs. Jee = 1 Hz), ~ 124.65 (m, CF) ppm. SHC at
`75.5 MHz (Bruker AC-300, 'H decoupled, ext. Me,Si ref), “Not all the
`weaker lines of the multiplets were listed on the print-out, but quartets ( from
`coupling to CF,) were usually easily distinguished from doublets (from
`CF).
`
`case in the present work, since a cylinder fitted with a dip
`pipe was used}, we adhere strictly to the manufacturer's
`(ICT) instructions to heat AHF cylinders in a shroud supplied
`with warm air ducted from a small electric fan heater, so that
`if, a0 circumstances can the temperature exceed 45 °C. (35
`“C is the prescribed normal value) 191.
`No hazards data appear to be available for 2,3.3,3-tetra-
`fluoropropene CL), and hence care was taken to avoid inha-
`lauion of the pure gas [10]. Inhalation of crude material
`containing Cafter removal of HF) SOF,. SE, and CH,COF
`wasstrictly avoided.
`
`i2 Preparation of 2,33, 3-tetrafluorepropene
`
`324. Transfer ofANF and SF, from commercial storage
`cylinders fo secondary containers
`
`3.24.1. Anhydrous hydrogen Auoride (with AL. Laing)
`A steel AHFcylinder (capacity 22 kg: dip pipe type used
`regularly in our laboratories to fill Simons ECFcells} was
`connected via flexible copper tubing to 4 robust semi-trans-
`parent PE bottle (500 cm’) fitted with a screw top carrying
`copper inlet and outlet
`tubes (0.6 mm id.j: the former
`reached halfway down the length ofthe bottle, while the latter
`led from the lid into a fume hood to carry away HFfumes.
`Once the PE bottle had been chilled to approximately — 5 °C
`
`Page 3 of 4
`
`CH, 880
`
`}i
`
`|
`
`7 2703 ~'
`F See
`.&
`
`any.
`
`/ i N\ asap
`of
`
`Fr
`bas
`
`

`

`174
`
`RE Bunks et al. / fournal uf Fluorine Chemistry 82 62997) 172374
`
`References
`
`review of selective reactions of SF, with carbonyl
`Lt} For a recent
`compounds, see Ab Burmakov, B.V. Kunshenko, L.A. Alekseova and
`LM. Yagupolskii,
`in L. German and S. Zemskov (eds.), New
`Fluorinating Agents in Organic Synthesis, Springer-Verlag, Berhn.
`1989, p. 197,
`{2] UV. Stepanov, Al. Burmakov, BV. Kunshenko, L.A Alekseeva and
`LM. Yagupol’skii, 24. Ore. Khim.. 79 (1983) 272 (English
`uanslation J. Org. Chem. USSR. 19 (1983) 244).
`[3] For
`example: CF,CF,;CH,;OH (Aldrich) -»*CF,CP,CH,OTs -»>
`CFLCF,CH. ~* (with Zn in MeCO,Hj}CF,CF=CH, CE.T. McBee,
`D.H. Campbell and CW, Roberts, J. Am. Chem. Soc. 77 (1955)
`3149). CF,CF,CH,OH + H./C-+ Chet act. C}CF,CF=CH, (K. Ihara,
`F. Yamaguchi and S. Yamane, EP 328148 Ai 890816 (to Daikin Ind};
`Chem. Abs.. 1/2 (1989) 781566).
`[4] PLL. Corio, Structure of High-Resolution NMR Spectra, Academic
`Press, London, 1966, Chapter 7.
`{5] JW. Emsley, L. Phillips and V. Wray, Fluorine Coupling Constants,
`Pergamon, Oxford, 1977, p. 169.
`[6] G.A. Boswell, WC. Ripks, RM. Scribner and C.W. Tullock, Org.
`React.. 21 (1974) 1
`{7] R.E. Banks, in RE. Banks, D.W.A. Sharp and LC. Tatlow (eds.),
`Fluorine the First Hundred Years: 1886-/986, Elsevier Sequoia,
`Lausanne and New York, 1986, pp. 6-8 (reproduced in J. Fluor.
`Chem 33 (1986) 1).
`[8]M. Hudlick¥, Chemistry ef Organic Flavrine Compounds: A
`Laboratory Manual with Comprehensive Literature Coverage,
`Horwood, Chichester, 1976 (reprinted and re-issued in 1992), pp. 17~
`19.
`{9] Anhydrous Hydroven Fluoride, IC] Technical Service Note Na. TS/C?
`2353.
`10) The assumption was made that CF,CF=<CH, can be handled like its
`well-known perflaorinated analogue CF CF
`=<CP,, which is not very
`
`rosac on an acute inhalation basis (4 h LCS in rats, approximately
`3000 ppm: see GL. Kennedy, Cru, Rev. Toxtcet., 27 (2) (1990)
`149,
`in RH Simons (ed), Pluerine Chemistry, Val.
`Til PH. Symons,
`Academic Presse, New York, }950, p. 225.
`(2) ELE. Evans and G. Mani, in M. Howe-Grant (ed. 1, Flaorine Chemisrry:
`A Comprehensive Treatment, Wiley, New York,
`1993, p. 185
`(reprinted fromKirk-Orhmer Encyclopedia of Chemical Technology).
`3) RE. Banks and H. Sutcliffe, Chem.
`ind. (London}, (19624 979
`4) AL. Henne and TP. Waalkes, / Am Chem. Sac., 68 (1946) 496.
`
`1.
`
`!
`
`scrubbing sectian (a glass tube (40 cm X 2.5 cm) loosely
`packed with dry granular KF leading to three Drechsel bottles
`in series, the first being empty (‘‘suck-back’’ trap), the sec-
`ond containing H,O (200 cm’) andthe third charged with 2
`M aqueous NaOH (200 cm*)), a drying tube (CaCl; 40
`em X 2,5 em) and finally two cold Pyrex traps (— 78 °C)
`which were connected (via a protective — 196 °C trap} toa
`vacuum system. Gaseous products were bled slowly and com-
`pletely from the autoclave by careful use of the vessel’s nee-
`dle valve, in conjunction(in the later stages) with ajudicious
`reduction of the pressure (to minimize the transfer of water
`vapour from the scrubbing train). The volatile material (33.0
`&) collected at ~ 78 °C was shown by GC-IR analysis (1.5
`m Porapak Q, 80 °C, N, carrier, uncalibrated katharoreter
`detector} to comprise (in order of elution; percentage com-
`position estimated from peak areas) SOF, (6%), SF, (8%),
`CF,CF=CH, (1; 70%), CHCOF (12%) and three uniden-
`tified products (total, 4%).
`The results of other experiments, all carried out in the same
`manner using the same quantityof trifluoroacetylacetone, are
`given in Table 1, Products from these runs were combined
`and the material (85.0 @} was rectified using an adiabatic
`low-temperature column (1 m < 15 mmpacked with 2.5 mm
`Fenske glass rings) similar to that described previously [13]
`to provide 47.6 g (0.42 mol, 75%) of 2,3,3,3-tetrafluoropro-
`pene (GC purity, better than 99%), b.p. ~ 28.0°C (literature
`[14], -28.3 °C}; MS data (70 eV EL, VG 7070 EPinstru-
`ment} a/e bid (M* 90(accurate mass: found, 114.0088,
`
`calculated for C,Haka, 114.0092)), 113 COM-1)7, 22%),
`
`OS [(M-F)}", 349), 75 (CUPS 11% 69 (CFS, 100%},
`65 (8%), 64 (C,HLFY, 96%), 45 (CLHOF*. 48%), 44
`(CHP, $9), 31 (CP, 9%): IR (10 cm gas cell, Natl
`optics, Nicolet 250 FTIR spectrophotomer) A,,,,: 3050vw
`(2C~Hsir.}, 1700m (C=C sip.) 1306s, 1388s,
`(357m.
`[209vs, 119Ovs, 1] Sivs, 1 160vs,
`| £53vs (C-P str}, 048m,
`cyitij.
`892m (2:C-H out-of-plane bending} em7! (see Fig
`
`Page 4 of 4
`
`