- Email: [email protected]
Characterization of foam catalysts as packing for tubular reactors
Accepted Manuscript Title: Characterization of foam catalysts as packing for tubular reactors Author: Farzad Lali PII: DOI: Reference:
S0255-2701(16)30087-3 http://dx.doi.org/doi:10.1016/j.cep.2016.04.001 CEP 6784
To appear in:
Chemical Engineering and Processing
Received date: Accepted date:
18-3-2016 5-4-2016
Please cite this article as: Farzad Lali, Characterization of foam catalysts as packing for tubular reactors, Chemical Engineering and Processing http://dx.doi.org/10.1016/j.cep.2016.04.001 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
*Manuscript
Characterization of foam catalysts as packing for tubular reactors Farzad Lali Institute of chemical process fundamentals of the Czech academy of sciences, Rozvojova 2/135 165 02 Prague, Czech republic, phone: +420 220 390 233, fax: 220920661,220920649, E-Mail: [email protected]
Abstract This study deals with the characterization of catalytically active foams as packing for continuous tubular reactors in terms of reactor performance and overall mass transfer. The tubular reactor was operated in a vertical direction such that gas and liquid entered the reactor in co-current upward flow. The hydrogenation reaction of α-methylstyrene was applied for characterization of the Pd/Al2O3 foam catalysts with a bed length of 50 cm and a diameter of 1.8 cm. Two pore densities for foams namely 30 and 45 PPI were investigated. The foam catalyst showed an overall mass transfer rate that ranged between 0.2 and 14 s-1 for 30 PPI and 0.3 to 18 s-1 for 45 PPI foam catalyst, whereas the energy dissipation was below 1 Wm-3. Keywords: overall mass transfer, foam catalyst, tubular reactor, hydrogenation
1. Introduction The chemical industry is a major consumer of energy worldwide and yet the majority of chemical processes involve catalytic reactors. Therefore, the energy efficiency of chemical reactors is a very crucial issue. For the heterogeneously catalyzed multiphase reactions, several reactor technologies have been already applied such as bubble columns, trickle bed reactors, CSTRs or there are technologies under development for instance wall-coated monolith reactors, monolith reactors with dumped catalysts and last but not least microreactors. The application of bubble columns and CSTRs involves suspended catalysts that demand a subsequent separation of catalyst particles. The pelleted catalysts are usually applied as random packed bed in trickle bed reactors. Yet the disadvantages of the random packed beds are high pressure drop and formation of hot-spots. Microreactor is an innovative reactor concept but the fabrication of microractors is associated with high production costs and the pressure drop inside such reactors is relatively high. Usually a trickle
regime is favored in conventional reactor design for continuous three phase reactions, where the gas side mass transfer is limiting due to low gas-liquid solubility. The advantage of the catalytically active foam packings is that, because of the high voidage (up to 97%) and low pressure drop, it is possible to run three phase reactions in an upward flow direction. A challenging issue of the design of trickle bed reactors is dealing with “hot spots” in the reactor, that causes a non-uniform temperature profile. Such hot spots are known to be undesired and to cause side products and shorten the life of the catalyst. Application of catalytically active foam packings in a co-current upward direction can increase the uniformity of the temperature profile due to very low energy dissipation, high void volume, and increased liquid holdup. Measurement of overall volumetric mass transfer can be utilized as a criterion to evaluate the performance of a structured packing. Several studies used the hydrogenation of AMS using Pd/Al2O3 catalysts to evaluate the performance of various reactor types. Turek and Lange [1] investigated a trickle bed reactor with an inner diameter of 3.4 cm with liquid flow rates ranging between 0 and 1.5 L h-1 and gas flow rates from 0 to 100 L h-1. Cini et al. [2] adopted the hydrogenation of AMS for tubular supported and Pd/γ-Al2O3 impregnated ceramic membranes. Kreutzer et al. [3] used this hydrogenation reaction to characterize the performance of a tubular reactor with wall-coated monolith packing. Purnama et al. [4] investigated the performance of a flow-through ceramic membrane reactor at temperatures ranging between 45-50 °C and pressures between 1-40 bar, and compared the performance of this reactor with such other reactor types as trickle bed and bubble column reactors. Haase [5] characterized wall-coated and composite (packed with pellets) minichannels for gas-liquid-solid-reactions in various multiphase flow regimes. Langsch [6] studied mass transfer in miniaturized packed bed reactors at different multiphase flow regimes using the hydrogenation of AMS to cumene. Tourvieille et al. [7] investigated minichannels packed with foam catalysts by applying a pulsing flow regime and obtained overall mass transfer values that ranged from 0.25 s-1 to 1.9 s-1. The present study deals firstly with characterization of the reactor performance foam catalysts as tubular packings. Subsequently, the overall mass transfer of 30 PPI and 45 PPI foams were measured and compared with the data from literature. The overall mass transfer is referred to the external mass transfer which is mainly influenced by the foam structure, whereas the internal mass transfer in the washcoat also plays a role. Thus, internal mass transfer was considered in the mathematical model used to evaluate the observed reaction rates. Although there is some effect
from internal mass transfer, the influence of the internal mass transfer was although available however, the major concentration of the palladium was detected in the vicinity of the washcoat surface as shown in a previous paper [10] by applying EDX analysis. Therefore, the influence of the internal mass transfer for different washcoat thicknesses between 17.8 µm (for 30 PPI) and 6 µm (for 45 PPI foams) plays a less important role.
2. Experimental 2.1.1.
Catalyst preparation
The foam catalysts with pore densities of 30 PPI and 45 PPI were prepared using the methods in a previous publication by Lali et al. [8] and Leon et al. [9].
Fig. 1 : Pictures of foam samples after different preparation steps 0) pristine aluminum foam 1) anodized foam 2) coated with alumina 3) impregnated with palladium acetate 4) reduced Pd/Al2O3 foam catalyst 5) spent foam catalyst
The preparation steps of a 30 PPI foam catalyst are shown in Fig. 1, where the image number 0 represents an uncoated aluminum foam, number 1 was an anodized foam, number 2 was washcoated anodized foam with alumina slurry, number 3 was impregnated with palladium acetate and number 4 was reduced foam catalyst. The image 5 in Fig.1 shows a spent foam catalyst.
2.2 Pretreatment of α-Methylstyrene
The purities of the α-methylstyrene and cumene supplied by SIGMA-ALDRICH were 99% and 98%, respectively. In order to prevent the AMS from polymerizing, 15 ppm of the inhibitor tert-butylcatechol (TBC) was added by the supplier. Earlier experiments showed that the TBC can build residues on the catalyst and decrease the catalytic activity drastically. The authors have several views about the deactivation of Pd/Al2O3 catalyst by TBC. While some studies [10] confirm a negative impact of TBC on catalytic activity, Meille et al. [11] found neither a negative impact on a Pd/Al2O3 catalyst nor any reaction inhibition. Nevertheless, our own preliminary experiments confirm the inhibitory effect of TBC. The low concentration of Pd in the foam catalysts in the present study also increases the vulnerability of the catalyst to being inhibited. For these reasons, AMS was distilled before preparation of the AMS solution with cumene by using a rotary evaporator (BÜCHI R-114). The vacuum distillation of AMS was carried out at 50 mbar pressure and a bath temperature of 85 °C. The cumene was also distilled in order to reduce its impurities, which can also build residues on active catalytic sites. 2.3 Experimental Setup
The hydrogenation experiments were carried out in a tubular reactor with a diameter of 1.8 cm and a height of 50 cm. The gas and liquid streams were continuously introduced into the reactor from the bottom. The liquid (20% by weight in cumene) was pumped with an Alphachrom pump. The flow rate of the pump could be adjusted over a range of 1 to 200 mL min-1. The gas flow was regulated by a Bronkhorst F231M mass flow controller. The foam catalysts were prepared as cylinders with a height of 2 cm and a diameter of 1.8 cm. In order to avoid bypassing the fluids and channeling, the foam catalysts were wrapped in PTFE strips with a thickness of ~0.5 mm. The AMS/Cumene mixture was preheated to the desired temperature prior to entering the reactor by a Horst HAT MC1 heating cord. The reactor was jacketed, and its temperature was controlled by a Julabo H-2 heater thermostat. The temperature gradient through the reactor was measured by four thermocouples. The experiments for the measurement of the overall mass transfer were carried out at atmospheric pressure and the expermients for testing the catalytic performance were carried out at 10 bar. The pressure inside the reactor was monitored by a WIKA pressure transducer. The product stream from the reactor was cooled before taking samples by a LAUDA WK500 recirculating chiller. The cooling water flowed through a helical copper pipe with a diameter of 3
mm that was wrapped around the sampling vessel K-01 to cool the liquid sample to the desired temperature of about 30 °C. Fig. 2 is a system diagram for the experimental setup. For the experiments at elevated pressures (10 bar), the valve V-6 was replaced with a relief valve, and a pressure transducer was installed immediately after the reactor to monitor the pressure variations. The relief valve
Fig. 2 system diagram for reaction studies on a continuous reactor with catalytic foam packing
(Swagelok) was used to regulate the desired pressure manually by means of a spring with a specific hardness for pressures ranging between 3.4 and 24.1 bar. It should be noted that the use of a spring-loaded relief valve is not the most accurate way to regulate pressure. Consequently, some pressure fluctuations were observed. Furthermore, the reaction studies at an elevated pressure of 10 bar were carried out polytropically, in order to study the temperature effects along the fixed bed of foam catalysts. The mass transfer measurements were based on the experiments at atmospheric pressure to avoid inaccuracy.
For taking samples, the valve V-07 was briefly opened to drain the liquid sample for refractive analysis. The temperature was monitored using type-K thermocouples at 5 spots along the reactor. The specifications of the foam catalysts, which were used in this study are listed in table 1.
Table 1: Specifications of the catalytically active foam packings used for reaction studies. Pore density Pd-loading based on washcoat mass (%) Washcoat mass (g) Catalyst concentration per reactor volume (gPd m-³react) Washcoats thickness (μm)
30 PPI
45PPI
2.082
1.881
2.5854
2.3435
575
372.3
17.8
6.63
1321.6
2892.8
25x2
33x1.5
0.0835
0.1013
Geometrical surface area per packing volume -1
(m ) Packing length (cm) Solid Holdup with Washcoat (m³S m-³)
2.4 Mass Transfer Model In this work the overall mass transfer coefficient was determined by means of a hydrogenation reaction of α-methylstyrene. The determination of overall reaction coefficient was done based on the film theory. For a first order reaction the relationship between reaction rate and mass transfer resistances are shown in Eq. 1. 1 1
1
1
1
∙
,
Eq. 1
It was assumed that the mass transfer resistance on the liquid side is much larger than the gas side. For this reason the gas side mass transfer resistance was negligible as shown in Eq. 2 . 1
1
1
Eq. 2
The first term in Eq. 2 represents the liquid side gas-liquid mass transfer resistance, the second term is the liquid solid mass transfer, and the third term represents the intrinsic kinetic together with the diffusion limitation in the catalyst support. The overall mass transfer coefficient, as shown in Eq. 3 , is the sum of the liquid side gas-liquid and liquid solid mass transfer resistances. 1
1
1
Eq. 3
This mass transfer model is valid for a reactor with almost plug flow, which was the case for the foam catalyst packing according to the obtained results in the preliminary hydrodynamic experiments, in which the Bodenstein number was much greater than 7. 2.5 Hydrogen Solubility and Intrinsic Kinetics The mass transfer model presented in the previous section shows that the solubility of the hydrogen in the AMS/cumene mixture plays an important role in evaluating the results from the observed reaction rate using Eq. 1. Several studies in the literature have dealt with the solubility of hydrogen in the AMS/cumene mixture. These studies are carried out for different reaction conditions. Herskowitz [12] correlated the solubility of hydrogen in AMS at temperatures ranging between 15 and 74 °C and ambient pressure with a maximum deviation of 3%. Löwe et al. [13] developed correlations for the hydrogen solubility in pure AMS and cumene. Stefoglo [14] determined empirically the Henry coefficient for hydrogen in AMS or cumene at temperatures ranging between 20 and 80 °C and pressures up to 5 bar. Herskowitz [15], in a later paper, investigated a range of temperatures from 30 to 100 °C and pressures from 1 to 69 bar.The solubilities corresponding to Löwe et al. [13] and Herskowitz [15] are relatively similar, whereas the correlation proposed by Meille et al. [11] exhibits the largest deviation. This discrepancy can be explained by different underlying assumptions. While, Löwe et al. [13] Stefoglo [14] and Herskowitz assumed different solubilities for hydrogen in AMS and cumene, Meille et al. [11] assumed that the solubility of hydrogen in AMS and cumene is approximately equal. This assumption allowed Meille to describe the
solubility of hydrogen in both liquids by applying the correlation proposed in Herskowitz’s first publication [12]. For evaluation of the observed reaction rate and calculation of the overall volumetric mass transfer, the correlation proposed by Herskowitz [15] in his later publication in 1983, who investigated among other conditions also the solubility of hydrogen in AMS/cumene mixture at 100 °C and ambient pressure was favored. The intrinsic kinetic for the hydrogenation reaction of AMS to cumene using a Pd/Al2O3 catalyst was taken from Meille et al. [11].
3. Results and Discussion
3.1 Catalyst Durability Catalyst durability can be affected by chemical residues blocking on the active sites. Therefore the activity of the foam catalysts with a pore density of 45 PPI were monitored over a period of 8 hours. The experiments were carried out at a liquid (20 wt. % AMS in cumene) flow rate of 10 ml min-1 , and a gas flow rate (hydrogen) of 0.05 L min-1. At these flow rates the liquid residence time was relatively long. The hydrogenation reaction was carried out at ambient pressure and 120 °C. The catalyst activity was monitored by measuring the conversion of AMS to cumene at 30 minutes intervals. A slight loss of activity was observed after 7 hours of time on stream, but thereafter, catalytic activity remained almost constant for the next 50 hours. 3.2 Experiments at Elevated Pressure The aim of the polytropical hydrogenation experiments at 10 bar was to compare the results with the available literature in terms of space time yield and productivity. The temperature profile in the reactor was also investigated. For the reaction studies at 10 bar and a liquid inlet temperature of 40 °C, the liquid flow rates were 10 and 20 mL min-1 respectively. Gas flow ranged from 0.1 to 1.2 L min-1.
80
0.9
70
0.8
60
0.7 0.6
50
0.5
40
0.4
30
0.3
temperature (°C)
Coonversion (-)
1.0
20
0.2
10
0.1 0.0
0
0
2
4
6
8
10
uG (mm s-1)
Fig. 3: outlet temperature (○) and total conversion (♦) in the reactor for 30 PPI foam catalysts for a packing length of 50 cm, Liquid flow rate of 10 mL min-1, 10 bar H2, polytropical conditions and an inlet temperature of 40 °C [16]
0,1
0,01
-1
-3
0,1
0,01
1E-3
1E-3
1E-4
1E-4 stirred tank with pellets
tubular reactor with foam catalysts
-3
10
1
CPd = 1557 g m
10 bar
-3
45.5 °C
activity space time yield
40 °C,10 bar CPd = 180 g m
-3
1
CPd = 590 g m
Pd
10
40 °C,10 bar 20
space time yield (mol s m )
100
100
trickle bed reactor
Fig. 4: Comparison of the reactor performance of continuous tubular reactor with foam catalysts with conventional reactors; Stirred tank with a fine slurry catalyst (T=45.5 °C, P=1.51 bar) Meille et al.[11]; tubular reactor with catalytically active foam packing in co-current upward flow (40 °C, 10 bar,VL=20 mL min-1) F. Lali [17];trickle bed reactor (40 °C, 10 bar, VL=25 mL min-1) Schubert [18]
Fig. 3 shows that the outlet temperature and total conversion increase as the gas flow rate increases due to the polytropical nature of the process. A comparison with a conventional trickle
bed reactor at 40 °C and liquid flow rate of 25 mL min-1 in Fig. 4, shows the catalyst activity and space time yield of foam catalysts for the same conditions are 10 times larger and yet the amount of Pd was 90% less and the hydrogen gas demand was 30% less. The stirred tank study in Fig. 4 from reference [11] was carried out at nearly intrinsic reaction conditions. 3.3 Pseudo-Arrhenius Plots To ensure that the mass transfer measurements were carried out in the mass transfer limited region, Pseudo-Arrhenius plots were made showing the temperature dependence of “observed” reaction coefficients for each flow rate. The observed reaction rate coefficient was calculated by applying Eq. 4. For this reason, external mass transfer effects were not ruled out. The observed (apparent) activation energy can be calculated from the plot from regions that were distinguished by different slopes as shown in Fig. 5. For the reaction studies a parameter window on the basis of the preliminary studies particularly, two phase pressure drop that was consisting of liquid flow rates ranging from 10 mL min-1 to 40 mL min-1 and the highest gas rates for every liquid rate were chosen slightly higher than the values obtained in the preliminary study for air-water system in order to the differences between the physical properties of both systems such as viscosity and surface tension.
Eq. 4
∗
2 Ea=8.56 kJ mol-1
1.8
L=20 mL/min G=0,23 G=0.23mL/min L min-1 L=30 mL/min G=0,35 G=0.35mL/min L min-1
lnkobs (s-1)
1.6
L=40 mL/min G=0,35 G=0.35mL/min L min-1
1.4 1.2
Ea=26.47 kJ mol-1
1 0.8 0.6 0.4 0.0025
0.0026
0.0027
0.0028
0.0029
0.003
0.0031
1/T (K-1)
Fig. 5: Pseudo-Arrhenius plot for 30 PPI foam packing 50 cm, atmospheric hydrogen pressure
The foam catalysts of 30 and 45 PPI pore density were investigated at temperatures ranging from 60 to 120° C at ambient pressure. The purpose of measuring at the atmospheric pressure was to avoid any pressure fluctuations which could influence the GL-mass transfer at elevated pressures. The reaction rate constant is depicted against the reciprocal temperature in Fig. 5. The determination of the kinetically limited region focused on the higher gas flow rates from the parameter window, at which a kinetic limitation was more likely due to the increasing gas holdup. Fig. 5 shows how the observed reaction rate increases with temperature in a pseudo-Arrhenius plot. The region between 333 and 393 K for the chosen parameter window was most likely to be kinetically limited. After this region, the effect of external mass transfer limitations was investigated. On the basis of these results a reaction temperature of 120 °C was chosen to rule out kinetic limitations. The values of the intrinsic activation energy in the literature range from 19.7 kJ mol-1 (Babcock et al.[19]) to 43.5 kJ mol-1 (Cini et al.[20, 21]). Cini et al. [20] also gives an overview of the results reported in the literature. Moreover, provide an extensive discussion of the various limitation factors, such as pore diffusion, that can influence apparent activation energy. The value reported by Cini et al. [21] was considered the most precise value for the intrinsic activation energy of AMS hydrogenation using Pd/Al2O3 catalyst. The majority of later studies also reported an intrinsic activation energy of about 40 kJ mol-1. Meille et al. [11], for instance, reported 38.7 kJ mol-1. The apparent activation energy for the flow rates in Fig. 6 was about 26 kJ mol-1.The discrepancy compared to the intrinsic activation energy can be elucidated by the reaction limitation by pore diffusion in the catalytic layer as Eq. 5 demonstrates:
,
,
,
Eq. 5
2
The apparent activation energy was smaller than the reported intrinsic activation energy and remarkably higher than the activation energy for the hydrogen diffusion in 20 wt. % AMS (EA,diff =12.5 kJ mol-1 [6]). By applying the Eq. 5 proposed by Satterfield et al. [22] for the pore diffusion limited area, an intrinsic activation energy of 40 kJ mol-1, which is in good agreement with the data reported in the literature. The majority of other authors who published after the work by Cini et al. [2] was published, reported values close to 40 kJ min-1.
The intrinsic kinetics of the hydrogenation reaction of AMS was calculated based on the reaction studies in the rotating foam reactor. These results were in good agreement with the intrinsic kinetics reported by Meille et al. [11]. As shown in the Fig. 5, the observed reaction rate coefficient depends on residence time in the reactor. For this reason it is crucial to calculate residence time. To obtain a more accurate estimate of residence times in the reactor, the following method was adopted. The residence time was calculated based on the amount of the AMS converted to cumene
. It was assumed that
a certain amount of liquid containing AMS, at steady state conditions, remains in the reactor as long as the hydrogen needed to convert
is consumed.
Eq. 6
1
Eq. 7
Eq. 8
Eq. 9
Eq. 10
Eq. 11
The gas holdup for the hydrogen in the reaction εΔG was calculated from Eq. 7 while the volume of the hydrogen in the reactor was calculated using Eq. 6. With the known amount of hydrogen in the reactor at steady state conditions, the moles of hydrogen consumed in the reactor were calculated from Eq. 9. The subsequent residence time was calculated as the ratio of the moles of hydrogen consumed to the molar flow of hydrogen calculated from Eq. 11. 3.4 Measurement of Overall Mass Transfer
The overall mass transfer measurements were carried out in a vertical tubular reactor with co-current upflow. The experiments were carried out according to information taken from
Arrhenius plots in the region reflecting external mass transfer limitation. Therefore the experiments were carried out isothermally at 120 °C. Different combinations of gas and liquid flow rates were investigated for 30 and 45 PPI foam catalysts. Eq. 12
The observed reaction rate was calculated using the Eq. 12, per reactor volume. To determine the reaction rate constant,
stands for the mass of Pd
, the observed reaction rate, must
be determined. Knowledge of the solubility of hydrogen in the AMS and cumene mixture at different temperatures is needed to calculate the reaction rate constant from the observed reaction rate by using the solubility data from Herskowitz [15]. The observed reaction rate was measured at different gas and liquid flow rates from the parameter window as depicted in Fig. 6. The gas and liquid flow rates were chosen so that a flow regime in the range of bubbly or slug flow could be obtained. At higher gas flow rates than the range given in Fig. 6, the difference pressure drop decreases drastically and the flow tends toward slug flow as obtained results from preliminary hydrodynamic measurements. In the following diagrams, the observed reaction rate per a) 30 PPI 0.05 10ml/min r' (mol s-1gPd-1)
20ml/min 30ml/min 40ml/min
0.025
0 0
5
10
b) 4515 PPI uG (mm s-1)
20
25
30
r' (mols-1gPd-1)
0.1 0.075 0.05
10ml/min 20ml/min
0.025
30ml/min 40ml/min
0 0
5
10
15 uG (mm s-1)
20
25
30
Fig. 6: Observed reaction rate for the parameter window at 120 °C and atmospheric hydrogen pressure a) 30 PPI b) 45 PPI
Pd amount foam catalysts at different gas flow rates is depicted. As was shown by Lali et al. [10], the geometric surface area of the 45 PPI foams is much higher than the surface area of the 30 PPI foam, whereas the void volume for both pore densities are the same (94 vol.%). Based on this, fluid flow faces a denser network of struts in a 45 PPI foam, which improves the mixing via mass transfer by imposing local turbulences. Another explanation centers on internal mass transfer in the catalytic layer. The thickness of the 30 PPI washcoat was 17.8 µm, whereas the washcoat thickness of the 45 PPI foam catalysts was about 6 µm. The smaller layer thickness of the 45 PPI foam led to a lower internal diffusion resistance. The reason for different washcoat thicknesses for 30 PPI and 45 PPI is that the foams were coated with one layer of alumina. The single layer catalytic coating was reported to be more active than double or triple layer coatings [9]. It also produces a more reproducible thickness distribution because the risk of washcoat clogging by a further coating was eliminated. Single layer coating is also advantageous because coating quality for each of 25 foam samples for a packing can be easily controlled. It was also observed that liquid flow rate influences reaction rate by more than one mechanism (changing the residence time). The effect of liquid flow rate on the reaction rate for a specific gas flow rate can be investigated by applying Eq. 12. Fig. 7.a and Fig. 7.b show that 45 PPI foam catalysts showed a reaction rate increase that increased with liquid flow rate, but that for 30 PPI foam catalysts there was no noticeable impact. The increase of reaction rate with liquid flow rate can be related to the structural properties of the foam catalysts. Specifically, local turbulences are formed that improve the liquid solid mass transfer. The overall mass transfer coefficient for the 30 PPI and 45 PPI foams was calculated using model equations Eq. 1through Eq. 3. The plots of Fig. 7.a and Fig. 7.b show the overall mass transfer for two liquid flow rates of 40 mL min-1 and 20 mL min-1. They also demonstrate clearly
that the charts follow a trend similar to the observed reaction rates, such that the overall mass transfer of the 45 PPI foams was closer to the values for 30 PPI foams. 10 9 8 7 6 5 4 3 2 1 0
14 30PPI 45PPI
10
30PPI 45PPI
kovaov (s-1)
kovaov (s-1)
12 8 6 4 2
a
0 0
10 uG (mm s-1)
20
30
b 0
5
10 uG (mm s-1)
15
20
Fig. 7: Comparison of overall mass transfer for 30 PPI and 45 PPI packings, 50 cm, 120 °C and atmospheric pressure at 40 mL min-1 (a) or 20 mL min-1(b).
The higher overall mass transfer for 45 PPI foams was due to higher energy dissipation in the 45 PPI foam catalyst packing. However, the difference is smaller than the observed reaction rate because of overall mass transfer. Since overall mass transfer describes the external mass transfer effects, the internal mass transfer effects were remarkable due to different catalytic layer thicknesses.
10
45 PPI foam catalysts (Present study)
µ-packed bed Losey et al. 2001
CR Haase et al. 2013 1
Tourvielle et al. 2014 Milli foam reactor
Kov (s-1)
Upflow foam reactor Stemmet et al. 2008 High interaction TBR Larachi et al. 2010
WCMR Kreutzer et al. 2001 0.1
Downflow foam reactor Stemmet et al. 2008 Low interaction TBR Larachi et al. 2010 0.01 0.04
0.4
4
40
400
4000
40000
Energy dissipation (W/m³) Fig. 8 Comparison of overall mass transfer and energy dissipation for various studies
Fig. 8 demonstrates the overall mass transfer values versus energy dissipation reported in the literature for various reactor types. The diagram indicates that overall mass transfers of 30 and 45 PPI foam catalysts in co-current upward flow are very good compared to conventional reactor types. These types include the stirred tank reactors studied by Schlüter et al. [23], the conventional trickle bed studied by Losey et al. [24], and the bubble column reactors studied by Linek et al. [25]. All studies mentioned in Fig. 8 employed the hydrogenation of AMS to characterize a specific reactor type. Therefore the values reflect comparative reactor performances. Losey et al. [24], in his study of packed microstructures, showed that the overall mass transfer for hydrogenation of
cyclohexene in a micro packed bed was very high (up to 15 s-1) for liquid flow rates ranging from 15 to 55 mg min-1and gas flow rates between 5 to 10 cm³ min-1. Losey et al. also reported a dynamic pressure drop below 1 bar cm-1, whereas the dynamic pressure drop of a two phase flow in the investigated parameter range was below 2 mbar m-1 for 30 PPI foams and below 3 mbar m-1for 45 PPI foams resulting in energy dissipations below 1 Wm-3. Thus, the power consumption to obtain a comparable overall mass transfer was significantly less for the foam systems. The fact that a high overall mass transfer can be realized using foam catalysts with a much lower pressure drop than other systems is a significant advantage compared to other reactor types. Furthermore, investment costs such as apparatus material and catalyst are less in a catalytically active foam packing with high overall mass transfer by maintaining the same productivity [5]. Moreover, operational costs also drop due to lower power consumption because of low energy dissipation in catalytically active foam packings.
4. Summary and Conclusions This study showed that the removal of TBC from AMS by applying vacuum distillation was quite effective in increasing durability of the catalyst. After 7 hours a small decrease in the catalyst activity was observed, but activity remained constant after that for the next 50 hours. The 30 PPI foam catalysts were first tested in a polytropical operation mode for hydrogenation of AMS to cumene. A 50 cm foam catalyst packing with a diameter of 18 mm was investigated. The productivity and space time yield was significantly higher than the values reported in the literature for a conventional trickle bed [18]. The overall mass transfer coefficient for a 30 PPI foam catalyst was compared to the coefficient for a 45 PPI foam catalyst. The parameter window of the liquid and gas flow rates was set by the flow regimes and for the two phase pressure drop measurements. In the parameter window the overall mass transfers ranging between 0.2 s-1 and 14 s-1 were measured for 30 PPI foam catalysts and values ranging between 0.3 s-1 to 18 s-1 for 45 PPI. Compared to the literature, the foam catalysts had a high overall mass transfer with a very low pressure drop (2-3 mbar m-1).
Acknowledgement The Authors gratefully acknowledge Prof. Dr.-Ing. habil. Rüdiger Lange for his scientific and technical guidance during our work and for giving us the opportunity to accomplish the present study. Furthermore, we would like to thank Dr. Volker Schmidt (TEC++ GmbH) for providing a flow-through refractometer for preliminary hydrodynamic measurements.
Nomenclature aGL
[m²] gas liquid interfacial area
aLS
[m²] liquid solid interfacial area
aov
[m²] gas liquid solid interfacial area
cj
[mol m-3] concentration of species j
EA,obs
[kJ mol-1 K-1] observed activation energy
EA,int
[kJ mol-1 K-1] intrinsic activation energy
EA,diff
[kJ mol-1 K-1] diffusion activation energy (heat of adsorption)
H
[-] Henry coefficient
kL
[m² s-1] liquid side mass transfer coefficient
kLS
[m² s-1] liquid-solid mass transfer coefficient
kint
[s-1] intrinsic reaction rate of AMS hydrogenation
kobs
[s-1] observed reaction rate of AMS hydrogenation
nH
[mol] moles of reacted hydrogen
ptot
[Pa] total reactor pressure
rH
[mol m-3 s-1] observed reaction rate
TR
[K] reactor temperature
VH
[m³] volume of hydrogen in reactor
VR
[m³] reactor volume
Greek letters
εΔG
[-] gas holdup for reacted hydrogen
η
[-] effectiveness factor
ρPd
[g m-3] mass of Pd per reactor volume
τ
[s] residence time
References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18.
Turek, F. and R. Lange, Mass transfer in trickle-bed reactors at low reynolds number. Chemical Engineering Science, 1981. 36(3): p. 569-579. Cini, P. and M.P. Harold, Experimental study of the tubular multiphase catalyst. AIChE Journal, 1991. 37(7): p. 997-1008. Kreutzer, M.T., et al., Mass transfer characteristics of three-phase monolith reactors. Chemical Engineering Science, 2001. 56(21-22): p. 6015-6023. Purnama, H., et al., α-methylstyrene hydrogenation in a flow-through membrane reactor. AIChE Journal, 2006. 52(8): p. 2805-2811. Haase, S., Minichannel flow reactors for gas liquid solid reactions, in Professur für chemische Verfahrenstechnik und Anlagentechnik. 2012, Technische Universität Dresden: Dresden. Langsch, R., Miniaturised packed bed reactors for gas liquid solid reactions, in Professur für chemische Verfahrrenstechnik und Anlagentechnik. 2014, Technische Universität Dresden. Tourvieille, J.N., R. Philippe, and C. de Bellefon, Milli-channel with metal foams under an applied gas-liquid periodic flow: External mass transfer performance and pressure drop. Chemical Engineering Journal, 2015. 267: p. 332-346. Lali, F., et al., Preparation and characterization of Pd/Al2O3 catalysts on aluminum foam supports for multiphase hydrogenation reactions in rotating foam reactors. Chemical Engineering Research & Design, 2015. 94: p. 365-374. Leon, M.A., et al., Rotating Foam Stirrer Reactor: Effect of Catalyst Coating Characteristics on Reactor Performance. Industrial & Engineering Chemistry Research, 2011. 50(6): p. 3184-3193. Jakobitz, C., Reaktionskinetische Untersuchungen zur Hydrierung von α-Methylstyrol an oberflächenmodifizierten Pd/γ-Al2O3-Trägerkatalysatoren, 2009, TU Dresden. Meille, V., C. de Bellefon, and D. Schweich, Kinetics of α-Methylstyrene Hydrogenation on Pd/Al2O3. Industrial & Engineering Chemistry Research, 2002. 41(7): p. 1711-1715. Herskowitz, M., Morita, S., and J.M. Smith, Solubility of Hydrogen in alpha-Methylstyrene. Chemical engineering, 1978. 23: p. 474-477. Löwe, R. and A. Busch, Beitrag zur experimentellen Untersuchung Katalytischer Dreiphasenreaktionen unter besonderer Berücksichtigung ausgewählter Teilprozesse. 1977, Technische Hochschule Merseburg. Stefoglo, E.F., Midoux, N., and J.C. Charpentier, Mesurede la concentration de l'hydrogene libre dissous dans le liquide lors de l'hydrogenation de l'alpha methylstyrene sur palladium supporte en suspension. Hungarian journal of industrial chemistry veszprem, 1980. 8: p. 243-256. Herskowitz, M., Wisniak, J., and L. Skladman, Hydrogen Solubility in Organic Liquids. Chemical engineering, 1983. 28: p. 164-166. Böttcher, G., Kontinuierliche Hydrierreaktion im Rohrreaktor mit katalytisch aktiver (Pd/Al2O3) Schaumpackung, 2012, TU Dresden. Lali, F., Catalytic active foams for gas liquid solid reaction. 2015, Technische Universität Dresden: Dresden. Schubert, M., Performance enhancement of trickle bed reaktors. 2007, TU Dresden.
19. 20. 21. 22. 23. 24. 25.
Babcock, B.D., G.T. Mejdell, and O.A. Hougen, Catalyzed Gas-Liquid Reactions in Trickling-Bed Reactors .1.Hydrogenation of a-Methylstyrene Catalyzed by Palladium. Aiche Journal, 1957. 3(3): p. 366-369. Cini, P. and M.P. Harold, Experimental-Study of the Tubular Multiphase Catalyst. Aiche Journal, 1991. 37(7): p. 997-1008. Cini, P., et al., Preparation and Characterization of Modified Tubular Ceramic Membranes for Use as Catalyst Supports. Journal of Membrane Science, 1991. 55(1-2): p. 199-225. Satterfi.Cn, A.A. Pelossof, and T.K. Sherwood, Mass Transfer Limitations in a Trickle-Bed Reactor. Aiche Journal, 1969. 15(2): p. 226-&. Schluter, V. and W.D. Deckwer, Gas-Liquid Mass-Transfer in Stirred Vessels. Chemical Engineering Science, 1992. 47(9-11): p. 2357-2362. Losey, M.W., M.A. Schmidt, and K.F. Jensen, Microfabricated multiphase packed-bed reactors: Characterization of mass transfer and reactions. Industrial & Engineering Chemistry Research, 2001. 40(12): p. 2555-2562. Linek, V., M. Kordac, and T. Moucha, Mechanism of mass transfer from bubbles in dispersions Part II: Mass transfer coefficients in stirred gas-liquid reactor and bubble column. Chemical Engineering and Processing, 2005. 44(1): p. 121-130.
S0255-2701(16)30087-3 http://dx.doi.org/doi:10.1016/j.cep.2016.04.001 CEP 6784
To appear in:
Chemical Engineering and Processing
Received date: Accepted date:
18-3-2016 5-4-2016
Please cite this article as: Farzad Lali, Characterization of foam catalysts as packing for tubular reactors, Chemical Engineering and Processing http://dx.doi.org/10.1016/j.cep.2016.04.001 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
*Manuscript
Characterization of foam catalysts as packing for tubular reactors Farzad Lali Institute of chemical process fundamentals of the Czech academy of sciences, Rozvojova 2/135 165 02 Prague, Czech republic, phone: +420 220 390 233, fax: 220920661,220920649, E-Mail: [email protected]
Abstract This study deals with the characterization of catalytically active foams as packing for continuous tubular reactors in terms of reactor performance and overall mass transfer. The tubular reactor was operated in a vertical direction such that gas and liquid entered the reactor in co-current upward flow. The hydrogenation reaction of α-methylstyrene was applied for characterization of the Pd/Al2O3 foam catalysts with a bed length of 50 cm and a diameter of 1.8 cm. Two pore densities for foams namely 30 and 45 PPI were investigated. The foam catalyst showed an overall mass transfer rate that ranged between 0.2 and 14 s-1 for 30 PPI and 0.3 to 18 s-1 for 45 PPI foam catalyst, whereas the energy dissipation was below 1 Wm-3. Keywords: overall mass transfer, foam catalyst, tubular reactor, hydrogenation
1. Introduction The chemical industry is a major consumer of energy worldwide and yet the majority of chemical processes involve catalytic reactors. Therefore, the energy efficiency of chemical reactors is a very crucial issue. For the heterogeneously catalyzed multiphase reactions, several reactor technologies have been already applied such as bubble columns, trickle bed reactors, CSTRs or there are technologies under development for instance wall-coated monolith reactors, monolith reactors with dumped catalysts and last but not least microreactors. The application of bubble columns and CSTRs involves suspended catalysts that demand a subsequent separation of catalyst particles. The pelleted catalysts are usually applied as random packed bed in trickle bed reactors. Yet the disadvantages of the random packed beds are high pressure drop and formation of hot-spots. Microreactor is an innovative reactor concept but the fabrication of microractors is associated with high production costs and the pressure drop inside such reactors is relatively high. Usually a trickle
regime is favored in conventional reactor design for continuous three phase reactions, where the gas side mass transfer is limiting due to low gas-liquid solubility. The advantage of the catalytically active foam packings is that, because of the high voidage (up to 97%) and low pressure drop, it is possible to run three phase reactions in an upward flow direction. A challenging issue of the design of trickle bed reactors is dealing with “hot spots” in the reactor, that causes a non-uniform temperature profile. Such hot spots are known to be undesired and to cause side products and shorten the life of the catalyst. Application of catalytically active foam packings in a co-current upward direction can increase the uniformity of the temperature profile due to very low energy dissipation, high void volume, and increased liquid holdup. Measurement of overall volumetric mass transfer can be utilized as a criterion to evaluate the performance of a structured packing. Several studies used the hydrogenation of AMS using Pd/Al2O3 catalysts to evaluate the performance of various reactor types. Turek and Lange [1] investigated a trickle bed reactor with an inner diameter of 3.4 cm with liquid flow rates ranging between 0 and 1.5 L h-1 and gas flow rates from 0 to 100 L h-1. Cini et al. [2] adopted the hydrogenation of AMS for tubular supported and Pd/γ-Al2O3 impregnated ceramic membranes. Kreutzer et al. [3] used this hydrogenation reaction to characterize the performance of a tubular reactor with wall-coated monolith packing. Purnama et al. [4] investigated the performance of a flow-through ceramic membrane reactor at temperatures ranging between 45-50 °C and pressures between 1-40 bar, and compared the performance of this reactor with such other reactor types as trickle bed and bubble column reactors. Haase [5] characterized wall-coated and composite (packed with pellets) minichannels for gas-liquid-solid-reactions in various multiphase flow regimes. Langsch [6] studied mass transfer in miniaturized packed bed reactors at different multiphase flow regimes using the hydrogenation of AMS to cumene. Tourvieille et al. [7] investigated minichannels packed with foam catalysts by applying a pulsing flow regime and obtained overall mass transfer values that ranged from 0.25 s-1 to 1.9 s-1. The present study deals firstly with characterization of the reactor performance foam catalysts as tubular packings. Subsequently, the overall mass transfer of 30 PPI and 45 PPI foams were measured and compared with the data from literature. The overall mass transfer is referred to the external mass transfer which is mainly influenced by the foam structure, whereas the internal mass transfer in the washcoat also plays a role. Thus, internal mass transfer was considered in the mathematical model used to evaluate the observed reaction rates. Although there is some effect
from internal mass transfer, the influence of the internal mass transfer was although available however, the major concentration of the palladium was detected in the vicinity of the washcoat surface as shown in a previous paper [10] by applying EDX analysis. Therefore, the influence of the internal mass transfer for different washcoat thicknesses between 17.8 µm (for 30 PPI) and 6 µm (for 45 PPI foams) plays a less important role.
2. Experimental 2.1.1.
Catalyst preparation
The foam catalysts with pore densities of 30 PPI and 45 PPI were prepared using the methods in a previous publication by Lali et al. [8] and Leon et al. [9].
Fig. 1 : Pictures of foam samples after different preparation steps 0) pristine aluminum foam 1) anodized foam 2) coated with alumina 3) impregnated with palladium acetate 4) reduced Pd/Al2O3 foam catalyst 5) spent foam catalyst
The preparation steps of a 30 PPI foam catalyst are shown in Fig. 1, where the image number 0 represents an uncoated aluminum foam, number 1 was an anodized foam, number 2 was washcoated anodized foam with alumina slurry, number 3 was impregnated with palladium acetate and number 4 was reduced foam catalyst. The image 5 in Fig.1 shows a spent foam catalyst.
2.2 Pretreatment of α-Methylstyrene
The purities of the α-methylstyrene and cumene supplied by SIGMA-ALDRICH were 99% and 98%, respectively. In order to prevent the AMS from polymerizing, 15 ppm of the inhibitor tert-butylcatechol (TBC) was added by the supplier. Earlier experiments showed that the TBC can build residues on the catalyst and decrease the catalytic activity drastically. The authors have several views about the deactivation of Pd/Al2O3 catalyst by TBC. While some studies [10] confirm a negative impact of TBC on catalytic activity, Meille et al. [11] found neither a negative impact on a Pd/Al2O3 catalyst nor any reaction inhibition. Nevertheless, our own preliminary experiments confirm the inhibitory effect of TBC. The low concentration of Pd in the foam catalysts in the present study also increases the vulnerability of the catalyst to being inhibited. For these reasons, AMS was distilled before preparation of the AMS solution with cumene by using a rotary evaporator (BÜCHI R-114). The vacuum distillation of AMS was carried out at 50 mbar pressure and a bath temperature of 85 °C. The cumene was also distilled in order to reduce its impurities, which can also build residues on active catalytic sites. 2.3 Experimental Setup
The hydrogenation experiments were carried out in a tubular reactor with a diameter of 1.8 cm and a height of 50 cm. The gas and liquid streams were continuously introduced into the reactor from the bottom. The liquid (20% by weight in cumene) was pumped with an Alphachrom pump. The flow rate of the pump could be adjusted over a range of 1 to 200 mL min-1. The gas flow was regulated by a Bronkhorst F231M mass flow controller. The foam catalysts were prepared as cylinders with a height of 2 cm and a diameter of 1.8 cm. In order to avoid bypassing the fluids and channeling, the foam catalysts were wrapped in PTFE strips with a thickness of ~0.5 mm. The AMS/Cumene mixture was preheated to the desired temperature prior to entering the reactor by a Horst HAT MC1 heating cord. The reactor was jacketed, and its temperature was controlled by a Julabo H-2 heater thermostat. The temperature gradient through the reactor was measured by four thermocouples. The experiments for the measurement of the overall mass transfer were carried out at atmospheric pressure and the expermients for testing the catalytic performance were carried out at 10 bar. The pressure inside the reactor was monitored by a WIKA pressure transducer. The product stream from the reactor was cooled before taking samples by a LAUDA WK500 recirculating chiller. The cooling water flowed through a helical copper pipe with a diameter of 3
mm that was wrapped around the sampling vessel K-01 to cool the liquid sample to the desired temperature of about 30 °C. Fig. 2 is a system diagram for the experimental setup. For the experiments at elevated pressures (10 bar), the valve V-6 was replaced with a relief valve, and a pressure transducer was installed immediately after the reactor to monitor the pressure variations. The relief valve
Fig. 2 system diagram for reaction studies on a continuous reactor with catalytic foam packing
(Swagelok) was used to regulate the desired pressure manually by means of a spring with a specific hardness for pressures ranging between 3.4 and 24.1 bar. It should be noted that the use of a spring-loaded relief valve is not the most accurate way to regulate pressure. Consequently, some pressure fluctuations were observed. Furthermore, the reaction studies at an elevated pressure of 10 bar were carried out polytropically, in order to study the temperature effects along the fixed bed of foam catalysts. The mass transfer measurements were based on the experiments at atmospheric pressure to avoid inaccuracy.
For taking samples, the valve V-07 was briefly opened to drain the liquid sample for refractive analysis. The temperature was monitored using type-K thermocouples at 5 spots along the reactor. The specifications of the foam catalysts, which were used in this study are listed in table 1.
Table 1: Specifications of the catalytically active foam packings used for reaction studies. Pore density Pd-loading based on washcoat mass (%) Washcoat mass (g) Catalyst concentration per reactor volume (gPd m-³react) Washcoats thickness (μm)
30 PPI
45PPI
2.082
1.881
2.5854
2.3435
575
372.3
17.8
6.63
1321.6
2892.8
25x2
33x1.5
0.0835
0.1013
Geometrical surface area per packing volume -1
(m ) Packing length (cm) Solid Holdup with Washcoat (m³S m-³)
2.4 Mass Transfer Model In this work the overall mass transfer coefficient was determined by means of a hydrogenation reaction of α-methylstyrene. The determination of overall reaction coefficient was done based on the film theory. For a first order reaction the relationship between reaction rate and mass transfer resistances are shown in Eq. 1. 1 1
1
1
1
∙
,
Eq. 1
It was assumed that the mass transfer resistance on the liquid side is much larger than the gas side. For this reason the gas side mass transfer resistance was negligible as shown in Eq. 2 . 1
1
1
Eq. 2
The first term in Eq. 2 represents the liquid side gas-liquid mass transfer resistance, the second term is the liquid solid mass transfer, and the third term represents the intrinsic kinetic together with the diffusion limitation in the catalyst support. The overall mass transfer coefficient, as shown in Eq. 3 , is the sum of the liquid side gas-liquid and liquid solid mass transfer resistances. 1
1
1
Eq. 3
This mass transfer model is valid for a reactor with almost plug flow, which was the case for the foam catalyst packing according to the obtained results in the preliminary hydrodynamic experiments, in which the Bodenstein number was much greater than 7. 2.5 Hydrogen Solubility and Intrinsic Kinetics The mass transfer model presented in the previous section shows that the solubility of the hydrogen in the AMS/cumene mixture plays an important role in evaluating the results from the observed reaction rate using Eq. 1. Several studies in the literature have dealt with the solubility of hydrogen in the AMS/cumene mixture. These studies are carried out for different reaction conditions. Herskowitz [12] correlated the solubility of hydrogen in AMS at temperatures ranging between 15 and 74 °C and ambient pressure with a maximum deviation of 3%. Löwe et al. [13] developed correlations for the hydrogen solubility in pure AMS and cumene. Stefoglo [14] determined empirically the Henry coefficient for hydrogen in AMS or cumene at temperatures ranging between 20 and 80 °C and pressures up to 5 bar. Herskowitz [15], in a later paper, investigated a range of temperatures from 30 to 100 °C and pressures from 1 to 69 bar.The solubilities corresponding to Löwe et al. [13] and Herskowitz [15] are relatively similar, whereas the correlation proposed by Meille et al. [11] exhibits the largest deviation. This discrepancy can be explained by different underlying assumptions. While, Löwe et al. [13] Stefoglo [14] and Herskowitz assumed different solubilities for hydrogen in AMS and cumene, Meille et al. [11] assumed that the solubility of hydrogen in AMS and cumene is approximately equal. This assumption allowed Meille to describe the
solubility of hydrogen in both liquids by applying the correlation proposed in Herskowitz’s first publication [12]. For evaluation of the observed reaction rate and calculation of the overall volumetric mass transfer, the correlation proposed by Herskowitz [15] in his later publication in 1983, who investigated among other conditions also the solubility of hydrogen in AMS/cumene mixture at 100 °C and ambient pressure was favored. The intrinsic kinetic for the hydrogenation reaction of AMS to cumene using a Pd/Al2O3 catalyst was taken from Meille et al. [11].
3. Results and Discussion
3.1 Catalyst Durability Catalyst durability can be affected by chemical residues blocking on the active sites. Therefore the activity of the foam catalysts with a pore density of 45 PPI were monitored over a period of 8 hours. The experiments were carried out at a liquid (20 wt. % AMS in cumene) flow rate of 10 ml min-1 , and a gas flow rate (hydrogen) of 0.05 L min-1. At these flow rates the liquid residence time was relatively long. The hydrogenation reaction was carried out at ambient pressure and 120 °C. The catalyst activity was monitored by measuring the conversion of AMS to cumene at 30 minutes intervals. A slight loss of activity was observed after 7 hours of time on stream, but thereafter, catalytic activity remained almost constant for the next 50 hours. 3.2 Experiments at Elevated Pressure The aim of the polytropical hydrogenation experiments at 10 bar was to compare the results with the available literature in terms of space time yield and productivity. The temperature profile in the reactor was also investigated. For the reaction studies at 10 bar and a liquid inlet temperature of 40 °C, the liquid flow rates were 10 and 20 mL min-1 respectively. Gas flow ranged from 0.1 to 1.2 L min-1.
80
0.9
70
0.8
60
0.7 0.6
50
0.5
40
0.4
30
0.3
temperature (°C)
Coonversion (-)
1.0
20
0.2
10
0.1 0.0
0
0
2
4
6
8
10
uG (mm s-1)
Fig. 3: outlet temperature (○) and total conversion (♦) in the reactor for 30 PPI foam catalysts for a packing length of 50 cm, Liquid flow rate of 10 mL min-1, 10 bar H2, polytropical conditions and an inlet temperature of 40 °C [16]
0,1
0,01
-1
-3
0,1
0,01
1E-3
1E-3
1E-4
1E-4 stirred tank with pellets
tubular reactor with foam catalysts
-3
10
1
CPd = 1557 g m
10 bar
-3
45.5 °C
activity space time yield
40 °C,10 bar CPd = 180 g m
-3
1
CPd = 590 g m
Pd
10
40 °C,10 bar 20
space time yield (mol s m )
100
100
trickle bed reactor
Fig. 4: Comparison of the reactor performance of continuous tubular reactor with foam catalysts with conventional reactors; Stirred tank with a fine slurry catalyst (T=45.5 °C, P=1.51 bar) Meille et al.[11]; tubular reactor with catalytically active foam packing in co-current upward flow (40 °C, 10 bar,VL=20 mL min-1) F. Lali [17];trickle bed reactor (40 °C, 10 bar, VL=25 mL min-1) Schubert [18]
Fig. 3 shows that the outlet temperature and total conversion increase as the gas flow rate increases due to the polytropical nature of the process. A comparison with a conventional trickle
bed reactor at 40 °C and liquid flow rate of 25 mL min-1 in Fig. 4, shows the catalyst activity and space time yield of foam catalysts for the same conditions are 10 times larger and yet the amount of Pd was 90% less and the hydrogen gas demand was 30% less. The stirred tank study in Fig. 4 from reference [11] was carried out at nearly intrinsic reaction conditions. 3.3 Pseudo-Arrhenius Plots To ensure that the mass transfer measurements were carried out in the mass transfer limited region, Pseudo-Arrhenius plots were made showing the temperature dependence of “observed” reaction coefficients for each flow rate. The observed reaction rate coefficient was calculated by applying Eq. 4. For this reason, external mass transfer effects were not ruled out. The observed (apparent) activation energy can be calculated from the plot from regions that were distinguished by different slopes as shown in Fig. 5. For the reaction studies a parameter window on the basis of the preliminary studies particularly, two phase pressure drop that was consisting of liquid flow rates ranging from 10 mL min-1 to 40 mL min-1 and the highest gas rates for every liquid rate were chosen slightly higher than the values obtained in the preliminary study for air-water system in order to the differences between the physical properties of both systems such as viscosity and surface tension.
Eq. 4
∗
2 Ea=8.56 kJ mol-1
1.8
L=20 mL/min G=0,23 G=0.23mL/min L min-1 L=30 mL/min G=0,35 G=0.35mL/min L min-1
lnkobs (s-1)
1.6
L=40 mL/min G=0,35 G=0.35mL/min L min-1
1.4 1.2
Ea=26.47 kJ mol-1
1 0.8 0.6 0.4 0.0025
0.0026
0.0027
0.0028
0.0029
0.003
0.0031
1/T (K-1)
Fig. 5: Pseudo-Arrhenius plot for 30 PPI foam packing 50 cm, atmospheric hydrogen pressure
The foam catalysts of 30 and 45 PPI pore density were investigated at temperatures ranging from 60 to 120° C at ambient pressure. The purpose of measuring at the atmospheric pressure was to avoid any pressure fluctuations which could influence the GL-mass transfer at elevated pressures. The reaction rate constant is depicted against the reciprocal temperature in Fig. 5. The determination of the kinetically limited region focused on the higher gas flow rates from the parameter window, at which a kinetic limitation was more likely due to the increasing gas holdup. Fig. 5 shows how the observed reaction rate increases with temperature in a pseudo-Arrhenius plot. The region between 333 and 393 K for the chosen parameter window was most likely to be kinetically limited. After this region, the effect of external mass transfer limitations was investigated. On the basis of these results a reaction temperature of 120 °C was chosen to rule out kinetic limitations. The values of the intrinsic activation energy in the literature range from 19.7 kJ mol-1 (Babcock et al.[19]) to 43.5 kJ mol-1 (Cini et al.[20, 21]). Cini et al. [20] also gives an overview of the results reported in the literature. Moreover, provide an extensive discussion of the various limitation factors, such as pore diffusion, that can influence apparent activation energy. The value reported by Cini et al. [21] was considered the most precise value for the intrinsic activation energy of AMS hydrogenation using Pd/Al2O3 catalyst. The majority of later studies also reported an intrinsic activation energy of about 40 kJ mol-1. Meille et al. [11], for instance, reported 38.7 kJ mol-1. The apparent activation energy for the flow rates in Fig. 6 was about 26 kJ mol-1.The discrepancy compared to the intrinsic activation energy can be elucidated by the reaction limitation by pore diffusion in the catalytic layer as Eq. 5 demonstrates:
,
,
,
Eq. 5
2
The apparent activation energy was smaller than the reported intrinsic activation energy and remarkably higher than the activation energy for the hydrogen diffusion in 20 wt. % AMS (EA,diff =12.5 kJ mol-1 [6]). By applying the Eq. 5 proposed by Satterfield et al. [22] for the pore diffusion limited area, an intrinsic activation energy of 40 kJ mol-1, which is in good agreement with the data reported in the literature. The majority of other authors who published after the work by Cini et al. [2] was published, reported values close to 40 kJ min-1.
The intrinsic kinetics of the hydrogenation reaction of AMS was calculated based on the reaction studies in the rotating foam reactor. These results were in good agreement with the intrinsic kinetics reported by Meille et al. [11]. As shown in the Fig. 5, the observed reaction rate coefficient depends on residence time in the reactor. For this reason it is crucial to calculate residence time. To obtain a more accurate estimate of residence times in the reactor, the following method was adopted. The residence time was calculated based on the amount of the AMS converted to cumene
. It was assumed that
a certain amount of liquid containing AMS, at steady state conditions, remains in the reactor as long as the hydrogen needed to convert
is consumed.
Eq. 6
1
Eq. 7
Eq. 8
Eq. 9
Eq. 10
Eq. 11
The gas holdup for the hydrogen in the reaction εΔG was calculated from Eq. 7 while the volume of the hydrogen in the reactor was calculated using Eq. 6. With the known amount of hydrogen in the reactor at steady state conditions, the moles of hydrogen consumed in the reactor were calculated from Eq. 9. The subsequent residence time was calculated as the ratio of the moles of hydrogen consumed to the molar flow of hydrogen calculated from Eq. 11. 3.4 Measurement of Overall Mass Transfer
The overall mass transfer measurements were carried out in a vertical tubular reactor with co-current upflow. The experiments were carried out according to information taken from
Arrhenius plots in the region reflecting external mass transfer limitation. Therefore the experiments were carried out isothermally at 120 °C. Different combinations of gas and liquid flow rates were investigated for 30 and 45 PPI foam catalysts. Eq. 12
The observed reaction rate was calculated using the Eq. 12, per reactor volume. To determine the reaction rate constant,
stands for the mass of Pd
, the observed reaction rate, must
be determined. Knowledge of the solubility of hydrogen in the AMS and cumene mixture at different temperatures is needed to calculate the reaction rate constant from the observed reaction rate by using the solubility data from Herskowitz [15]. The observed reaction rate was measured at different gas and liquid flow rates from the parameter window as depicted in Fig. 6. The gas and liquid flow rates were chosen so that a flow regime in the range of bubbly or slug flow could be obtained. At higher gas flow rates than the range given in Fig. 6, the difference pressure drop decreases drastically and the flow tends toward slug flow as obtained results from preliminary hydrodynamic measurements. In the following diagrams, the observed reaction rate per a) 30 PPI 0.05 10ml/min r' (mol s-1gPd-1)
20ml/min 30ml/min 40ml/min
0.025
0 0
5
10
b) 4515 PPI uG (mm s-1)
20
25
30
r' (mols-1gPd-1)
0.1 0.075 0.05
10ml/min 20ml/min
0.025
30ml/min 40ml/min
0 0
5
10
15 uG (mm s-1)
20
25
30
Fig. 6: Observed reaction rate for the parameter window at 120 °C and atmospheric hydrogen pressure a) 30 PPI b) 45 PPI
Pd amount foam catalysts at different gas flow rates is depicted. As was shown by Lali et al. [10], the geometric surface area of the 45 PPI foams is much higher than the surface area of the 30 PPI foam, whereas the void volume for both pore densities are the same (94 vol.%). Based on this, fluid flow faces a denser network of struts in a 45 PPI foam, which improves the mixing via mass transfer by imposing local turbulences. Another explanation centers on internal mass transfer in the catalytic layer. The thickness of the 30 PPI washcoat was 17.8 µm, whereas the washcoat thickness of the 45 PPI foam catalysts was about 6 µm. The smaller layer thickness of the 45 PPI foam led to a lower internal diffusion resistance. The reason for different washcoat thicknesses for 30 PPI and 45 PPI is that the foams were coated with one layer of alumina. The single layer catalytic coating was reported to be more active than double or triple layer coatings [9]. It also produces a more reproducible thickness distribution because the risk of washcoat clogging by a further coating was eliminated. Single layer coating is also advantageous because coating quality for each of 25 foam samples for a packing can be easily controlled. It was also observed that liquid flow rate influences reaction rate by more than one mechanism (changing the residence time). The effect of liquid flow rate on the reaction rate for a specific gas flow rate can be investigated by applying Eq. 12. Fig. 7.a and Fig. 7.b show that 45 PPI foam catalysts showed a reaction rate increase that increased with liquid flow rate, but that for 30 PPI foam catalysts there was no noticeable impact. The increase of reaction rate with liquid flow rate can be related to the structural properties of the foam catalysts. Specifically, local turbulences are formed that improve the liquid solid mass transfer. The overall mass transfer coefficient for the 30 PPI and 45 PPI foams was calculated using model equations Eq. 1through Eq. 3. The plots of Fig. 7.a and Fig. 7.b show the overall mass transfer for two liquid flow rates of 40 mL min-1 and 20 mL min-1. They also demonstrate clearly
that the charts follow a trend similar to the observed reaction rates, such that the overall mass transfer of the 45 PPI foams was closer to the values for 30 PPI foams. 10 9 8 7 6 5 4 3 2 1 0
14 30PPI 45PPI
10
30PPI 45PPI
kovaov (s-1)
kovaov (s-1)
12 8 6 4 2
a
0 0
10 uG (mm s-1)
20
30
b 0
5
10 uG (mm s-1)
15
20
Fig. 7: Comparison of overall mass transfer for 30 PPI and 45 PPI packings, 50 cm, 120 °C and atmospheric pressure at 40 mL min-1 (a) or 20 mL min-1(b).
The higher overall mass transfer for 45 PPI foams was due to higher energy dissipation in the 45 PPI foam catalyst packing. However, the difference is smaller than the observed reaction rate because of overall mass transfer. Since overall mass transfer describes the external mass transfer effects, the internal mass transfer effects were remarkable due to different catalytic layer thicknesses.
10
45 PPI foam catalysts (Present study)
µ-packed bed Losey et al. 2001
CR Haase et al. 2013 1
Tourvielle et al. 2014 Milli foam reactor
Kov (s-1)
Upflow foam reactor Stemmet et al. 2008 High interaction TBR Larachi et al. 2010
WCMR Kreutzer et al. 2001 0.1
Downflow foam reactor Stemmet et al. 2008 Low interaction TBR Larachi et al. 2010 0.01 0.04
0.4
4
40
400
4000
40000
Energy dissipation (W/m³) Fig. 8 Comparison of overall mass transfer and energy dissipation for various studies
Fig. 8 demonstrates the overall mass transfer values versus energy dissipation reported in the literature for various reactor types. The diagram indicates that overall mass transfers of 30 and 45 PPI foam catalysts in co-current upward flow are very good compared to conventional reactor types. These types include the stirred tank reactors studied by Schlüter et al. [23], the conventional trickle bed studied by Losey et al. [24], and the bubble column reactors studied by Linek et al. [25]. All studies mentioned in Fig. 8 employed the hydrogenation of AMS to characterize a specific reactor type. Therefore the values reflect comparative reactor performances. Losey et al. [24], in his study of packed microstructures, showed that the overall mass transfer for hydrogenation of
cyclohexene in a micro packed bed was very high (up to 15 s-1) for liquid flow rates ranging from 15 to 55 mg min-1and gas flow rates between 5 to 10 cm³ min-1. Losey et al. also reported a dynamic pressure drop below 1 bar cm-1, whereas the dynamic pressure drop of a two phase flow in the investigated parameter range was below 2 mbar m-1 for 30 PPI foams and below 3 mbar m-1for 45 PPI foams resulting in energy dissipations below 1 Wm-3. Thus, the power consumption to obtain a comparable overall mass transfer was significantly less for the foam systems. The fact that a high overall mass transfer can be realized using foam catalysts with a much lower pressure drop than other systems is a significant advantage compared to other reactor types. Furthermore, investment costs such as apparatus material and catalyst are less in a catalytically active foam packing with high overall mass transfer by maintaining the same productivity [5]. Moreover, operational costs also drop due to lower power consumption because of low energy dissipation in catalytically active foam packings.
4. Summary and Conclusions This study showed that the removal of TBC from AMS by applying vacuum distillation was quite effective in increasing durability of the catalyst. After 7 hours a small decrease in the catalyst activity was observed, but activity remained constant after that for the next 50 hours. The 30 PPI foam catalysts were first tested in a polytropical operation mode for hydrogenation of AMS to cumene. A 50 cm foam catalyst packing with a diameter of 18 mm was investigated. The productivity and space time yield was significantly higher than the values reported in the literature for a conventional trickle bed [18]. The overall mass transfer coefficient for a 30 PPI foam catalyst was compared to the coefficient for a 45 PPI foam catalyst. The parameter window of the liquid and gas flow rates was set by the flow regimes and for the two phase pressure drop measurements. In the parameter window the overall mass transfers ranging between 0.2 s-1 and 14 s-1 were measured for 30 PPI foam catalysts and values ranging between 0.3 s-1 to 18 s-1 for 45 PPI. Compared to the literature, the foam catalysts had a high overall mass transfer with a very low pressure drop (2-3 mbar m-1).
Acknowledgement The Authors gratefully acknowledge Prof. Dr.-Ing. habil. Rüdiger Lange for his scientific and technical guidance during our work and for giving us the opportunity to accomplish the present study. Furthermore, we would like to thank Dr. Volker Schmidt (TEC++ GmbH) for providing a flow-through refractometer for preliminary hydrodynamic measurements.
Nomenclature aGL
[m²] gas liquid interfacial area
aLS
[m²] liquid solid interfacial area
aov
[m²] gas liquid solid interfacial area
cj
[mol m-3] concentration of species j
EA,obs
[kJ mol-1 K-1] observed activation energy
EA,int
[kJ mol-1 K-1] intrinsic activation energy
EA,diff
[kJ mol-1 K-1] diffusion activation energy (heat of adsorption)
H
[-] Henry coefficient
kL
[m² s-1] liquid side mass transfer coefficient
kLS
[m² s-1] liquid-solid mass transfer coefficient
kint
[s-1] intrinsic reaction rate of AMS hydrogenation
kobs
[s-1] observed reaction rate of AMS hydrogenation
nH
[mol] moles of reacted hydrogen
ptot
[Pa] total reactor pressure
rH
[mol m-3 s-1] observed reaction rate
TR
[K] reactor temperature
VH
[m³] volume of hydrogen in reactor
VR
[m³] reactor volume
Greek letters
εΔG
[-] gas holdup for reacted hydrogen
η
[-] effectiveness factor
ρPd
[g m-3] mass of Pd per reactor volume
τ
[s] residence time
References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18.
Turek, F. and R. Lange, Mass transfer in trickle-bed reactors at low reynolds number. Chemical Engineering Science, 1981. 36(3): p. 569-579. Cini, P. and M.P. Harold, Experimental study of the tubular multiphase catalyst. AIChE Journal, 1991. 37(7): p. 997-1008. Kreutzer, M.T., et al., Mass transfer characteristics of three-phase monolith reactors. Chemical Engineering Science, 2001. 56(21-22): p. 6015-6023. Purnama, H., et al., α-methylstyrene hydrogenation in a flow-through membrane reactor. AIChE Journal, 2006. 52(8): p. 2805-2811. Haase, S., Minichannel flow reactors for gas liquid solid reactions, in Professur für chemische Verfahrenstechnik und Anlagentechnik. 2012, Technische Universität Dresden: Dresden. Langsch, R., Miniaturised packed bed reactors for gas liquid solid reactions, in Professur für chemische Verfahrrenstechnik und Anlagentechnik. 2014, Technische Universität Dresden. Tourvieille, J.N., R. Philippe, and C. de Bellefon, Milli-channel with metal foams under an applied gas-liquid periodic flow: External mass transfer performance and pressure drop. Chemical Engineering Journal, 2015. 267: p. 332-346. Lali, F., et al., Preparation and characterization of Pd/Al2O3 catalysts on aluminum foam supports for multiphase hydrogenation reactions in rotating foam reactors. Chemical Engineering Research & Design, 2015. 94: p. 365-374. Leon, M.A., et al., Rotating Foam Stirrer Reactor: Effect of Catalyst Coating Characteristics on Reactor Performance. Industrial & Engineering Chemistry Research, 2011. 50(6): p. 3184-3193. Jakobitz, C., Reaktionskinetische Untersuchungen zur Hydrierung von α-Methylstyrol an oberflächenmodifizierten Pd/γ-Al2O3-Trägerkatalysatoren, 2009, TU Dresden. Meille, V., C. de Bellefon, and D. Schweich, Kinetics of α-Methylstyrene Hydrogenation on Pd/Al2O3. Industrial & Engineering Chemistry Research, 2002. 41(7): p. 1711-1715. Herskowitz, M., Morita, S., and J.M. Smith, Solubility of Hydrogen in alpha-Methylstyrene. Chemical engineering, 1978. 23: p. 474-477. Löwe, R. and A. Busch, Beitrag zur experimentellen Untersuchung Katalytischer Dreiphasenreaktionen unter besonderer Berücksichtigung ausgewählter Teilprozesse. 1977, Technische Hochschule Merseburg. Stefoglo, E.F., Midoux, N., and J.C. Charpentier, Mesurede la concentration de l'hydrogene libre dissous dans le liquide lors de l'hydrogenation de l'alpha methylstyrene sur palladium supporte en suspension. Hungarian journal of industrial chemistry veszprem, 1980. 8: p. 243-256. Herskowitz, M., Wisniak, J., and L. Skladman, Hydrogen Solubility in Organic Liquids. Chemical engineering, 1983. 28: p. 164-166. Böttcher, G., Kontinuierliche Hydrierreaktion im Rohrreaktor mit katalytisch aktiver (Pd/Al2O3) Schaumpackung, 2012, TU Dresden. Lali, F., Catalytic active foams for gas liquid solid reaction. 2015, Technische Universität Dresden: Dresden. Schubert, M., Performance enhancement of trickle bed reaktors. 2007, TU Dresden.
19. 20. 21. 22. 23. 24. 25.
Babcock, B.D., G.T. Mejdell, and O.A. Hougen, Catalyzed Gas-Liquid Reactions in Trickling-Bed Reactors .1.Hydrogenation of a-Methylstyrene Catalyzed by Palladium. Aiche Journal, 1957. 3(3): p. 366-369. Cini, P. and M.P. Harold, Experimental-Study of the Tubular Multiphase Catalyst. Aiche Journal, 1991. 37(7): p. 997-1008. Cini, P., et al., Preparation and Characterization of Modified Tubular Ceramic Membranes for Use as Catalyst Supports. Journal of Membrane Science, 1991. 55(1-2): p. 199-225. Satterfi.Cn, A.A. Pelossof, and T.K. Sherwood, Mass Transfer Limitations in a Trickle-Bed Reactor. Aiche Journal, 1969. 15(2): p. 226-&. Schluter, V. and W.D. Deckwer, Gas-Liquid Mass-Transfer in Stirred Vessels. Chemical Engineering Science, 1992. 47(9-11): p. 2357-2362. Losey, M.W., M.A. Schmidt, and K.F. Jensen, Microfabricated multiphase packed-bed reactors: Characterization of mass transfer and reactions. Industrial & Engineering Chemistry Research, 2001. 40(12): p. 2555-2562. Linek, V., M. Kordac, and T. Moucha, Mechanism of mass transfer from bubbles in dispersions Part II: Mass transfer coefficients in stirred gas-liquid reactor and bubble column. Chemical Engineering and Processing, 2005. 44(1): p. 121-130.