C3H8. C2H6. Figure 11. Performance at 510 K of Pd (2 wt%) catalyst on a ...... about 12,000 m ozone concentrations become significant in polar regions. This.
CONTRIBUTION OF CATALYSIS TOWARDS THE REDUCTION OF ATMOSPHERIC AIR POLLUTION: CO2, CFCs, N2O, OZONE A.E. VAN DIEPEN, F. KAPTEIJN, M. MAKKEE and J.A. MOULIJN Section Industrial Catalysis, Department of Chemical Process Technology, Delft University of Technology, Julianalaan 136, 2628 BL, Delft, Netherlands In this article a number of problems concerning the environment are addressed. The enhanced greenhouse effect, which is believed to be responsible for global warming, is caused by the accumulation in the lower atmosphere of mainly carbon dioxide, methane, nitrous oxide, and chlorofluorohydrocarbons, emitted as a result of human activity. The latter two are also the main contributors to ozone depletion in the stratosphere. Reducing emissions of these gases is the main topic of this article. Solutions for the CO2 problem can be found in better energy efficiency, capture and storage of CO2, chemical conversion into useful products, and the use of non-fossil fuel alternatives such as biomass. Recovery and conversion of stored chlorofluorohydrocarbons into less harmful refrigerants is a viable option for tackling both the greenhouse and the ozone depletion problem. An integrated approach based on a kinetic/mechanistic study combined with process development is presented. N2O is present in off-gases from fluid-bed combustors. Reducing emissions from this source is problematic since this often leads to enhanced NOx emissions, while the concentration of N2 O in these off-gases is low. N2 O in the off-gas from adipic acid plants, on the other hand, is a relatively easy target because it is very much concentrated. This article briefly discusses the mechanistic aspects of N2 O destruction, together with some process opportunities. Other issues touched briefly are the formation of ozone at ground level, leading to smog, and its prevention, and the reduction of the ozone concentration in aircraft cabins through a catalytic process.
1.
Introduction
Many gases contribute to the greenhouse effect, the major ones being CO2, CFCs (chlorofluorocarbons), N2O, and CH4. CO2 is held responsible for about 55% of the greenhouse effect, caused by human activity. The other gases, although emitted in much smaller quantities, are also important (CFCs 17%, CH4 15%, N2O 6%) because they are much more harmful on a per kg basis. This is due to their longer lifetime in the atmosphere and their stronger greenhouse effect as expressed in terms of the “relative radioactive forcing” and the “global warming potential (GWP)”. Figs. 1 and 2 show the total quantities emitted and the global warming potentials of some important greenhouse gases. The GWP defines the time-integrated warming effect due to the instantaneous release of unit mass of a given gas in today’s atmosphere relative to that of CO2 (Xiaoding and Moulijn, 1996). Therefore, the contribution of each greenhouse gas to global warming may be evaluated as the product of GWP and the amount of gas emitted.
25000
10000
Currently yearly emitted
8000
20000
GWP
Individual Global Warming Effect (Million tons)
30000
15000
4000
Stored waste
10000
6000
2000
5000
1
6
0
CO2
N2O
0
CFC-11
CFC-12
Figure 1. Individual global warming effects of some greenhouse gases
CO2
N2O HFC-32 CFC-11 CFC-12
Figure 2. Global Warming Potential (GWP) of some greenhouse gases; GWP for CO2 = 1.
As can be seen from Fig. 1, CO2 is the most abundant by far, and a reduction of this gas in the atmosphere could contribute much to solving the global warming problem. However, CO2 is often present in diluted streams from for instance power plants, and therefore, its capture and subsequent disposal or conversion is difficult. For CO2 reduction the most promising options are a more efficient production and usage of energy, and the use of biomass, solar energy, or fuel cells as an alternative for fossil fuel combustion. The conversion of CO2 to useful chemicals and fuels can only deliver a limited contribution to CO2 mitigation, but may be advantageous in some special cases. N2O has only recently been identified as a greenhouse gas. Some chemical plants, such as adipic acid production plants produce concentrated N2O streams. This N2O can be converted to nitrogen and oxygen. Good results have been obtained. The conversion of the CFCs (CFC-11 (CCl3F) and CFC-12 (CCl2F2)) seems most promising with respect to reducing global warming. These CFCs can be converted to HFC-32 (CH2F2), which is much less harmful. Both N2O and CFCs are also major contributors to depletion of the ozone layer in the stratosphere. Stratospheric ozone constitutes a natural shield, protecting mankind from excess UV radiation. Therefore, reduction of N2O and CFC emissions would tackle two important environmental problems at once. As much as ozone is desired in the stratosphere, as badly it is undesired in other places. Issues addressed briefly are the problems of the formation of ozone at ground level, giving rise to smog, and the ozone problem in aircraft.
2.
Mitigation of CO2 by Chemical Conversion
CO2 is a natural constituent of the atmosphere. However, the CO2 content in the atmosphere is increasing as a result of human activities. Fig. 3 shows the worldwide mass balance of CO2 (Aresta, 1987). Atmosphere 720 Gt 100 Gt/y Oceans
60 Gt/y
Biological respiration
60 Gt/y
6-8 Gt/y
Photosynthesis (land and water)
Biomass
New fossil fuels
560 Gt
0.1 Gt/y
Human activities 5 Gt/y Fossil fuels 40000 Gt
Figure 3. Worldwide mass balance of CO2 (as carbon); inventories (in rectangles) and flow rates (net values, except quasi-equilibrium atmosphere ↔ ocean); 1 Gt = 109 tons. Adapted from (Aresta, 1987)
A higher atmospheric CO2 concentration promotes plant photosynthesis, while oceans, aquifers, etc. are a large sink for CO2, so not all the CO2 released by human activities accumulates. Still, the net man-made emission of CO2 to the atmosphere is about 6-8 Gt (1 Gt = 109 tons) carbon equivalents. 5 Gt of this amount is attributed to direct or indirect combustion of fossil fuels, and the remainder to land use conversions and deforestation (Xiaoding and Moulijn, 1996). In principle, three strategies can be used for the reduction of CO2 levels in the atmosphere, viz., reduction of the net amount of CO2 produced, storage of CO2, and usage of CO2. In the latter two cases, CO2 has to be captured and separated. 2.1. Reduction of the Amount of CO2 produced The amount of CO2 emitted as a result of combustion processes depends amongst others on the fuel used. Natural gas and oil have larger energy efficiencies than coal. Hence, replacement of coal by the less carbon-rich fossil fuels reduces CO2 emissions relatively easy. However, from a practical viewpoint, for obvious reasons it is wise policy to diversify the use of energy sources. Currently electricity is produced at 30-40% efficiency. With the best available technologies the efficiency could be increased to about 60%, thus halving the
amount of CO2 emitted (Kram and Okken, 1989). Fuel cells for electricity production (from electrochemical reaction between hydrogen and oxygen) have high efficiencies, ranging from 40 to 60%. The use of renewable energy sources such as biomass, solar energy, wind power, geothermal energy, etc. could significantly decrease the amount of CO2 emitted.
2.2. Capture and Separation of CO2 For CO2 from flue gases to be converted or disposed of, it first has to be captured and separated or at least concentrated. This considerably adds to the cost of the process. Naturally, separation costs depend on the source. For instance, in ammonia production, CO2 is available at relatively high concentration, whereas flue gas from coal-fired plants is very diluted. Furthermore, SOx and NOx are present in this flue gas, and these are often catalyst poisons in downstream chemical processes. Hence, separation is more costly for flue gases. Several separation technologies are available, viz., gas/liquid scrubbing systems, gas/solid adsorption systems, cryogenic techniques, and membrane separation. Absorption systems using chemical solvents (e.g., MEA, monoethanolamine) are most frequently used by far (Barer and Stern, 1988), but membrane techniques also seem promising, although the technology is still in the early stage of development (Xiaoding and Moulijn, 1996). CO2 can be removed from flue gas using polymeric gas separation membranes. However, the selectivity for CO2-N2 separation of commercially available membranes is generally small (10 – 35 %). Furthermore, as flue gas of current power generation plants is at atmospheric pressure, it has to be compressed in order to have sufficient driving force for the membrane separation process. The energy consumption for compression appears to be a severe limitation to the process (Feron, 1994). The generating efficiency of coal-fired power plants has been calculated to decrease from 35% to 9-18% at 80% CO2 removal, and from 40% to 31.2% at 61% CO2 removal (Feron, 1994). Improvements can be made if more selective membranes become available.
2.3. Storage or Disposal of CO2 Aquifers and oceans constitute a large CO2 sink and buffer. In principle, vast amounts of CO2 could be stored, in the order of 2x107 Gt carbon (IEA, 1993). However, the ecological impact of injecting large quantities of CO2 into sea and ocean water is still uncertain, while insufficient information is available still about the location and size of suitable aquifers.
Storage of CO2 in depleted oil and gas reservoirs is another possibility. The total capacity has been estimated to be 80-300 Gt C in exhausted gas wells and 40-200 Gt C in exhausted oil wells (IEA, 1993). However, whether this storage will provide long-term fixation of CO2 remains to be seen. 2.4. Chemical Usage of CO2 CO2 is one of the cheapest and most abundant carbon-containing raw materials in the world. Therefore, it could provide a building block for carbon-carbon chains or a competitive carbon source in the chemical industry. However, CO2 is rather inert and most of its reactions are energetically highly unfavourable. The development of active catalysts can solve the former, but the latter is determined by thermodynamics. Practical processes are feasible only under special conditions (high temperature or pressure, CO2 required for chemical reasons). 2.4.1.
Present and Promising Applications of CO2
CO2 usage can be divided into two groups: those using the phyisical properties and those using the chemical properties. The former does not contribute to the reduction of CO2 emissions directly but consumption of other materials might decrease, often indirectly leading to CO2 reduction. At present, only about 1% of the CO2 produced is used, and its consumption in the manufacture of chemicals is an order of magnitude lower: about 0.1% (Xiaoding and Moulijn, 1996). Fig. 4 shows the major applications of CO2 in the USA (Keim, 1987). 50 Refrigeration
% CO used
40
2
30 Beverage Carbonation
20
Enhanced Oil Metal Chemicals Recovery Fabrication Manufacture 10
0
Refrieration
Beverage carbonation
Enhanced oil recovery
Metal fabrication
Chemicals manufacture
Figure 4. Major uses of CO2 in the USA. Data from (Keim, 1987)
Examples of the use of the physical properties of CO2 are refrigeration, the beverage industry, enhanced oil recovery (EOR), and supercritical CO2 extraction. EOR represents one of the largest potential CO2 applications, and is expected to grow in the near future (Xiaoding and Moulijn, 1996). In refrigeration applications CO2 can substitute CFCs, which is favourable in combating the greenhouse effect as well as the ozone depletion problem. The chemical properties of CO2 are currently not exploited on a large scale. Examples of products from CO2 are urea, methanol, cyclic organic carbonates, and inorganic carbonates. Furthermore, CO2 is consumed in biomass production and in greenhouses to stimulate the growth of plants. Another application is in water purification and in neutralisation processes. In principle, CO2 can be used in many reactions, leading to myriad products. However, for the chemical usage of CO2 to be effective in reduction of atmospheric CO2 levels, it must be applied in the production of large volume chemicals, such as methanol. Even then, the contribution is small, as demonstrated later on.
2.4.2.
Applications of CO2 involving Syngas
Chemicals such as methanol and acetic acid are based on synthesis gas, a mixture of hydrogen and carbon monoxide, which can be produced through several routes (see Table 1).
Table 1. Reactions in the production and conversion of syngas.
∆Hr0 (kJ/mol) Steam reforming of methane CH4
+
H2O
↔ 3 H2
+
CO
206
(1)
+
2 CO
247
(2)
+
CO
41
(3)
- 91
(4)
- 123
(5)
CO2 reforming of methane (“dry reforming”) CH4
+
CO2
↔ 2 H2
Reversed water-gas shift (WSG) CO2
+
H2
↔ H2O
Methanol synthesis 2 H2
+
CO
↔ CH3OH
Acetic acid synthesis CH3OH +
CO
↔ CH3COOH
At present, steam reforming of natural gas (methane) is the most common technology. An interesting alternative is dry reforming. In steam reforming of methane a syngas mixture with a H2/CO ratio of 3 is produced, while dry reforming produces a mixture with equimolar amounts of H2 and CO. For methanol synthesis, the ratio should be about 2. With conventional steam reforming this ratio is usually achieved in a separate reactor via the reverse water-gas shift (WGS) reaction. This reaction converts expensive hydrogen into water, and is, therefore, not very economical. However, one can circumvent this problem by two ways. A combination of steam reforming and dry reforming can be used. Alternatively, dry reforming can be applied on its self, with separation of excess CO and usage in, for instance, acetic acid production. These process options lead to CO2 consumption, avoidance of a separate WGS reaction, and decreased fossil fuel requirements since CO2 is a carbon source. A problem that might occur is enhanced carbon deposition due to the decreased amount of H2O in the reaction mixture. Carbon formation occurs by decomposition of CH4 into C and H2, or by the so-called Boudouard reaction in which CO is converted to C and CO2. The situation can be improved by adding more steam (or oxygen), or alternatively, by developing catalysts which do not catalyse the carbon forming reactions. Fig. 5 shows a typical flow diagram of a process combining dry reforming with steam reforming, the Sparg process. This process solves the carbon deposition problem by partially deactivating the reforming catalyst by coverage with sulfur, which is always present in natural gas at ppm level. The sites for carbon formation are blocked leaving sufficient sites for the reforming reactions. Table 2 shows typical operating conditions of the Sparg process (Udengaard, 1992), with operation ranges between brackets.
Table 2. Operating conditions of the Sparg process (Udengaard ,1992).
Catalyst Inlet temperature Exit temperature Inlet pressure Exit pressure CO2/CH4 (molar, feed) H2O/CH4 (molar, feed) H2/CH4 (molar, feed) H/C (atomic, feed) O/C (atomic, feed) H2/CO (molar, product)
Partially sulfur-poisoned Ni 770 K (760 - 840 K) 1170 K (1150 – 1220 K) 0.84 Mpa 0.63 Mpa 0.54 (0.07-2.5) 0.9 (0-1) 0.1 (0.1-0.26) 3.7 (1.2-6.1) 1.3 (1.05-1.93) 1.8 (0.55-3.2)
CO2 import Fuel Air
Steam Natural gas feed
Radiation section
Flue gas to stack
Convection section Steam
Reformer
BFW
Cooling water
Waste heat recovery
CO2 wash
CO rich gas
CO2 recovery
Figure 5. Sparg process for the production of synthesis gas.
Steam and import or recycle carbon dioxide are added to the natural gas feed. After preheating to 770-920 K, the feed gas is passed through the tubular reformer. The heat for the endothermic reactions is supplied by external heating through combustion of natural gas. The flue gas from the reformer is used for preheating the feed gas and combustion air (not shown). The heat of the product gas is used to raise the process steam. Excess CO2 is removed in a CO2 wash system (amine solution). The CO2 present in the flue gas could also be recovered in an amine-absorption unit. This is done in the Calcor process (Kurz and Teuner, 1990). Then an almost complete conversion of carbon to CO is possible, however, at increased investment and operating cost. Operation of the Sparg reformer without steam is also possible, provided that the methane is free of higher hydrocarbons (Udengaard, 1991). The Sparg process may be an attractive alternative for revamp of existing plants. Only minor modifications are needed and part of the natural gas and steam can be replaced with less expensive CO2 (Dibbern et al., 1986; Udengaard et al., 1991). A comparison of the Sparg process with the conventional steam reforming process (see Fig. 6) shows that in the production of a CO-rich syngas, about 50% more CO2 is consumed (includes CO2 in exhaust) in the Sparg process. In the production of syngas with higher H2 content CO2 emissions are much lower when using the Sparg process instead of conventional steam reforming.
CO2 usage (ton CO2 / ton syngas)
0.3
H2/CO = 1
H2/CO = 1.8
0.24 0.2
0.15
0.1
Sparg 0
Sparg
Conv.
Conv. -0.06
-0.1
-0.2 -0.22 -0.3
Figure 6. CO2 consumption and emission in Sparg and conventional syngas process. Positive values correspond to consumption and negative values to emission of CO2.
CH4 requirement, CO2 emission (ton / ton methanol)
Besides the positive influence on CO2 emissions, the Sparg process also reduces energy requirements by about 20% and capital costs by 25% (van Diepen and Moulijn, 1994). As apparent from Fig. 7, application of the Sparg process in the production of methanol leads to considerable reductions in CO2 emissions and CH4 consumption
0.7
Sparg H2/CO = 2.05
Conventional H2/CO = 3
0.6
CH4 0.5
CH4
0.4
CO2
0.3
0.2
0.1
CO2
0
Figure 7. CH4 consumption and CO2 emission in the production of methanol.
Although no CO2 is actually consumed, methanol production via the Sparg process can still be beneficial from a CO2-mitigation perspective, if it replaces the conventional route.
Methanol has many outlets, either as fuel or as base chemical. The main commodity chemical produced from methanol is acetic acid. Dimethyl carbonate (DMC) is currently produced on a small scale by reaction of methanol with phosgene, (Battelle, 1993), but is a possible large-scale future product (as replacement of phosgene in the manufacture of isocyanates, polycarbonate plastics, etc.). A process for the production of DMC via a route involving CO2, thus avoiding the use of the hazardous phosgene, is feasible (Battelle, 1993; van Diepen en Moulijn, 1994). Table 3 quantifies the effect of usage of CO2 in the production of methanol and its main derivatives on emission reduction. Clearly, compared to the man-made CO2 emissions of approximately 25 Gt/year (6-8 Gt carbon), chemical conversion contributes only marginally (< 0.1%). Even when we would consider all possible uses, this figure would still not exceed 1%.
Table 3. Operating conditions of the Sparg process.
Chemical Methanol Acetic acid DMC a b
Demand (million ton/year) 23 5 2
CO2 consumption (ton/ton product) 0.31a 0.33b 0.43b
CO2 consumption (million ton/year) 7.1 1.7 1.1
CO2 emission avoided by using Sparg instead of conventional process. From methane via syngas and methanol.
2.4.3.
Applications of CO2 involving Biomass and Carbon Storage
Several processes based on renewable energy carriers have been suggested as part of the solution to the CO2 problem. One such process is the HYDROCARB process, (Steinberg et al., 1994). This process is based on the simultaneous hydrogenolyis of biomass and fossil fuels according to the general reaction: CHxOy +
(2 + y – x/2) H2
→
CH4
+
y H2O
(6)
Hydrogen is produced by catalytic decomposition of methane at 1273-1373 K (Steinberg et al., 1994): CH4
→ C +
2 H2
(7)
Another part of the methane reacts with the steam formed in reaction (6) to form CO and H2 according to reaction (1). The syngas is used in the synthesis of methanol, and, after separation, the remaining hydrogen can be used in the hydrogenolysis processes, while carbon can be used as a fuel or stored permanently or for future use. Fig. 8 shows a block diagram of the HYDROCARB process.
Fossil fuels Gas, oil or coal
Photosynthesis H2O + CO2 + solar energy
Biomass
Process gas recycle (H2, CH4)
Hydrogenolysis Process gas (CH4, H2, H2O)
Carbon to storage
Methanol synthesis
Methane decomposition
Methanol Liquid fuel
Syngas (H2, CO, CH4)
Figure 8. HYDROCARB process.
The HYDROCARB process realises an essentially zero CO2 emission by removing carbon from the cycle. However, storage of carbon, of course, leads to an energy penalty. Another process that makes use of carbon storage is the CARNOL process (Steinberg et al., 1993), see Fig. 9. Natural gas for heat production
CH4 (natural gas)
CO2 H2
Methane Gasification decomposition 50% C
50% C to storage
Natural gas for heat production
CO
Methanol synthesis
Methanol
Figure 9. CARNOL process.
This process is similar to the HYDROCARB process in that they both use methane decomposition and methanol synthesis. The feedstock in the CARNOL process,
however, is methane instead of biomass, while the Boudouard reaction is applied instead of hydrogenolysis: CO2
+
C
↔ 2 CO
(8)
50% of the carbon is stored and, as a consequence, prevented from entering the atmosphere.
3.
Conversion of CFCs
Until recently, chlorofluorocarbons (CFCs) were produced for use in amongst others refrigeration applications. Technically, they possess very favourable properties, and, moreover, they are low-priced chemicals. However, it is now generally accepted (Brune, 1996) that fully halogenated CFCs are responsible for the depletion of the ozone layer, and that they contribute to the greenhouse effect. Worldwide, production and consumption of CFCs has been terminated, but considerable amounts are still present. Recovery and subsequent destruction of these substances is a logical step. Many destruction techniques have been proposed, but only combustion has been applied on a commercial scale. Obviously, the conversion of CFCs into valuable chemicals is much more desirable. At Delft University of Technology, a catalytic process has been developed in which CCl2F2 (CFC-12) is converted into CH2F 2 (HFC-32), which is an ozone-friendly refrigerant and much less harmful greenhouse gas.
3.1. Thermodynamics CCl2F2 belongs to the family of halogenated methanes. Fig. 10 (van de Sandt et al., 1996b) shows all the possible C1 derivatives containing Cl, F, and H. The arrows indicate the thermodynamic stability (e.g., CCl4 least stable, CH4 most stable), and not necessarily a reaction sequence. All hydrogenolysis reactions starting from CCl2F 2 are exothermic irreversible reactions. Although thermodynamically the reaction of CCl2F2 would proceed via CHClF2, CH2ClF, CH3F, and finally methane, a reaction towards CH2F2 might be hoped for (provided that a selective catalyst is available), because the carbonfluorine bond is much stronger than the carbon-chlorine bond (Lacher et al., 1956).
C C l4 10
C C l3 F
C H C l3
11
C C l2F 2
20
C H C l2 F
12
C C lF 3 13
C H 2 C l2
21
30
C H C lF 2
C H 2 C lF
C H 3C l
22
31
40
C F4
CHF3
C H 2F 2
C H 3F
14
23
32
41
C 2H 6 170
C 3H 8
CH4 50
290
Figure 10. Thermodynamic relation between chlorinated and fluorinated methanes. Direction of arrows indicates the thermodynamic stability at 298 K and atmospheric pressure (van de Sandt et al., 1996b).
The reaction to be carried out is the hydrogenolysis of CCl2F2: CCl2F2 + 2 H2 → CH2F2 + 2 HCl
∆Hr0 = - 156 kJ/mol
(9)
The reaction enthalpy shows that the reaction is very exothermic. An important side reaction is the formation of methane, which is even more exothermic: CCl2F2 + 4 H2 → CH4 + 2 HCl + 2 HF
∆Hr0 = - 319 kJ/mol
(10)
3.2. Experimental Results Several noble metals on activated carbon supports have been tested for their catalytic activity. Palladium was found to be the most suitable metal for the CCl2F2 hydrogenolysis reaction (Wiersma, 1997). Pretreatment of the activated carbon support material is important. Washing with aqueous sodium hydroxide, aqueous hydrochloric acid, and water serves to remove impurities (van de Sandt et al., 1997; Wiersma et al., 1994). The main product formed in the catalytic hydrogenolysis of CCl2F2 is CH2F 2, although significant amounts of CHClF2 and methane are also formed. The results reported here are steady-state values. The effect of washing of the activated carbon support is evident from Fig. 11 (van de Sandt et al., 1997). The formation of CHF 3 and the lower alkanes C2H6 and C3H8 is strongly suppressed. Washing of the activated carbon removes accessible
100
5 non-washed washed
80
4
60
3
40
2
20
1
0
selectivity (mol%)
conversion/selectivity (mol%)
impurities such as iron and aluminium, which can act as Friedel-Crafts catalysts. These metals catalyse the undesired chlorine/fluorine exchange reactions, such as CHClF2 → CHF3, and CH2F 2 → CH2ClF. Iron also catalyses the formation of CH4, and other lower alkanes (van de Sandt et al., 1997).
0 conv. 12
sel. 32
sel. 22
CCl2F2 CH2F2 CHClF2
sel. 50
CH4
sel. 23
sel. 41
sel. 31
sel. 40
sel. 30
sel. 170+290
CHF3 CH3F CH2ClF CH3Cl CH2Cl2 C2H6 C3H8
Figure 11. Performance at 510 K of Pd (2 wt%) catalyst on a non-washed and a washed activated carbon in the hydrogenolysis of CCl2 F2 (van de Sandt et al., 1997).
The results reported below have all been obtained with palladium on a washed activated carbon support. Fig. 12 shows the conversion of CCl2F2 and the selectivities towards the most abundant products as a function of temperature (Wiersma, 1997). The main product is CH2F 2 but CHClF2 and methane are also formed. Other possible products such as CH2ClF and CH3F are hardly formed. HCl and HF are also formed during hydrogenolysis. HF does not affect the conversion and the selectivity towards CH2F2 (van de Sandt et al., 1996b), but addition of HCl results in lower selectivity towards CH2F 2 as demonstrated in Fig. 13. Interestingly, both Figs. 12 and 13 show that the selectivity towards methane remains nearly constant as a function of conversion, and is not affected by addition of HCl.
conversion/selectivity (mol%)
100 80 60 40 20 0 400
450
500
550
600
temperature (K) Figure 12. Conversion (triangles) and selectivity in the hydrogenolysis of CCl2 F2 over 1 wt% Pd/C to CH2F2 (solid circles), CHClF2 (squares), and CH4 (open circles) as a function of temperature. Conditions: WHSV=1 g/(g h), H2/CCl2 F2=3 mol/mol, p=0.3 Mpa (Wiersma, 1997).
Conversion/selectivity (mol%)
80
H2 H2/HCl
70 60 50 40 30 20 10 0
CCl 2F2 Conv.
CH 2F2
CHClF 2 Selectivity
CH4
Figure 13. Influence of HCl on hydrogenolysis reaction. Conditions: WHSV=1 g/(gh), H2/CCl 2F2=3 mol/mol, H2/HCl = 1 mol/mol, p=0.2 MPa (van de Sandt et al., 1996b).
3.3. Development of a Kinetic Model Many possible pathways can be formulated for the formation of the observed reaction products, but the experimental results and thermodynamic data make a plausible choice possible. CCl2F2 is a rather inert molecule. Therefore, the first step
in its conversion must be dissociative adsorption. In this step, two surface species can be formed, *CClF2 and *CCl2F. Since CH2F 2 is the main reaction product, obviously *CClF2 is formed preferentially, which indicates that the C-Cl bond is much weaker than the C-F bond. This is in accordance with the dissociation energies (at 298 K), which are 318 and 460 kJ/mol, respectively (Weast, 1983). Furthermore, CClF2 is thermodynamically more stable (see Table 4, which gives a survey of possible gasphase reaction intermediates with their Gibbs free energy of formation (Barin, 1995). Table 4. Gibbs energy of radicals that can be intermediates in hydrogenolysis of CCl2F2 (Barin, 1995).
Radical CF3 CClF2 CHF2 CCl2F CH2F CHCl2 CCl3 CH2Cl CH3 CF2 CClF CHF CCl2 CHCl CH2
∆G0 (kJ/mol) -549 -361 -331 -194 -103 -9 -9 44 88 -254 -51 35 159 238 329
Radical CF CCl CH
F Cl H C
∆G0 (kJ/mol) 192 435 540
32 72 184 670
The adsorbed CClF2 may either react with adsorbed hydrogen to yield CHClF 2, loose a fluorine atom to give *CClF, or loose the second Cl atom to form *CF 2. CH2F2 being the main product, the formation of *CF 2 is most likely, also because it is a relatively stable species as apparent from Table 4. Its conversion to CH2F2 must proceed via *CHF 2, which is even more stable. The effect of HCl on the conversion of CCl2F2 and the relative selectivity towards CH2F2 and CHClF2 can be explained by strong adsorption of HCl and the increased concentration of chlorine on the catalyst surface at higher HCl partial pressure (*CF2 ↔ *CClF2 equilibrium shifts to right). Therefore, adsorption of HCl should be taken into account in a realistic kinetic model.
Dissociative re-adsorption of CHClF2 and CH2F 2 is neglected on the basis of their low reactivity (less than 3% of CCl2F2 reactivity, see Fig. 14 (van de Sandt, 1996b). The low reactivity of CHClF2 proves that CH2F 2 is not formed via a reaction network of the type: CCl2F2
H2 ↔
CHClF2
H2 ↔
CH2F2
H2 ↔
etc.
(11)
If this had been the case, the conversion of CHClF2 would have been of the same order as that of CCl2F2.
conversion (mol%)
100 80 60 40 20 0
CCl2F2
CHClF2
CH2F2
Figure 14. Comparison of the reactivity of CCl 2F2, CHClF2, and CH2F2 in the catalytic hydrogenolysis over 2 g of 2wt% Pd/C. Conditions: T=510 K, H2/CFC=3, p=0.3 MPa, feed=16.5 mmol/h [35].
It is not entirely clear which surface intermediates lead to the formation of methane, but the experimental results clearly indicate that they are different from those leading to CH2F2, i.e. methane is formed via a parallel rather than a sequential pathway. One possible route is through dissociative adsorption of CCl2F2 to *CCl2F, a kind of “accident”. The further pathway could proceed via *CClF, *CClFH, *CHCl, *CH2Cl, *CH3, and finally CH4. Because of the low stability and hence high reactivity of these intermediates, they will react to the thermodynamically most stable product, methane. Therefore, kinetically only the dissociative adsorption of CCl2F2 to *CCl2F is important for this route. In order to derive a kinetic model, the following assumptions were made by Wiersma (1997):
-
Total number of active sites is constant; All surface species occupy only one catalyst site; All reactions take place at the catalyst surface: No gas phase reactions occur; CHClF2 is formed via reaction of *CClF2 with *H. The possible reaction of *CHF2 with *Cl is neglected for practical reasons; The sequential reaction of *CHF 2 to methane is neglected. Methane is formed via route 2 (see Fig. 15); *F is neglected. Although fluorine is present on the catalyst surface, it plays no role in the proposed mechanism.
These assumptions together with the above discussion lead to the simplified mechanism of Fig. 15, which includes the most important reaction products and the surface species through which they are most likely formed. The intermediates in methane formation are not important kinetically. CCl2F2
CHClF2
CH2F2
CH4
Route 1 Route 2
*CClF2
*CF2
*CHF2
*CCl2F Figure 15. Simplified mechanism of CCl2F2 hydrogenolysis (Wiersma, 1997).
Because on the basis of the experimental results and gas-phase thermodynamic data, a large number of possible surface species have been eliminated, leaving only CClF2, CF2, CHF2, CCl2F, H, and Cl, the kinetic model can be reduced to the 8 kinetically important equations shown in Table 5. The adsorbed CCl2F (formed via reaction 19) reacts through a series of surface intermediates to *CH3, which desorbs to form methane: *CH3 + *H → CH4 + 2*
(20)
This step and all the possible intermediate steps are much faster than adsorption of CCl2F2 (Eq. 19). Therefore, adsorbed CCl2F reacts immediately, and its surface concentration is negligible. As stated above, adsorbed F is also neglected.
Table 5. Kinetically relevant elementary steps in the hydrogenolysis of CCl2F2 (Wiersma, 1997).
Route 1: formation CH2 F2 and CHClF2 Dissociative adsorption: r1 r2
CCl2F2 + 2 * H2 + 2 *
→ * CClF2 + *Cl ↔ 2 *H
rate and equilibrium constants k1 K2 = k2/k-2
(12) (13)
Surface reactions: r3 r4
*CClF2 + * *CF2 + *H
↔ *CF2 + *Cl → *CHF2 + *
K3 k4
(14) (15)
k5 k6 K7
(16) (17) (18)
k8
(19)
Associative desorption: r5 r6 r7
*CHF2 + *H → CH2F2 + 2 * *CClF2 + *H → CHClF2 + 2 * *Cl + *H ↔ HCl + 2 *
Route 2: formation of methane r8
CCl2F2 + 2 *
→ *CCl2F + *F
3.4. Rate Expressions Using only the most abundant reaction intermediates (called “mari”), it can be concluded that only reactions 12 through 19 are kinetically significant. Reactions 12, 15, 16, 17, and 19 are assumed to be rate determining, while reactions 13, 14, and 18 are assumed to be in quasi-equilibrium. The rate equation is found by using the quasi-equilibrium relationships, the site-balance (Eq. 21), and coupling of the reaction rates of reactions 12, 15, 16, 17, and 19.
1 = θ * + θ Cl + θ H + θ CClF2 + θ CF2 + θ CHF2
(21)
The following relations hold (refer to Table 5):
r = r1 + r8;
r4 = r5;
r1 = r5 + r6
(22)
The rate expressions for the formation of CHClF2, CH2F2, and methane are shown in Table 6.
Table 6. Reaction rate expressions derived from the kinetic model (Wiersma, 1997).
k1 sN T p CCl2 F2
CHClF2 production:
1 + 1 p H 2 S p HCl r6 = ( ADS ) 2
(23)
k1 sN T p CCl 2 F2
CH2F2 production:
With:
1 + S p HCl pH2 r5 = ( ADS ) 2 S=
k6
(24)
(25)
k4 K 3K 7 K 2
CH4 production:
With:
r9 =
k 8 sN T p CCl 2 F2
( ADS )2
p HCl ADS = 1 + + K 2 p H 2 + ADS 2 K 7 K 2 pH 2
k 1 p CCl 2 F2 ADS 2 = k 5 k 6 K 3 p HCl k5 K 2 p H2 + k4 K7
k 5 K 3 k 5 p HCl + 1+ k 4 k 4 K 7 K 2 p H 2
(26)
(27)
(28)
Although at first the mechanistic models appears to be rather complicated, neglecting of the combined adsorption of the carbon containing surface intermediates (ADS2) results in a useful model. This is justified when step 5 and step 6 (Eqs. 16 and 17) are fast, resulting in low surface coverage with CClF2, CF2, and CHF2. From Eqs. 23 and 24 it follows that increasing the hydrogen partial pressure enhances the selectivity towards CH2F2, the desired product, while addition of HCl leads to more CHClF2. The combined rate of formation of CH2F2 and CHClF 2 is equal to: r1 =
k 1 sN T p CCl 2 F2
(29)
( ADS )2
conversion/selectivity(mol%)
Hence, the rate equations predict a constant ratio of formation of CH2F2 and CHClF 2 together relative to the formation of methane, consistent with the experimental results (Fig. 12). The kinetic model describes the experimental results well as shown in Figs. 16 and 17 below, while it also has predictive value (Wiersma, 1997).
100 80 60 40 20 0 0
1
2
3
4
5
W (g) Figure 16. Prediction of conversion and selectivities as a function of catalyst amount (W) by kinetic model (conditions: T=490, P=0.28 MPa, H2/CCl2F2=10 mol/mol), (measured values: =conv. ℓ =sel. CH2F2 , =sel. CHClF2, =sel. CH4). Adapted from (Wiersma, 1997)
These results show that the use of rate expressions based on kinetic models consisting of elementary steps is very useful. Even though a large number of elementary steps are involved, practical and not too complicated rate expressions result by eliminating unimportant steps in advance.
conversion/selectivity (mol%)
100 80 60 40 20 0 440
460
480
500
520
540
temperature (K) Figure 17. Prediction of conversion and selectivities as a function of temperature by kinetic model (conditions: P=0.5 MPa, H2/CCl2F2=10 mol/mol, WHSV=1 g/(g.h)), (measured values: =conv. ℓ =sel. CH2F2 , =sel. CHClF2, =sel. CH4). Adapted from (Wiersma, 1997)
3.5. Process Design A preliminary process design has shown that a continuous CFC-destruction process based on the Pd catalyst is technically and economically feasible (van de Sandt et al., 1996b). Fig. 18 shows a possible configuration. hydrogen recycle purge
Hydrogen
product CH2F2
raw feed CCl2F2
oil, water CHClF2, R-115
acids CCl2F2, CHClF2 recycle
Feed pretreatment
Reactor
Acid removal
Lights separation
Distillation
Figure 18. Process flow sheet for the conversion of CCl 2F2 to CH2F2 (van de Sandt et al., 1996a).
Oil, water, and CClF2-CF3 (R-115), which may have been introduced in the CCl2F2 feed during its recovery, have to be removed. Oil and water might cause problems with the catalyst and corrosion of process equipment, while CHF 2-CF3 (R125), the hydrogenolysis product of CClF2-CF 3, complicates separation because it forms an azeotrope with CH2F 2 (van de Sandt et al., 1996a). CHClF2, which is introduced with the recycle stream from the distillation section, is also removed. Approximately 80 % of the CCl2F2 fed to the reactor is converted. Acids, mainly HCl, but also some HF which is formed simultaneously with methane, are removed in the acid removal section prior to light ends separation. It has been shown (van de Sandt et al., 1996b) that a high hydrogen/CCl2F2 ratio increases the conversion, the selectivity towards CH2F2, and the catalyst stability (life time). The presence of methane in principle deactivates the catalyst, mainly due to some additional coke formation (van de Sandt et al., 1996a). However, if the excess hydrogen is high enough (≈ 8 - 16 mol/mol CCl2F2), some methane is allowed (van de Sandt et al., 1996a). The excess hydrogen is recycled, together with light products such as methane and ethane which inevitably end up in this recycle stream. In order to control their concentrations, a purge stream is required. From the experimental results it is apparent that CHClF2 should not be recycled to the reactor. The product CH2F 2 is separated from CCl2F2 and CHClF2, which are recycled to the pretreatment section. When the process is to be applied in an industrial plant,a suitable type of reactor is a multi-tubular fixed-bed reactor (6600 tubes of ≈ 20-30 mm) because of the considerable exothermicity of the reaction (standard heat of reaction -156 kJ/mol, standard heat of reaction for the formation of methane - 319 kJ/mol). Cooling is performed by means of boiling water. Temperature control is very important to prevent runaways and to protect the catalyst. For this reason a fixed bed reactor with intermediate quench cooling, as used in for instance ammonia synthesis, is not suitable; Due to the high conversion (80 %), which is desired to keep equipment small, a lot of heat evolves, and hot spots might easily occur. An alternative design might be based on a reactor in which the reactant and products are in the liquid phase. In that case temperature control is much easier. However, no performance data are available for al liquid-phase process.
4.
N2O
Nitrous oxide (N2O) has only recently been identified as a harmful compound, and is presently getting increasing attention. It is a relatively strong greenhouse gas (see Figs. 1 and 2) and a contributor to the ozone depletion in the stratosphere.
4.1. Sources of N2O Important man-made sources of N2O include land cultivation, adipic acid and nitric acid production, and combustion processes (see Table 7).
Table 7. Estimated amounts of N2 O emitted as a result of human activities. Based on (Wojtowicz et al. , 1993; Choe et al., 1993; Reimer et al., 1993; Soete (1993); Brem (1990)).
Source Land cultivation, fertilizers Adipic acid production Nitric acid production Fossil fuels (stationary) Fossil fuels (mobile) Biomass burning a
kton / year 1000 – 2200 371 280 – 370 190 – 520 400 – 850 500 – 1000
Point sources 23 255 > 1000 > 2⋅108
% Man madea 14 – 45 5–8 4–8 4 – 10 4 – 15 10 – 20
Total global man made emissions assumed 4.7 - 7⋅103 kton per year (Choe et al., 1993).
Many processes using nitric acid (HNO3) as oxidising agent inevitably produce N2O as by-product, e.g., Eq. 30 for the production of adipic acid, O
OH + 2 HNO3
HO
OH
+ N2O + 2 H2O
(30)
O
Nitric acid itself is produced by oxidation of ammonia at high temperature over Pt-Rh wire gauzes. Traces of N2O are formed as by-product. In combustion processes, N2O is formed as a result of fuel nitrogen or nitrogen from the air with oxygen. Besides the sources of N2O shown in Table 7, there are also other sources, but quantitative data are not available. An example is the formation of N2O during etching of steel with HNO3. A sour message is that N2O is not only formed as a by-product of production processes, but also as a result of measures taken to control emissions of other environmentally harmful species. Examples are the nonselective catalytic reduction of NOx in power plants, and the removal of NOx , CO, and hydrocarbons from otto exhaust gases in a three-way catalyst (Kapteijn et al., 1996). Emission levels for N2O are expected to become regulated within a couple of years. Therefore, there is a strong incentive towards emission control. In principle, two methods are available, i.e., reducing N2O formation or after-treatment (end-ofpipe solution).
4.2. Reduction of N2O Formation Emissions from stationary combustion sources are relatively small, but fluid-bed combustors (FBC), which are currently gaining importance, release more N2O per unit of energy than traditional coal combustors. Therefore, much attention has been paid recently to N2O formation and destruction mechanisms in FBC and control strategies (Kapteijn et al., 1996). Two approaches for the reduction of N2O formation can be used, viz., innovative combustor design to produce low-emission systems, and minimisation of emissions through improvements in operating conditions and process control of boilers. The first option seems relatively expensive at present, although future legislation might make it more attractive. Changing the combustor operating conditions to decrease N2O formation will often result in more NOx. If not, the combustor efficiency will be lower. Therefore, although recommended, the second option is difficult to apply (Kapteijn et al., 1996). In plants for the production of chemicals the formation of N2O could be reduced by making the process more selective. In processes that traditionally use HNO3 one could opt for an alternative oxidising agent. An example is the use of air instead of nitric acid in the oxidation of cyclohexanol to produce adipic acid. Although the N2O problem is solved and corrosion problems associated with nitric acid are absent, this alternative has the disadvantage of lower adipic acid yields (Chauvel and Lefebvre, 1989). Another (long-term) option is to develop entirely new process routes.
4.3. Removal of N2O from Off-gases Catalysis plays a vital role in N2O removal technologies. In the case of chemical processes N2O abatement is tackled most efficiently in processes producing offgases with high N2O concentration. Adipic acid plants are the most suitable candidates in this respect, because they produce an off-gas containing 30-50% N2O, while in other processes the concentration ranges from 0 to 3000 ppm (Kapteijn et al., 1996). The limited number of point sources (see Table 7) is an additional argument.
4.3.1
Reactions
The reaction aimed at is the decomposition of nitrous oxide into nitrogen and oxygen: 2 N2O → 2 N2
+
O2
∆Hr0 = - 163 kJ/mol
(31)
Due to the high activation energy for thermal fission of the N-O bonds this reaction requires temperatures of over 900 K (Kapteijn et al., 1996). Depending on the catalyst, off-gas composition, and addition of reducing agents like H2, CO, hydrocarbons, and C, other reactions may also occur, see Table 8. Table 8. Reactions in the conversion of N2 O.
∆Hr0 (kJ/mol)
Reaction N2O
+
CO
→ N2
+
CO2
- 365
(32)
2 N2O +
C
→ 2 N2
+
CO2
- 557
(33)
N2O
+
NO
→ N2
+
NO2
- 139
(34)
N2O
+
SO2
→ N2
+
SO3
- 181
(35)
N2O
+
O2
↔ 2 NO2
- 114
(36)
Many catalysts have been tested for the decomposition of N2O. Among these are metal catalysts, pure and mixed metal oxides, supported metal oxides, and zeolite catalysts. The latter, transition metal ion-exchanged zeolites, have shown high activities for the decomposition reactions (Kapteijn et al., 1997).
4.3.2
Mechanism and Kinetics
Classically, the reaction over metal oxides has been described by adsorption of N2O followed by oxidation of the active sites, and subsequent removal of the adsorbed oxygen by recombination (see Table 9). The adsorption and desorption steps are generally assumed to be in quasi-equilibrium, while the reaction of adsorbed N2O is considered rate determining. For steady-state conditions and assuming a constant number of active sites this model yields rate expression (Eq. 40). Depending on the parameter values, Eq. 40 can be simplified (Kapteijn et al., 1997). Alternatively, on metal containing zeolites the mechanism often consists of two irreversible rate-determining steps (see Table 9). The ratio k1/k2 in Eq. 43 equals NO*/N* (the concentration of sites occupied by oxygen relative to the concentration of empty sites) and hence determines the state of the active sites. When k1/k2 >> 1 the difficult step is irreversible removal of oxygen from the catalyst surface by reaction with N2O, Eq. 42, which means that the sites are oxidised. The catalyst sites
are in reduced state when k1/k2
