Ind. Eng. Chem. Res. 1996,34, 3392-3398
3392
Kinetics of the Liquid Phase Hydrogenation of Furan Amines David R Holm and Charles G. Hill,Jr.* Department of Chemical Engineering, University of Wisconsin, 1415 Johnson Dr., Madison, Wisconsin 53706
Anthony H.Comer USDA-Forest Service, Forest Products Laboratory, One Gifford Pinchot Dr., Madison, Wisconsin 53705
The liquid phase hydrogenation reactions of both furfurylamine and 5,5’-ethylidenedifurfurylamine to the corresponding tetrahydrofuran compounds have been investigated over a rhodium on alumina catalyst suspended in a methanol solution. The reactions were studied at temperatures from 42.5 to 80.1 “Cand hydrogen pressures from 231 to 1380 kPa (33.5 to 200 psi). The effects of the amount of catalyst and the concentrations of both reactants and products on the rates of these reactions were also investigated. Analysis of the data from these investigations using the method of initial rates indicates that the rates of hydrogenation of both furfurylamine and 5,5’-ethylidenedifrylamine are consistent with Hougen-Watson-Langmuir-Hinshelwood reaction mechanisms. The overall activation energies for the hydrogenation of furfurylamine and 5,5’-ethylidenedifurfurylamine are 49.4 f 2.5 and 63 f 4 kJ/mol, respectively.
Introduction Difurfuryl diamines and d i e l diisocyanates are compounds that have potential applications in a number of polymer resin systems (Cawse et al., 1984;Holfinger, 1992). For example, difurfuryl diamines have been shown t o be good curing agents for epoxies (He et al., 1992), and difurfuryl diisocyanates exhibit excellent properties for use as exterior grade wood adhesives (Holfinger, 1992;Holfinger et al., 1992). The composite wood products industry is particularly interested in developing economic methods of synthesizing these compounds because they are derived from renewable resources (e.g., corn cobs, oat hulls, and wood), rather than petroleum. The hydrogenation of 5,5’-ethylidenedifurfurylamine (EDFA) to 5,5’-ethylideneditetrahydrofurfurylamine (EDTHFA) (see Figure 1) is of interest because it is believed that the tetrahydrofuran compound wiU broaden the scope of potential uses of difurfuryl diamines and diisocyanates, thus leading t o a concomitant increase in demand. For example, the color of pure EDFA is light yellow which becomes darker over time during storage, whereas EDTHFA is colorless and remains colorless during storage. Thus, the use of EDTHFA in a product would be acceptable if color stability is of primary importance; by contrast, the use of EDFA would be unacceptable in this situation. The hydrogenation of furfurylamine (FA) is used as a model reaction for subsequent studies of the hydrogenation of EDFA. FA is an ideal model compound for EDFA because FA both contains structural elements present in EDFA (see Figure 1) and is readily available commercially.
Experimental Materials Catalyst. The catalyst used in the hydrogenation experiments is a 3.65% (w/w) rhodium on alumina (Rh/ A l 2 0 3 ) powder that was obtained from Aldrich Chemical Co. (Milwaukee, WI).This catalyst was pulverized and passed through a 350 mesh (44pm) sieve prior to use. A particle size analysis of the catalyst powder by Coulter
* To whom correspondence should be addressed.
H z N C H a C kHH -3 o C H a N H a
EDTHFA m C H z N H 2
o C H z N H a
FA
THFA
Figure 1. Structures of 5,5’-ethylidenedifurfurylamine(EDFA), 5-[l-[2-(aminomethyl)tetrahydrofuran-5-yllethyllfurfurylamine (1/ 2-EDFA), 5,5’-ethylideneditetrahydrofurfurylamine(EDTHFA), hrfurylamine (FA), and tetrahydrofurfurylamine (THFA).
Scientific Instruments (Hialeah, FL) indicated that the volume average diameter of the catalyst powder was 9.2 pm. Galbraith Laboratory (Knoxville, !t”)determined the amount of rhodium on the alumina support using plasma emission spectroscopy. Furfurylamine. Furfurylamine (FA) was obtained from QO Chemicals (Memphis, TN) and was purified via vacuum distillation using a water aspirator. The fraction which boiled a t 47-48 “C was collected. FA was stored in a freezer under nitrogen when not in use. Tetrahydrofurfurylamine. Tetrahydrofurfurylamine (THFA)was obtained from QO Chemicals (Memphis, TN)and was purified via vacuum distillation using a water aspirator. The fraction which boiled at 51 “C was collected. THFA was stored in a freezer under nitrogen when not in use. 5,5’-Ethylidenedifwfurylamine. B,B’-Ethylidenedifurfurylamine (EDFA) was obtained from Richman Chemical Co. (Ambler, PA) who performed a custom synthesis using the method of Holfinger (Holfinger, 1992). EDFA was purified via vacuum distillation using a vacuum pump. The fraction which boiled at 126-131
Q888-5885l95I2634-3392$09.0QlQ 0 1995 American Chemical Society
Ind. Eng. Chem. Res., Vol. 34, No. 10, 1995 3393 by hydrogen consumption. The pressure in the sample cylinder is monitored using a test gauge (Ashcroft type 1082).
Experiments Performed
I
I
Furfurylamine. The effects of the amount of catalyst, the concentrations of both FA and THFA, the activity of hydrogen in solution, and the temperature on the rate of hydrogenation of FA have been investigated (see Table 1). In addition, the effects of external mass transfer limitations and intraparticle diffusion limitations on the rate of this reaction were investigated by changing the agitation rate, incorporating baffles into the reactor to enhance turbulent mixing, and changing the gross external dimensions of the catalyst powder. The activity of hydrogen in solution ( U H J was calculated using the relation
Figure 2. Schematic representation of hydrogenation reactor (TC = thermocouple, CW = cooling water, IW = ice water, SC = sample cylinder, PG = pressure gauge).
"C was collected. EDFA was stored in a refrigerator under nitrogen when not in use. 5,B'-Ethylideneditetrahydrofurfurylamine. 5,5'Ethylideneditetrahydrofurfurylamine (EDTHFA) was synthesized in the laboratory using reaction conditions that are presented in this paper. EDTHFA was purified via evaporation and then vacuum distillation using a vacuum pump. The fraction which boiled at 125-129 "C was collected. EDTHFA was stored in a refrigerator under nitrogen when not in use. The identities and purities of FA, THFA, EDFA, and EDTHFA were verified using both gas chromatography and lH-NMR and 13C-NMRspectroscopies in each case. Methanol. Reagent grade methanol was obtained from J. T. Baker Inc. (Phillipsburg, NJ) and was used without further purification. Hydrogen. Hydrogen gas (purity > 99.96%) was obtained from Badger Welding (Madison, WI)and was used without further purification.
Experimental Apparatus A schematic of the hydrogenation reactor is shown in Figure 2. The reactor is a 50 mL stainless steel cylinder (internal depth = 5.715 cm, i.d. = 3.302 cm) with a screw-on reactor head. This reactor was purchased from Parr Instrument Co. (Moline, IL). Mixing of the reactor contents is accomplished using a stirring shaft mounted in the reactor head. A stirring motor, belt, and pulley provide agitation rates of 0-2000 rpm. Four custom made baffles can be incorporated into the reactor to aid in mixing the reactor contents. The reaction temperature is monitored using a type J thermocouple (1/8 in. stainless steel sheath) that is mounted through the reactor head. The reaction temperature is maintained constant by partially submerging the reactor in a constant temperature circulating oil bath (Haake model N3-B). This oil bath is equipped with a proportional temperature controller. The reactor pressure is monitored using a test gauge (Ashcroft type 1082; Dresser Industries, Inc., Stratford, CT) and is controlled by use of a two-stage gas regulator. Samples are withdrawn during the reaction with the aid of a dip tube (1/16 in. tubing, 0.02 in. wall). A stainless steel sample cylinder of known volume is used t o feed hydrogen t o the reactor when monitoring the reaction
where P H ~is the partial pressure of hydrogen and f is the hypothetical liquid fugacity of hydrogen a t standard pressure. The hypothetical liquid fugacity of hydrogen was calculated using the relation In f
= 14.637 844
+ 856.9404/T
(2)
where T i s the absolute temperature in degrees kelvin. Equation 2 is a correlation developed by Radhakrishnan et al. (1983) using the data of Lemcoff (1977) for purposes of calculating the solubility of hydrogen in a mixed solvent. Equation 1assumes that the vapor phase is ideal and that the hypothetical liquid fugacity of hydrogen is independent of pressure for the experimental conditions employed. 6 , S - E t h y l i d e n e ~ l a m i n eThe . effects of the amount of catalyst, the concentrations of both EDFA and EDTHFA, the activity of hydrogen in solution, and the temperature on the rate of hydrogenation of EDFA have been investigated (see Table 2). In addition, the effects of external mass transfer limitations and intraparticle diffusion limitations on the rate of this reaction were investigated by incorporating baffles into the reactor to enhance turbulent mixing and changing the gross external dimensions of the catalyst powder.
Experimental Procedure The catalyst was dried overnight a t approximately 110 "C in an oven and then placed in a desiccator over PzO5. The catalyst was weighed and quantitatively transferred to the reactor. The required amounts of either FA or EDFA and methanol were then added to the reactor. The reactor was then sealed and placed in an ice bath. The reactor was purged with hydrogen when the temperature of the slurry reached approximately 5 "C. Purging of the reactor consisted of slowly pressurizing the reactor with hydrogen to 584 kPa followed by slowly depressurizing the reactor to 170 kPa. This procedure was performed five times. The reactor was then placed in the oil bath at the desired temperature. Upon reaching the desired reaction temperature, the reactor was pressurized with hydrogen t o the desired pressure in the range from 231 to 1380 kPa. A
3394 Ind. Eng. Chem. Res., Vol. 34,No. 10, 1995 Table 1. Summary of Experiments Performed To Investigate the Effects of Several Variables on the Rate of Hydrogenation of FA (a~. = the Activity of Hydrogen in Solution; ro = Initial Rate of Reaction) temperature [FA10 [THFAlo amount of 3.65% ro x lo4 experiment ("C) (kmol/m3) (kmol/m3) aH9 RWAl203 (kg/m3) (kmol/(m3 8)) 0.00 0.0169 3.21 3.30 FA1 70.4 1.33 6.42 11.32 0.00 0.0330 1.33 FA2HP 70.6 0.00 0.0168 6.42 6.31 FA2B 70.6 1.33 0.00 0.0170 6.59 6.13 FA2D 2.67 70.6 0.0084 6.42 3.12 0.00 FA2LP 70.7 1.34 1.26 0.0172 6.27 4.88 FA2TH 1.30 70.5 0.00 12.84 0.0168 12.77 FA3 70.8 1.33 0.0168 6.42 3.53 FA5B 60.0 1.33 0.00 2.72 1.33 1.28 0.0168 6.43 FA5TH 60.0 0.00 0.0168 12.85 7.62 60.0 1.33 FA6 4.27 1.33 0.00 0.0427 6.46 FABHHP 50.0 3.47 0.0327 6.42 0.00 FA8HP 50.0 1.33 2.10 6.42 0.00 1.33 0.0170 FA8 50.0 0.00 4.17 0.0168 12.85 FA9 50.0 1.33 FA10 42.5 1.33 0.00 0.0168 12.85 2.75 FA11 1.33 0.0168 6.42 9.23 0.00 80.1 Table 2. Summary of Experiments Performed To Investigate the Effects of Several Variables on the Rate of Hydrogenation of EDFA ( a ~=,the Activity of Hydrogen in Solution; ro = Initial Rate of Reaction) temperature [EDFAlo [EDTHFAIo amount of 3.65% ro x 104 experiment ("0 (kmol/m3) (kmoWm3) aHz RWAl20 (kg/m3) (kmoW(m38 ) ) E l2HHP 79.1 0.65 0.00 0.0498 3.14 4.08 0.0399 El2HP 0.00 79.1 0.65 3.14 4.05 79.1 0.65 0.00 0.0169 3.14 3.80 E12 79.1 0.66 0.00 0.0098 3.14 2.62 El2LP 0.00 79.1 0.66 0.0169 6.27 6.97 Ell ElHHP 70.3 0.66 0.00 0.0484 3.16 3.20 0.00 ElHP 70.1 0.66 0.0330 3.17 2.82 0.00 El 70.3 0.66 0.0169 3.17 2.52 1.31 0.00 ED 70.6 0.0337 6.23 7.05 0.00 0.65 E2 70.3 0.0168 6.31 3.82 0.00 E2LP 0.65 70.3 0.0087 6.27 3.12 0.59 0.65 E2TH 70.3 0.0338 6.29 5.35 0.00 12.55 0.65 E3 70.5 0.0169 8.15 E4 60.0 0.66 0.00 0.0168 3.17 1.17 60.0 0.65 0.0168 6.31 0.00 E5 2.20 60.0 0.66 0.0168 0.00 E6 12.56 4.55 E9 50.0 0.66 0.00 0.0168 12.56 1.85
stirring speed of 2000 rpm was then used throughout the experiment.
Monitoring the Reactions The progress of each experiment was monitored using gas chromatography (GC) except for experiment FA2D (see Table l),experiments for which the agitation rate was varied, and experiments for which the gross external dimensions of the catalyst were varied. The latter experiments were monitored by observing the decrease in hydrogen pressure in the sample cylinder of the experimental apparatus. It should be noted that the results obtained using the hydrogen consumption measurements and the results obtained using GC analyses were in excellent agreement up to approximately 60% conversion. The protocol for the GC analyses is indicated below. Analysis Using Gas Chromatography. k General Procedure. The reactions were monitored using gas chromatography (GC) by taking samples during the reaction, working up the samples, and then injecting them into a gas chromatograph for analysis. An internal standardization technique was used for quantitation. The internal standards used for FA and EDFA were n-decane and n-hexadecane, respectively. A Hewlet-Packard Model 5890A gas chromatograph
equipped with a flame ionization detector was used in analyses of the samples obtained from the hydrogenation experiments. Sampling of the reaction mixture consisted of purging the dip tube line with approximately 0.5 mL of reaction slurry and then withdrawing approximately an additional 0.5 mL of the reaction slurry into a separate test tube that was placed in an ice bath. The slurry in the test tube was then drawn into a 1.0 mL airtight syringe and filtered through a 0.22pm Teflon syringe filter into a vial that contained a known amount of internal standard. The vial was then weighed. The samples were then diluted with chloroform and injected into the gas chromatograph for analysis. Each sample was analyzed in triplicate. B. Analysis of FA Product Mixture. An HP-1 (Hewlett-Packard) wide bore capillary column (5 m, 0.53 mm i.d., 2.65 pm film thickness) was used for these analyses. The flow rate of carrier gas (helium) was 15 mumin, and the split ratio was 11:l. No make-up gas was used. The initial column temperature was 40 "C. After 3 min, the column was heated at a rate of 20 "C/ min t o 200 "C. The column was then held at 200 "C for 5 min. C. Analysis of EDFA Product Mixture. An HP-1 wide bore capillary column (25 m, 0.32 mm id., 1.05
Ind. Eng. Chem. Res., Vol. 34, No. 10, 1995 3395 pm film thickness) was used for these analyses. The flow rate of carrier gas (helium) was 2 mumin, and the split ratio was 50:l. 28 mUmin of nitrogen was used as the make-up gas. The initial column temperature was 180 "C. After 15 min, the column was heated at a rate of 20 "C/min to 220 "C. The column was then held at 220 "C for 5 min.
Results and Discussion The data were analyzed using the method of initial rates (see Tables 1 and 2 for the data pertinent to FA and EDFA, respectively). Initial rates were determined by performing linear least squares fits of the data for conversions from 0 to approximately 30%. Analysis of the data using a numerical integration method will be the subject of a future paper. The effects of the amount of catalyst, the concentrations of both reactants and products, the activity of hydrogen in solution, and the temperature on the rates of hydrogenation of both FA and EDFA are discussed below. The yields of the desired products (i.e., THFA or EDTHFA) were approximately 95% for the experimental conditions employed. Hydrogenation of Furfurylamine. A. Effect of the Amount of Catalyst. The initial rate of hydrogenation of FA is directly proportional t o the amount of catalyst in solution. This result indicates both that the catalyst in adequately suspended during an experimental run and that the rate of mass transfer of hydrogen from the gas phase to the liquid phase does not limit the rate of reaction. B. Effect of the Concentrationof FA. The initial rate of hydrogenation of FA is independent of the concentration of FA for concentrations in the range from 1.33to 2.67 km0Ym3(see experiments FA2B and FA2D). This result indicates that for this range of reactant concentrations, substantially all of the catalytic sites on which FA adsorbs are occupied by this species. C. Effect of the Concentration of THFA on the Rate of Hydrogenation of FA. The initial rate of hydrogenation of FA decreases approximately 23% when THFA is present in equimolar amounts (see experiments FA2B and FA2TH and experiments FA5B and FA5TH). This result indicates that FA and THFA compete for adsorption on the catalytically active sites, i.e., that the product THFA acts as an inhibitor of the reaction. D. Effect of the Activity of Hydrogen in Solution. At 70.6 "C, the initial rate of hydrogenation of FA is approximately first order with respect to the activity of hydrogen in solution. However, at 50.0 "C, the order of the reaction with respect to hydrogen is approximately 0.75. This result suggests that at the lower reaction temperatures, the catalytic sites available for adsorption of hydrogen are occupied to a greater degree as the activity of hydrogen in solution increases. E. Effect of Temperature. The effect of temperature on the initial rate of hydrogenation of FA is shown in Figure 3 in the form of an Arrhenius plot. The average activation energy corresponding t o the best fit straight lines is 49.4 & 2.5 kJ/mol. F. Mass Transfer Limitations. For heterogeneous catalytic reactions, mass transfer andor intraparticle diffusion effects often influence the experimentally observed kinetics. Hence it is important to assess the extent to which these transport processes influence the rate of hydrogenation of FA. The effects of external mass transfer and intraparticle diffusion on the rate of
0.11
,
,
O B
1
0.88
,
,
I
0.80
0.31
0.32
lrr(1K) xloe Figure 3. Arrhenius plot based on initial rate data for the hydrogenation of FA (0 = 12.84 kg 3.65% RWA1203/m3, and 0 = 6.42kg 3.65% RWA1203/m3). 11.0
'0
106
-
10.0
-
9.6
-
4
n
2 111
"a
5
8
-
bo 8.6
-
8.01
I
,
,
'
,
,
,
,
,
3398 Ind. Eng. Chem. Res., Vol. 34, No. 10, 1995 Table 3. Preliminary Estimates (Based on Initial Rate Data) of the Adsorption Equilibrium Constant &, the Ratio k&, and the Ratio RdR1 for the Hydrogenation of Furfurylamine
:1 7.0
70.5 60.0 50.0 h0 4.0
4
3.0 0.m
0.00
0.08
0.04
Yd, (lun-9
Figure 8. Effect of the gross external dimensions of the 3.65% RWAl203 catalyst on the rate of hydrogenation of FA (experimental conditions: temperature = 70.6"C, [FA10 = 1.34 M, UH* = 0.0169, and w = 6.42 kg/m3).
of the reaction (Fogler, 1986). The results of the fourth set of experiments also confirm that intraparticle diffusion effects are negligible under the experimental conditions employed. G. Discussion. The results of the initial rate analysis indicate that the rate of hydrogenation of FA is consistent with a rate expression of the form d[FAldt
wk 1[FA1aHz
+
+
(1 K,[FAl+ K2[THFAI)(1 K3aHz)
(3)
where w is the weight of the catalyst per unit volume, k l is the reaction rate constant, KI, K2, and K3 are adsorption equilibrium constants, [FA] is the concentration of FA present in the liquid phase, and [THFAI is the concentration of THFA present in the liquid phase. This rate expression is consistent with a two-site Hougen-Watson-Langmuir-Hinshelwood reaction mechanism of the following form:
+ o1 FA + H2
-
0,
+ H2-al FAH2-02 + H2-a1 FA-u2
=*
H2-0,
(4)
FA-0,
(5)
FAH2-02
+ o1
(6)
THFA-0,
+
(7)
+
g1
THFA-0, THFA 0, (8) The assumptions made in deriving eq 3 from this model are (1)hydrogen adsorbs on a different type of catalytic site than does either FA or THFA (2) FA and THFA compete for adsorption on one type of catalytic site, (3) the rate-limiting step is reaction between an adsorbed FA molecule and an adsorbed hydrogen molecule (see eq 61,(4) the reactions shown in eqs 6 and 7 are irreversible, and (5)adsorption of hydrogen, FA, and THFA on their respective catalytic sites are a t equilibrium (see eqs 4, 5, and 8). For the temperature range from 42.5 t o 70.6 "C, the apparent overall activation energy for this reaction is 49.4 -+ 2.5 kJ/mol. Values of the adsorption equilibrium constant K3,the ratio kllK1, and the ratio KdKl are shown in Table 3. The values for K3 a t 50.0 and 70.5 "C were determined from both the slope and the intercept of plots of llro versus l/UH,. From these two values of K3, the enthalpy change accompanying the chemisorption of hydrogen
2.35 5.27 11.92
5.98 3.70 2.32
0.28 0.36
was determined to be -73 kJlmo1. This enthalpy change was then used to estimate the value of K3 at 60.0 "C. Values for the ratio k1K1 were determined using the values of K3 and the assumption that Kl[FAI is much greater than unity for experiments in which THFA was not present initially. The values for the ratio KdK1 were determined from the initial rate data for experiments FA2TH and FABTH, the previously determined values of K3, and the assumption that the sum (Kl[FAI K2[THFAI) is much greater than unity. A Hougen- Watson-Langmuir-Hinshelwood model based on an irreversible bimolecular reaction between an adsorbed FA molecule and an adsorbed hydrogen molecule on the same type of site was not chosen to represent the hydrogenation of FA because the rate of reaction does not exhibit a maximum value when either the activity of hydrogen in solution is increased or the concentration of FA is increased. Similarly, a HougenWatson-Langmuir-Hinshelwood model based on an Ely-Rideal type reaction mechanism was not chosen to represent the hydrogenation of FA because the order of the reaction is less than first order with respect to hydrogen and zeroth order with respect to FA. A Hougen- Watson-Langmuir-Hinshelwood model based on an irreversible bimolecular reaction between an adsorbed FA molecule and an adsorbed hydrogen atom was also rejected as a model for the kinetics of the hydrogenation of FA because the observed order of the reaction with respect to hydrogen is greater than 0.5. Hydrogenation of 5 , 6 ' - E t h y l i d e n mlamine. The hydrogenation of EDFA to EDTHFA proceeds through the half hydrogenated intermediate, 5-1142(aminomethyl)tetrahydrofuran-5-yllethyllfurfurylamine (1l2-EDFA). This intermediate is analagous t o the intermediate that is formed during the hydrogenation of methylenedianiline to 4,4'-methylenebiscyclohexylamine (Allen, 1988). A. Effect of the Amount of Catalyst. The initial rate of hydrogenation of EDFA is directly proportional to the amount of catalyst in solution. This result indicates both that the catalyst is adequately suspended during an experimental run and that the rate of mass transfer of hydrogen from the gas phase to the liquid phase does not limit the rate of reaction. B. Effect of the Concentration of EDFA. The initial rate of hydrogenation of EDFA increases approximately 22% when the concentration of EDFA is increased from 0.66 to 1.31 km01/m3 (see experiments ED and ElHP). This result suggests that at the higher concentration of EDFA, virtually all of the sites available for adsorption of EDFA are occupied. C. Effect of the Concentration of EDTHFA. The initial rate of hydrogenation of EDFA decreases by approximately 4% when equimolar amounts of EDTHFA are present in the reactor (see experiments E2TH and ElHP). This result suggests that EDTHFA is not adsorbed on the catalyst surface to a significant extent. The effect of EDTHFA on the rate of hydrogenation of
+
Ind. Eng. Chem. Res., Vol. 34, No. 10, 1995 3397
0.0
1.0
8.0
3.0
4.0
6.0
Activity of Hydrogen x 1oP Figure 6. Effect of the activity of hydrogen in solution on the initial rate of hydrogenation of EDFA (0 = 79.1 "C, and 0 = 70.4 "C).
rate of this reaction (1)does not change significantly when baffles are present in the reactor (the rate of reaction at 79.1 "C decreased approximately 8% when baffles are not used in the reactor), (2) is directly proportional to the amount of catalyst used (see subheading entitled Effect of the Amount of Catalyst), and (3) is essentially independent of the gross external dimensions of the catalyst powder if the volume average diameter of the particles is less than approximately 18 pm. The results of the second set of experiments combined with the results shown in Figure 6 confirm that the rate of mass transfer of hydrogen from the gas phase t o the bulk liquid does not influence the kinetics of the reaction under the conditions investigated. The results of the set of experiments in which the gross external diameter of the catalyst particles was varied confirm that if baffles are used in the reactor the rate of mass transfer of the reactants from the bulk liquid to the exterior surface of the catalyst does not influence the kinetics of the reaction under the conditions investigated (Fogler, 1986). These results also confirm that intraparticle diffision effects are negligible under the experimental conditions employed (Fogler, 1986). G. Discussion. The results from the initial rate analysis indicate that the hydrogenation of EDFA is consistent with rate expressions of the following form:
- d[EDFAI dt
wkl[EDFAlaH2
wk ,[EDFA1aH2
- d[EDFAl -
0.1 0.88
0.m
030
0.~1
om
lrr(lni) x1oa Figure 7. Arrhenius plot based on initial rate data for the 0= hydrogenation of EDFA (0 = 12.56 kg 3.65% RWAlzOs/m3, 6.31 kg 3.65% RWA120s/m3, and A = 3.17 kg 3.65% RWA1203/m3). EDFA differs from the previously described effect of THFA on the rate of hydrogenation of FA. One plausible explanation for this difference is that steric effects hinder the ability of EDTHFA to adsorb on the catalyst surface to a significant extent. D. Effect of the Activity of Hydrogen in Solution. The effect of the activity of hydrogen in solution on the initial rate of hydrogenation of EDFA is illustrated in Figure 6. These data demonstrate that the rate of hydrogenation of EDFA reaches an asymptotic value as the activity of hydrogen in solution increases. This result indicates that as the activity of hydrogen in solution increases from 0.0087 to 0.0498, virtually all of the sites on which hydrogen adsorbs are occupied. E. Effect of Temperature. The effect of temperature on the initial rate of hydrogenation of EDFA is shown in Figure 7 in the form of an Arrhenius plot. The average activation energy corresponding t o the best fit straight lines is 63 f 4 kJ/mol. F. Mass Transfer Limitations. The effects of external mass transfer and intraparticle diffusion on the rate of hydrogenation of EDFA were investigated by incorporating baffles into the reactor to enhance turbulent mixing, changing the amount of catalyst used, and changing the gross external dimensions of the catalyst powder. It can be concluded that external mass transfer effects and intraparticle diffision effects do not influence the kinetics of the hydrogenation of EDFA t o a significant extent a t temperatures of 79.1 "C or lower because the
(9)
(1 + Ki[EDFAI)(1 +
dt
+
(1 Ki[EDFA]
+ K ~ u H ~(10) )~
where [EDFA] is the concentration of EDFA present in the reactor. Equation 9 is consistent with the following Hougen- Watson-Langmuir-Hinshelwood mechanism in which hydrogen and EDFA are adsorbed on different types of sites:
+ EDFA + EDFA-(T, + H2-a1 H2
-
( T ~
H2-a1
(11)
(T,
EDFA-u,
(12)
EDFAH2-a2 + ( T ~ (13)
The remaining steps for the formation of EDTHFA are not shown for this mechanism because they are not required for the derivation of eq 9. The assumptions made in deriving eq 9 from the indicated mechanism are (1)hydrogen adsorbs on a different type of catalytic site than does EDFA, (2) the rate-limiting step is the reaction between an adsorbed hydrogen molecule and an adsorbed EDFA molecule (see eq 12), and (3) the rates of adsorption and desorption of hydrogen and EDFA on their respective catalytic sites are sufficiently rapid that these processes may be presumed to be at equilibrium. Equation 10 is consistent with a two site HougenWatson-Langmuir-Hinshelwood mechanism in which all species are adsorbed on the same type of site.
H2
+
(T
EDFA
+
(T
EDFA-(T
H,-u
(14)
EDFA-(T
(15)
6 . )
+ H,-(T-*EDFAH~-(T +
(T
(16)
Again, the remaining steps for the formation of EDTHFA via this mechanism are not shown because
3398 Ind. Eng. Chem. Res., Vol. 34, No. 10,1995 they are not required for the derivation of eq 10. The assumptions made in deriving eq 10 are identical t o those used to derive eq 9. The sole difference between the mechanisms associated with eqs 9 and 10 is that in the latter case hydrogen and EDFA are assumed to adsorb on the same type of catalytic site rather than on unlike sites. Because of the paucity of the initial rate data, one cannot employ standard statistical methods to clearly differentiate whether eq 9 or eq 10 best represents the rate of hydrogenation of EDFA. However, a future paper based on analysis of the data using a numerical integration procedure will aid in determining which of these two rate expressions best describes the hydrogenation of EDFA. In addition, values of the reaction rate constant and the adsorption equilibrium constants present in the rate expressions shown in eqs 9 and 10 are not presented here because of the paucity of the data. Work is in progress to address this issue. For the temperature range from 50.0 to 79.1 “C, the apparent overall activation energy for the hydrogenation of EDFA is 63 f 4 kJ/mol. A Hougen-Watson-Langmuir-Hinshelwood model based on an Ely-Rideal type reaction mechanism was not chosen t o represent the hydrogenation of EDFA because the order of the reaction is not first order with respect to either EDFA or hydrogen. Conclusions Results from investigations of the liquid phase hydrogenation of FA and EDFA to the corresponding tetrahydrofuran compounds over a W A l 2 0 3 catalyst at moderate temperatures (42.5-80.1 “C) and moderate hydrogen pressures (231- 1380 kPa) indicate that both of these reactions proceed via Hougen-Watson-Langmuir-Hinshelwood reaction mechanisms. The ratelimiting step for these two systems is the reaction between an adsorbed hydrogen molecule and an adsorbed furfurylamino compound (i-e., either FA or EDFA). The overall activation energies for the hydrogenation of FA and EDFA are 49.4 f 2.5 and 63 f 4 kJ/mol, respectively. Acknowledgments The authors wish to express their appreciation for financial support provided by the Wisconsin legislature for studies relating to economic development and administration through the University-Industry Research program. Additional support was provided by QO
Chemicals, Inc., Eureka Trading Ltd., and the USDA Forest Service Cost Share Program. Nomenclature a H p = activity of hydrogen in solution d, = diameter of catalyst powder, pm EDFA = 5,5’-ethylidenedifurfurylamine EDTHFA = 5,5’-ethylideneditetrahydrofurfixylamine FA = furfurylamine f = hypothetical liquid fugacity of hydrogen, Pa K 1 = reaction rate constant, m3/(skg 3.65% WAl203) K1,K2 = adsorption equilibrium constants for furan amines, m3flun01 K3 = adsorption equilibrium constant for hydrogen p~~ = partial pressure of hydrogen, Pa ro = initial rate of reaction (kmol/(m38)) T = temperature, K THFA = tetrahydrofurfurylamine w = concentration of catalyst, kg/m3 l/P-EDFA = 5-[ 1424aminomethyl)tetrahydrofuran-5-y1]ethyllfurfurylamine
Literature Cited w e n , G. F. The hydrogenation of methylenedianiline. Chem. Znd. (Dekker), 1988,33 (Catal. Org. React.), 323. Cawse, J. L.; Stanford, J. L.; Still, R. H. Polymers from renewable sources, 1 Diamines and diisocyanates containing difurylalkane moieties. Makronol. Chem. 1984,185, 697. Fogler, H. S. Elements of Chemical Reaction Engineering; Prentice-Hall: Engelwood Cliffs, NJ, 1986, pp 604-607. He, X.; Conner, A. H.; Koutsky, J. A. Evaluation of furfuryl amines as curing agents for epoxy - - resins. J . Polyrn. Sci., Chem. Ed. 1992,30;533. Holfinger, M. S.Wood Adhesives from Furfural-Based Diamines and Diisocvanates. PhD thesis, Universitv of Wisconsin, Madison, WI, 1992. Holfinger, M. S.;Conner, A. H.; Lorenz, L. F.; Hill, C. G., Jr. Difurfurvl diisocvanates: New adhesives derived from renewable res&rces. ). Appl. Polym. Sci. 1993,49,337. Lemcoff, N.0.Liquid phase catalytic hydrogenation of acetone. J . Catal. 1977,46, 356. Radhakrishnan, K.; Ramachandran, P. A,; Brahme, P. H.; Chaudhare, R. V. Solubility of hydrogen in methanol, nitrobenzene, and their mixtures. Experimental data and correlation. J . Chem. Eng. Data. l963,28,1.
Received for review January 4, 1995 Accepted June 2, 1995* IE9500190
* Abstract published in Advance ACS Abstracts, September 1, 1995.
