Olefin Oxidation-Palladium Charcoal Catalysis
Salt-Active
Kaoru Fujimoto, Yasuhiko Negami, Tadashi Takahashi, a n d Taiseki K u n u g i ' Department of Synthetic Chemistry,Faculty of Engineering, Cna'versifyof Tokyo, Hongo, Bunkyo-ku, Tokyo, Japan
A new catalyst system for the Wacker-type reaction was discovered, which is composed of palladium salts and active (charcool without cupric or ferric salts. Olefins were oxidized in the presence of water to yield carbonyl compounds with high selectivity in the temperature range of 60-1 30°C. Reaction rate increased markedly in the vapor phase with an increase in the partial pressure of steam but decreased with a rise in temperature!. The best carrier for palladium salts was active charcoal made from coal, followed by wood and coconut shell sources. Catalyst activity increased remarkably with an increase in the specific surface area of active charcoal. Among palladium salts, palladium chloride gave the best result. Activities of chlorides RhC13 RuCh IrCh of platinum group metals on active charcoal were in the following order: PdCI? > PtCI,. The selectivity of acetaldehyde was usually about 99% when ethylene was oxidized over the palladium chloride-active charcoal catalyst. Palladium reduced through oxidation of olefin was reoxidized with oxygeri b y the catalytic action of active charcoal.
>
O x i d a t i o n of olefins b v metal ions is a novel method for synthesis of carbonyl compounds. Especially palladium ion oxidation is generally represented by the following equation:
'
CzC
R2
+ Hk' + PdX2
"
-
>
>
with olefins and water were reoxidized by oxygen with the catalytic action of active charcoal; subsequently, a new catalyst system was constituted. Several catalytic features of active charcoal as a reoxidation catalyst aiid parameters which will affect this reaction system were studied in a vaporphase conveiitional flow-type reactor and with the cyclic oxidation-reduction method. This type of catalyst is also effective for the catalytic synthesis of vinylacetate from ethylene, acetic acid, and oxygen (Fujimoto and Kunugi, 1969). Experimental
R2/
Y'
wherein H-Y is a nucleophilic reagent such as water, carboxylic acids, alcohols, amines, nitriles, olefins, and aromatic hydrocarbons. These reactions, however, are st,oichiometric, and generally the yield of products is less than the amount of palladium reduced through the above reaction. Smidt et a1 (1959) established a iieiv catalyst system for the oxidation of olefins into carbonyl compounds by use of a PdC12-CuC1? redox ,system. The reoxidation reagents or cocatalysts of palladium salts, ot'lier than cupric chloride, are ferric chloride (Smidt et al., 1969), benzoquinone derivatives (Moiseey et al., 196013), met,al nitrates (Tamura arid Yasui, 1969), and hydrogen peroxide (Moiseev et al., 1960a). O n the other hand, t,here are several examples whereby metallic ions are oxidized by oxygen with the catalytic action of act,ive charcoal to 'nigher oxidation states. These are F e z + (Posner, 1953), C o Z T (Larseii and JTTaltoii, 1940)) and Sn2+ (Bjerruni and lIcRej-nold, 1946). However, these react'ions have been conducted in isolated systems, not in the presence of reducing reagents such as olefins and alcohols. Palladium salt,s adsorbed on active charcoal were excellent catalysts for the reaction of olefins by oxygen in the presence of steam (Fujimoto et al., 1970). 111this paper we assumed that the palladium salts which were reduced through the reaction To whom correqpondence should be addressed.
Apparatus and Procedure. The apparat,us was a coiiveiitional tubular f l o ~ react,or equipped with devices for measuring and controlling the temperature or flow rates. An experimental apparatus with a gas circulating system was also used to measure the adsorption rate of oxygen. The reactor was a glass tube (500 mm long, 16 mm i d . ) with a thermocouple sheath along the central axis. It was heated by mi electric furnace or oil bath. The catalyst weight charged as the fixed bed was 1-20 grams. Above and below the catalyst zone, glass spheres of 1-2 mm diameter were packed. Water was fed with a microfeeder and was vaporized a t the upper part of the reactor. The mixture of gaseous reactants was then passed through the catalyst bed. The liquid products were collected in traps cooled with water. Kheii the apparatus equipped with the circulation system was used, the effluent gas mixture was circulated a t the flow rate of about 210 ml/min. The decrease in pressure of the system was balanced by the addition of the mixture gas of ethylene and nitrogen. Figure 1 shows the flow sheet of the apparatus. Analytical. Organic products were analyzed by a gas chromatograph attached with a flame-ionization detector by the use of a 3-m column packed with 20% diiionyl phthalate on celite. A two-stage gas chromatograph with columns packed with Xolecular Sieve 5-1 aiid silica gel was used to en, nitrogen, carbon monoxide, carbon dioxide, and ethylene. -\cetaldehyde was analyzed by volumetry by Ind. Eng. Chem. Prod. Res. Develop., Vol. 1 1 , No. 3, 1972
303
f
Figure 1 . Flow sheet of apparatus: (a) water feeder, (b) temperature controller, (c) reactor, (d) water jacket, (e) liquid holder, (f) bypass, (9) drying device for gaseous reactants, (h) soap film meter, (i) recirculating pump, (i) acetaldehyde absorber, (k) gas sampling device, ( 1 ) orifice flow meter, (m) gas holder
3
20-
0
-a
t/-*/*-
0
c
e 5*-*-a4
I-*
Active Chorcoai
~
L
0
.I
Ot v
0
I
2
4
Reaction
'I
Time
46
48
50
(hrl
Figure 2. Effect of carriers on ethylene oxidation over Temperature 110°C; molar PdClz-carrier (Pd 1 wt ratio, C2H4:Oz : HzO = 4 : 1 : 6.5; 0, PdCIP-active charcoal; 0, PdClz-silica gel; 0 , PdCI2-yalumina
70).
use of hydroxylamine hydrochloride after adsorption with water . Materials. Ethylene and propylene (99.9yo purity) were fed to the reactor after passing through calcium chloride and active charcoal. Commercially available oxygen and nitrogen mere also treated by calcium chloride and active charcoal before feeding. The ion-eschanged water was used as one of the raw materials. Active charcoals were commercially available, made of wood, coal and coconut shell. Catalysts. PdCl?-Active Charcoal Catalgst. Palladium dichloride (0.84 gram) was dissolved in 300 ml of I N hydrochloric acid (Solution .\). Active charcoal, which was boiled with dilute nitric acid and then washed by pure mater, was mixed with 200 nil of 1.1- hydrochloric acid (Mixture A) and boiled for 10 minutes. -\fter Mixture -1was cooled to room temperature, Solution A was added n i t h stirring (Mixture B). Xixture B was left to stand for 24-48 hr a t room temperature. Palladium chloride in the solution was adsorbed completely on the active charcoal. The active charcoal supporting palladium chloride nab washed with 200 ml of pure water and then dried a t 150°C for 6 hr in vacuo. 304
Ind. Eng. Chem. Prod. Res. Develop., Vol. 1 1 , No. 3, 1972
The palladium chloride-active charcoal catalyst prepared contains little free hydrochloric acid, and the atomic ratio of chlorine to palladium is about 2.0 to 2.2. Pd (OSc)~-dctive Charcoal Catalyst. Palladium diacetate (1.01 gram) was dissolved in 300 ml of pure acetic acid (Solution B). Active charcoal was boiled in 100 ml of acetic acid for a short time and cooled to room temperature (Mixture C). Solution B was poured into Mixture C and allowed to stand for 48 hr. In some cases the misture was heated a t 100-110°C for several hours to complete the adsorption. Thus, palladium diacetate was completely adsorbed on the active charcoal. The mother liquid was then filtered off, and the resulting catalyst was dried a t 130°C for 5 hr in vacuo. Other palladium salts were adsorbed on active charcoal in the corresponding acid solution to the anion of the salt and were treated in a similar way. Results and Discussion
The main product was acetaldehyde. Small amounts of butenes, carbon dioxide, and methyl ethyl ketone were formed as by-products. Effect of Carriers. Oxidation of ethylene was performed over three types of catalysts, namely, palladium chloride-active charcoal, palladium chloride-silica gel, and palladium chloride-y-alumina. The results are shown in Figure 2. Palladium chloride is catalytically active only when supported on active charcoal. The activity and selectivity of the catalyst were kept substantially constant over 50 hr. The selectivity of acetaldehyde n as in the range of 92-95%. The molal amount of acetaldehyde formed was more than 2000 times that of palladium chloride, and the production rate of acetaldehyde was more than 40 mol/mol of palladium chloride per hour. This rate can be improved up to about 200, even under normal pressure by selecting the proper conditions of catalyst preparation and oxidation reaction, e.g., adsorption of PdC12 from dilute solution and reaction a t lower temperature. Other carriers have little catalytic action, but a somewhat larger molal amount of acetaldehyde per mole of palladium chloride was formed over the palladium chloride-silica gel catalyst. Finely dispersed metallic palladium dissolves slowly into aqueous acidic solutions in the presence of oxygen as the palladium salts. Metallic palladium on the silica gel seems to be an extremely fine particulate. Possibly, the olefin oxidation (Equation 1) and reosidation reactions (Equation 2) occur simultaneously on active charcoal in this catalyst system. CZH4
+ H20 + PdC12 Pd"
+
+ 2HC1 +
+ Pd" + HC1 PdClz + H20
CHsCHO
' 1 ' 2 0 2 -+
(I) (2)
The amount of hydrogen chloride which came out of the catalyst during reaction was about 0.1 mol 70 of supported palladium chloride per 1 hr. Palladium Salts and Platinum-Group M e t a l Salts.
Figure 3 shows the results of osidation of ethylene over palladium chloride and palladium sulfate supported on active charcoal. Palladium sulfate on active charcoal gave 350 moles of acetaldehyde per mole of palladium sulfate over 12 hr and proved to be a fairly excellent catalyst. Other palladium salts also have catalytic activity M hen supported on active charcoal. Table I shows the results on several palladium salt catalysts. Palladium chloride is most active, but its optimum temperature is under 100°C. The activity of palladium chloride catalyst decreases markedly with a rise in the temperature. We have understood this phenomenon as follows: the reaction (Equation 1) proceeds
Table I. Oxidation of Ethylene over Palladium Salt-Active Charcoal Catalysts P d content of catalyst, 0.5 m t % Molar ratio, CzH4:02:Hz0 = 3.75.1.0.6.0 Contact time, 7.4 g-cat hr/mol Products, 1 0-3rnol/g-cat hr 95OC llO°C 125'C
Catalyst
PdClz-active charcoal -0
4
i?
iReactlon
6 Time
8
10
/2
(hr)
Figure 3. Oxidation of ethylene over palladium saltactive charcoal (Pd 1 wt %). Temperature, 110°C; molar ratio, CsH4: 0 2 : H?O == 4 : 1 : 7
CH3CH0 COZ CiHs PdS04-active CHEHO charcoal COZ CeHs Pd(OAc)2-active CH8CHO charcoal COZ CiHs Pd(S08)Z-active CH3CHO charcoal C4Hs
coz
8 48 0 042 0 012 1 61 0 015 Trace 0 486 0 029 Tiace 0 524 0 013 Trace
6 10 0 036 0 025 4 28 0 056 0 005 1 046 0 033 0 003 1 478 0 032 Trace
1 0 0 4 0 0 0
41 016 041 44 046 024 492 0 026 0 005 0 750 0 025 0 008
140'C
0 0 0 3 0 0 0 0 0 0 0 0
68 025 036 30 052 038 065 025 012 536 028 012
Table II. Catalyst Activity Vs. Various Treatments Catalyst, PdClz-active charcoal, P d 1 wt Molar ratio, C2H4:02:Hz0 = 4: 1: 10 Contact time, 22 g-cat/mol
Treatment
0
3
6
0
re
Reaction Ttme (hr)
Figure 4. Reaction of ethylene over PdCI2-active charcoal (Pd 0.5 wt %). Temperature, 110°C; molar ratio, CzH4: O?(N,): HZ0 = 3 : 1 : 8
according to the mechanisms similar to t,hose proposed by Henry (1964), Jira et al. (1966), or Dozono and Shiba (1963) in this gas-solid reaction. Accordingly, t'wo or three water molecules participate in the formation of one molecule of acetaldehyde. In this reaction system, water molecules should be adsorbed on the #catalyst before reaction, while bhe adsorbed water decreases wit'h a rise in tlie temperature; t'herefore, the apparent reiiction rate decreases with a rise in the temperature. Other platinum-group metal chlorides on active charcoal could also catalyze t h e Wacker-type reaction, but their activities were low, aiid the activit'ies decreased rapidly. The order of act'ivity v a s PdClz > RhC18 > RuC13 > IrC13 > PtC14 N 0. Behavior of Palladium Compounds. In Table 11, we tried the oxidation of ethylene over the preliminary reduced catalyst. Runs (a-e) mid (d-h) were conducted over the same cat,alysts, respectively. The original catalyst gave 21.6% coiiversioii of ethylene, and the addition of hydrogen chloride to the catalyst resulted in the inhibition of the reaction as shown in (b) and (c). In tlie esperimeiits of runs (d-h) , several pretreatments were made successively with the same catalyst. The catalyst act,ivity decreased markedly as the reduction advanced, but the activity of this deactivated catalyst is easily restored by adding hydrogen chloride in the feed gas, followed by removal of escess hydrogen chloride with nitrogen stream. From t'hese results, the active species of the catalyst is palladium(11) ion, aiid the catalytic action is based on a'kind of redos system composed of palladium, acid, and active charcoal.
a) Original catalyst b) HC1 addition: HCl, 0.42 mmol/g-cat c) HC1 addition: HC1, 0.52 mmoljg-cat d) Hz reduction: 180"C, 4 hr e) Hz reduction: 28OoC, 4 hr f ) HC1 addition: PECI, 0.006 atm g) S, purge: 130"C, 3 hr h) Kzpurge: 13OoC, 3 hr
Temp, OC
Conversion of &Ha,
Selectivity of CHsCHO,
%
%
120
21.6
90
110
6.5
85
110 121 118
1.2 13.6 0.8
90 95 68
115 120 110
0.1 14.0 19.1
85 75 95
In Wacker-type reactions the olefin osidation and reosidatioii proceed independently (Smidt et al., 1959). The same phenomenon was also observed with our catalyst system. Several esperiments with and without oxygen were tested as shown in Figures 4-8. In Figure 4, for example, SSin (I) and (111) columns means that the feed gas did not contain osygen and was composed of ethylene, nitrogen, and water. X o r e than 4 moles of acetaldehyde were produced per 1 mole of palladium chloride even when no oxygen was fed; this seems to be caused by the reoxidation of reduced palladium with oxygen adsorbed on active charcoal. The amount of hydrogen chloride which came from the catalyst bed in these experiments is listed in Table 111. A considerable amount of hydrogen chloride came out when the reaction was carried out in the absence of oxygen. When oxygen was added in place of nitrogen, the catalytic activity increased rapidly, but the yield of acetaldehyde in steady state was 60-7070 a t most', as compared with the last oxidation experiment. The palladium chloride-silica gel catalyst showed similar behavior as seen in Figure 5 , but, its restored activity is low. The palladium chloride-cupric chloride-silica gel catalyst (Figure 6) exhibited high activity similar to that of the palladium chloride-act'ive charcoal catalyst, but its activity Ind. Eng. Chern. Prod. Res. Develop., Vol. 1 1 , No. 3, 1 9 7 2
305
Table 111. Effluence of Chloride Ion During Various Reaction Conditions Effluence of chloride ion, mmol Experiment
Figure Figure Figure Figure Figure
4 6
7 8 8
Catolyst
Temp, OC
PdCL-active charcoal PdClp-CuClZ-Si02 NaPdCld-active charcoal PdCla-active charcoal PdCla-active charcoal
110 120 120 85 125
(11)
1111)
(IV)
0.34
0.10
0.03
0.20 0.005 0.50
0.39 0.012 0.31
0.07 0.022 0.30 0.005 0.30
(1)
0 0
c
I
I
E c
Y
U c 0
P
F 6 Reaotlon Time
9
IO
Reaction of ethylene over PdClz-silica gel (Pd 115°C; contact time, 8.5 g-cat. hr/mol; molar ratio, CaH4: 0 2 ( N a ): H a O = 3 : 1 : 8
"0
3 Reaction
6 Time
It
I
3
6
9
Reaotlon Tlme
%). Temperature,
9
0,0008
0 20
(hrj
Figure 5.
0.5 wt
:L
(V)
I2
(hr)
Figure 7. Reaction of ethylene over NazPdC14-active charcoal (Pd 0.5 wt %). Temperature, 120°C; contact time, 8.5 g-cat. hr/mol; molar ratio, C?Hd: Oa(N2): HzO = 3 : 1 : 8; a ) HCI 0 2 treatment, 1 20°C, 1 hr
+
12 Reaction Tlme
(hr)
(hr)
Figure 6. Reaction of ethylene over PdCl2-CuCIz-silica gel (Pd 0.5 wt %, Cu/Pd molar ratio 5/1). Temperature, 1 20°C, contact time, 8.5 g-cat.hr/mol; molar ratio, C2H4: Os(Na): HzO = 3 : 1 : 8
Figure 8. Reaction of ethylene over PdClz-active charcoal (Pd 0.5 wt yo);contact time, 8.5 g-cat.hr/mol; 0 : temperature, 85°C; molar ralio, C?HI : 02(Ns) : H 2 0: N2 = 3 : 1 : 4 : 4; 0:temperature, 125°C; molar ratio, C2H4: Oz(Nz) : H20 = 3 : 1 : 8; a) HCI 0 2 treatment, 125"C, 1 hr
decreased more rapidly than that of the active charcoal catalyst. I n the experiment shown in Figure 7 , Na2PdClr was used as the source of palladium chloride to distinguish the total amount of chloride ion in the catalyst. I n this case NasPdCl4 was dissolved in pure r a t e r and was adsorbed on active charcoal. After the adsorption of palladium chloride was completed, the amount of chloride ion (in the form of NaC1) in aqueous solution was determined. The amount of chloride ion fed as SaZPdC14minus the amount of chloride ion present in aqueous solution, which was determined above, equalled the amount of chloride ion on active charcoal. These results show that the reduced palladium is fairly stable and is reoxidized easily with oxygen when present on active charcoal as well as on silica gel rTith cupric chloride. But the reduced palladium on silica gel without cupric chloride rapidly loses its anion as hydrogen chloride (the data are not
shown), aggregates to palladium particles (which has been observed lvith the electron micrograph), and consequently loses its catalytic activity irreversibly. Under a reduced state, the free chloride ion seems to be extremely unstable, perhaps because of the absence of Coulombic attraction, and consequently, it migrates from the active sites on the catalyst surface and is effluxed from the catalyst bed. Palladium losing anion can no longer be reoxidized because of the high barrier of oxidation potential, but e\ en after three hours of reduction with ethylene aiid steam, the catalyst still has an activity of 60-65% as compared with that of the original catalyst. The deactivated catalyst by cyclic oxidation and reduction is easily restored by treatment with oxygen aiid hydrogen chloride (Figure 8). I t seems reasonable to consider that active charcoal holds hydrogen chloride near the palladium atom and prevents the aggregation of the reduced palladium compound.
306
Ind. Eng. Chem. Prod. Res. Develop., Vol. 1 1 , No. 3, 1972
+
~~
~~~
~
Table IV. Effect of Active Charcoals on Activity PdCL-active charcoal, metallic P d 0.5 wt % Temp, 10l°C Molar ratio, C2H4:02:H20= 4:1:6 Contact time, 10.2 g-cat.hr/mol Raw material Run
"
50 0
0
Specific
1500
1000
Surface Area
110 111 092 091 097 095 100 090 094 108
[m'/g)
098 Figure 9. Effect of amctive charcoal on oxidation of ethylene with PdClz-active charcoal (Pd 0.5 wt yo).Temperature, 100°C; contact time, 7.5 g-cat. hr/mol; molar ratio C?H, : O Z : H 2 0 = 4 : 1 15.5 ,.-
.
0
O
L
10
'
0.10
0.05
0.15
PO, (atm)
Figure 10. Effect of oxygen pressure on reaction rate. Catalyst, PdCI*-active charcoal (Pd 0.5 w t %); 0 : tempera, atm; PCIHa, 0.23 atm; 0: temperature, 105°C; P H ? ~ 0.57 ture, 105°C; P H ~ 0.57 ~ , atm; PCIHa, 0.10 atrn; a: ternperature, 100°C; P H ? ~0.'; , atm; PC?H~,0.3 atm
The experimeiits shown in Figure 6 indicate that the palladium on silica gel is slowly oxidized by oxygen in the presence of hydrogel1 chloride, and the Wacker-type reaction is greatly accelerated by cupric chloride. Effect of Hydrogen Chloride. Escess hydrogen chloride resulted in a marked inhibition on the homogeneous Kackertype reaction (Smidt et al., 1959). In this reaction, a small amount of free hydrogen chloride on active charcoal also inhibited the reaction markedly, as shown in Table 11. Howerer, t,he catalyst poisoned b y hydrogen chloride is easily reactivated by purging it with iiit'rogen stream a t about 130°C; therefore, it' seems that the affinity of hydrogen chloride to palladium chloride on active charcoal is not so strong and is easily remoyed by nitrogen purge. The mechanism of inhibition owing to hydrogen chloride is attributed to the coordination of chloride ion to palladium chloride and, consequently, t o the disturbance of the coordination of olefins and water as shown previously (Henry, 1964; Jira et ai., 1966).
109 099 093
of active charcoal
Coal Coal Coal Coal Coal Coal Wood Wood Kood Coconut shell Coconut shell Coconut shell Coconut shell Coconut shell
Speciflc surface area, m2/g
Rate o f formation, 1 OW3 mol/g-cat. hr C,Hs
coz
277 318 790 925 1070 1220 1240 1280 1520 785
1.15 1.16 3.87 2.92 7.18 11.63 8.79 8.42 6.25 2.32
Trace Trace 0.004 0,025 0.008 0.019 0,015 0,012 0,002 0.453
Trace Trace 0.020 0.055 0 077 0,055 0,075 0.056 0.056 0.045
952
6.15
0.020
0.014
1199
4.28
0.720
0.050
1220
3.67
0.020
0,036
1310
6.50
0.039
0.012
CHICHO
On the other hand, hydrogen chloride is the necessary reagent of reoxidation. Accordingly, the catalyst deactivated through losing chloride ion as hydrogen chloride is restored under the reaction condition by adding free hydrogen chloride to the feed materials and removing the excess afterward [Figure 8, a) and Table I1 (f, g)]. Effect of Active Charcoal. Various active charcoals which adsorbed the same amount of palladium chloride \\ere tested for the oxidation of ethylene, as shown in Table IT. Among them, active charcoal made of coal had the best activity. Though the active charcoal made of coal contained more inorganic impurities such as S O z , Fe203, - 2 1 2 0 3 , and CaO than those made of wood or coconut shells, these inorganic substances seemed to have no special influence upon the synthetic reaction. On the other hand, the activity increased with the specific surface area as shown in Figure 9. We assumed that the palladium chloride was well dispersed on active charcoal made of coal or on that having a larger surface area, but the details m-ere not clear. Effect of Oxygen Pressure on Reaction Rate. T h e effect of oxygen pressure on the reaction rate was studied by a floa-type recirculation system. I n this experiment the rate of gas absorption corresponded to the rate of reaction. The partial pressures of ethylene and steam were kept constant without regard to oxygen pressure. Results are shown in Figure 10. T h e reaction in t h e recirculation system x a s in a steady state experimentally. Oxygeii pressure affected the reaction rate only n h e n it was low. From Figure 10, it seems reasonable that Reaction 2 proceeded more rapidly then Reaction 1. Therefore, almost all of the palladium on active charcoal was 111 the oxidized state under t h e usual reaction conditions, and a considerable amount of reduced palladium was present only when the oxygen pressure was low. This observation and consideration caii be analyzed qualitatively n ith the theory of dynamic equilibrium between Reaction 1 and 2, which is scheduled to be reported in the near future. Ind. Eng. Chem. Prod. Res. Develop., Vol. 11, No. 3, 1972
307
As a generalization, the catalytic action of active charcoal is concluded to be based O n the holding of hydrogen chloride formed through Reaction 1 near the palladium, the stabilizati011 of t,he reduced palladium compound to prevent aggregation, and the rapid reoxidation of the reduced palladium compound by the activation of oxygen. Literature Cited
Bjerrum, J., MeReynold, J. P., “Inorganic Synthesis,” Vol 2, p 216, 3IcGraw-Hill, Xew York, KY, 1946. Dozono, T., Shiba, T., Bull. Jap. Petrol. Inst., 5 , 8 (1963). Fujimoto, K., Kunugi, T., Kogyo Kagaku Zasshi, 72, 1760 (1969).
Olefin Oxidation-Mercuric Charcoal Catalysis
Fujimoto, K., Negami, Y., Kunugi, T., ibid., 73, 1822 (1970). J . Amer. Chem. Sac., 86, 3246 (1964). Henry, P. M., ~ iR., ~sedlrneier, ~ , J , , srnidt, J., Chem,,693, 99 (1966). Larsen, E. C., Walton, J. H., J . Phys. Chem., 44,70 (1940). lloiseev, I. I., Vargaftik, 11. S., Syrkin, Ya. K., Dokl. Akad. A-auk, USSR,130, 820 (1960a). 1\loiseev, 1, I,, vargaftik, 11,x., syrkin, ya,K,, ibid,,133, 377 (1960b). Posner, A. AI.,Trans. Faraday Soc., 49, 389 (1953). Smidt, J., Hafner, W., Jira, R., Sedlmeier, J., Sieber, R., Ruttinger, R., Kojer, H., Angew. Chem., 71, 176 (1959). Yasui, A , , Kogyo Kagaku Zasshi, 72, 528 (1969). Tamura, M,, RECEIVED for review November 15, 1971 ACCEPTEDSlay 22, 1972
Salt-Active
Hiromichi Arai, Katsuya Uehara, Shin-ichi Kinoshita, and Taiseki Kunugil Department of Synthetic Chemistry, Faculty of Engineering, Cniuersity of Tokyo, Hongo, Bunkyo-ku, Tokyo, J a p a n
Catalytic reaction of isobutene with mercuric chloride supported on active charcoal in the vapor phase yielded metacrolein as a major product. The kinetics of the oxidation were determined in a fixed-bed reactor. The rate of metacrolein formation was R = k [ 0 2 ][iso-CaHs], where the value of k was 0.1 48 mol/hr,atm2, g-cat (at 140°C, Hg: 5 wt %). Reaction rates of other olefins were measured. Propylene, 1 -butene, and 2butene underwent an analogous reaction more slowly.
O r g a n i c compounds have been subjected to a number of reactions with mercuric salt’s. The addition, substitution, and oxidation reactions in the liquid phase have been reported (Kitching, 1968). The addition of oxy salts of mercury to olefins to give p-oxy organomereuric compounds is wellknown (Hoffmann and Sand, 1900). The so-called hydroxymercuration reaction is expressed as Hgz+
+ CH2=CHR + H20
+
[HgCHzCHROH]+
+ H+
where R is H, alkyl, or allyl. Stoichiometric, stereochemical, and kinetic studies in the aqueous or organic solvents for the formation of the u-mercury complex have been made extensively (Kitching, 1968; Chatt, 1951; Halpern and Tinker, 1967). Further reaction of the u-complex x i t h mercuric ion gave oxidized olefinic compounds as major products. For example, propylene gave acrolein (Fielding and Roberts, 1966). CH3CH(OH)CH*Hg+
+ 3Hgz+
+
CH,=CHCHO
+ 3H+
Although the stoichiometric reactions of mercuric ions with olefins in the liquid phase are known, no study has been reported on the catalytic reoxidation process b y use of mercurous ion or mercury. In this paper the catalytic oxidation of olefins with mercuric salts supported 011 active charcoal catalysts in the vapor phase, by use of a fixed-bed reactor, is reported. To whom correspondence should be addressed.
308
Ind. Eng. Chem. Prod. Res. Develop., Vol. 1 1 , No. 3, 1972
Ethylene is catalytically oxidized by molecular oxygen and water to give acetaldehyde over palladium salt-active charcoal catalysts without the use of redox substances such as cupric chloride (Fujimoto et al., 1972). Mercuric salts supported on active charcoal also had the catalytic activity to oxidize propylene or isobutene to acrolein or metacrolein, respectively, a t the reaction temperature of 140-200°C. Bismuth molybdate is an excellent catalyst for the oxidation of propylene and isobutene a t higher temperatures, Le., 460°C (Adams, 1965). Experimental
Materials. Ethylene, propylene, 1-butene, cis-2-butene, trans-2-butene, and isobutene were 99.9% pure. Preparation of Catalysts. Active charcoal (made from coconut; surface area, 952 m2/g, 10-20 mesh) was boiled with 10% aqueous nitric acid for 3 hr and washed with boiling water repeatedly until the final p H of rinsed water became 4. Mercuric chloride-active charcoal catalyst was prepared by dipping active charcoal in the prescribed solution of mercuric chloride in 0.LV aqueous hydrogen chloride. After adsorption the catalyst was water-washed to a p H of 4 and dried over calcium chloride. The mercuric chloride supported on active charcoal is usually 5 wt yo as metallic mercury. Prior to use, the catalyst was evacuated a t 150°C for 4 hr. Other mercuric salt-active charcoal catalysts were prepared in the same way with the acid solution corresponding to the anion of each salt. Apparatus. T h e reactor IS made from a glass tube 20 m m in diameter with a concentric thermowell. F i r e grams of
