J. H. RALEY, R. D. MULLINEAUX, AND C. W. BITTNER
3174
[CONTRIBUTIOS FROM
THE SHELL
DEVELOPMEWT CO., EMERYVILLE, CALIF.]
High Temperature Reactions of Iodine with Hydrocarbons. BY
JOHN
H.
RALEY,
RICHARD D.
VOl. 85
h I U L L I N E A U X , AND
I. Dehydrogenation
CLARENCE w.BITTNER
RECEIVEDFEBRUARY 18, 1963 T h e stoichiometric, reversible reaction of elemental iodine with CP t o CS paraffins, cyclopentane, and inethylcyclopentane a t 2 500" produces monoolefinic, diolefinic, or acetylenic hydrocarbons with the same carbon skeleton as the reactant. Hydrogen iodide is formed in amounts equivalent t o the reacted iodine. Oxygen added to the reaction mixture reacts primarily with hydrogen iodide and the resulting iodine reacts with additional hydrocarbon. A free radical mechanism is proposed.
Introduction Xlthough the reaction of elemental iodine with terpene or hydroaromatic hydrocarbons to produce H I has long been known, iodine generally has been regarded as otherwise unreactive toward saturated hydrocarbons. We have found, however, t h a t a t high temperatures ( 2 3 3 0 ° ) , iodine reacts essentially stoichiometrically with hydrocarbons, including the lower paraffins, to form hydrogen iodide and a variety of unsaturated hydrocarbons. These reactions are described in a series of three papers. In part I the dehydrogenation of Cr-C5 paraffins, cyclopentane, and methylcyclopentane is discussed. Aromatization and rearrangement reactions are described in parts I13aand III.3b Experimental Apparatus and Procedure.-.% schematic diagram of the apparatus is shown in Fig. 1. \'apor from molten iodine, held a t about 250' (PI, -4 a t m . ) in a specially designed, Hastelloy C pressure vessel, was passed through a metering valve, rotameter, and connecting tubing (all constructed of Hastelloy C and heated ~:lectrically to 250-300') t o the reaction tube. A white heat lamp mounted near the rotameter permitted visual observation of the platinum float. Connection to the reaction tube was made by a spherical joint, the Hastelloy C ball being ground slightly under size t o allow for expansion. The hydrocarbon was also passed through a rotameter or, if liquid, alternatively delivered by a suitable pump.
sorption of radiant energy from the furnace elements by the thermocouples. The reaction times listed are those calculated from the volume of the tube, average temperature, and total moles of reactants and diluents introduced per unit time but neglecting volume changes due t o reaction or tionideal behavior of the gases. For reactions incorporating oxygen, tubes with two or three reaction sections separated by mixing zones were used. Oxygen could then be introduced either with the other reactants or after some dehydrogenation had occurred. In order to isolate the unsaturated hydrocarbon products and prevent their reacting with H I or residual I? after emergence from the reaction tube, the hot, gaseous effluent was quenched with a solution which preferentially absorbed and reacted with H I and 1 2 . In some experiments the gases, after passage through a small-volume, heated (400') Vycor tube, were bubbled through a solution of sodium acetate in ethylene glycol (weight ratio, 0 . 5 : l ) held a t 150'. In others, as shown in Fig. 2 , the gases were mixed rapidly, first with a large quantity ( e . g . , 10 volumes) of steam and then with a spray of a q . S a O H or a q . H I solution. Temperature reduction and neutralization were thus accomplished within a few milliseconds. The latter procedure was necessary for satisfactory recovery of diolefinic products. A series of condensers, separatory receivers, and traps then served t o separate the quench solution from the organic reaction products. Gaseous and liquid products were analyzed by mass spectrometric (m.s.) and gas-liquid chromatographic (g.1.c.) methods. Occasionally the liquids were distilled and the fractions analyzed by infrared. The g.1.c. analyses were normally carried out using a 50 f t . by 0.25 in. column packed with 2 i ) x , dimethylsulfolane on 40-60 mesh firebrick operated with a helium flow of 50 cc.
TABLE I REACTION OF IODIXEWITH ETHANE, PROPANE, A N D ISOBUTANE Flow Experiments a t One Atmosphere CZH~-----
C s HI.-------
7--
i-C~Hio
Temperature, "C. 600 685 525 550 685 495 121CwHzn+ z 0.64 4.6 0.66 0.86 1.73" 0 4 Reaction time, sec. 18 1.3 40 36 0.28 25 6 d b C h Quench Is reacted, yo ... 24 -80 -60 68 -90 Xfoles/100 moles C,Hzn+z introduced 56.0 50.1 2.9 66.5' 59.0 1 9 CnHzn+P 12 8 48.4 87.0 28.8" 37.7 72 0 CtzHzn CnHzw-z g o % p-xylene) csH16.
+
The quartz (\.pcor) reaction tube (1.2-4.0 c m . diameter, 43-l8D cc. reaction volume; Fig. 2 ) provided for separate preheating, followed by rapid and thorough mixing of the reactant vapots and inert diluent (SZ or H e ) , if present. The reported reaction temperature is an average of temperatures over the length of the reaction zone. The latter values varied considerably because of the endothermicity of the reactions and the ab11) ( a ) A E t a r d and H . Moissan, Bull. BUG c h i m , [ 2 ] 34, 69 (1880); B e l , 13, 1862 (1880), (b) F J . Nellensteyn, C h e m U'rekblad, 21, 102 (1924); C ' h r m . A b r t i . , 13, 1976 (1924); ( c ) K. Shishido a n d H . Pu'ozaki, J . SOL i'htm I x d . J a p a n , 47, 819 (19441, Chem A b s i r . , 42, 63Rlh (1948). ( 2 ) I H. Raley. U S . P a t e n t 2,901,518(1959); J . H.Raley and R D S l o l l i n e a u x , U. S P a t e n t 2,880,252(1959). 13) ( a ) R D . Mullineaux and J H . Raley, J . A m . C h e m . Soc., 86, 3178 ' 1 9 6 3 ) . ( b ) L.H Slaugh, R . D. Mullineaux, and J. H. Raley, ibid., 86, 3180 I
1967:
per minute a t 30 or 0'. T h e H I and I 2 contents of the gaseous product were inferred from analysis of the quench solution. With the glycol solution this determination was approximate because of side reactions. The reactant iodine was also determined from the rotameter setting or occasionally from the loss of weight from the pressure vessel. Materials.-Reagent grade (resublimed, Baker and Xdamson) iodine was used without further purification. Most of the hydrocarbons were commercially available "pure" grade materials of 99.5+yc purity. In general, these were andlyzed by the same procedures used for the reaction products. Cylinder oxygen, 99.9 Yc,was used a s received.
+
Results Typical data for the reaction of elemental iodine with ethane, propane or isobutane in a flow system are given
Oct. 20, 1963
DEHYDROGENATION OF HYDROCARBONS B Y IODINE DILUENT
1-
0
0H Y D R O C A R B O N
INSULATED C H A M I I R
t
fURNACE
IODINE VESSEL
3175
LIQUID HYDROCARBON
GAS E OU S PRODUClS
(H A S l E L LOY C)
U
Fig. 1.-Apparatus.
in Table I. If the reaction is suddenly halted by "quenching" the effluent vapors (see Experimental) the predominant organic product is the corresponding monoolefin. Under these conditions hydrogen iodide accounts fur 90% or illore of the abstracted hydrogen and nearly all of the reacted iodine. If the vapors are allowed to cool before neutralization, however, substantial amounts of organic iodides are formed. At very high conversion, ethane produces acetylene and propane gives methylacetylene and allene in significant quantities. Small amounts of dimeric products were also observed. The principal dimer from propane was benzene. From isobutane an aromatic fraction (>goyo p:xylene) and one or more octenes were found. Dimer yields were essentially independent of temperature, iodine mole fraction, or reaction time. Reaction of iodine with propylene or isobutylene, or pyrolysis of a 2:1 mixture of isobutylene and isobutyl iodide, gave only slightly larger amounts of the corresponding dimers. With n-butane and n- or isopentane, the monoolefinic products are accompanied by substantial amounts of the corresponding conjugated dienes (Table 11). At
I
19 cm.
i
THERMOCOUPLES
70 crn.
L
DIAMETER
TABLEI1 REACTIONO F IODINE
WITH TZ-BUTANE, %-PENTANE, A N D I SOPENTAN E
n-CIHia
Temperature, OC. 12/'CnHzn+ 2 Reaction time, sec. 1 2 reacted, 7 0 Moles/100 moles C,Hzn+2 introduced" Cn&+ 2b CnHznb CnHzn-zb Skeletal isomers C < n hydrocarbonse C > n hydrocarbonse Organic iodides
575 1.87 1.5 52
n-CaH)t
i-CaHiz
525 1.0 4.0 77
548 1.5 2.6 75
12.8 35.5 9.8 53.8 33.9 63.2 25.4 15.3 17.5 2.5" l.Od Xot detcd. 3.6 3.4 2.7 3.71 1.2 3.28 Trace 0.4 0.6 Hz 50.5 H?,.+ 2 converted, Tc CnH>,,2 T n € i z u C,,H.,,- 9 yield. SC
0 38
~
"
Calculated.
51 14
*
1 3 76 0 28 16
Equil 82
0 24 16
1sopentane ( I ~ C S H I1Z5 ,)
1 3' 69 0 42 19
2 6 90 0 28 18
Equil." 92
0.27 20
I Z / C ~ H I1.7. ?,
Effect of Oxygen.---The effect of adding oxygen to an iodine hydrocarbon system has also been determined. Typical results are given in Table lr for a mixture of n-butane and 1-butene. The oxygen reacts rapidly, and additional butane and n-butenes are consumed, principally to form butadiene. The major effect is that expected from ( I ) a selective oxidation of di) "Selected Values of Properties of Hydrocarbons and Related Comp o u n d i . ' A P I R r i e a r c h Project 44, Carnegie Institute of Technology. P i t t s b u r y h P e n n a , 1)ecernher 3 1 . j 9 . 2 ( 7 ) "Selected values of Chemical Thermodynamic P r o p e r t i e s , ' ' Series 111, l J S S a t 1 Bur Standards, Washington, D . C . , June 30, 1948. (8) H H Vuge and S C . M a y , J . A m C h e m . S o c . , 68, 550 (1946)
HI to form I?, and (2) reaction of the latter with the hydrocarbons toward establishment of a new eyuilibrium position. This view is supported by the agreement between the observed oxygen consumptions and those calculated for the examples of Table V from the measuredg rate of the vapor phase reaction %HI
+
'/202
+HzO
+ 12
Carbon monoxide was the only organic, oxygencontaining product found which could be attributed to hydrocarbon oxidation. Decomposition products, various C1-C3 hydrocarbons, increased slightly. The ratio of this increased decomposition to the CO formed suggests that, for each Cq species which reacts with oxygen, one carbon atom a m e a r s as CO and three as C1-C3 hydrocarbons TABLE l' EFFECTOF OXYGEN AT 550
+ 5'
React ants n-CdHIOIl-C4H8 I?/h? drocarbonb O?/hydrocarbons Reaction time, sec 0 2 reacted,
075 0 65 0 1 1
1 0 0 76 0 23 0 9 91 0
1 0 1 3
0 1 4
1 0
1 5 0 12 10" 86 4
Moles/100 moles hydrocarbon introducedb
167 9 9 21 0 12 6 n-C4HI; z-C4HgC 55 0 42 6 51 2 41 1 20 2 34 3 22 4 36 7 1,3-C1Hs C1-C3 h l drocarbonsd 1 6 5 8 1 2 3 2 0 1 2 CO 0 5 5 Condensationandoxidn prod 2 2 1 0 1 9 1 2 O2 admitted after 0.3 sec. * Steatn-SaOH quench. So skeletal isomers detected. Expressed as equivalents of reactant hydrocarbons.
lYith decreasing iodine or increasing oxygen initial concentrations, the iodine-catalyzed oxidation of the hydrocarbons increases continuously, culminating in excessively high reaction temperature. With n-butane the equilibria A and B were shifted in the expected manner also by the introduction of (9) G. G. Bay16 a n d C R G u m , private communication.
3ct. 20, 1963
DEHYDROGENATION OF HYDROCARBONS BY IODINE
tthylene. Ethane was produced and a corresponding idditional amount of n-butane and n-butenes were lehydrogenated. The HI-C2H4 reaction was suffi:iently slower than the I Z - C ~ Hreaction, ~Q however, t h a t rreversible decomposition reactions became important Jefore over-all equilibrium was established.
Discussion A free radical mechanism consisting of successive iydrogen abstraction reactions involving both atomic tnd molecular iodine encompasses all of the experinental observations. The formation of free radicals irom the hydrocarbon is attributed, as in iodine-cata.yzed pyrolyses,'O to attack by iodine atoms produced sy the equilibrium dissociation of I?. At 800'K. and 01, = 0.5 atm., the equilibrium pressure of iodine atoms s 3 mm." Reaction 2 is endothermic by 27 f 3 kcal./ IZ I
+ M 21'
+M 1 +H i + R-C-C-
+ R-C-C-l~ I I I
R-C-C-
I
H H
H
'
1
+ H i --+ R-C-C-I
I
1
H
+I
1
'
+R .
I I + C=C 1 H
H
(2')
(3)
1
With the high initial iodine concentrations in this study, iowever, another reaction leading to dehydrogenation supersedes 3. There are two paths which might be suggested
I 1
1
R-C-C-
+ I +R-C=C-I
1
+ Hi
(4)
+ HI + I
(5)
H
ind the over-all process
I
R-C-C-
1
+ Iz
I
R-C=C-
1
1 .
H
H
AH for reaction 5 for isopropyl and sec-butyl radicals -+5 kcal./mole. Although 4 should have a high specific rate in view of its exothermicity (-30 kcal./ nole for the isopropyl or sec-butyl radical), 5 must also 3e considered in view of the known rapidity of reactions setween alkyl radicals and 1215and the high ratio of molecular to atomic iodine. (10) ( a ) A. A Frost and R . G . Pearson, "Kinetics a n d Mechanism," John Uiley and Sons,Inc., New York, N.Y . , 1961, p. 204; (b) H . J . Schumacher, 'Chemische Gasreaktionen," Theodor Steinkopff. Leipzig, 1938, p p . 388-395; c ) G. K . Rollefson and R . F. F a d , J . A m Chem. S O L . ,69, 6 2 5 (1937); (d) j. Bairstow and C N. Hinshelwood, J . Chem S O L . 1155 , (1933). (11) Calculated from d a t a in ref 7. (12) Using DHI = 71 kcal./molelz and D R - H = 98 i 3 kcal./mole for i = CzHa, n-CaH?, n-CIHs 12.14 (13) T. I.. Cottrell, "The Strengths of Chemical Bonds," Academic Press, 'nc., New Y o r k , N Y , 1954, p. 272. (14) E . W. R. Steacie, "Atomic and Free Radical Reactions," Reinhold 'ublishing Corp., New York, N .Y . , 1954, pp. 95-98.
I I
+ 12
1 1
R-C-CH
+I
(6)
i
but, since these, a t 2300", are known'j l 6 to decompose rapidly, especially in the presence of iodine, or to react with HI, their role under the present conditions can be a t most that of an intermediate. The most plausible steps for 5 , including 6, are those suggested by Benson,lGwho has analyzed the original data of Jones and Ogg" for the Iz-catalyzed pyrolysis of n-CaH71. They are 7 and 8
,
H
1
I
I
'
(2)
H H
I
I I
R-C-C-
R-C-C-
mole for ethane and for attack a t the primary positions Jf higher alkanes. l 2 For secondary and tertiary positions in alkanes the corresponding values are estimated 3s 23 and 18 k ~ a l . / m o l e , 'respectively, ~ while for allylic positions the endothermicity should be -10 kcal./ mole. One should expect, therefore, a substantial selectivity for attack by atomic iodine on the various :arbon-hydrogen bonds and a reflection of this selectivity in temperature requirement. At very low initial I2 concentrations, radical decomposition by cleavage of P-carbon-carbon bonds prejominates3aS"h R-C-C-
At lower temperatures, alkyl iodides would be the expected products from 6
(1)
1
3177
I
I I
I If R-C=C-
t
+I
(8)
I
The sum of 6, 7 , and 8 is 5 . Furthermore, combination of the reverse reactions for 8, 7, 6, and 2 provides a route for the reverse of the over-all dehydrogenation reaction
I
R-C=C-
I
+ 2Hi +R-C-C-I I
1 1
+ iz
H H
BensonI6 has pointed out that 8 may participate in lowering the activation energy of 7. An additional possibility is that 6 and 7 are combined in a single reaction (9)
I I
R-C-CH
1 .
+ Iz T' R-C-C-I
1
H '
I .I. '
1
1
'
I
R-C-C-
I
+ HI (9)
A path for the reverse of the over-all reaction is also provided by the formation of alkyl iodide from H I and olefin,I8 followed by the reverse of 6 and of 2.'jh Alternatively, formation of an alkyl radical may occur directly from reaction of H I and an olefin by t h e reverse of 4, as suggested by Bose and Benson. The sequence of 2 and 4 or 5 also may be applied to the conversion of monoolefins to the corresponding CnH2n--2 products. Step 2 is faster for attack at an allylic position and, where conjugated diolefins can be formed, hydrogen atom abstraction from the allylic radical also is rapid. Therefore, k h j k , >> 1. Since both calculated and experimental results have established that KB/KAfor n-butane and n- and isopentane is in the range of 0.025 to 0.5 a t 475-575', kh,!ka >> K B / K A . Consequently, the diolefin/monoolefin ratio must increase continuously as the system is moved from an equilibrium- toward a rate-controlled composition. Also, since k b , >> kat, the partial equilibrium B is maintained a t much shorter reaction times than are required for establishment of A. Under such circumstances, the diolefin yield can equal or exceed (and the monoolefin yield will be less than) that corresponding to over-all equilibrium. With n-butane at 500' and 12/C4H10 = 1.3, the observed compositions of the system a t 2-7 sec. reaction time are in good agreement with those calculated from the measured I? conversions assuming attainment of the partial equilibrium B.
+
(15) ( a ) See ref 14, pp. 263-274, 733-747, for discussion of R . 12 + R1 I and thermal reactions of R I ; ( h ) S W Renson and E . O ' N e a l , J Chem. P h y s . , 34, 514 (1961) (16) S. W. Benson, ibid., 34, 521 (1961) (17) J . L. Jones and R. A. Ogg, J r . , J . A m . Chem. Soc., 69, 1931 (1937). (18) A. N . Bose and S W Benson, J Chem P h y s . , 37, 1081 (1962).
+
3178
RICHARD D. MULLINEAUX AND
The proposed mechanism also is qualitatively consistent with the results given in Table V. At high initial 1 2 concentration, the oxidation of hydrocarbon radicals v i a a sequence such as 10 should be minimized
+
.C~HQ
0 2
+ [CIHQOZ.] + .C2Hj
f [C,HsO.l
+ ,CH3 + CO (10)
1.C3H7 + CO
[CONTRIBUTION FROM
THE
JOHN
H. RALEY
VOl. 85
by 5 or 4 and by 2' and the oxygen principally consumed by reaction with HI. With 1 2 (and, hence, H I ) in smaller amounts, or at higher oxygen concentrations, 10 becomes more important. Acknowledgment.-The authors are indebted to L. H. Slaugh, E. F. Magoon, and J. J. Madison for performance of some of the experiments, and to 2 . V. Jasaitis for aid with the g.1.c. analyses.
SHELL DEVELOPMENT CO., EMERYVILLE, CALIF.]
High Temperature Reactions of Iodine with Hydrocarbons. BY RICHARD D. MULLINEAUX AND
JOHN
11. Aromatization
H. RALEY
RECEIVED FEBRUARY 18, 1963 Aromatic hydrocarbons and hydrogen iodide are the predominant products from the reaction of elemental iodine with hydrocarbons containing a chain of six or more adjoining, nonquaternary carbon atoms. T h e structures of the aromatic products are defined by the structure of the reactant hydrocarbon and a specific mode of decomposition of a cyclic free radical intermediate. Ring closure is attributed t o cyclization of a triene t o a cyclohexadiene. If the hydrogen iodide decomposes in situ, further reaction of the resultant iodine ensues.
Introduction In part I of this series,' the reversible, free radical reactions of elemental iodine with Cz-C5 paraffins a t high temperatures were discussed. This paper describes the reactions of iodine with straight and branched chain, higher members of the aliphatic series. Results The type of product obtained from the reaction of a higher alkane with elemental iodine a t ca. 500° is markedly influenced by the mole fraction of reactant iodine. Typical results are illustrated in Fig. 1 for nhexane. At the conditions used, n-hexane in the absence of iodine undergoes negligible reaction. The addition of -1% iodine brings about an iodine-sensitized decomposition similar to that observed with many organic characterized by carbon-carbon scission reactions yielding methane and C2-C6 hydrocarbons, mainly olefins, and interpretable by a RiceHerzfeld4 type of mechanism. With further additions of iodine, degradation reactions become less important and c g compounds account for more than 80% of the products a t the highest 12,/C6H14 value shown. Benzene becomes the major product, accounting for i5yoof the reacted hexane and corresponding to a 96% yield on iodine from reaction 1 C&,r
+ 412
--f
C&
+ 8HI
(1)
Coincident with these increases is a rise in the saturate: olefin ratio of the C2-C6 products to a value of -6. The products from the reaction of several Cs-Clo aliphatic hydrocarbons with one or more moles of iodine are given in Table I. Except in the case of 2,2,5-trimethylhexane, the products formed in highest yield are aromatic hydrocarbons containing the same number of carbon atoms as the reactant. With n-hydrocarbons containing seven or more carbon atoms, however, aromatic products of lower carbon number are also observed, their relative importance increasing with increasing chain length of the original compound. Significantly, each lower (C,2-z)aromatic is accompanied by a nearly equivalent yield of a C , paraffin. These C, paraffins constitute the other major group of de(1) J . H.R a i e y . R D hlullineaux, and C 86, 3171 (106:O ( 2 ) S Bairstow and C N Hinshelwood, J
a;
Bittner, J A m Chem. Soc.,
Chem. Soc.. 11.55 (1933). (3) (a! See H J Schumacher, "Chemische Gasreaktionen," Theodor Steinkapff. I.eipzig, 1938, pp 388-395, for a review, (b) G . K. Rollefson and R . F. Faull, J . A m Chem. Sor , 69, 625 (1937). (4) F. 0 Rice and K F. Herzfeld, ihid., 66, 284 (1934).
composition products. These data indicate a specific cleavage path rather than random or successive decomposition reactions. A single, specific cleavage process is also inferred from the structures of the aromatic products from n-decane. The infrared spectra showed that the Clo fraction contained both monosubstituted and o-disubstituted benzenes, whereas the Cg and C9 aromatics consisted of only monosubstituted compounds. A very small yield of naphthalene, but no methylindanes nor methylindenes, was found. The results for 2,s-dimethylhexane are in distinct contrast. Although the yield of total aromatics is comparable to that from n-octane and 1-octene, benzene and toluene account for less than 2% (instead of 4050'%) of the aromatic material. Other pertinent observations are: (a) the high ( 2_950jo) o-content of the xylenes from the straightchain compounds and the high p-content of the xylenes from 2,5-dimethylhexane, (b) the formation of n-octane from 1-octene, and (c) the greater reactivity of an olefin compared to the corresponding paraffin. These observations demonstrate the absence of methyl migration and confirm other conclusions drawn in the preceding paper. Cyclohexane is converted almost exclusively to benzene. With a 30-fold excess of cyclohexane a t 350°, however, a product containing a cyclohexene/benzene ratio of 2 was obtained. Cyclohexadiene or fivemembered ring structures were not detected under any reaction conditions investigated. The small amount of H2 formed in these reactions in Vycor tubes is consistent with the well known rate of the homogeneous, vapor-phase decomposition of HI.5.6 Since this decomposition is subject to surface catalysis, especially by noble metals,',* the reaction tube in some experiments was partially filled with platinum-containing solids. As illustrated in Table 11, both H2 formation and the combined yield of toluene and benzene increased markedly, the latter exceeding that corresponding to reactions 2 and 3
+
+
C7H16 412 +C7Hs 8HI 3Iz --+ CsHs CHa 6HI
-____ C7Hie
+
+
+
(2) (3)
( 5 ) J . H . Sullivan, J . Chem. P h y s . , SO, 1292 (1959), and references cited therein. (6) L. S Kassel, "Kinetics of Homogeneous Gas Reactions," Reinhold Publishing Corp., New York, N. Y , 1932, p p . 148-156 (7) C. S . Hinshelwood a n d C R Prichard. J . C h e m . Soc., 137, 1552 (1925). (8) C. N. Hinshelwood and R . E . B u r k , ibid., 137, 2896 (1925)
