Enhanced decoking of a coked zeolite catalyst using a glow discharge in Ar–O2 gas mixture

Applied Catalysis A: General 272 (2004) 141–149

Enhanced decoking of a coked zeolite catalyst using a glow discharge in Ar–O2 gas mixture M.A. Khan a,∗ , A.A. Al-Jalal b a

Laser Research Section, Center for Applied Physical Sciences, The Research Institute, King Fahd University of Petroleum and Minerals, Box: 1947, Dhahran 31261, Saudi Arabia b Department of Physics, King Fahd University of Petroleum and Minerals, Box: 1947, Dhahran 31261, Saudi Arabia Received in revised form 6 May 2004; accepted 17 May 2004 Available online 25 June 2004

Abstract We have recently reported decoking of a coked zeolite catalyst in a glow discharge in O2 gas. Spectroscopic monitoring of the atomization of O2 and the accompanying conversion of coke to CO, CO2 and other gaseous products was used in conjunction with X-ray photoelectron spectroscopy (XPS) to determine the rate and extent of decoking. We have now investigated several different gas mixtures, including N2 –O2 , He–O2 and Ar–O2 mixtures with different compositions, to study possible enhancements in the decoking process. The He–O2 and Ar–O2 gas mixtures provided significantly large enhancements in the decoking process. However, the best results were from the Ar–O2 gas mixture: over 600% increase in the atomization of O2 and up to 350% increase in the CO yield, associated with the decoking process, were recorded. © 2004 Elsevier B.V. All rights reserved. Keywords: Catalyst regeneration; Glow discharge; CO monitoring; Visible region spectroscopy; Enhanced decoking

1. Introduction The active life of a catalyst used in the oil or petrochemical industry can be quite short, the most common reason being deposition of carbonaceous over-layers usually called ‘coke’. Other reasons include metal agglomeration and catalyst poisoning. In any case, it becomes necessary to replace the catalyst. However, the high costs associated with frequent use of fresh catalysts necessitate the search for efficient methods of restoring the activity of the deactivated catalysts. In the case of ‘coked’ catalysts, this involves removing the coke, but without causing any change in the internal structure of the catalyst. In this context, the conventional method employing thermal regeneration at 400–800 ◦ C is effective in removing the coke, but can be accompanied by some irreversible changes in the internal structure of the catalyst. This highlights the need for alternative techniques that could solve the problem associated with thermal regeneration. An important emerging technology uses oxygen plasma for regeneration of catalyst [1–5]. We have recently re∗

Corresponding author. Tel.: +966 3 860 4364; fax: +966 3 860 4281. E-mail address: [email protected] (M.A. Khan).

0926-860X/$ – see front matter © 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.apcata.2004.05.027

ported the use of a glow discharge in oxygen to investigate ‘decoking’ of a zeolite catalyst [6]. The process involves dissociation of O2 in a glow discharge, thereby creating O-atoms. Subsequent reactions of O-atoms with coke lead to the formation of CO, CO2 and other gaseous products that could be easily pumped out. One of the main channels for production of atomic oxygen is through simple dissociation of an oxygen molecule by electrons, i.e., e + O2 ⇒ 2O + e

(1)

However, atomic oxygen can also be produced by electron collisions through some other channels as explained below: e + O 2 ⇒ O + O−

(2)

e + O2 ⇒ O2 + + 2e

(3)

e + O2 + ⇒ 2O

(4)

Indeed, several other collision processes are also possible (see references [7,8]). For example, at high electron energies, it may be possible to produce O+ ions. e + O 2 ⇒ O+ + O − + e

(5)

e + O2 ⇒ O+ + O + 2e

(6)

142

M.A. Khan, A.A. Al-Jalal / Applied Catalysis A: General 272 (2004) 141–149

In some cases, even higher ionization stages of O may be created, but that is not of interest in the context of catalyst regeneration and will not be discussed in this paper. At low electron energies (average energy ∼3 eV), the rate constant for channel (1) is of the order of 2 × 10−9 cm3 s−1 [7,8]; that may be a factor of 10 higher than the corresponding ionization rate constant (channel 3), depending on the experimental conditions. Likewise, channels 2, 4-6 also have substantially lower rate constants compared with channel (1). In fact, atomic oxygen can be a dominant species in the discharge, if the experimental conditions are carefully controlled. In our experiments, atomization of O2 was monitored spectroscopically using the 3s 3 S–3p 3 P multiplet of O-atoms around 772–775 nm [6,9]. For CO, the band heads at 451.1 and 519.8 nm belonging to the B1 Σ–A1 Π bands of the Angstrom system [6,10] were used. For CO2 however, the band head at 353.4 nm belonging to the A2 Π 3/2u –X2 Π 3/2g system of CO2 + was used [10,11]. Furthermore, XPS analysis was used to confirm that decoking actually occurred and to determine the extent of decoking. The present paper reports the effects of different gas mixtures including N2 –O2 , He–O2 and Ar–O2 mixtures with different compositions on possible enhancements in the decoking process. The He–O2 and Ar–O2 gas mixtures provided significantly large enhancements in the decoking process. However, the best results were from the Ar–O2 gas mixture: over 600% increase in the atomization of O2 and up to 350% increase in the CO yield associated with the decoking process were recorded.

2. Experimental details The experimental set-up has been described in a previous paper [6]. Briefly, it consisted of a cross-shaped discharge cell with an adjustable gap between the electrodes. The light emitted by the discharge could be collected through one of the two side ports of the cell for the purpose of monitoring and analysis. The cell was first evacuated to pressures of the order of 10−3 mbar and then filled to the appropriate pressure of oxygen or oxygen-containing gas mixture under carefully controlled conditions. A periodic-purge mode of discharge was used, where after a chosen time (usually three minutes), the gaseous contents of the cell were pumped out, and fresh gas or gas mixture was added. A small amount of coked catalyst was spread evenly inside a ceramic boat essentially as a single layer and placed in the discharge cell, before evacuating the cell and then filling it with oxygen or an oxygen-containing gas mixture. This was exposed to the discharge for the selected period of time, followed by the cycles of evacuation of gaseous products, addition of fresh gas mixture and further exposure to the discharge for the selected period of time. The population densities of O-atoms, and CO and CO2 molecules were monitored through spectrally resolved light

emitted by the discharge using a scanning monochromator coupled to a thermoelectrically cooled photomultiplier. A compudrive unit provided precise scanning of the monochromator in the selected wavelength region. The signal from the photomultiplier was transferred to a boxcar averager and signal processing system linked to a PC. The background levels of CO and CO2 from the discharge cell, with emptied ceramic boat (no coked catalyst) inside the discharge, were also recorded for 3 min with fresh gas mixtures. The possible sources of the background CO and CO2 were the ceramic boat, the glass tube and the stainless steel electrodes. These background signals were subtracted from the respective signals obtained in the presence of the catalyst to get the net CO or CO2 yields. Records of the variations of the emission intensity of selected atomic and molecular transitions were obtained as a function of time. Indeed, the spectral emission intensity is directly proportional to the instantaneous population density of a particular species. Other parameters influencing the measured intensity include the appropriate transition probability, some geometrical factors (solid angle and field of view etc), detector sensitivity at the chosen wavelength and the detector bias. When one studies the time history of changes in the intensity of a particular atomic or molecular transition, the geometric and instrumental factors as well as detector sensitivity and detector bias should remain the same. Thus, the changes in intensity occurring over time are only dependent on the changes in the population of the particular species. Accordingly, it is possible to measure the total emission (area under the curve) in a given period of time and to relate it with the total number of those atoms or molecules present in the volume under observation during that time. This can subsequently be related to the rates of consumption of O-atoms or the evolution of CO and CO2 gases and hence to the rates of decoking. The related time histories of consumption of O-atoms as well as the evolution of CO and CO2 gases in our experiments were investigated using the above-noted spectral emissions. For this, the monochromator was set at the peak of the spectral emission corresponding to the particular atomic transition or the particular band head of the molecule and the boxcar was used in the static-gate mode. The actual cumulative amount of CO produced during the exposure to the discharge was recorded as a function of time, and the area under the curve was used as an index of decoking. Further confirmation of decoking was obtained through XPS analysis of the carbon (C) content of the catalyst samples. For this, the catalyst samples were irradiated in ultra high vacuum (10−9 mbar) with Al K-alpha photons (1486.6 eV) that ejected the 1s electron of carbon [6]. An electron energy analyzer was used to record the energy spectra of these ejected electrons (VG Scientific ESCALAB MKII Spectrometer). The area under the C 1s spectrum was read as the carbon content of the particular coked or decoked catalysts. Our experiments were conducted on a platforming refinery catalyst (Pt–Re/Al2 O3 ). It consisted of spherical gran-

M.A. Khan, A.A. Al-Jalal / Applied Catalysis A: General 272 (2004) 141–149

ules each about 1.6 mm in diameter. However, the basic processes discussed in this paper apply equally to any coked catalyst and are not dependent on the nature or the internal structure of the catalyst.

3. Results and discussion We first present some results on decoking in a glow discharge in O2 . Next, we will discuss enhanced decoking in a glow discharge containing various gas mixtures, particularly the Ar–O2 mixture. 3.1. Decoking in a glow discharge in O2 As noted earlier, decoking in a glow discharge in O2 proceeds in two steps: namely, atomization of O2 and subsequent reactions of O-atoms with the coke, thereby creating some gaseous products including CO and CO2 . The atomization of O2 and the accompanying creation of O-atoms in the discharge was monitored through a detailed

143

study of the radiated emission on the 3s 5 S–3p 5 P transitions around 777.2–777.5 nm mentioned earlier. The next step was to confirm that these O-atoms actually react with the coke to create CO and CO2 gases. The evolution of CO and CO2 gases resulting from this decoking process was also monitored spectroscopically. For CO, spectral emission on the band heads at 451.1 and 519.8 nm belonging to the B1 Σ–A1 Π bands of the Angstrom system [6,10] were investigated. However, as confirmed in our earlier work [6], the evolution of CO2 essentially follows the pattern of evolution of CO. As such, the behavior of evolution of CO2 is not discussed in the present paper. We first investigated the time history of consumption of O-atoms, produced in the glow discharge to correlate it with the evolution of CO in time. The population of O-atoms as well as CO molecules was monitored as a function of time as noted above. In the absence of the coked catalyst, the population of O-atoms stayed essentially constant in time (Fig. 1a). However, the situation was quite different when the coked catalyst was present. Fig. 1b summarizes the experimental data on the time history of consumption of O-atoms

Fig. 1. (a) Cumulative emission from O-atoms on the 3s 5 S–3p 5 P multiplet during the first 3-min exposure in the absence of the catalyst. (b) Time histories of consumption of O-atoms and the accompanying generation of CO gas during the decoking of a catalyst in a glow discharge in O2 .

144

M.A. Khan, A.A. Al-Jalal / Applied Catalysis A: General 272 (2004) 141–149 O-Atoms Yield vs. Pressureand Composition: Ar+ O 2

O-Atoms Yield (relative)

6

14% O2

17% O2 4

20% O2

50% O2 2

100% O2 0 0

1

(a)

2

Pressure (mbar)

O-Atoms Yield for Different Mixtures of O2 with Arat P = 1.4 mbar

O-Atoms Yield (relative)

12

8

4 0

5

10

(b)

15

20

25

30

% of O2 in the Mixture

Fig. 2. (a) Yield of O-atoms with different compositions of Ar–O2 gas mixtures as a function of pressure. (b) Yield of O-atoms with different compositions of Ar–O2 gas mixtures at a fixed total pressure of 1.4 mbar.

O-Atoms Yield vs. Pressure: He-O 2 Mixture 7

6% O2 O-Atoms Yield (relative)

6

10% O2

5

25% O2 4

3

50% O2

2

75% O2 1 0

1

2

3

4

5

Pressu re (m bar)

Fig. 3. Yield of O-atoms with different compositions of He–O2 gas mixtures as a function of pressure.

and the evolution of CO in the decoking process. Here, the population of oxygen atoms in the absence of the coked catalyst is taken as our reference. The numerical difference between the reference population (area under the curve when no coked catalyst was present) and the measured population of the O-atoms present at a particular time (area under the curve when the catalyst was being decoked) is taken as the net consumption of O-atoms. In fact, care is needed in reading the data presented in Fig. 1b. In particular, the values shown are normalized values and are not absolute. For instance, the consumption of O-atoms shown by solid triangles is in percent (%) of the total number of O-atoms available in the absence of the coked catalyst. Likewise, the CO yield shown by solid circles in this figure is in percent (%) of the total cumulative yield recorded in 720 s (12 min) in four gas purges. As seen in Fig. 1b, the O-atoms were rapidly consumed in the first 3 min, while the evolution of CO showed a corresponding rapid increase. More specifically, in the first 3-min interval (0–3rd minute) corresponding to a time of

M.A. Khan, A.A. Al-Jalal / Applied Catalysis A: General 272 (2004) 141–149

145

O-Atoms Yield vs. Pressure: N 2-O2 Mixture

O-Atoms Yield (relative)

3

99% O 2

90% O 2 2

96.7% O2

50% O 2 1 10% O 2

0 0

0.5

1

1.5

2

2.5

3

3.5

4

Pressure (mbar) Fig. 4. Yield of O-atoms with different compositions of the N2 –O2 gas mixture as a function of pressure.

180 s in Fig. 1b, nearly 66% of the produced O-atoms were consumed. On the other hand, about 64% of the total CO was produced in the same period of time. In the 9th–12th minute interval corresponding to the time of 720 s, only about 20% of the O-atoms were consumed, while only about 6% of the total CO was produced in the same 3-min time interval. This ties up rather nicely with our understanding of the on-going processes, where initially a lot of coke has to be removed, and we see a large consumption of O-atoms and a correspondingly large production of CO. On the other hand, when most of the coke has been removed, the rate of consumption of O-atoms slows down considerably after three or four 3-min exposures and gas purges, as does the production of CO.

At first sight, it might appear that the rate of creation of O-atoms through atomization of O2 in the discharge considerably exceeds the rate of its consumption in the decoking process (Fig. 1b), except in the first 3 min of exposure to the discharge. But, it has to be kept in view that the data on CO2 evolution are not included in this figure, and that could account for the difference. Furthermore, the CO2 formed in the discharge also dissociates to CO and O and adds to the net population of O-atoms and CO molecules, albeit to a small extent. It is to be noted that this dissociation of CO2 is a process dominated by collisions with electrons. However, the reverse process of conversion of CO back to CO2 is expected to proceed through collisions with O-atoms that are usually much slower than the electrons.

Maximum Yields of O-Atoms using Different Gas Mixtures with O2 12

O-Atom Yields

10

8

6

4

2

0 97%Ar

93%He

1%N2

100%O2

Best Gas Mixtures

Fig. 5. Bar graph showing the maximum yield of O-atoms recorded in 3-min interval exposures to discharge. Each bar represents the area under the corresponding 3-min emission record from the stated gas mixture.

146

M.A. Khan, A.A. Al-Jalal / Applied Catalysis A: General 272 (2004) 141–149

So far, we have not checked if the activity of the catalyst for the actual reaction in the refining process has been restored. Our main concern so far has been decoking of the coked catalysts. However, we are working at considerably lower temperatures, typically ≤300 ◦ C, compared to over 600 ◦ C in the thermal regeneration process. It is therefore anticipated that the internal structure will not be significantly affected in this process at such lower temperatures. Accordingly, the activity of the catalyst should be revived after the decoking process discussed here. We intend to investigate these aspects of the process in our future work. 3.2. Enhanced atomization of O2 in a glow discharge using different gas mixtures The strong correlation between the population of O-atoms and the accompanying generation of CO and CO2 gases as a result of decoking reactions has now been established in our work. For achieving enhanced decoking, the first step would be to increase the production of O-atoms in the discharge using different gas mixtures, for example. We investigated several different gas mixtures including N2 –O2 , He–O2 and Ar–O2 mixtures with different compositions. Interestingly, only a small amount of O2 in a large quantity of Ar or He produced high yields of O-atoms. On the other hand, only a small amount of N2 in a large volume of O2 was needed to produce some measurable enhancement in the yield of O-atoms, while large quantities of N2 added to O2 resulted in the opposite effect. Subsequently, it was confirmed that the higher number densities of O-atoms actually lead to higher yields of CO and CO2 gases, thus resulting in a higher rate of decoking. 3.2.1. Ar–O2 mixture Fig. 2a shows the effect of gas composition on the overall yield of O-atoms measured from the intensity of emission on the reference transition using different Ar–O2 mixtures in the glow discharge. The discharge current in all these experiments was kept constant at 100 mA. The voltage sustaining the discharge was between 490 and 500 V for pure O2 gas and between 430 and 450 V for the Ar–O2 (9:1) mixture at a pressure of 0.3 mbar. Experimental data showing the corresponding yield of O-atoms in a pure O2 gas are also included in this figure. While all compositions show their respective maximum atomizations in the pressure region between 1 and 1.5 mbar (Fig. 2a), higher yields of O-atoms are recorded when the partial pressure of O2 is much smaller (≤3%) than that of Ar, as shown in Fig. 2b. This is quite interesting because we are looking at emission from O-atoms and not from Ar atoms. As expected, we do not see any emission on this transition of O-atoms if pure Ar is used. 3.2.2. He–O2 mixture Fig. 3 shows the corresponding effect of gas composition on the overall yield of O-atoms measured from the intensity of the reference transition from the glow discharge in differ-

ent He–O2 mixtures. As in the case of Ar–O2 mixture, the yield is at its maximum when the O2 component in the mixture is much smaller (≤3%). The O-atom yields in Figs. 2b and 3 can be compared directly, since all the geometric and instrumental factors as well as the detector bias are kept constant in this series of experiments. The total yield in the case of He–O2 mixture remains somewhat lower than that of the corresponding Ar–O2 mixture for similar O2 contents of the gas mixture. 3.2.3. N2 –O2 mixture The overall yield of O-atoms measured from the intensity of the reference transition from the glow discharge is shown in Fig. 4 as a function of the gas composition using different N2 –O2 mixtures. Unlike the cases of Ar–O2 and He–O2 mixtures, the maximum yield is obtained, when the N2 component in the mixture is much smaller (≤3%), consistent with the report of Kaufman and Kelso [12]. However, the total yield of O-atoms in the N2 –O2 mixture remains much lower than the case of Ar–O2 and He–O2 mixtures (see Figs. 2–4 where the data along the y-axis can be directly compared). This may be due to the fact that Ar and He do not react with O-atoms, while both N2 and N could react with O atoms forming complexes such as NO, NO2 , N2 O, N2 O5 etc. Fig. 5 is a bar graph showing a comparison of the influence of different gas mixtures including N2 –O2 , He–O2 and Ar–O2 mixtures with different compositions, on the overall yield of O-atoms. Here, the areas under the intensity curves for the reference transition of O-atom were measured for a 3-min period. From this bar graph, the mixture of choice under our experimental conditions clearly seems to be Ar–O2 mixture with only a few percent of O2 , as noted earlier. The role of any inert or noble gas in increasing the population of O-atoms can be understood from the following two considerations. They could play a role in increased atomization of O2 , and equally importantly, a role in inhibiting the reverse processes including recombination. We will consider the specific case of Ar in our discussion here, since it produces the best results, but the arguments apply equally to any other noble gas. The dissociation of O2 and the creation of different excited states of Ar and O-atoms (represented by Ar∗ and O∗ , respectively) through collisions with energetic electrons present in the discharge may be described as: e + O2 ⇒ e + 2O

(1)

e + O ⇒ e + O∗

(7)

e + Ar ⇒ e + Ar∗ .

(8)

The excited atoms Ar∗ usually end up in one of the metastable states of Ar described by ArM , after some radiative decays [13,14]. The ArM atoms have a very small transition probability for radiative decay and are expected to transfer their energy to O2 molecules through collisions

M.A. Khan, A.A. Al-Jalal / Applied Catalysis A: General 272 (2004) 141–149

147

CO-Yield vs Pressure: He + O2 and Ar + O2 with 10% O2 50

CO-Yield (relative)

40

Ar +O2

30

20 He +O2

10

0 0

1

2

3

4

Pressure (mbar)

Fig. 6. Comparative CO yields as a function of total pressure from discharges in He–O2 and Ar–O2 gas mixtures with 10% of O2 . Maximum CO-Yield for Different Gas Mixtures 50 45 40

CO-Yield (a.u)

35 30 90% Ar 90% He 100% O2

25 20 15 10 5 0 90% Ar

90% He

100% O2

Fig. 7. Bar graph showing the maximum yield of CO for different gas mixtures in the first 3-min exposure at 0.3 mbar pressure.

yielding O-atoms or even to O-atoms yielding O∗ as follows: ArM + O2 ⇒ Ar + O + O Ar M + O ⇒ Ar + O∗

Time History of Cummulative CO Yield

(9) (10)

O + O− ⇒ e + O2

(11)

CO Yield (Relativbe)

200

On the other hand, the presence of such a large number of Ar atoms (≥90%) could effectively prevent some of the reverse processes such as recombination that might otherwise cause a decrease in the number density of O-atoms. Two examples of such recombination processes are:

Ar + O2 y = 888.16e-0.0083x

100

O2 only

+

−

O + O ⇒ O2

y = 121.6e-0.003x

(12)

These reverse processes could be effectively inhibited by the increasingly bigger and thicker “walls of Ar” between the O, O+ and O− atoms or ions. We have recently carried out detailed investigations on the possible pathways of energy transfer between Ar∗ and O2 as well as O-atoms and leading to O∗ . These will be reported in a forthcoming paper.

0

0

100

200

300

400

500

600

Time (S)

Fig. 8. Comparative time histories of cumulative CO emission in 3-min interval exposures to the discharges in pure O2 gas (solid triangles) and in Ar–O2 gas mixture (solid circles). CO band head at 519.8 nm was used here. Each point corresponds to the end position of the particular 3-min interval and represents the area under the curve showing the total CO generated in that interval. Exponential fits to the data are also shown.

148

M.A. Khan, A.A. Al-Jalal / Applied Catalysis A: General 272 (2004) 141–149

3.3. Enhanced decoking of the coked catalyst The most important question that must be addressed in the context of decoking of a catalyst in a glow discharge in O2 -containing gas mixture is whether or not enhanced atomization of O2 actually leads to more efficient decoking. For this, any one or more of the techniques available for characterization of decoked catalyst could be used to monitor and quantify the extent of decoking. As noted in our earlier paper [6], the results based on the XPS analysis correlate well with the record of cumulative CO and CO2 emissions as far as decoking is concerned. In the present paper, we used the CO yield monitored from the intensity of the band heads at 451.1 and 519.8 nm belonging to the B1 Σ–A1 Π bands of the Angstrom system as indicators of enhanced decoking. Additionally, XPS analysis of the carbon content of the coked and decoked catalyst samples was also used for further confirmation of the decoking. 3.3.1. Enhanced CO yield We have investigated the cumulative CO emission from catalyst samples exposed to a glow discharge in pure O2 and in two different gas mixtures: Ar–O2 and He–O2 . The corresponding CO yields from the glow discharges in a He–O2 and Ar–O2 gas mixtures with 10% O2 are shown in Fig. 6 as a function of the total pressure in the region where maximum yield of O-atoms in the mixture was observed. Here, the CO band at 451.1 nm was used. Indeed, similar results were obtained with the CO band at 519.8 nm. The discharge current was again maintained at 100 mA in these experiments. It is apparent from this figure that lower pressures are more appropriate for our set-up to get the maximum yield of CO and hence the maximum decoking. Apparently, the working pressure of 0.3 mbar is more suitable in our case (specific discharge geometry and location of the catalyst). The reason for this can be understood from the following considerations. (i) It is the collisions of O-atoms with the coke, which generate the CO. So, the collision-free path for O-atoms becomes more important over here. This should be contrasted with the maximum yield of O-atoms where collisions of electrons with O2 and other atoms and molecules and hence the collision-free path of electrons is really important, even though collisions between excited atoms and molecules also contribute there. (ii) It should be remembered that the geometry, i.e., the location of the catalyst with respect to the discharge, plays an important role here. For the dissociation of O2 that is in gaseous form filling the entire space between the electrodes, the electrons see the O2 in all the directions they travel. On the other hand, the catalyst sample is located at the bottom of the sample holder sitting on the floor of the discharge cell, so only the fraction of O-atoms travelling in the downward direction will find

it. Indeed, the best conditions may be quite different for a different experimental set-up. It is clear from Fig. 6 that for higher CO yields, the Ar–O2 gas mixture provides better results compared with the He–O2 , as indeed was the case for higher O-atom yields (Fig. 5). Fig. 7 is a bar graph showing a comparison of the overall yield of CO gas in the first 3-min exposure, as a measure of decoking. Thus, the gas mixture of choice in our experimental conditions is Ar–O2 with only a few percent O2 . Fig. 8 presents the time histories of the net CO yields in the process of decoking in glow discharges in pure O2 (solid triangles) and in Ar–O2 gas mixture (solid circles). Here, CO band head at 519.8 nm was used. The cumulative CO emission in each 3-min period appears as a point corresponding to the end position of the particular time interval, but actually representing the area under the curve showing the total CO signal in the 3-min interval. It is noted that in the absence of the discharge (t = 0), there is no CO emission and therefore, the curve should start from zero. Furthermore, as reported in our earlier paper [6], the points showing the CO yield at 1-min (60 s) and 2-min (120 s) exposures lie essentially on a straight line joining the zero emission (t = 0 s) and the point representing the CO yield in the first 3-min (0–3rd minute) exposure. From the exponential fits to the data, the rates of CO emission are read as 3 × 10−3 s−1 for decoking in a discharge in pure O2 gas and 8.3 × 10−3 s−1 for decoking in a discharge in the Ar–O2 gas mixture. However, the discontinuous nature (periodic purges) of the experiment has to CPS

2500

1 2000

1500

2 1000

3

287.5

282.5

277.5

Binding Energy (eV) Fig. 9. XPS spectra of C 1s showing the carbon content of the catalyst samples: (1) untreated coked catalyst; (2) catalyst sample de-coked in a glow discharge in pure O2 gas for a total of 15 min exposure; and (3) catalyst sample de-coked in a glow discharge in Ar–O2 gas mixture for a total of 15 min exposure.

M.A. Khan, A.A. Al-Jalal / Applied Catalysis A: General 272 (2004) 141–149

be kept in mind. Due to these inherent discontinuities, the rates calculated can only be considered as “averaged rates” in the context of the specific experimental details. 3.3.2. XPS analysis The actual carbon content of the catalyst samples before and after exposure in the discharge was measured using XPS analysis. Fig. 9 shows the measured carbon content in the reference coked catalyst (curve 1), and in two identical catalyst samples after exposure in different glow discharges in pure O2 gas (curve 2) and in Ar–O2 gas mixture (curve 3). The samples corresponding to the curves 2 and 3 were exposed to discharges in the respective gas mixtures for a total of 15 min (five 3-min exposures and the accompanying five purges with fresh gases). The horizontal scale in Fig. 9 shows the binding energy of the 1s electron in carbon at 284.5 eV, while the vertical scale gives the signal counts per second (CPS). The sample exposed to a glow discharge in pure O2 (curve 2) shows an unexpected second peak at slightly lower binding energy of 281.6 eV. This may correspond to a carbide peak, as discussed in our earlier paper [6]. We intend to carry out further investigations to establish the origins of this second peak. Interestingly, this peak does not appear in the sample exposed to a glow discharge in Ar–O2 gas mixture, and that may be an added advantage. The extent of decoking was determined from the area under the peaks above the background and corresponds to 65% in the case of discharge in pure O2 (if carbide peak is included) and 76% in the case of discharge in the Ar–O2 gas mixture. We believe that some improvements in the discharge geometry can be incorporated in our experiments to raise the level of decoking to over 90%. It has to be remarked that the initial rate of decoking is considerably higher, when Ar–O2 gas mixture is used as seen in Fig. 8. However, if longer exposure times are used (e.g., ≥15 min), all removable carbon from the coked sample could be oxidized even in the case of a discharge in pure O2 . 4. Conclusions We have demonstrated that a significant increase in the yield of O-atoms and the accompanying yield of CO is

149

possible using different gas mixtures including N2 –O2 , He–O2 and Ar–O2 in a glow discharge. The Ar–O2 gas mixtures with small amounts of O2 were seen to give the best results for decoking; over 600% increase in the atomization of O2 and up to 350% increase in the CO yield were recorded.

Acknowledgements This work is a part of the Research Project SABIC-2000-11. The support from King Fahd University of Petroleum and Minerals and the Research Institute is gratefully acknowledged. Dr. Tabet and Mr. Iftikhar Khattak of the Physics Department, KFUPM carried out the XPS analysis for us and we deeply appreciate their cooperation. The help of Mr. I.A. Bakhtiari and Mr. Pillai in some decoking experiments is also gratefully acknowledged.

References [1] E.A. Dadashova, T.V. Yagodovskaya, V.V. Lunin, V.A. Plakhotnik, Kinet. Katal. 32 (1992) 1511. [2] H. Meissner, K.-P. Franck, H. Rudolp, T. Hammer, Catal. Today 72 (2002) 325. [3] G.P. Vissokov, Catal. Today 72 (2002) 197. [4] G.P. Vissokov, M.I. Panayotova, Catal. Today 72 (2002) 213. [5] S. Futamura, A. Zhang, H. Einaga, H. Kabashima, Catal. Today 72 (2002) 259. [6] M.A. Khan, A.A. Al-Jalal, I.A. Bakhtiari, Anal. Bioanal. Chem. 377 (2003) 89. [7] J.R. Hollahan, A.T. Bell, Techniques and Applications of Plasma Chemistry, John Wiley & Sons, New York, 1974. [8] B.D. Blaustein (Ed.), Chemical Reactions in Electrical Discharges, Advances in Chemistry Series (80), American Chemical Society, Washington, DC, 1969. [9] A. Striganov, N. Sventitski, Tables of Spectral Lines of Neutral and Ionized Atoms, Plenum Press, New York, 1968. [10] R.W.B. Pearse, A.G. Gaydon, The Identification of Molecular Spectra, fourth ed., Chapman & Hall, London, 1976. [11] D.J. McCAA, D.E. Rothe, AIAA J. 7 (1969) 1648. [12] F. Kaufman, J.R. Kelso, J. Chem. Phys. 32 (1960) 301. [13] S. Schohl, D. Klar, T. Kraft, H.A.J. Meijer, M.-W. Ruf, U. Schmitz, S.J. Smith, H. Hotop, Z. Phys. D: At., Mol. Clusters 21 (1991) 25. [14] C. Lin Chun, L.W. Anderson, in: Advances in Atomic Molecular and Optical Physics, vol. 29, Academic Press Inc., USA, 1992, pp. 1–32.