Chlorination and dehydrochlorination reactions relevant to the manufacture of trichloroethene and tetrachloroethene: Part 1. Reaction pathways

Applied Catalysis A: General 399 (2011) 1–11 Contents lists available at ScienceDirect Applied Catalysis A: General journa l homepage: www.e lsev ier .com Chlorin s r trichlo Part 1. Iain W. S ma David Le a School of Che b Ineos ChlorVi a r t i c l Article history: Received 9 No Received in re Accepted 28 F Available onlin Keywords: Chlorination Dehydrochlorination Attapulgite Copper(II) chloride Homogeneous catalysis Heterogeneous catalysis of t been tinuo m. He o tric oroet mum batch reactor), radical chlorination to form pentachloroethane is dominant. Above 573K and under flow conditions, free radical dehydrochlorination to form trichloroethene becomes dominant. Heterogeneous chlorination under flow conditions provides a route to pentachloroethane and thence tetrachloroethene. High conversions favour the formation of oligomeric products. 1. Introdu Large-sc and tetra-ch carbon indu as solvents a carbon refri atmospheri at least in lifetimes of carbon prod replacemen route to hyd nation proc is at a very Problem environmen and selectiv environmen tal study o of trichloro ∗ Correspon E-mail add 0926-860X/$ – doi:10.1016/j. © 2011 Elsevier B.V. All rights reserved. ction ale manufacture of the unsaturated chlorocarbons, tri- loroethenes has been an important part of the chloro- stry for many years. Traditionally they have been used ndas feedstocks for thefirst generationofhydrofluoro- gerants developed as a response to the depletion of the c ozone layer [1]. Although this latter use will change, Europe, as a result of concerns over the atmospheric the most widely used hydrofluorocarbons [2], chloro- uction will continue to be important, if only because tofC-ClbyC-Fbonds is still theonlypractical large scale rofluorocarbon production; development of oxyfluori- esses, whereby C-H bonds are converted directly to C-F, early stage [3]. s associated with chlorocarbon production centre on tal concerns, for example combustion reactions [4], ity issues, particularly unwanted side products in the t. For these reasons we have undertaken a fundamen- f the factors that determine the selective production ethene vs. tetrachloroethene. Both have been impor- ding author. Tel.: +44 141 330 4372. ress: [email protected] (D. Lennon). tant feed-stocks but their use will probably decline due to legislative changes [2]. However, reactions leading to these com- pounds are an important system for fundamental study, since they depend critically on a balance between thermodynamic vs. kinetic considerations and among different types of homoge- neous or heterogeneous chlorination and hydrochlorination or dehydrochlorination reactions. Studies of reaction fundamentals should yield information which is generic to hydrochlorocarbon and chlorocarbon syntheses and not simply limited to the produc- tion of CHCl CCl2 and CCl2 CCl2. To place the present results in context, some salient features of the most important types of reactions are first described. The possible chlorination reactions can be split into three categories: homogeneous chlorination of alkanes, homogeneous chlorination of alkenes and heterogeneous chlorinations. Homogeneous chlorination of alkanes is believed to take place viaa radical substitutionmechanism[5]. Forethaneandchlorinated ethanes reactions occur in tubular quartz reactors at 373–773K [6]. For a series of chloroalkanes, C2H6−xClx, chlorination activity decreases with increasing x, this being illustrated by a kinetic study of Cl atom attack on C2H6, C2H5Cl and CH2ClCH2Cl [7]. Carbon cen- tred radicals with a Cl in the �-position are thermally unstable and decay rapidly to yield a Cl atom and an alkene [7]. The radical intermediate formed during the formation of 1,2-dichloroethane from chloroethane is unstable, and breaks down to yield ethene see front matter © 2011 Elsevier B.V. All rights reserved. apcata.2011.02.035 ation and dehydrochlorination reaction roethene and tetrachloroethene Reaction pathways utherlanda, Neil G. Hamiltona, Christopher C. Dud nnona,∗, John M. Winfielda mistry, Joseph Black Building, University of Glasgow, Glasgow G12 8QQ, Scotland, UK nyls Ltd., Runcorn Site HQ, South Parade, PO Box 9, Runcorn, Cheshire, WA7 4JE, UK e i n f o vember 2010 vised form 16 February 2011 ebruary 2011 e 8 March 2011 a b s t r a c t Factors which affect the selectivity dehydrochlorination reactions have equilibrium measurements, and a con gas-phase chemistry within the syste rination of 1,1,2,2-tetrachloroethane t catalyst favours formation of pentachl The latter is the thermodynamic mini / locate /apcata elevant to the manufacture of nb, Peter Jonesb, he chlorination of 1,2-dichloroethane and the associated examined using approximate thermodynamic calculations, us flow micro-reactor. There is a balance between surface and terogeneous catalysis is not necessary to effect dehydrochlo- hloroethene but an attapulgite-supported copper(II) chloride hane and its dehydrochlorination product, tetrachloroethene. of the system. Below 473K and with long reaction times (2h, 2 I.W. Sutherland et al. / Applied Catalysis A: General 399 (2011) 1–11 and a chlorine atom; this process is termed radical dehydrochlori- nation. Homogeneous chlorination of alkenes is significantly slower than that of alkanes, for a given set of reaction conditions [6]. Chlo- rine additio [8]. An alter postulated chloronium Copper( here, is a w example tow CuCl2. Abov from the C determinin substitution from metha and the pro than a sub tent with e intermedia ing agent w might be ex tions [10]. a partition ways, as in [11,12]. The sec hydrochlor two possibl processes. phase proc mechanism dehydrochl and 1,1,2,2 small quan seems likel abstraction and H2O. N dichloroeth was observ other hydro Lewis a 2-chloropro prisingly, th one of the ascribed to as oligomer supported o nation of 2 chlorination rination ab particularly was 10% ( rial, which [20]. FTIR stud over �-alum dehydrochl to dehydro This surfac 1,1,1-trichl effectively l was minim largely unp Similar beh chloride on Although much is known about reactivities of chloro-alkanes and -alkenes, the overall reaction pathways from the reaction of 1,2-dichloroethane with Cl2 to yield CHCl CCl2 or CCl2 CCl2 can- not be written down in an unambiguous form. Furthermore, the the C ring gene rial [2 o pro first vity o ed, w , a ed he erim perim erim anif rease rom 2,2-T dis inde ion-r ani and e tra asse ted r andi ume he va rd fo ed, a ma ld a oced samp d, an ssur -liqu h an 5% P ldia lit ra onis tor a (BO se fa n dic con tely entr grad n. C to m tion to t to fo ver tion is. D sitio n to vinyl chloride proceeds via a radical mechanism native possibility is via an ionic route, as is commonly for liquid-phase reactions, thought to proceed via a ion intermediate [5]. II) chloride, present in the commercial catalyst used ell known heterogeneous chlorinating agent [9], for ards both alkanes and alkenes over pumice-supported e 673K, a significant quantity of chlorine was evolved uCl2 melt formed. The rate of Cl2 evolution was rate- g and the products were those expected from a radical reaction, specifically the formation of chloromethane ne [10]. In the range 493–603K, Cl2 was not evolved ducts formed with alkenes reflected an addition rather stitution reaction. Product distributions were consis- lectrophilic substitution involving a chloronium ion te. A lack of substitution implied that the chlorinat- as CuCl2 itself rather than adsorbed Cl, as the latter pected to result in some abstraction/substitution reac- In other studies however, it has been proposed that exists between radical-based and bridged-ion path- chlorination of alkenes with Cl2 under dinitrogen ond major reaction type is dehydrochlorination of oalkanes. Under homogeneous conditions, there are e pathways, radical chain or concerted E2 elimination The E2 mechanism is favoured generally for solution esses [13,14] but in the vapour phase, a radical chain , is believed to be more likely [15,16]. The rates of orination of 1,2-dichloroethane, 1,1,2-trichloroethane, -tetrachloroethane are enhanced by the presence of tities of O2 and Cl2 under flow conditions [16] and it y that this is the result of additional routes to hydrogen that are provided by Cl2 and O2, ultimately yielding HCl o rate enhancement for chloroalkanes other than 1,2- ane, 1,1,2-trichloroethane or 1,1,2,2-tetrachloroethane ed [16] (see also related aspects of this study involving chloroalkanes [17]). cid oxidic catalysts promote dehydrochlorination of pane and 1,2-dichloropropane [18,19]. Perhaps sur- e most acidic catalyst studied, zeolite ZSM-5, displayed lowest activities for dehydrochlorination. This was its high activity for other, competing reactions such isation, cracking and aromatization [18]. When CuCl2 n �-alumina was used to catalyse the dehydrochlori- -chloropropane, a mixture of dehydrochlorination and products was formed [18], consistent with the chlo- ility of CuCl2 and Lewis acidity of �-alumina. This is relevant to the present work, in which the catalyst w/w) CuCl2 supported on attapulgite, a clay mate- has been shown to exhibit Lewis acid properties ies of the dehydrochlorination of 1,1,1-trichloroethane ina confirm that Lewis acid sites are required for orination [21]. The material was activated at 1000K xylate the surface and generate Lewis acid sites. e was highly active for the dehydrochlorination of oroethene at 450K. Surface saturation with pyridine ocked Lewis acid sites and dehydrochlorination activity al. The OH stretching region of the spectrum remained erturbed during the dehydrochlorination process [21]. aviour has been reported for CH3CCl3 and for t-butyl their exposure to halide Lewis acids [22]. role of ufactu hetero indust of CO t As a selecti improv burden report 2. Exp 2.1. Ex Exp uum m with g cated f 1,1, vacuum rine (L corros of the m vessel remov Deg evacua by exp ple vol from t cupboa remov vacuum manifo this pr into a weighe the pre by gas tedwit linked interna (9:1 sp flame i integra wasHe respon tions i known separa of conc whose questio verted conven Due dency to reco extrac analys compo uCl2/K+ attapulgite catalyst used in the large scaleman- process was not clear. It should be noted that a related ous catalyst, CuCl2/K+ on silica, has attracted recent 3] and academic attention [24] for the oxy-chlorination duce phosgene. stage in a programme designed to understand how the f the 1,2-dichloroethane chlorination process could be ith a consequent improvement of the environmental detailed reaction scheme has been derived and is re. ental ents under static conditions ents under static conditions utilised a Pyrex high vac- old designed for use with dichlorine, mercury free and less stopcocks (J. Young). Reaction vessels were fabri- either stainless steel or Pyrex. etrachloroethane (Aldrich, 99.5%) was degassed, then tilled onto activated 5A molecular sieves (2 g). Dichlo- 99%) was dispensed via a lecture bottle fitted with a esistant regulator (Praxair) to avoid over-pressurisation fold. An aliquot was condensed into an evacuable Pyrex a series of freeze/pump/thaw cycles carried out to ce HCl and other very volatile contaminants. d 1,1,2,2-tetrachloroethane (1.5 g) was distilled into the eaction vessel. An equimolar quantity of Cl2 was added ng a pressure (5.3 kPa) of Cl2 into a pre-calibrated sam- and condensing the gas. The vesselwas sealed, removed cuum line and placed in a barrel furnace within a fume r 2h. After reaction at a specified temperature, it was llowed to cool to room temperature, returned to the nifold and the most volatile fraction expanded into the nd removed. Loss of the less volatile fraction during ure was minimal. The less volatile fraction was distilled ling vessel, fitted with a septum. The products were d the valve opened briefly to the atmosphere to allow e to equilibrate. An aliquot (1�L) was taken for analysis id chromatography (GLC) using a Chrompack C9000 fit- Agilent HP-5 capillary column (stationary phase – cross H ME siloxane, film thickness 0.25�m, length 30m, and meter0.32mm). Injectionwasperformedviaa split port tio) held at 523K. Detection and quantification were by ation (temperature 523K), an Hewlett Packard HP3396 nd an isothermal (353K) oven program. The carrier gas C99.9%) at aflowrate =1.5 cm3 min−1. ChlorocarbonFID ctors were determined by serial dilution of stock solu- hloromethane (Aldrich, 99%) to create four solutions of centrations. Five aliquots of each solution were injected into the GC and the peaks’ mean values noted. A plot ation against mean peak area yielded calibration curves ients gave the FID response factor for the compound in omponents of each weighed product sample were con- oles and product yields and conversions defined in the al way. he relative involatility of some materials and the ten- rm oligomeric products, it was not always possible a measurable mass of product. In these cases, solvent using CCl4 was carried out to obtain a sample for GLC ata from these reactions are presented as percentage ns of GLC samples. I.W. Sutherland et al. / Applied Catalysis A: General 399 (2011) 1–11 3 ow ex 2.2. Continu 2.2.1. Descr The con stainless st via 1/4 in. C to the toxic and produc cupboard, w Gas flows controllers ware via an (1,1,2,2-tetr as appropri bler; feed w maintain a fitted to the a large surf stream. Dic controller. A so that the prevented i in the line a lines of iden as required A mixing ensure com diately befo mercury-fil exceed 0.5 twin reacto Post-reacto 500 cm3 mi maintained the reactor manual spl analytical s gas-sample To ensur tions, the pa materials u attapulgite, inyls 64), 2±0 reac 3 pa lyst o requ . Sm r in p React pro eneo -tetr uartz d to ed a reac as m Dich wer 3 m Fig. 1. The apparatus for continuous fl ous flow experiments iption of the flow rig tinuous flow reactor (Fig. 1) was constructed in 1/8 in. eel tubing with 1/4 in. quartz reactor tubes connected ajon Ultra-Torr unions fitted with Viton O-rings. Due and corrosive nature of the reactions, both the feeds ts, the flow rig was housed entirely within a fume hich was maintained under permanent extraction. were regulated by a series of calibrated mass flow (Brooks 5850S) controlled by Brooks SmartControl soft- RS-485 multi-drop serial bus. Hydrochlorocarbon feed achloroethane, pentachloroethane or trichloroethene ate) was supplied to the reactor via a dinitrogen bub- as immersed in a temperature controlled water bath to constant vapour pressure. A sintered-glass diffuser was inlet of the bubbler to break up the gas flow, ensuring ace area of bubbles and eliminating pulsing of the feed hlorine was fed to the reactor directly from a mass flow N2 purge systemwasfitted to allmass-flowcontrollers system was purged with dry N2 when not in use. This ngress of H2O by maintaining a slight positive pressure ChlorV C500-4 and 0.5 neous 0.90 cm of cata quartz culated reacto 2.2.2. The homog 1,1,2,2 A q couple contain to the that w bath. 99.5%) 25.0 cm nd thus helped to minimise corrosion. Additional feed tical construction were used for anhydrous HCl and O2 . vessel, packed with quartz beads (diameter 2mm) to plete mixing of the reactant gasses, was fitted imme- re the reactor. Pressure was controlled via a pair of led lutes to ensure that the reactor pressure did not bar g. The oven (Shimadzu GC-14A) was fitted with r tubes, allowing flexibility in the catalyst charge used. r, the product stream was diluted with N2 (flow rate n−1), to ensure that highly chlorinated products were in the vapour phase. To prevent over-pressurisation of and to protect the analytical system from corrosion, a it was added; only 5% of the flow passed through the ystem. Product stream sampling used an air-actuated valve (Valco UW-type, 6 port). e that all reactions were kept under comparable condi- cked volume of the reactor was kept constant. Packing sed were ground quartz, the catalyst and the support, the latter two materials being obtained from INEOS defined by t reactor bed 1.7×103 h− by the volu inlet flow p (500 cm3 m vent valve recorded at titioned wi analyses. The reac tor channel concentrati Some varia result of sm pressure of consistent, centrations to [26], wer tionswere n periments. (catalyst sample Cu 0951, Lot 31; attapulgite sample whosemeasureddensitieswere 1.11±0.05, 0.84±0.04 .04g cm−3 respectively. The packing used for homoge- tions was 1000mg ground quartz, corresponding to a cking volume. For heterogeneous reactions, the mass r attapulgite was determined and the mass of ground ired to bring the total bed volume to 0.90 cm3 was cal- all quartz wool plugs held the solid sample within the lace. ion protocol tocol for all reactions was identical; that for the us chlorination of 1,1,2,2-tetrachloroethane with a 3:1 achloroethane:Cl2 mol ratio is described as an example. U-tube was packed with ground quartz (1000mg) and the flow system. The second channel of the system n unpacked U-tube. 1,1,2,2-tetrachloroethane was fed tor via a N2 fed bubbler (flow rate =20.0 cm3 min−1) aintained at 303K by a temperature-controlled water lorine (0.1 cm3 min−1) and N2 (4.9 cm3 min−1; BOC e co-fed to give an overall volumetric flow rate of in−1. In this manner the gas hourly space velocity, as he ratio of the total volumetric flow at the inlet and the volume [23], remained constant for all experiments at 1. These arrangements afford a space time, as defined me of the catalyst bed divided by the volume of the er unit time [25], of 2.2 s. Post-reactor, additional N2 in−1) was added to minimise condensation. The split- was closed slowly until a flow of 5.0 cm3 min−1 was the vent of the gas sample valve. Thus flows were par- th 520 cm3 min−1 flow to vent and 5.0 cm3 min−1 for tion mixture was passed through the unpacked reac- at 523K for 1h, or until GLC analysis showed a stable on of 1,1,2,2-tetrachloroethane in the sample stream. tion was observed between reactions; it was probably a all changes in the split ratio causing fluctuations in the the sample loop. However, all data sets were internally and as no direct comparisons were made between con- (yields, conversions and selectivities, definedaccording e used for comparisons among data sets) these fluctua- ot a significant problem.When a stable product stream 4 I.W. Sutherland et al. / Applied Catalysis A: General 399 (2011) 1–11 had been achieved, feed flows were diverted through the packed reactor and the products were thereafter sampled every 10min, usually for 1h. After 1h, the reaction temperature was increased to 573Kand the systemallowed to equilibrate for 15min. This process was repeate Followin cool to ro unpacked r flow contro and allowe use, the sy (30 cm3 min Dehydro conditions, continuous tion of 1,1,2 at 623K wi of quartz pa had little o in yield ob inhibited th of packing. To deter associated experiment detected w duced into in the proto percentage was determ presented f to have rea recording a particular s mass balan ues, with th that series exceeded th represent fa flow measu iments wer here are re reaction con 2.2.3. Produ Analyses described a sample valv less steel s (temperatu grator with an isotherm were prepa was injecte equilibratio (SGE 10-MD taking care then used t To comp were deter bubbled thr liberated w starch indic perature. Th water for 10 titration ag Table 1 BET surface areas and pore volumes. Material BET surface area (m2 g−1) Pore volume (cm3 g−1) lgite st (un st (cal cator re us tem befo oces l Cl2 a solu orine -tetr 0mg 2.2 s, ema s pro hin t Exper oug igom e sp as th d th valv ns p ers w d occ mpl ectab low dual ment pite som ller b ird s of th l. Ac exam gh th l ma ave talys com on t inyls 4h a a BET ined area -treatment has little effect on the catalyst structure. Slight ions in surface area and pore volume were observed for the d sample, suggesting the possible onset of a collapse of the t structure. However, the variations could equally be within ge expected from different sample batches. d for the temperatures 623 and 653K. g reaction completion, the oven was allowed to om temperature, flows were diverted through the eactor and the split-vent valve was opened. All mass llers were switched to N2, the bubbler by-passed d to cool to ambient temperature. When not in stem was maintained under a constant N2 purge −1). chlorination of hydrochloroalkanes may, under certain be catalysed by Pyrex [27]. To determine if the quartz flow system had any catalytic effect, dehydrochlorina- ,2-tetrachloroethane to trichloroethenewas performed th the protocol described above. Variation of the mass cking, for a constant residence time within the reactor, r no effect on trichloroethene yields. The slight drop served using an unpacked reactor was attributed to ermal transport properties of the reactor in the absence mine the magnitudes of errors from the flow system with the measurements, a series of multiple injection s was performed. A mixture of the four chlorocarbons as placed in the bubbler, and the mixture was intro- an unheated empty reactor. The mixture was diluted as col for the full reaction and analysed via on-line GC. The error associated with the analysis for each compound ined by analysis of these multiple-pulse data. Data are or flow measurements when the catalyst was deemed ched steady state operation. This was determined by number of repeat measurements (typically 4 or 5) at a et of conditions. The reaction parameters, conversions, ces, selectivities and yields, are presented as mean val- e errors presented as the standard deviations within of replicate measurements. Typically, those deviations e analytical errors (as described above) and therefore irly the experimental variations associated with these rements. To ensure reproducibility of results, all exper- e performed in at least duplicate. The data sets reported presentative of the results obtained for the individual ditions. ct analyses were carried out by on-line GLC using the equipment bove. Injection was through an air-actuated 6-port gas e (Valco UW) fitted with a pre-calibrated 25�l stain- ample loop (Valco) directly onto the column. An FID re 523K) was linked to a Hewlett Packard HP3396 inte- carrier N2 (flow rate 1.5 cm3 min−1); the oven utilised al program at 353K. Gas phase calibration samples red in vacuo. A small quantity (1�l) of chlorocarbon d into a sealed, evacuated vessel of knownvolume. After n, a sample was withdrawn using a gas-tight syringe R) and was introduced slowly into the sample loop, not to pressurise the loop. The gas sample valve was o inject the sample. lete the off gas analysis, unchanged Cl2 and product HCl mined in some experiments. The product gases were ough standard aqueous KI solution for 10min. Diiodine as titrated against standard sodium thiosulfate, with ator. Analysis was repeated four times for each tem- e product stream was then bubbled through distilled min to form a solution of HCl, which was quantified by ainst a standard NaOH solution using phenolphthalein Attapu Cataly Cataly as indi peratu The 30min This pr that al by, the Chl 1,1,2,2 and 10 time= chlorin specie for wit 2.2.4. data Alth and ol tion. Th as it w exit an tle the reactio oligom tion di For exa be det ever, a the gra experi Des above, contro ment. A th erties and HC by, for Althou from C could h 2.3. Ca The ported ChlorV N2 for tion vi uncalc surface Pre reduct calcine suppor the ran (uncalcined) 79.5 0.29 calcined) 80.8 0.34 cined) 71.6 0.25 . This procedurewas performedfive times for each tem- ed. perature was increased to 573K and equilibrated for re the product stream was re-analysed for Cl2 and HCl. s was repeated for 623 and 653K. The method assumes ndHCl in the product stream reactedwith/was trapped tion. The data obtained suggest that this was justified. mass balances (g atom) for the chlorination of achloroethane over ground quartz, 100mgbare support catalyst (total volumetric flow 25cm3 min−1, space GHSV=1.7×103 h−1) are presented in Section 3. All ss balances are effectively complete, confirming that all duced have been analysed, and all chlorine is accounted he system. imental issues leading to possible ambiguities in the h the reaction rig was designed to minimise corrosion er formation, these did arise during long-term opera- lit-vent valvewas themajor site of oligomer deposition, e major constriction in the line between the reactor e rig vent. It was necessary to remove and disman- e periodically for cleaning. Although the majority of erformed exhibited a closed carbon mass balance, some ere formed. Apparently, a low level of oligomer forma- ur but it was so low as to be undetectable by analysis. e, a 1–2% deficit in the carbon mass balance would not le, as it is within the error range of the analysis. How- level of oligomer formation (such as 2%) would lead to fouling of the split-vent valve as was observed in some s. the use of a dry N2 purge on the system as described e corrosion did occur. In particular, the Cl2 mass flow ecame corroded over time, necessitating its replace- ource of ambiguity is the result of changes in the prop- e support, particularly its acidity, from exposure to Cl2 idity changes due to chlorination of alumina surfaces ple anhydrous HCl, have been widely studied [28,29]. ere was no evidence for massive losses in Cl2 and HCl ss balance experiments, low level surface chlorination occurred. t preparation and characterization mercial catalyst used was copper(II) chloride sup- he clay mineral, attapulgite with a promoter, KCl (Ineos Ltd.). Prior to use, it was calcined at 673K in flowing nd was characterised by total surface area determina- isotherm studies, performed in both the calcined and forms, also on thebare attapulgite supportmaterial. The s and pore volumes are shown in Table 1. I.W. Sutherland et al. / Applied Catalysis A: General 399 (2011) 1–11 5 Fig. 2. The re re de tion/dechlorin To exam doped attap A known m round-botto concentrati nel, while entered the remove exc tion at 343 further 2h. 3. Results A therm possible rou of batch ex possible rou reaction-tes all reaction 3.1. Thermo The init vious stud ethene [30 chosen her trial feedst from 1,2-di passes thro ing to trich in the react in Fig. 2. Standard able [33] for of reactionw cces erim om t ients ere mpt s sim icm ,1,2,2 action scheme used to consider thermodynamic changes. Reactions 1 and 5 a ations. ine the role of the KCl promoter, an 8% (w/w) KCl ulgite was prepared by the incipient wetness method. ass of attapulgite was added to a previously-evacuated, med flask. An aqueous KCl solution of pre-calculated on was added via a two-way adaptor and dropping fun- under reduced pressure, to ensure that the solution pores of the support. The solution was decanted to ess liquid. The catalyst was dried by rotary evapora- K for 2h before final drying in an oven at 353K for a were a at exp [34]. Fr coeffic latter w no atte Thi dynam from 1 and discussion odynamic study was performed first to determine the tes of reaction within the system, followed by a series periments to substantiate these calculations. Once the tes had been established, a series of continuous flow ting experimentswas performed, to establish the over- scheme. dynamic aspects ial reaction framework developed is based on pre- ies of tri-and tetra-chloroethene syntheses from ] or 1,2-dichloroethene [31,32]. The starting point e is 1,1,2,2-tetrachloroethane rather than the indus- ock, 1,2-dichloroethane, since the reaction sequence chloroethane to trichloroethene and tetrachloroethene ugh 1,1,2,2-tetrachloroethane and the reaction branch- loroethene or tetrachloroethene occurs after this point ion sequence [30–32]. The simplified scheme is shown enthalpies of formation and entropy values are avail- all reactants andproducts involved. Standardentropies ere calculated directly, standard enthalpies of reaction pared with from 1,1,2, from 1,1,2, −77.4 kJmo trichloroeth 673K. Hexa ble at temp at above 5 Trichloroet temperatur Althoug plifications system. The the reaction 3.2. Reactio The rea Cl2 at 373 tachloroeth to pentachl ture and to a series of s formed also tion 6 in Fig CHCl2CCl3. hydrochlorination/hydrochlorinations; 2, 3, 4 and 6 are chlorina- sible through Hess’s Law [34] and reaction enthalpies ental temperatures were derived using Kirchoff’s Law hese, followGibbs free energies of reaction, equilibrium and the equilibrium distribution of the products. The calculated by considering the isolated reactions only; was made to model the system as a whole. plified approach indicates that CCl2 CCl2 is the thermo- inimum,witha total freeenergychange for its formation -tetrachloroethane of −112.4 kJmol−1 at 673K, com- −2.4 kJmol−1 for the formation of trichloroethene 2-tetrachloroethane. Formation of pentachloroethane 2-tetrachloroethane has a free energy of reaction of l−1 at 673K; formation of pentachloroethane from ene has a free energy of reaction of −64.1 kJmol−1 at chloroethane formation is thermodynamically unfeasi- eratures above 523K; any hexachloroethane present 73K will break down to form tetrachloroethene. hene formation is thermodynamically feasible at all es used. h these findings are not definitive because of the sim- made, they are a useful guide to the chemistry in the y demonstrate what reactions are possible at each of temperatures used within the study. ns under static conditions ction of an equi-molar mixture of CHCl2CHCl2 and K in a stainless steel vessel produces only pen- ane. Increasing the reaction temperature to 423K leads oroethane as a major component of the product mix- hexachloroethane as a minor product, consistent with equential chlorination reactions. At 473K CCl2 CCl2 is , possibly via the thermal decomposition of C2Cl6 (reac- . 2) or through a wall-catalysed dehydrochlorination of 6 I.W. Sutherland et al. / Applied Catalysis A: General 399 (2011) 1–11 Fig. 3. Chlorin experiments ( in Section 2). of shading: te grey], hydroge hatch] and pen Dehydro CHCl CCl2 chloride [30 not due to Pyrex glass of CHCl2CH broadly con important stainless ste tentwith th such as FeC stepwise ch anism, the CHCl2CHCl2 FTIR spectro The reac at 503K lea CCl2 CCl2 ( ca. 3%). As t tively high be dehydro and 5 respe The hom ground aga can be sens pulgite sup 3.3. Reactio Anoverv 3:1 mole ra under flow heterogene cial CuII/KC Fig. 3. These d information ent conditi in Cl2 from In principle dehydrochlorination [15,16] to form CHCl CCl2, the only product observed under these conditions (reaction 1 in Fig. 2). Reactions over either bare support or catalyst show small drops in Cl2 levels, believed to be due to the consumption of Cl2 by atio t som roeth lgite r at t ove iffere e te ave a of p thre -act abov n ac In ve on Cu ffere . The , for catio sions uent Cl co f the ion 3 avio bsen port in tur 1,1,2 sence ction me time ctio sence ete; u ion o actio e mass balances (g atom) for 3:1 CHCl2CHCl2:Cl2 continuous flow reactor packed with 100mg solid support or catalyst as described The molecules are distinguished by the following arrangements trachlroroethane [white], dichlorine [black], trichloroethene [light n chloride (calculated) [dark grey], tetrachloroethene [single diagonal tachloroethane [double diagonal hatch]. chlorination of CHCl2CHCl2 and CHCl2CCl3 to form and CCl2 CCl2 respectively can be catalysed by iron(III) ]. To ensure that the reaction chemistry observed was catalysis by a chlorided layer on the reaction vessel, a reactorwasused to carry out the reactions. Chlorination Cl2 in Pyrex, although producing product distributions sistent with those obtained in stainless steel, has one difference. The tetrachloroethene formed at 473K in el is absent when the reaction vessel is Pyrex, consis- e dehydrochlorination of CHCl2CCl3 requiring a catalyst l3 for the pathway to be facile at this temperature. The lorination observed is consistent with a radical mech- initial step being abstraction of a hydrogen atom from to form HCl. Formation of the latter was confirmed by scopy of the product mixtures. tion between equi-molar CHCl2CHCl2 and Cl2 in Pyrex ds to CHCl CCl2 as the major product (yield ca. 50%), yield ca. 35%) andCHCl2CCl3 as theminorproduct (yield chlorin rial bu trichlo attapu greate formed little d the sam to beh mation The a redox trated 3.3.3, a moter. based been o moters phases tion of disper conseq ever, K some o in Sect Beh in thea or sup detail 3.3.1. the pre Rea the com idence the rea the ab compl format The re he reaction produces CHCl CCl2 and CCl2 CCl2 in rela- yields, these newly accessible pathways are believed to chlorination of CHCl2CHCl2 and CHCl2CCl3 (reactions 1 ctively in Fig. 2). ogeneous batch reaction study thus provides the back- inst which reactions carried out under flow conditions ibly evaluated and catalysis involving CuII or the atta- port identified. ns under flow conditions iewof the chlorinemass balance results obtainedwhen tio mixtures of CHCl2CHCl2 and Cl2 are allowed to react conditions in the presence of packed quartz (i.e. with no ous catalyst present), calcined attapulgite or a commer- l/attapulgite catalyst at various temperatures is given in ata demonstrate chlorine balance and provide useful about chlorination reactions observed under differ- ons. For reactions over quartz there was no decrease 298 to 653K, indicating that dichlorine is regenerated. it could be used in the initiation step of the radical is 100% wit observed in This strong does not af to give CHC (yield =8%) role in con chlorinatio 3.3.2. 1,1,2 the presence Doublin the contact resulted in balance sho The deh 573K and observed a enhanced b perature at over 100m withCuCl2 ns. They are small in the case of the support mate- ewhat greater over the catalyst at 653K. Mixtures of ene and tetrachloroethene are formed over calcined at 623 and 653K, the extent of reaction being slightly he higher temperature. Although some CHCl CCl2 is r the catalyst, CuII/KCl/attapulgite at 623K, its yield is nt from that obtained from the uncatalysed reaction at mperature. At 653K, the ability of CuII/KCl/attapulgite t a catalyst for chlorination is demonstrated by the for- entachloroethane in addition to CHCl CCl2. e components of the commercial catalyst used here are: ive, CuII species,whose influenceonchlorination is illus- e and will be exemplified further in Sections 3.3.1 to idic support, attapulgite, and KCl, regarded as a pro- ry recent studies of the ethene oxychlorination catalyst Cl2/�-alumina, convincing experimental evidence has d for the roles of KCl and other ionic metal chloride pro- y may be agents for the formation of mixed chloride example KxCuCl2+x, competitors for Cu2+ for the satura- n vacancies in alumina and for the modification of Cu2+ . In all cases changes in CuII/CuI redox behaviour and changes in catalytic activity would result [35]. How- uld act additionally as a Lewis base and so neutralize attapulgite surface acidity. This property is illustrated .3.7. ur under these three sets of conditions, over the catalyst, ce of catalyst and in thepresenceof unmodified support modified by KCl impregnation are considered in more n below. ,2-Tetrachloroethane with dichlorine (3:1mol ratio) in of the CuII/KCl/attapulgite catalyst (100mg) of CHCl2CHCl2 with Cl2 (3:1mol ratio) over 100mg of rcial catalyst (total volumetric flow 25cm3 min−1, res- 14.6 s, contact time 2.2 s, WHSV 0.83h−1) resulted in n profile which was rather similar to that obtained in of catalyst. The carbon mass balance was effectively nder these conditions there was no evidence for the f oligomeric products. The onset of reaction is 573K. n selectivity towards trichloroethene at 573 and 623K h conversions that are only slightly reduced from those the homogeneous reaction (see below Section 3.3.4). ly suggests that, under these conditions, the catalyst fect materially the dehydrochlorination of CHCl2CHCl2 l CCl2. At 653K a small quantity of CHCl2CCl3 is formed , cf. Fig. 3. At this temperature, the catalyst does play a trolling reaction selectivity, its role being to promote n. ,2-Tetrachloroethane with dichlorine (3:1mol ratio) in of the CuII/KCl/attapulgite catalyst (200mg) g the mass of catalyst used to 200mg and thus doubling time,which should emphasise the catalytic component, the CHCl2CHCl2 conversion profile and carbon mass wn in Fig. 4(a) and (b). ydrochlorination product, CHCl CCl2 is observed at above, while the chlorination product, CHCl2CCl3 is t 623K and 653K, Table 2. Formation of CHCl2CCl3 is y the greater catalystmass, 200mgvs. 100mg. The tem- which its formation is observable decreases from653K, g catalyst to 623K. Formation of CHCl2CCl3 is consistent acting as the active catalytic species for chlorination, in a I.W. Sutherland et al. / Applied Catalysis A: General 399 (2011) 1–11 7 Fig. 4. Tempe (b) mean carb 25 cm3 min−1; are standard d 10min interva manner sim may regene However, e with respec TheCHC from ca. 57 Formation o being 46% a 653K, respe to a gas-ph explored fu The carb ability but t at the highe 95%, almost 3.3.3. Trich presence of The poss ence of the conditions CHCl CCl2: Table 2 Meanvaluesa o catalyst. Temperature 523 573 623 653 a Determine b Yield and s empe of cata n mas he ab tures lues o . ly, th ed a dout n of C prod ]. Th lised on the basis of a very rapid rate of dehydrochlorination rature relationships for (a) mean conversions of CHCl2CHCl2 and on mass balance, for 3:1 CHCl2CHCl2:Cl2, total volumetric flow 200mg catalyst; over the temperature range 523–653K. Error bars eviations on the mean values of five measurements conducted at ls. ilar to that observed previouslywith CuCl2 inwhich Cl2 rate the active species by re-oxidising CuI to CuII [9–11]. ven at 653K, the yield of CHCl2CCl3 and the selectivity t to its formation are relatively small, Table 2. l2CHCl2 feedstock conversion at 653K increases slightly % over 100mg of catalyst to ca. 63% over 200mg, Fig. 4. f CHCl CCl2 is not significantly altered however, yields Fig. 5. T absence of carbo both in t tempera mean va intervals pected produc be rule rinatio as the [10,11 rationa nd 45% over 100mg and 200mg catalyst (Table 2) at ctively. These observations suggest its formation is due ase radical dehydrochlorination process; this aspect is rther below in Section 3.3.4. on mass balances shown in Fig. 4(b) show some vari- here is no real evidence for carbon loss, except possibly st temperature, 653K, where mass balances drop to ca. within experimental error. loroethene with dichlorine (3:1mol ratio) in the the CuII/KCl/attapulgite catalyst (200mg) ibility that CHCl CCl2 could be chlorinated in the pres- CuII/KCl/attapulgite catalyst was investigated under similar to those used above (over 200mg catalyst, 3:1 Cl2, total volumetric flow 25mlmin−1). Rather unex- f product yields and selectivities,b 3CHCl2CHCl2:Cl2 feed over 200mg (K) Mean product yield (%) Mean selectivity (%) w.r.t. CHCl CCl2 CHCl2CCl3 CHCl CCl2 CHCl2CCl3 0 3.8±0.2 0 20.9±0.5 6.9±0.2 75.1±6.5 24.7±1.6 45.0±0.4 16.6±0.5 69.8±1.6 25.7±0.5 d from 4–9 measurements made at 10min intervals. electivity defined following [23]. under these CHCl CCl2 products, p 3.3.4. 1,1,2 under homo From th the chlorin but the con CHCl2CHCl2 vious work as a radical the Lewis a behaviour o similar con experiment Fig. 3. Convers 653K is sh CHCl2CHCl2 653K, there tion. Formati mean conv mean yield ysed case, T reactions o rature relationships for, (a) mean conversions of CHCl2CHCl2 in the lyst or support; feedmol ratio of 3:1 CHCl2CHCl2:Cl2, (b)mean values s balances corresponding to (a), (c) mean conversions of CHCl2CCl3, sence of catalyst or support; feed mol ratio of 3:1 CHCl2CCl3:Cl2. The are in the range 473–653K. Error bars are standard deviations on the f four (a and b) or five (for c) measurements conducted at 10min e only product observed was CCl2 CCl2, which was t 653K. Althoughadirect substitutionof Cl-for-H cannot , CCl2 CCl2 ismore likely tobe formedviadehydrochlo- HCl2Cl3 (reaction 5 in Fig. 2). This can be accounted for uct of, CuII mediated, addition of Cl2 to the C C bond e absence of CHCl2CCl3 in the product stream can be conditions. It seems likely therefore that chlorinationof is a major route in the formation of chlorinated reaction roviding a chlorination catalyst is present. ,2-Tetrachloroethane with dichlorine (3:1mol ratio) geneous conditions i.e. in the absence of CuII catalyst e experiments described above, the role of CuII in ation of CHCl2CHCl2 and CHCl CCl2 is implicated ditions required to promote dehydrochlorination of and CHCl2CCl3 are not so clearly defined. From pre- , dehydrochlorination could be promoted either by Cl2 initiator [15,16] or, heterogeneously [18,19,21,22], by cid sites present on the attapulgite support [20]. The f CHCl2CHCl2/Cl2 and CHCl2CCl3/Cl2 flows under very ditions to those used in Section 3.3.1 was studied; these s supplementing the data given for CHCl2CHCl2/Cl2 in ion of CHCl2CHCl2 at temperatures between 473 and own in Fig. 5(a); Carbon mass balance data for the case are shown in Fig. 5(b); at the highest temperature, is some evidence for loss, possibly due to oligomerisa- on of CHCl CCl2 is detected at 573K; at 623K and 653K ersions of CHCl2CHCl2 are enhanced and CHCl CCl2 s, Table 3, are rather comparable to those in the catal- able 2, bearing in mind that uncatalysed chlorination f CHCl2CHCl2 and CHCl CCl2 do not occur. 8 I.W. Sutherland et al. / Applied Catalysis A: General 399 (2011) 1–11 Fig. 6. Schematic of the mechanism proposed for the formation of trichloroethene and of pentachloroe The behaviour observedwith respect to formation of CHCl CCl is little diffe of CHCl2CH Cl2. Howev than that o level of deh observation ation via ho Cl•, will be (339kJmol CHCl2CHCl2 of CHCl2CH to yield HCl the latter w erated and dehydrochl competitive nationof CH kineticallyw In both cas postulated, quent step would be in favoured. C tion consum dehydrochl Compari catalysed d with the co heterogene conclusion important a Table 3 Mean yield va Temperatur 523 573 623 653 523 573 623 653 a Determine following [23] 3.3.5. Penta eneo vers tho c). I the c hrou sole l2CC in Fi urab uan nd o CCl l yiel mea Fig. 5 own simi sma Cl3 as a bove nific hlor Trich eneo ch st to l/att ed ov flow 2 rent from that observed in the thermal decomposition Cl2 under comparable conditions but in the absence of er, the onset temperature is significantly lower, 80K, f thermal decomposition. Thus Cl2 induces a greater ydrochlorination activity, in agreement with previous s in other, related reactions [15–17,27,30]. Radical initi- molytic Cl Cl bond (243kJmol−1) scission [33] to form feasible at a lower temperature than the C-Cl scission −1) [33]necessary to facilitate thermaldecompositionof . For radical initiation via Cl•, the dehydrochlorination Cl2 should occur via abstraction of an hydrogen atom and a carbon centred radical intermediate. Collapse of ill yield CHCl CCl2 and Cl•. The Cl radical is thus regen- is effectively a homogeneous catalytic species for the orination. The process is compared with the potentially chlorination to give CHCl2CCl3 in Fig. 6. Radical chlori- Cl2CHCl2 to formCHCl2CCl3 appears tobe less favoured ith respect to dehydrochlorination to formCHCl CCl2. es, a common, carbon-centred radical intermediate is Fig. 6; however, in the dehydrochlorination the subse- is intramolecular, while that required for chlorination termolecular. Thus dehydrochlorination is kinetically hlorinemass balance data in Fig. 3 confirm that the reac- es little or no Cl2, which supports the model of radical orination. ng the reaction characteristics for homogeneously ehydrochlorination of CHCl2CHCl2, Fig. 5(a) and Table 3, rresponding reaction performed in the presence of the ous catalyst, Fig. 4(a) and Table 2, also leads to the that a Cl2-catalysed dehydrochlorination pathway is nd possibly dominant in the latter case. homog Con ilar to Fig. 5( above, 100% t is the of CHC tion 5 unfavo small q front e to form overal The 653K, (not sh under atively CHCl2C of Cl2 (see a the sig hydroc 3.3.6. homog The contra CuII/KC observ metric luesa for uncatalysed reactions. e (K) Organic component of the feed Product Mean yield (%) CHCl2CHCl2 CHCl CCl2 0 CHCl2CHCl2 CHCl CCl2 3.7±0.1 CHCl2CHCl2 CHCl CCl2 25.9±2.0 CHCl2CHCl2 CHCl CCl2 57.3±1.4 CHCl2CCl3 CCl2 CCl2 0 CHCl2CCl3 CCl2 CCl2 5.1±0.4 CHCl2CCl3 CCl2 CCl2 22.4±1.8 CHCl2CCl3 CCl2 CCl2 33.8±1.4 d from 4–9 measurements made at 10min intervals; yield defined . It appears t tion 3 in Fig 3.3.7. 1,1,2 the presence support To comp of Lewis ac analogy wit an effect on prisingly, a also. The ef of the two contained i thane under homogeneous conditions, illustrating the roles of Cl• . chloroethane with dichlorine (3:1mol ratio) under us conditions i.e. in the absence of CuII catalyst ion data for CHCl2CCl3/Cl2 flows under conditions sim- se reported for CHCl2CHCl2/Cl2 above are shown in n contrast to the results for CHCl2CHCl2, described arbon mass balance for the CHCl2CCl3, not shown, was ghout the temperature range studied. That CCl2 CCl2 identified product from the homogeneous reaction l3 in the presence of Cl2 is to be expected (reac- g. 2). Chlorination of CHCl2CCl3 is thermodynamically le at these temperatures. It is, however, possible that tities of C2Cl6 could be formed in the cooler zone at the f the reactor; if formed, C2Cl6 would decompose readily 2 CCl2 (reaction 6 in Fig. 2) and thus contribute to the d. n value of the conversion of CHCl2CCl3 is ca. 34% at (c). This is approximately 10% higher than that found here) from the thermal decomposition of CHCl2CCl3 lar flow conditions but in the absence of Cl2. The rel- ll increase is attributed to the great susceptibility of to thermal decomposition; consequently, the effect radical initiator is less than in the CHCl2CHCl2 case ). This is an important observation as it illustrates ant difference between the otherwise closely related, ocarbons. loroethene with dichlorine (3:1mol ratio) under us conditions i.e. in the absence of CuII catalyst lorination of CHCl CCl2 was investigated but in the situation encountered in the presence of a apulgite catalyst (Section 3.3.3), no reaction was er the range 523–623K (3:1 CHCl CCl2:Cl2, total volu- 25 cm3 min−1, residence time14.6 s, contact time2.2 s). herefore that a CuII catalyst is required to observe reac- . 2. ,2-Tetrachloroethane with dichlorine (3:1mol ratio) in of the unmodified or KCl-modified attapulgite lete the study, the possible effect on reaction pathways id sites on the attapulgite surface was investigated. By h earlier work cited [18,19,21,22] such sites could have dehydrochlorination pathways; however, more sur- role for the support in a chlorination step was indicated fect of the presence of KCl is clear from a comparison reactions that involve CHCl2CHCl2 with Cl2. Details are n Fig. 7(a) and (b) and in Table 4. I.W. Sutherland et al. / Applied Catalysis A: General 399 (2011) 1–11 9 Fig. 7. Tempe 3:1mol ratio C range 523–653 the temperatu dard deviation nine measurem The two ditions, a 3 or KCl-imp balances ar 673K exam for oligome data are no Although files, Fig. 7( undoped- o which are g pulgite, for Table 4 Mean valuesa doped support Temperature 523 573 623 653 523 573 623 673 a Determine b Yield and s mean value of the yield is 61.3% at 653K. These observations are not too dissimilar to those described above for reaction in the absence of a catalyst (Table 3). In view of the lack of a clear enhancement by attapulgite, the most important pathway for the formation of CHCl CCl2 from CHCl2CHCl2 (reaction 1 in Fig. 2) is considered to be homogeneous rather than heterogeneous. The unexpected feature of the reaction of CHCl2CHCl2 +Cl2 over undoped attapulgite is the formation of CCl2 CCl2. This reaction is well established at 653K. The mean values of the yields in relation to the mean conversion are in line with what might be expected, the mean value of the selectivity with respect to CCl2 CCl2 being 20.5%, Table 4. The most obvious route to CCl2 CCl2 is by dehy- drochlorination of CHCl2CCl3 (reaction 5 in Fig. 2). This implies that attapulgite is a catalyst for the chlorination of CHCl2CHCl2 to CHCl2CCl3 (reaction 2 in Fig. 2) as well as for CHCl2CCl3 dehy- drochlorination. In marked contrast, impregnation of attapulgite with KCl results in CHCl CCl2 as the sole product, Table 4. The observed behaviour over undoped-attapulgite is therefore consis- tent with the sequence, reaction 2 followed by reaction 5 such that e of consumption of CHCl2CCl3 exceeds its rate of formation. s basis CHCl2CCl3 is an unobserved intermediate in the for- of CCl2 CCl2. In contrast, there is no evidence for CCl2 CCl2 ion when the reaction is carried out over attapulgite that en impregnated with KCl. The profile and carbon mass bal- e both very similar to those obtained from the reaction of in sence of a catalyst (Fig. 5 and Table 3). Comparing the data l-impregnated attapulgite in Table 4 with the data for the lysed reaction, Table 3, indicates that, although mean values Cl2CHCl2 conversions are broadly comparable, mean yields l CCl2 are rather higher in the former case, particularly at er temperatures examined. the rat On thi mation format has be ance ar the ab for KC uncata for CH of CHC the low rature relationships for (a) mean conversions of CHCl2CHCl2 for a HCl2CHCl2:Cl2 over 100mg bare support and over the temperature K; (b) mean conversions of CHCl2CHCl2 under identical conditions in re range 523–673K but over KCl-doped support. Error bars are stan- s on the mean values of five (or for the highest temperature in (a) ents conducted at 10min intervals. reactions were carried out under identical flow con- :1mol ratio CHCl2CHCl2:Cl2 over 100mg attapulgite regnated attapulgite. In both cases 100% carbon mass e observed over the temperature range, 523–653 or ined. In neither case therefore is there any evidence risation and, for this reason, the carbon mass balance t presented. theCHCl2CHCl2 meanconversionvs. temperaturepro- a) and (b), are very similar, whether the solid phase is r KCl-doped attapulgite, the products formed, details of iven in Table 4, are very different. Over undoped atta- mation of CHCl CCl2 is observed from 573K and the of yield and selectivities,b 3CHCl2CHCl2: Cl2 feed over bare- or KCl- , 100mg in each case. (K) Dopant Mean yield (%) Mean selectivity w.r.t. (%) CHCl CCl2 CCl2 CCl2 CHCl CCl2 CCl2 CCl2 – 0 0 – 6.2±0.3 0 82.2±1.7 – 31.5±2.2 7.8±0.4 – 61.3±1.3 14.7±0.3 85.4±3.6 20.5±1.1 KCl 0 0 KCl 8.4±0.2 0 KCl 41.4±1.4 0 KCl 75.6±1.5 0 d from 4–9 measurements made at 10min intervals. electivity defined following [23]. Fig. 8. Tempe mean values o bare support, deviations on surements con rature relationships for, (a) mean conversions of CHCl2CCl3 and (b) f carbon mass balances, for feed 3:1mol ratio CHCl2CCl3:Cl2; 100mg over the temperature range 473–653K. Error bars are standard the mean values of five (or for the highest temperature) seven mea- ducted at 10min intervals. 10 I.W. Sutherland et al. / Applied Catalysis A: General 399 (2011) 1–11 Fig. 9. Pathways for the chlorination and dehydrochl It is pr KCl-impreg material, is formation o to give CHC (reactions 2 acid sites, w basic chlori the ability o tion, for exa [36], canbe or KCl [37]. chlorination catalyst, as are neutrali 3.3.8. Penta presence of It was s the dehydr is enhance dehydrochl over attap data for CHCl2CCl3: Fig. 8(a) and Values o specified in In thepr to form CC concluded t nation cata sites on the substantial rination ov acids are ca [18,19,21], attapulgite. Table 5 Mean values o Temperatur 473 523 573 623 653 a Determine to [23]. carb from sion s 10 ass b eric wnst serv ed, in reas esen s wit A pro orina in Fi ction ce o KCl/a in th with hydro erall lysed orina Cl3 a oure le th CuC l2CH , sug ing p by d le rou ere c ata s mati oposed that the different behaviour observed over nated attapulgite, compared with that over the bare consistent with the ability of attapulgite to catalyse f CCl2 CCl2 via catalytic chlorination of CHCl2CHCl2 l2CCl3 followed by dehydrochlorination of CHCl2CCl3 and 5 in Fig. 2); evidently chlorination occurs at Lewis hich are lost when attapulgite is impregnated with the de, KCl. In related work it has been demonstrated that f a bare�-alumina surface to catalyse dehydrochlorina- mple of 1,2-dichloroethane to vinyl chloride monomer attenuatedbydopingwith ionic chlorides suchasMgCl2 However, this pathway is unlikely to be important for processes in the presence of the CuCl2/KCl/attapulgite the strongest Lewis acid sites that would be required zed by KCl. chloroethane with dichlorine (3:1mol ratio) in the unmodified attapulgite support hown in Section 3.3.5, that CCl2 CCl2 is formed by ochlorination of CHCl2CCl3. Conversion of CHCl2CCl3 d significantly compared with an homogeneous orination, when CHCl2CCl3/Cl2 mixtures are flowed ulgite. The conversion and carbon mass balance a reaction carried out under the conditions of Cl2 = 3:1mol ratio over attapulgite (100mg) is given in (b). f the mean yields of CCl2 CCl2 at the temperatures Fig. 8 can be found in Table 5. esenceof attapulgite, dehydrochlorinationofCHCl2CCl3 l2 CCl2 (reaction 5 in Fig. 2) is substantial and it is hat the bare support material acts as a dehydrochlori- lyst in this case, consistent with catalysis by Lewis acid support [18,19,21]. This is in contrast with the lack of evidence for the catalysis of CHCl2CHCl2 dehydrochlo- er attapulgite discussed above. Although strong Lewis pable of acting as generic dehydrochlorination catalysts this is not always the case for the clay mineral, calcined The ferent conver reache bon m oligom the do was ob achiev the inc ates pr specie 3.3.9. Chl shown Rea presen CuCl2/ those nated the de cess ov is cata Chl CHCl2C are fav plausib ported of CHC pulgite involv lowed possib Wh ance d the for f CCl2 CCl2 yields,a 3CHCl2CCl3:Cl2 feed over 100mg bare support. e (K) Mean yielda (%) 0 30.8±2.1 67.7±1.3 101.4±1.1 85.0±0.9 d from 5–7 values determined at 10min intervals; defined according the flow of arises unde steel. Lewis the putativ nations and Reaction tionswhere to be comp tions and th including t of HCl to C [40]. orination reactions. on mass balance shown in Fig. 8(b) is significantly dif- those obtained from the other reactions; once the of CHCl2CCl3 to give the volatile product CCl2 CCl2 0%, observed at 623K, Fig. 8(a), a decrease in the car- alance is observed, suggesting formation of involatile products. These were observable as a purple solid in ream arm of the U-tube reactor. The mass imbalance ed only after 100% conversion to CCl2 CCl2 had been dicating that oligomer formation is a result, either of ed concentration of carbon centred radical intermedi- t at high conversions or of the presence of unsaturated hin the reactor. posed reaction scheme tion anddehydrochlorinationpathways established are g. 9. s that involve CHCl2CHCl2/Cl2 mixtures in the f calcined attapulgite or the commercial catalyst ttapulgite are rather similar in their characteristics to eir absence or in the presence of attapulgite impreg- KCl. This indicates that an heterogeneous pathway for chlorination is, at best, a minor contributor to the pro- . In marked contrast, dehydrochlorination of CHCl2CCl3 strongly by the Lewis acid attapulgite. tion under flow conditions, of CHCl2CHCl2 to give nd HCl and of CHCl CCl2 also to give CHCl2CCl3, Fig. 9, d by the catalyst. Following previous work [10,11], it is at the heterogeneous chlorination route involves sup- l2 as the active catalytic species. However the behaviour Cl2/Cl2 when flowed over calcined clay mineral, atta- gests that Lewis acid site induced chlorination, perhaps olarization of Cl-Cl at very strong Lewis acid sites, fol- ehydrochlorination of the intermediate CHCl2CCl3 is a te from CHCl2CHCl2 to CCl2 CCl2. onversion of the feed is substantial, carbon mass bal- uggest that loss of unsaturated products occurs with on of oligomeric material. This is most pronounced in CHCl2CCl3 over calcined attapulgite. A similar situation r static conditions for reactions carried out in stainless acid catalysis is implicated. Though not reported here, e role of iron(III) chloride in hydrochlorocarbon chlori- subsequent reactions will be reported elsewhere [38]. s reported here have been investigated under condi- the dichlorine in the feed is insufficient for chlorination lete. The effects of operation under chlorine-rich condi- e effect of alteration of the chlorine supply generally, he intervention of the Deacon reaction, i.e. conversion l2 [39], will be reported in a forthcoming publication I.W. Sutherland et al. / Applied Catalysis A: General 399 (2011) 1–11 11 Acknowledgements We thank EPSRC, Ineos ChlorVinyls and the CRYSTAL Faraday Partnership for support of this work. References [1] R.L. Powell, J. Fluorine Chem. 114 (2002) 237–250. [2] A.A. Lindley, A. McCulloch, J. Fluorine Chem. 126 (2005) 1457–1462. [3] M.A. Subramanian, L.E. Manzer, Science, 297 (2002) 1665; M.A. Subramanian, T.G. Calvarese, ‘Fluorineand theEnvironment’, Adv. FluorineSci., in:A. Tressaud (Ed.), Elsevier, Amsterdam, 2 (2006) 203. [4] J. Kanters, R. Louw, Chemosphere 32 (1996) 89–97. [5] J. Clayden, N. Greeves, S. Warren, P. Wothers, Organic Chemistry, OUP, Oxford, 2001, pp. 504–505, 1035–1039. [6] I.M. Dahl, E.M. Myhrvold, U. Olsbye, R. Rohr, O.A. Rokstad, O. Swang, Ind. Eng. Chem. Res. 40 (2001) 2226–2235. [7] M.G. Bryukov, I.R. Slagle, V.D. Knyazev, J. Phys. Chem. A 107 (2003) 6565–6573. [8] R.S. Iyer, F.S. Rowland, J. Phys. Chem. 89 (1985) 3730. [9] C.M. Fontana, E. Gorin, G.A. Kidder, C.S. Meredith, Ind. Eng. Chem. 44 (1952) 363–368. [10] R.P. Arganbright, W.F. Yates, J. Org. Chem. 27 (1962) 1205–1208. [11] P.P. Nicholas, R.T. Carroll, J. Org. Chem. 33 (1968) 2345–2349. [12] M. Poutsma, J. Am. Chem. Soc. 87 (1965) 2172–2183. [13] K.P.C. Vollhardt, Organic Chemistry, W.H. Freeman, New York, 1987. [14] A.L. Roberts, P.M. Gschwend, Environ. Sci. Technol. 25 (1991) 76–86. [15] D.H.R. Barton, K.E. Howlett, J. Chem. Soc. (1951) 2033–2038. [16] D.H.R. Barton, J. Chem. Soc. (1949) 148–155. [17] D.H.R. Barton, K.E. Howlett, J. Chem. Soc. (1949) 155–164; D.H.R. Barton, K.E. Howlett, J. Chem. Soc. (1949) 165–169; D.H.R. Barton, P.F. Onyon, Trans. Farad. Soc. 45 (1949) 725–735; D.H.R. Barton, P.F. Onyon, J. Amer. Chem. Soc. 72 (1950) 988–995. [18] C. Pistarino, E. Finnochio, G. Romezzano, F. Brichese, R. di Felice, G. Busca, M. Baldi, Ind. Eng. Chem. Res. 39 (2000) 2752–2760. [19] C. Pistarino, E. Finnochio, M.A. Larrubia, B. Serra, S. Braggio, G. Busca, M. Baldi, Ind. Eng. Chem. Res. 40 (2001) 3262–3269. [20] Q.G.Wang,M. Tu,W.D.Wang, J.Y. Shen, Chem. J. ChineseUni. Chinese 18 (1997) 625–627. [21] T.H. Ballin T.H. Ball 2473–247 [22] D.G. McB (1990) 67 J. Thomso J. Thomso B. Adamc C.H. Barc Skapin, J. 40–47. [23] Chem. Eng.News, 83, July 11 (2005) 67; ICIS Chemical Business source, Septem- ber 15–21 (2008) 16–19. [24] T. Zhang, C. Troll, B. Rieger, J. Kintrup, O.F.-K. Schlüter, R. Weber, Appl. Catal. A 357 (2009) 51–57; T. Zhang, C. Troll, B. Rieger, J. Kintrup, O.F.-K. Schlüter, R. Weber, Appl. Catal. A 365 (2009) 20–27. [25] C.H. Bartholomew,R.J. Farrauto, Fundamentals of Industrial Catalytic Processes, second ed., John Wiley & Sons, Inc., Hoboken, New Jersey, 2006, p. 206. [26] K. Weissermel, H.-J. Arpe, Industrial Organic Chemistry, fourth ed., Wiley-VCH, Weinheim, 2003, pp. 449–450. [27] D.H.R. Barton, A.J. Head, Trans. Farad. Soc. 46 (1950) 114–122. [28] G. Clet, J.-M. Goupil, D. Cornet, Bull. Soc. Chim. Fr. 134 (1997) 223–233; G. Clet, J.M. Goupil, G. Szabo, D. Cornet, J. Mol. Catal. A: Chem. 148 (1999) 253–264; G. Clet, J.M. Goupil, G. Szabo, D. Cornet, Appl. Catal. A: Gen. 202 (2000) 37–47. [29] A.R. McInroy, D.T. Lundie, J.M. Winfield, C.C. Dudman, P. Jones, S.F. Parker, D. Lennon, and refs. therein, Catal. Today 114 (2006) 403–411. [30] T Kawaguchi, Y. Suzuki, R. Saito, Ind. Eng. Chem. 62 (1970) 36–41. [31] A. Arcoya, A. Cortés, X.L. Seoane, Ind. Eng. Chem. Prod. Res. Dev. 19 (1980) 77–82. [32] A. Arcoya, A. Cortés, X.L. Seoane, Ind. Eng. Chem. Prod. Res. Dev. 19 (1980) 82–86. [33] CRC, Handbook of Chemistry & Physics, 86th ed., CRC, Cleveland, 1998. [34] P. Atkins, Physical Chemistry, sixth ed., OUP, Oxford, 1999, pp. 65–72. [35] C. Lamberti, C. Prestipino, F. Bonino, L. Capello, S. Bordiga, G. Spoto, A. Zecchina, S. Diaz Moreno, B. Cremaschi, M. Garilli, A. Marsella, D. Carmello, S. Vidotto, G. Leofanti, Angew. Chem. Int. Ed. 41 (2002) 2341–2344; N.B. Muddada, U. Olsbye, L. Caccialupi, F. Cavani, G. Leofanti, D. Gianolio, S. Bor- diga, C. Lamberti, Phys. Chem. Chem. Phys. 12 (2010) 5605–5618; N.B. Muddada, U. Olsbye, G. Leofanti, D. Gianolio, F. Bonino, S. Bordiga, T. Fuglerud, S. Vidotto, A. Marsella, C. Lamberti, Dalton Trans. 39 (2010) 8437–8449. [36] E. Finocchio, N. Rossi, G. Busca, M. Padovan, G. Leofanti, B. Cremaschi, A. Marsella, D. Carmello, J. Catal. 179 (1998) 606–618. [37] D. Carmello, E. Finnochio, A. Marsella, B. Cremaschi, G. Leofanti, M. Padovan, G. Busca, J. Catal. 191 (2000) 354–363. [38] D.S.C. Taylor, I.W. Sutherland, D. Armstrong, D. Lennon, G. Milne, J.M. Winfield, submitted for publication. [39] C.N. Kenney, Catal. Rev. Sci. Eng. 11 (1975) 197–224; J. Villadsen, H. Livbjerg, Catal. Rev. Sci. Eng. 17 (1978) 203–272; . Pan, 6–30 Rouco .M. H Morte 1–21 ieken ueju, 05) 15 . Suth mitte ger, J.T. Yates Jr., J. Phys. Chem. 96 (1992) 1417–1423; inger, R.S. Smith, S.D. Colson, J.T. Yates Jr., Langmuir 8 (1992) 8. eth, J.M. Winfield, B. Cooke, N. Winterton, J. Chem. Soc., Dalton Trans. 1–676; n, G. Webb, J.M. Winfield, J. Mol. Catal. 67 (1991) 117–126; n, G. Webb, J.M. Winfield, J. Mol. Catal. 68 (1991) 347–354; zyk, A. Hess, E. Kemnitz, J. Mater. Chem. 6 (1996) 1731–1735; lay, H. Borzorgzadeh, E. Kemnitz, M. Nickkho-Amiry, D.E.M. Ross, T. Thomson, G. Webb, J.M. Winfield, J. Chem. Soc., Dalton Trans. (2002) H.Y 299 A.J. M.W M. 213 U. N L. X (20 [40] I.W sub R.G. Minet, S.W. Benson, T.T. Tsotsis, Ind. Eng. Chem. Res. 33 (1994) 03; , J. Catal. 157 (1995) 380–387; isham, S.W. Benson, J. Phys. Chem. 99 (1995) 6194–6198; nsen, R.G. Minet, T.T. Tsotsis, S.W. Benson, Chem. Eng. Sci. 54 (1999) 39; , O. Watzenberger, Chem. Eng. Sci. 54 (1999) 2619–2626; L. Jie, Z. Guangdong, Z. Kaiji, L. Wenxing, C. Tiexin, Catal. Lett. 100 3–159. erland, N.G. Hamilton, C.C. Dudman, P. Jones, D. Lennon, J.M. Winfield, d for publication. Chlorination and dehydrochlorination reactions relevant to the manufacture of trichloroethene and tetrachloroethene Introduction Experimental Experiments under static conditions Continuous flow experiments Description of the flow rig Reaction protocol Product analyses Experimental issues leading to possible ambiguities in the data Catalyst preparation and characterization Results and discussion Thermodynamic aspects Reactions under static conditions Reactions under flow conditions 1,1,2,2-Tetrachloroethane with dichlorine (3:1mol ratio) in the presence of the CuII/KCl/attapulgite catalyst (100mg) 1,1,2,2-Tetrachloroethane with dichlorine (3:1mol ratio) in the presence of the CuII/KCl/attapulgite catalyst (200mg) Trichloroethene with dichlorine (3:1mol ratio) in the presence of the CuII/KCl/attapulgite catalyst (200mg) 1,1,2,2-Tetrachloroethane with dichlorine (3:1mol ratio) under homogeneous conditions i.e. in the absence of CuII catalyst Pentachloroethane with dichlorine (3:1mol ratio) under homogeneous conditions i.e. in the absence of CuII catalyst Trichloroethene with dichlorine (3:1mol ratio) under homogeneous conditions i.e. in the absence of CuII catalyst 1,1,2,2-Tetrachloroethane with dichlorine (3:1mol ratio) in the presence of the unmodified or KCl-modified attapulgite sup... Pentachloroethane with dichlorine (3:1mol ratio) in the presence of unmodified attapulgite support A proposed reaction scheme Acknowledgements References