`
` icine
`ER
`
`Mi
`
`Fi
`ELSEVI
`
`Available online at www.sciencedirect.com
`
`science()oinecr:
`
`Applied Catalysis B: Environmental 59 (2005) 205-211
`
`
`
`FEET
`
`¥
`
`APPLIED
`CATALYSIS
`
`B ENVIRONMENTAL
`
`www.elsevier.com/locate/apcatb
`
`Nature of nitrogen specie in coke andtheir role in NO,
`formation during FCC catalyst regeneration
`
`I.V. Babich, K. Seshan “, L. Lefferts
`
`University of Twente, Catalytic Processes and Materials, Faculty of Science and Technology,
`IMPACT, P.O. Box 217, 7500AE Enschede, The Netherlands
`
`Received 15 September 2004; received in revised form 7 February 2005; accepted 9 February 2005
`Available online 9 March 2005
`
`Abstract
`
`NO, emission during the regeneration of coked fluid catalytic cracking (FCC)catalysts is an environmental problem. In order to follow the
`route to NO, formation and try to find ways to suppress it, a coked industrial FCC catalyst has been prepared using model N-containing
`compounds, e.g., pyridine, pyrrole, aniline and hexadecane—pyridine mixture. Nitrogen present in the FCC feed is incorporated as
`polyaromatic compounds in the coke deposited on the catalyst during cracking. Its functionality has been characterized using XPS.
`Nitrogen specie of different types, namely, pyridine, pyrrolic or quaternary-nitrogen (Q-N) have been discriminated. Decomposition of the
`coke during the catalyst regeneration (temperature programmed oxidation (TPO) and isothermal oxidation) has been monitored by GC and
`MS measurements of the gaseous products formed. The pyrrolic- and pyridinic-type N specie, present more in the outer coke layers, are
`oxidized under conditions when still large amount of C or CO is available from coke to reduced NO, formed to N>. ““Q-N” type species are
`present in the inner layer, strongly adsorbed on the acidsites on the catalyst. They are combusted last during regeneration. As most of the coke
`is already combusted at this point, lack of reductants (C, CO, etc.) results in the presence of NO, in the tail gas.
`© 2005 Elsevier B.V. All rights reserved.
`
`Keywords: NO,; FCC; Regenerator, XPS
`
`1. Introduction
`
`Nitrogen oxides (NO,) are harmful atmospheric pollu-
`tants [1]. Legislations to limit NO, emissions are continually
`being introduced; 70-77% cut in NO, emissions is targeted
`by the EU before 2009 [2]. NO, emissions from a typical
`refinery are about 2000 tons/year, the fluid catalytic cracking
`(FCC) section contributes about 50% of this [3,4], and is
`thus an area of environmental concern.
`
`Hydrocarbon feedstock for FCC units usually contains a
`variety of aromatic, N-heterocyclic compounds. During
`cracking, part of this N present is included in the cokethatis
`deposited on the catalyst [3]. During regeneration of the
`catalyst, nitrogen in cokeis released to the atmosphere as N
`and NO,. Typical levels of NO, in the flue gas from fluid
`catalytic cracking regeneration units vary between 50 and
`
`* Corresponding author. Tel.: +31 53 4893554; fax: +31 53 4894683.
`E-mail address: k.seshan@utwente.nl (K. Seshan).
`
`0926-3373/$ — see front matter © 2005 Elsevier B.V. All rights reserved.
`doi: 10.1016/j.apcatb.2005.02.008
`
`500 ppm, depending upon feed nitrogen content and rege-
`nerator conditions [5,6]. The major component (>95%) of
`NO, is NO. Formation of N20 and NOz is negligible under
`FCC regenerator conditions [3,6]. Reduction ofNO,levels in
`the exhaust gases is an essential part of the technology for
`a “green”refinery.
`Real-life data from refineries show that correlations exist
`
`between nitrogen contents, basicity of nitrogen precursors
`present in the feed and NO, emissions during regeneration
`[3,5]. Typically, N is present as (substituted poly-) aromatic
`compounds
`such as pyrroles (—C—NH-C-), pyridines
`(-C=N-C=) or anilines (~C—NH;). Of these, pyridine is
`a strong base, aniline and pyrrole are weaker bases because
`of the delocalization of the electron pair on N to the aromatic
`ring. Higher the basicity of the molecule, stronger is its
`adsorption on the acidic FCC catalyst and larger is the chance
`to be deposited in the coke during cracking.
`The nature of N in coke and the mechanism ofits
`conversion to NO, or N> in the FCC regeneratorare not very
`
`WRG-1012
`
`1
`
`WRG-1012
`
`

`

`206
`
`LV. Babich et al./Applied Catalysis B: Environmental 59 (2005) 205-211
`
`well understood.It is generally suggested, in analogy with
`coal combustion [7,8], that NO, is formed via two routes:
`directly from nitrogen in coke (fuel NO,) and through
`volatile nitrogen intermediates such as HCN and NH;
`(prompt NO,) [5,6]. The mechanism of nitrogen transfor-
`mation during coke or coal/char combustion is complex, and
`attempts to correlate nature of nitrogen functionality (its
`chemical environment) with the amount of NO, released
`[9,10], have not been very successful. In the case of FCC
`catalysts such relationships are even less available.
`Due to the very low nitrogen contents in coked FCC
`catalysts, the discrimination of nitrogen specie in coke is
`complicated. Presence of polar or non-polar type nitrogen
`has been reported, but no attempts have been madetorelate
`this to NO, formation during coke combustion in the FCC
`regenerator [11].
`Aim of this work is to understand (i) if and how, the
`variations in the coking conditions and different types of N
`compounds present in FCC feed influence the nature of
`nitrogen in the coke and (ii) the transformation of the N
`specie in coke to Nz and NO, during oxidative regeneration.
`In order to establish this, reactions occurring in a FCC unit
`(cracking and regeneration) were carried out on lab scale
`using a commercial FCC catalyst and model N-heterocyclic
`aromatic compounds. The knowledge generated should be
`helpful in the developmentof catalyst systems to minimize
`NO,, emission from FCC regenerators.
`
`2. Experimental
`
`2.1, Materials
`
`FCC catalysts were obtained from Grace Chemicals,
`Germany. Reagent grade (>99%) pyridine, aniline, pyrrole
`and hexadecane
`(Aldrich Chemicals) were used as
`precursors for coking experiments.
`
`coking experiments, catalysts were either cooled to RT or
`allowed to age at 550 °C in He for varying durations. Catalyst
`regeneration experiments were carried out in a fixed-bed
`reactor (~0.2 ¢ samples) in a temperature programmed
`oxidation (TPO) mode. TPO was carried out between 30 and
`850 °C,using a heating rate of 10 °C/min and a gas mixture
`containing 2-4 vol.% Oz; 0-1 vol.% CO and balance He.
`Total flow rates were between 50 and 100 ml/min.
`
`2.3. Analytical techniques
`
`Elemental analyses (C, N and H) of coked catalysts were
`carried out on a Elemental Analyzer Carlo Erba EA1108.
`Thermo-gravimetric measurements were performed using a
`TGA/DSC (Mettler-Toledo)set-up. Typically, 50 mg sample
`and 30 ml/min flow ofeither 5% O./Heor pure He wasused.
`During TPO, gas composition at the outlet was constantly
`monitored. Nx, O2, CO, CO2 and N2»O were analyzed by
`Varian 3300 GC, using a Valco 16 position ST valve, two
`columns (Heysep Q and MS 5A) and a TCD detector. The
`amounts of NO and NO, were determined by a NO, analyzer
`(Thermo Environmental Instruments Inc. 42C). In addition,
`a Quadruple MS(Pfeiffer Vacuum) was used for continuous
`monitoring of the gas composition.
`Coke formed on the catalyst was characterized (N1s,
`Cls) by XPS (® Quantera Scanning ESCA Microprobe
`spectrometer). Cls binding energy (BE) of 284.8 eV was
`used to calibrate the spectra. In order to differentiate the
`types of N specie present in coke, the non-symmetric N1s
`experimental envelopes were subjected to a de-convolution
`procedure using a minimal number of peaks, varying
`FWHM,position andintensities of the peaks to result in the
`best NLLS fit (x? = 0.98). A mix of Gaussian (80%) and
`Lorenzian (20%) functions, typical for XPS [12], was used.
`
`3. Results
`
`2.2. Coking and regeneration experiments
`
`3.1. Coke formation and characterization
`
`Coking experiments were carried out at 550°C. 0.4g
`catalyst was placed between two quartz pugs in a quartz
`tubular reactor (d = 4 mm). The catalyst was preheated first in
`He (50 ml/min) at 550°C for 30 min. Subsequently, coke
`precursors were introduced into the gas stream using a syringe
`pump. Flow of the liquid precursor was varied between 0.2
`and 180 ml/h, coking time was varied from 10 s to 4 h. After
`
`Table 1 gives the details of the ““N”’ containing precursors
`used for coking experiments (see Section 2). pKpase values of
`the precursors listed in the table show that acidity decreases
`in the order pyridine < aniline < pyrrole,i.e., pyridineis the
`most basic of the three components used. The table also
`gives the amount/composition of coke deposited on the
`catalysts after 4 h time on stream. No relation was observed
`
`Table 1
`
`Details of the coke precursors and coke formed on the catalysts*
`
`Coke precursor™
`
`N/C atomicratio for
`precursor
`
`Pyridine
`Aniline
`Pyrrole
`
`0.20
`0.16
`0.25
`
`PRobase
`
`8.64
`9.38
`14.00
`
`Coke composition related to FCC catalyst (wt.%)
`
`C (wt.%)
`49+ 0.1
`73404
`6440.1
`
`N (wt.%)
`0.90 + 0.05
`0.85 + 0.05
`0.85 + 0.05
`
`N/C atomicratio
`0.15
`0.10
`0.11
`
`* Total feed flow 100 ml/min containing 10 vol.% of coke precursor and He balance; coking time 4 h.
`
`2
`
`2
`
`

`

`LV. Babichet al./Applied Catalysis B: Environmental 59 (2005) 205-211
`
`207
`
`0.2
`
`
`240
`
`015 2
`©
`2
`§&
`
`bot
`
`aQ
`
`s
`_
`=
`z
`(
`
`—>
`
`{
`—_+—__
`
`Lo.05 2
`
`ernege meets
`
`-
`
`S
`
`T
`
`0
`
`Coking time (min)
`
`Fig. 1. Amount and composition of coke over the FCC catalyst as a function
`of time. Pyridine was used as the coke precursor; total feed/catalyst ratio is
`2 ml/g; T= 550 °C. Left X-axis: (@) carbon content; (@) nitrogen content
`in the coke formed; right axis: (x) N/C atomic ratio.
`
`in coke varied with coking time, XPS spectra in the region of
`Nils BE were recorded for catalyst samples coked with
`varying times on stream. Typical N1s experimental spectra
`are shownin Fig.2.
`It can be seen from the figure that position and shape of
`the N1s peaks changed with coking time. Thus, increase of
`coking time from 10s to 4h caused a shift in the spectra
`position from 400.8 to 400.0 eV as well as broadening of the
`peak. In order to probe the changesin the spectra in terms of
`different types of “N”’’ specie, experimental Nls spectra
`were de-convoluted with multiple peaks as described in the
`experimental section. The details of the fit results, viz.
`positions of the individual peak maxima andtheir relative
`intensities are summarized in Table 2. The bestfit for the
`
`experimental N1s spectra, in all cases, was obtained with
`two individual components (Fig. 2a—d). BE of these two
`peaks did not depend on way coke was formed (coking time,
`precursor). Furthermore, FWHM of each peak varied only
`within the accuracy of the binding energy determination.
`Theresults indicate, probably, two types of nitrogen specie
`to be present in the coke and characterized by BE at 399.9
`and 402eV. Relative amounts of a nitrogen species
`corresponding to these were estimated from the integral
`intensity of the peaks and are presented in Table 2.
`In the case of coking experiments with pyridine, the
`contribution of the peak corresponding to the “N” specie
`characterized by the higher BE (402 eV) decreased from 58
`to 20% with increasing the coking time from 10 to 600s.
`Further, independent of the precursor used, longer coking
`(4h)or further ageing times(6 h) resulted in low intensities
`of this peak, i.e., between 10 and 17% of total integral
`intensity of Nls experimental peak (Table 2).
`
`betweenthe relative amount of N (indicated by N/C) in the
`coke precursor andin the coke deposited on the catalyst. As
`expected, the more basic compound,pyridine, causes higher
`percentage of N in the coke and detailed studies were
`undertaken with pyridine.
`Fig. 1 showsthe extent of coking with pyridine at 550 °C
`as a function of time. It can be seen from the figure that
`majority of the carbon and nitrogen is deposited on the
`catalyst at short
`times on stream.
`In the 4h coking
`experiment shown, 70% of the total coke was deposited
`during the first 60s. No significant deviation in the coke
`composition (N/C ratio) was observed during this coking
`period. Further heating at the coking temperature (550 °C) in
`an inert atmosphere (He) up to 6h did not cause any
`additional changes. In order to checkifthe nature of the “N”
`
`
`
`Intensity(a.u.)
`
`
`
`406 404 402 400 398 396
`
`406 404 402 400 398 396
`
`Fig. 2. Ns experimental XPS spectra of coked FCC catalyst. Experimental curves are fitted with two individual peaks. Shown spectra are after catalyst coked
`for 10s (a), 60s (b), 600s (c) and 4h (d).
`
`Ep(eV)
`
`Ep(eV)
`
`3
`
`

`

`208
`
`Table 2
`
`LV. Babich et al. /Applied Catalysis B: Environmental 59 (2005) 205-211
`
`Details of coking procedures, Nls BE of “N” specie and composition of gaseous products formed during oxidative regeneration
`
`Coke precursor
`
`Coking time/ageing
`time in He
`
`Nls BE (eV)*
`
`Peak area
`ratio (%)
`
`Distribution of C and N in gaseous products
`during regeneration of the cokedcatalyst (vol.%)
`
`Carbon
`As CO,
`
`As CO
`
`Nitrogen
`As N>
`
`As NO
`
`Pyridine
`Pyridine
`Pyridine
`Pyridine
`Pyridine
`Pyridine
`Pyrrole
`Aniline
`Hexadecane:pyridine (3.6:1 v/v)
`
`10 s/0s
`60 s/0s
`600 s/0 s
`4W0h
`4Wlh
`4h/6h
`4W0h
`4W0h
`4h/0h
`
`58:42
`399.9
`401.9
`32:68
`400.0
`402.1
`20:80
`400.1
`402.2
`16:84
`399.9
`402.0
`19:81
`400.0
`402.3
`12:88
`400.0
`402.0
`10:90
`399.9
`401.9
`14:86
`400.0
`401.9
`Low intensity peaks, not quantified
`
`85.4
`69.7
`60.3
`65.4
`79.4
`75.1
`62.2
`53.7
`50.0
`
`14.6
`30.3
`39.7
`34.6
`20.6
`24.9
`37.8
`46.3
`50.0
`
`61.7
`de
`85.6
`78.0
`76.6
`80
`87.0
`89.1
`81.6
`
`38.3
`24.8
`14.4
`22.0
`23.4
`20
`13.0
`10.9
`18.4
`
`* Reproducibility in the individual N1s peak position is +0.2 eV.
`
`3.2. Catalyst regeneration
`
`Coke combustion was followed by TGA(catalyst coked
`with pyridine at 550 °C for 4 h) in 5% O./He(data are not
`shown). Complete coke removal
`(by oxidation) was
`observed within the temperature interval 450-750 °C for
`different samples. The observed weightloss corresponded to
`the amount of coke determined by chemical analysis.
`Evolution of gaseous products during oxidative regen-
`eration (TPO)of catalyst coked with pyridine at 550 °C for
`4h is shown in Fig. 3. The trendis typical for other samples
`also; the main gaseous products observed during the TPO
`were CO;, CO, N; and NO. Only very small amounts NO,,
`NH:;, HCN or N2O were observed.
`Tt can be seen form Fig. 3 that the order of appearance of
`the gases with respect to temperature is CO2 = CO < N2
`< NO. This trend was the same for all samples studied.
`Further, in all cases, coke combustion started between 450
`and 500 °C and was accompanied by oxygen consumption
`and CO, and CO formation. Between 550 and 600°C in
`addition to CO/CO, formation of molecular N> was also
`observed. Maximum in N> release was observed around
`
`
`
`(ppm)
`
`N.,NO,
`
`
`
`CO2,CO(ppm)
`
`675 °C. Formation of NO occurred at higher temperatures
`(650-750 °C) coinciding with the decrease in CO and CO3.
`Quantitative data on the carbon (CO and CO>)and nitrogen
`(N2 and NO)containing gaseous products formed during coke
`combustion (TPO)are presented in Table 2.It is interesting to
`note that the relative amount of NO formed decreases from 38
`to 22% if coking time is increased from 10 s to 4 h. This is in
`line with the changes in N1s peak areas in XPS spectra of the
`coked catalysts and will be discussed in the next section.
`It can be seen from Table 2 that the relative amounts of
`CO vary inversely with those of NO. In order to check the
`influence of CO on the formation of NO/N, during
`combustion, TPO experiments were carried out with CO
`in the feed. Results of these experiments are shown in
`Table 3. It is observed that in the presence of 1 vol.% CO the
`amountof NOis about two times lowerthan if only O2 were
`present in the regeneration feed.
`In order to simulate conditions in an FCC regeneratorin the
`laboratory, isothermal coke combustion experiments were
`performed. Results of these experiments are shownin Fig. 4.
`Contacting the catalyst with 2% O2/He at 600 °C resulted in a
`rapid evolution of CO and was complete by 90-100 min. At
`this time removalofcoke from the catalyst was also complete.
`The NO peak followed the CO, peak with a time delay.
`
`4, Discussions
`
`The aim of the current study is to follow (i) the nature of
`“N”specie present in coke, (ii) the factors that influence
`
`Table 3
`
`Effect of CO on the formation of Nz and NO during combustion coked
`catalyst followed by TPO
`
`Regeneration conditions
`
`Temperature (°C)
`
`Fig. 3. Evolution of gases during temperature programmed oxidative
`regeneration of the FCC catalyst coked with pyridine (4h). Heating rate
`10 °C/min, amountof catalyst 100 mg. (@) CO; (0K) CO; (@) N2 and (+)
`NO concentration.
`
`2 vol.% O» + He (balance)
`2 vol.% O, + 1 vol.% CO + He(balance)
`
`The catalyst was coked with pyridine during 60 s.
`
`Total N products
`released (%)
`
`NO
`
`24.8
`13.9
`
`N>
`
`75.2
`86.1
`
`4
`
`

`

`209
`LV. Babichet al./Applied Catalysis B: Environmental 59 (2005) 205-211
`
`
`100
`
`Ee
`&
`Oo
`=
`
` 20
`
`o
`
`°
`2
`=
`“a
`oO
`4 6
`
`0
`
`Time (min)
`
`Ne
`
`”
`
`=
`eae
`
`= 80
`uao
`E 60
`
`22S
`
`s
`uo
`
`2>
`
`40
`
`2
`
`0
`
`« &
`. *
`NO
`
`T
`T
`T
`T
`T
`T
`T
`0.1
`0.2
`03
`0.4
`0.5
`0.6
`O7
`Ratio "Q" N / total N
`
`0
`
`Fig. 6. Correlation between the type of nitrogen specie and the selectivity
`towards NO, formation under oxidative regeneration of the FCC coked
`catalyst.
`
`is represented by nitrogen which substitutes for carbon in the
`“graphene” type structure. For FCC catalysts, coked under
`industrial conditions, such discrimination has been difficult
`dueto the low nitrogen contentin the coke. In ourexperiments,
`the N contents in coke are high because we use model
`compounds. For the sake of argument, we assume thatthis
`(large amountofhydrocarbon coke) does notaffect the nature
`of “N”specie in coke. Extensive characterization ofcatalysts
`prepared in this study, have been carried ourby Barth etal. [4].
`Using IR, MAS NMR and MALDOTOPF
`they show that N in
`coke is present as polyaromatic species such as carbazole,
`imidazole or porphyrine derivatives, adsorbed on acid sites.
`This is in agreementwith what has been suggested in literature
`[13,16], that the nature of chars and FCC coke are similar. We
`will thus use the modelused for ““N”’ specie in model chars to
`characterize catalysts coked in this study.
`In accordance with the BE observed in our case and those
`
`Fig. 4. Isothermal oxidative regeneration of the coked catalyst at 600 °C.
`Catalyst coked with pyridine at 550 °C for 4 h;gas flow 5% O2/He, switched
`at t= 0.
`
`their formation and(iii) if any correlations exist between the
`nature of the ““N” in coke and formation of N2/NO, during
`regenerative combustion of coke. This information will help
`in the modification of FCC catalyst and in optimization of
`the regeneration procedure in order to minimize NO,
`emissions during catalyst regeneration.
`Firstofall, in agreementwith whatis reported, coke formed
`from the more basic compounds such as pyridine, retains a
`higherrelative amountofnitrogen. This is significant, asall the
`NO, formed during regeneration originates from the N in coke
`(no thermal NO,is formed during regeneration). Analysis of
`cokedcatalysts by XPS allowed us todiscriminate two types of
`nitrogen specie based onthe different N1s BE (Fig. 2, Table 2)
`observed. The XPS technique has been used before to
`discriminate differenttypes ofnitrogen functionalities in coals
`and model chars. Kapteijn et al.
`[13] reported nitrogen-
`containing species in char-type materials schematically to be
`as in Fig. 5. Based on N1s BEs,different authors [10,11,13—
`16] have suggested the following ““N”specie to be present in
`chars,
`ie., pyridinic (“Py-N”)
`(Ege = 398.7 + 0.2 eV);
`pyrrolic (“Pr-N”) (Ege = 400.3 + 0.2 eV); quaternary-nitro-
`gen (Q-N,Ege = 401.4 + 0.2 eV). As shown in Fig. 5, “Q-N”
`
`
`
`reported, we ascribe the ““N”species with BE around 399.9/
`400.1 as similar to pyridinic/pyrrolic types located at the
`edges of the carbon structure shown in Fig. 5 [13]. During
`catalytic cracking,
`feed molecules or
`their cracked
`fragments are adsorbed at the catalyst surface, especially
`via the strong interaction of nitrogen with the surface acid
`(Lewis and Bronsted) sites [17]. The shift in electron density
`awayfromNto the acidsite will result in an increase in N1s
`BE.Further, based on BE, these specie are similar to the “Q-
`N”nitrogen species shown in Fig. 5. We, therefore, propose
`that the acid sites on the catalyst are responsible for the
`
`+HO7I co
`|e (1)
`OY
`
`Cc
`
`co,
`
`+HO7 NO
`[+0 (2)
`OY
`
`N
`
`NO
`
`NO+CO>%N,+CO,
`
`NO+ C>%N,+CO
`
`(3)
`
`(4)
`
`Fig. 5. Schematic representation of nitrogen functionalities in char [13].
`
`Scheme 1. Simplified reaction pathways during coke combustion.
`
`5
`
`

`

`210
`
`LV, Babich et al./Applied Catalysis B: Environmental 59 (2005) 205-211
`
`
`
`FCC catalyst
`
`initial, low T
`
`(a)
`
`final, high T
`
`(b)
`
`Scheme 2. Schematic representation of coke combustion over coked FCCcatalyst at initial (low temperature) (a) and final (high temperature) (b) stages. Main
`products are shown in black; products, which are formed to a less extent, are shown in grey.
`
`the acid sites and limited to the inner layers of coke, their
`formation of the “Q-N” type species. With increase of
`combustion at the earlier stages of the catalyst regeneration
`coking time, nitrogen-containing groups, with N1s BE close
`is limited most probably due to lack of oxygen availability.
`to pyridinic/pyrrolic structures and located at the edge of
`Atlater stages of the coke oxidation,i.e., in the deeper
`coke particle are continually formed. However, at the same
`layers, the availability of possible reductants, C and CO
`time,
`the acid sites on the catalyst surface sites are
`progressively consumed and are unavailable after a while.
`becomelimited and feasibility of the reactions (3) and (4) is
`Thus, the “Q-N” type N specie showsarelative decrease
`diminished. Further, we notice that oxidation of C in coke is
`with time in the XPS spectra. This would explain our
`more facile than N (Fig. 3, CO,, N2 and NO patterns). This
`experimental observation that, with increase in coking time
`is in agreement with data of chemical analysis of partially
`the relative intensity of this “Q-N” type specie (Nls
`regenerated industrially coked FCC catalyst [3,6], ie.,
`Eg = 401.9) actually decreases (Table 2).
`partial coke removal results in the increase nitrogen content
`In Fig.6, we see that there is a strong correlation between
`in the coke remaining. Thus, nitrogen specie, which is
`the amount of “Q-N”’
`type species and the selectivity
`combusted last during regeneration are those strongly
`towards NO,. This implies that the presence of large amount
`bound to the catalyst surface, i.e., “Q-N” type. As most of
`of “Q-N” type specie can be related to higher NO,
`the coke is already combusted at
`this point,
`lack of
`formation during regenerative combustion. Despite a large
`reductants such as CO,results in the appearance of NO, in
`number of publications regarding regeneration of coked
`the gas phase as this is not completely reduced to No
`FCC catalysts, the actual reaction mechanism is not clear
`(Scheme2b). Since coking or ageing time doesnotstrongly
`yet, but it is generally accepted that in a simplified way the
`influence the extent of formation of ““Q-N”type specie, the
`coke combustion reactions can be represented by Scheme 1
`only way to reduce NO, formed during combustionis its
`[7]. A similar mechanism was also proposed for combustion
`sequential reduction.
`of coal and model chars [8,10]. We have not included in the
`scheme, the formation of volatile nitrogen compounds like
`NH;and HCN (asdefinite intermediates), these are believed
`to be intermediates for N» and NO, formation [8,10]. We
`only observed traces of NH; and HCN in our experiments.
`In order to explain the formation ofNO, andits relation to
`“N”specie nature,it is essential to understand what happens
`during coke combustion.
`It is generally agreed that, during combustion of coal,
`chars and diesel particulates [8,13], oxidation begins from
`the outside of a coke particle. Oxidation of nitrogen-
`containing specie in the coke particle starts at the outer
`surface, which is easily accessible for oxygen (see Scheme
`2a). The pyrrolic- and pyridinic-type N specie present more
`in the outer layers are oxidized under conditions when still
`large amountof C or COis available from coke. As shown
`earlier (Scheme 1) sequential reduction of NO to N2
`(reactions (3) and (4)) by CO is feasible under these
`conditions. As a result, N, is seen in the earlier stages of
`combustion (Fig. 3, N2 pattern). As the “Q-N”type specie
`are those which are directly interacting with the catalyst via
`
`Using model compounds, it is possible to follow the
`reaction pathwaysof N present in FCC feed, as heterocyclic
`compounds, to its incorporation in coke as polyaromatic
`specie. Coke formed from the more basic compounds,e.g.,
`pyridne,
`retains a higher relative amount of nitrogen.
`Analysis of coked catalysts by XPS allows characterization
`of the N specie present in coke, namely pyridinic, pyrrolic or
`quaternary-nitrogen. The pyrrolic- and pyridinic-type N
`specie, present more in the outer layers, are oxidized under
`conditions whenstill large amount of C or CO is available
`from coke and N>is the main product. “Q-N”’ type species
`are those strongly adsorbed on the acid sites on the catalyst
`and limited to the inner layer of deposited coke. Thus, “Q-
`N” type specie is combusted last during regeneration. As
`most of the coke is already combusted at this point, lack of
`reductants, such as CO,results in the appearance of NO, in
`the gas phase. In this way, larger amount of “Q-N”’ type
`
`5. Conclusions
`
`6
`
`

`

`LV. Babichet al./Applied Catalysis B: Environmental 59 (2005) 205-211
`
`211
`
`specie present on the coked FCCcatalystis related to higher
`NO, formed during regenerative combustion of the coke.
`
`Acknowledgements
`
`The authors wish to thank Dr. Robert Ukropec for
`carrying out coking experiments. Financial support form the
`EU in the framework of “Competitive and Sustainable
`Growth Program” under contract GIRD-CT97-0065 is
`gratefully acknowledged.
`
`References
`
`[1] H. Bosch, F. Janssen, Catal. Today 2 (1988) 369.
`[2] R.H. Harding, A.W.Peters, I.R.D. Nee, Appl. Catal. A 221 (2001) 389.
`[3] X. Zhao, A.W. Peters, G.W. Weatherbee, Ind. Eng. Chem. Res. 36
`(1997) 4535.
`[4] J.-O. Barth, A. Jentys, J.A. Lercher, Ind. Eng. Chem. Res. 43 (2004)
`3097.
`
`[5] E.F. Iliopoulou, E.A. Efthimiadis, LA. Vasalos, J.-O. Barth, J.A.
`Lercher, Appl. Catal. B 47 (2004) 165.
`[6] K.L. Dishman, P.K. Doolin, L.D. Tullock, Ind. Eng. Chem. Res. 37
`(1998) 4631.
`[7] E.A. Efthimiadis, E.F. Iliopoulou, A.A. Lappas, D.K.Iatridis, LA.
`Vasalos, Ind. Eng. Chem. Res. 41 (2002) 5401.
`[8] K.M. Thomas, Fuel 76 (1997) 457.
`[9] S. Kambara, T. Takarada, M. Toyoshima, K. Kato, Fuel 74 (1995)
`1247.
`
`[10] M.A. Wojtowicz, J.R. Pels, J.A. Moulijn, Fuel 74 (1995) 507.
`[11] K. Qian, D.C. Tomezak, E.F. Rakiewicz, R.H. Harding, G. Yaluris,
`W.C. Cheng, X. Zhao, A.W. Peters, Energy Fuels 11 (1997) 596.
`[12] P.M.A. Sherwood, in: D. Briggs, M.P. Seah (Eds.), Practical Surface
`Analysis: Auger and X-ray Photoelectron Spectroscopy, vol. 1, John
`Wiley & Sons, Chichester, 1990, p. 555.
`[13] F. Kapteijn, J.A. Moulijn, 8. Matzner, H.P. Boehm, Carbon 37 (1999)
`1143.
`
`[14] D. Cagniant, R. Gruber, J.P. Boudou, C. Bilem, J. Bimer, P.D. Salbut,
`Energy Fuels 12 (1998) 672.
`[15] B. Gong, A.N. Buckley, R.N. Lamb, P.F. Nelson, Surf. Interface Anal.
`28 (1999) 126.
`[16] LR.Pels, F. Kapteijn, J.A. Moulijn, Q. Zhu, K.M. Thomas, Carbon 33
`(1995) 1641.
`[17] A.A. Khaleel, K.J. Klabunde, Chem. Eur. J. 8 (2002) 3991.
`
`7
`
`