Sie sind auf Seite 1von 8

Chemical Engineering Journal 230 (2013) 439–446

Contents lists available at SciVerse ScienceDirect

Chemical Engineering Journal


journal homepage: www.elsevier.com/locate/cej

Model diesel denitrogenation by modified activated carbon with iron


nanoparticles: Sulfur compounds effect
Javier A. Arcibar-Orozco, J. Rene Rangel-Mendez ⇑
Instituto Potosino de Investigación Científica y Tecnológica A.C., División de Ciencias Ambientales, Camino a la Presa San José 2055, Col. Lomas 4a sección, San Luis Potosí,
SLP 78216, Mexico

h i g h l i g h t s g r a p h i c a l a b s t r a c t

 Iron nanoparticles in AC enhance 15%


the adsorbents denitrogenation
capacity.
 The improvement in denitrogenation
capacity is mainly due to indoline
uptake.
 The higher indoline uptake was
related to its non-planar shape.
 Modified carbons are more selective
for nitrogen than for sulfur
compounds.
 Iron–nitrogen interactions are
stronger than iron–sulfur.

a r t i c l e i n f o a b s t r a c t

Article history: In this work, iron-modified activated carbon was studied for the removal of nitrogen compounds (indo-
Received 3 May 2013 line and quinoline) naturally present in diesel. Model diesel was simulated using a mixture of decane and
Received in revised form 24 June 2013 hexadecane that also contained the sulfur compounds dibenzothiophene and 4,6 dimethyldibenzothi-
Accepted 25 June 2013
ophene. The results showed that iron enhances the indoline uptake of carbon by 30%, while quinoline
Available online 3 July 2013
adsorption also increased, but to a lesser extent. Increased iron denitrogenation capacity was also
observed during the quinoline–indoline competition experiments; the iron-modified sample showed a
Keywords:
30% and 9% increase in indoline and quinoline uptake, respectively. When nitrogen and sulfur compounds
Adsorption
Denitrogenation
compete for adsorption sites, the indoline and quinoline uptake in the iron materials increased. In con-
Diesel trast, carbon desulfurization performance decreased in iron-modified materials, indicating that these
Activated carbon adsorbents are more selective for nitrogen than for sulfur compounds. The improved performance of
Iron nanoparticles the modified activated carbon was attributed to p–p interactions between quinoline and activated car-
bon and to the nitrogen in indoline and iron oxyhydroxides.
Ó 2013 Elsevier B.V. All rights reserved.

1. Introduction presence of nitrogen compounds [4–6]. Recent studies on the ul-


tra-deep desulfurization of model diesel have demonstrated that
Diesel fuel denitrogenation and desulfurization have garnered the presence of even small amounts of nitrogen aromatic com-
interest in recent years. One of the reasons is because ultra-low pounds has an inhibitory effect on sulfur removal [7–10]. Adsorp-
nitrogen levels in commercial diesel fuel are required by environ- tion under normal conditions has been proposed as a suitable
mental organizations around the world [1] to abate the impact of process for both deep denitrogenation and desulfurization of diesel
nitrogen oxides (NOx) released to the environment [2,3]. On the fuel [11–15]. However, in order to achieve efficient desulfurization
other hand, desulfurization technologies are often affected by the it is necessary to develop materials that can either remove nitrogen
and sulfur in a single stage or that are capable of removing nitrogen
⇑ Corresponding author. Tel.: +52 4448342000; fax: +52 4448342010. compounds selectively, prior to desulfurization. Adsorptive deni-
E-mail address: rene@ipicyt.edu.mx (J.R. Rangel-Mendez). trogenation by activated carbon is a process that has attracted

1385-8947/$ - see front matter Ó 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.cej.2013.06.102
440 J.A. Arcibar-Orozco, J.R. Rangel-Mendez / Chemical Engineering Journal 230 (2013) 439–446

attention recently [16–18], owing to the fact that carbon is gener- 2.2. Material characterization
ally cheaper than catalysts, can operate under normal temperature
and pressure conditions, do not need hydrogen, and may be easily 2.2.1. Microscope images
regenerated. However, the higher limitation of activated carbon Images of materials were obtained on a Tecnai FEI 300 Trans-
lies in its lack of selectivity for specific compounds. Almarri et al. mission Electron Microscope (TEM) operated at 300 keV. Elemental
[17] studied the interactions among quinoline, indole, and oxygen- mapping was done on the same equipment operating in Scanning-
ated groups present in carbon. It was found that the type and Transmission (STEM) mode.
amount of oxygen groups play a decisive role in nitrogen com-
pound uptake. Stronger acidic moieties were more relevant for 2.2.2. Surface area and porosity
quinoline adsorption, whereas carbonyl and quinone contributed Porosity and specific surface area of AC and AC-I were deter-
to the indole uptake. mined from the corresponding nitrogen adsorption isotherm at
Recently, Xiao et al. [4] found an interesting relationship be- 77 K obtained on an ASAP 2020 (Micromeritics) surface area and
tween total oxygen concentration in carbon and nitrogen removal porosity analyzer. Samples were degassed at 120 °C until reaching
capacity. The authors modified commercial activated carbon with 105 Torr. The specific surface area was determined using the BET
ammonium persulfate and H2SO4 at different times and tempera- method. Pore size distribution and volume of pores of different
tures, in order to selectively anchor carboxylic groups without sig- sizes (micropores Vmicro, mesopores Vmeso and Vtotal) as well as
nificant change to its material porosity. Their results showed that available surface area of pores (Smicro and Smeso) were calculated
carboxylic and anhydride groups bear the greatest responsibility by using the density functional theory (DFT).
for nitrogen uptake. These two studies illustrate the importance
of the chemical surface of carbons and the nature of nitrogen mol- 2.2.3. Ash content and inorganic composition
ecules for fuel denitrogenation. The ash content was obtained in a Versa Therm HS thermo-
Activated carbon modified with iron oxyhydroxide nanoparti- gravimetric analyzer from Thermo Khan. 30 mg of material were
cles is a very interesting material that might promote nitrogen heated from room temperature to 900 °C at a heating rate of
interactions with carbon, due to the increase in oxygen group den- 10 °C/min. The weight loss was recorded as a function of tempera-
sity after iron addition. ture. At the final temperature, all organic components in materials
In our previous work we demonstrated that it is feasible to an- were volatilized, and the ash percentage was calculated as the ratio
chor iron oxyhydroxide nanoparticles on the surface of carbon by of the initial and final weights.
using a capping agent [19]. The high surface coverage by nanopar- The inorganic composition of materials was obtained after
ticles and their very small size give special surface features to acti- digesting their ashes in a microwave (Milestone, Ethos 1). 40 mg
vated carbon that promote the adsorption of diverse molecules. of ashes were mixed with 20 mL of a HNO3:H2SO4 (5:1) solution.
The purpose of this work is to evaluate the influence of iron The microwave port was hermetically closed and heated at 10 °C/
nanoparticles on the model diesel denitrogenation when they are min from room temperature up to 150 °C for 1 h. Next, the solution
anchored on the surface of activated carbon, considering the was diluted to 50 mL with double deionized water and analyzed in
importance of both carbon surface acidity and nitrogen’s ability an ICP-AES (Varian spectrometer 730-ES).
to bond with specific metal oxide groups created on the carbon
surface. The effect of sulfur compounds on denitrogenation was 2.2.4. Surface charge distribution
also investigated. The surface charge distributions as well as the point of zero
charge of materials were obtained after a potentiometric titration
in a Metler Toledo P70 automatic titrator [20]. The experimental
procedure was as follows: 100 mg of material were placed into
2. Materials and methods contact with 50 mL of 0.01 M NaNO3. Equilibrium was reached
by stirring the solution overnight. Next, a certain amount of
All chemicals were reagent grade with the following purities: 0.1 M HCl was added until the solution reached a pH of 3. Once
quinoline (P98%), indoline (P99%), decane (P99%), n-hexadecane the pH was constant, the solution was titrated with 0.1 M NaOH.
(P99%), dibenzothiophene (P99%), and 4,6-dimehtyldibenzothi- A blank containing only the electrolyte and the same amount of
ophene (P99%), all provided by sigma–Aldrich. FeCl36H2O HCl was also titrated. During titration, the solution was continu-
(P98%) was provided by Fermont and H3PO4 (85.6%) was acquired ously purged with N2 in order to avoid interferences by atmo-
from J.T. Baker. spheric CO2. A difference in the volumes needed to reach a pH
The modified activated carbon (250–500 lm) was Filtrasorb certain value in the sample and blank yields the following charge
400 (F-400), manufactured by Calgon. This material was rinsed distribution (Eq. (1))
with double deionized water (R > 18 MX/cm until constant pH
IR ¼ C  ðV m  V b Þ=m ð1Þ
was reached, then dried overnight at 100 °C, and finally stored un-
til use. where IR is the ions released from the surface of materials; C is the
electrolyte concentration; Vm is the sample volume necessary to
achieve a certain pH value, mL; Vb is the blank volume necessary
to achieve the same pH value as the sample, mL; and m is the ti-
2.1. Material preparation trated sample mass, g.

The activated carbon (AC) was modified according to the meth- 2.3. Adsorption isotherms
odology proposed by Arcibar-Orozco and co-workers [19]: 500 mg
of AC were placed into contact with 10 mL of a solution containing Model diesel fuel was simulated by a mixture of 1:1 w/w of dec-
3 M of FeCl3 and 3 M of H3PO4 in a tightly closed flask and then ane–hexadecane. Nitrogen concentration was obtained by adding
heated to 80 °C for 56 h. Afterwards, the carbon was rinsed with several concentrations of quinoline [11,21] and indoline, which
double deionized water until no iron, chloride, or phosphorous are nitrogen compounds that can inhibit desulfurization to a great
was detected in the remaining solution. This material was identi- extent [22]. Competition experiments of quinoline and indoline
fied as AC-I. were achieved by mixing these compounds at equimassic
J.A. Arcibar-Orozco, J.R. Rangel-Mendez / Chemical Engineering Journal 230 (2013) 439–446 441

concentrations. The effect of sulfur compounds on nitrogen com-


pound removal in model diesel was simulated by adding 25 ppm S
of dibenzothiophene (DBT) and 25 ppm S of 4,6-dimethyldibenzo-
thiophene (DMDBT).
The procedure for obtaining the adsorption isotherms was as
follows: in a 25 mL flask, several masses of AC or AC-I were placed
into contact with 10 mL of model diesel containing the adsorbates
at the desired concentration. Flasks were closed and placed in a
stirred incubator at 25 °C. After equilibrium was reached (10–
15 days), the solutions were filtered and analyzed by HPLC as de-
scribed in Section 2.4.
Through a mass balance the amount of mass adsorbed by the
materials is obtained, as follows (Eq. (2)).

Q i ¼ V  ðC 0i  C eqi Þ=mad ð2Þ

where Qi is the adsorption capacity at equilibrium, mg/g, V is the


volume of solution (mL) mad is the adsorbent mass, C0i is the initial
concentration of the compound i, and Ceqi is the equilibrium concen-
tration of the compound i.
Fig. 1. Surface area distribution of studied materials described by DFT.

2.4. Compound determination should be taken with caution because the use of N2 as adsorbate
could have some disadvantages when measuring ultra-micropores
Compound analysis was conducted by high performance liquid with reference to other gases like Ar [23] or CO2 [24,25]. Since iron
chromatography (HPLC, 1200 series, Agilent Technologies) is present as very small nanoparticles, it is believed that the active
equipped with a diode array detector (DAD). The column for com- adsorption sites related to iron are mainly present in the activated
pound separations was a Purosphere STAR, Hibar 250-4,6, RP-18e carbon pores.
(5 lm). The following isocratic method was used: 10 lL of sample Moreover, the slight difference in porosity between AC and AC-I
were injected into the column at a constant flow of 0.6 mL/min. (see Table 1) allows us to suggest that the differences in adsorption
The mobile phase concentration was 10% water and 90% methanol. capacities between these two adsorbent materials can be attrib-
The compound retention time and wavelength were as follows: uted to the presence of iron nanoparticles on the surface of AC-I.
indoline (5.7 min, 254 nm), quinoline (6.2 min, 254 nm), DBT As is widely known, adsorption of compounds depends on the
(14.5 min, 235 nm), and DMDBT (26 min, 235 nm). At the end, textural properties of the adsorbent and on the chemical composi-
the composition was changed to 100% methanol to elute decane tion of the surface on which the adsorbate–adsorbent interactions
and hexadecane. are carried out. For this reason the chemical composition of mate-
rials was evaluated.
3. Results and discussion The ash content of activated carbon changed from 7.20% to
9.45% after iron deposition. A detailed analysis of raw (bituminous)
3.1. Material characterization activated carbon revealed that most of the natural inorganic impu-
rities consist of Al, Ca, Fe, and Na (Table 2). After modification, the
Micrograph images (Fig. S1 in Supplementary information) iron content increased from 0.20% to 2.35% (Table 2). The decrease
show very small nanoparticles across the surface of the AC-I. Such in the percentages of Al and Ca indicates that most impurities were
nanoparticles are in the range of 1–10 nm and are associated with washed during iron modification. Because of this, it is expected
oxygen as shown in the elemental mapping. that the contribution of those inorganic elements to the adsorption
Surface area and pore volumes of materials are summarized in of aromatic compounds will be negligible.
Table 1, and the cumulative pore distribution is reported in Supple- The charge distribution of the materials studied (Fig. 2) is rep-
mentary information, Fig. S2. The specific surface area of raw acti- resented as the sum of protons released or adsorbed from the car-
vated carbon increased by 4% after iron modification, which is bon surface. The surface of activated carbon F-400 (AC) is basic,
linked to a slight increase in micropore volume of activated carbon. with a point of zero charge at pH 11. After modification, the mate-
This increase can be attributed to the porosity created among the rial became more acidic as a result of the anchorage of iron acidic
nanoparticles incorporated onto the carbon surface. This effect particles: the point of zero charge of AC-I is 8.3. The change in the
was observed when very small iron nanoparticles were incorpo- material surface charge will affect the degree to which molecules
rated onto the surface of carbons [19]. Due to the nanosize of an- are adsorbed depending on their acidity. Basic moieties will be
chored particles, the porosity generated in carbon is reflected in more attracted by the AC-I than by the AC.
an increase of micropore surface. The incremental surface area dis-
tribution in Fig. 1 shows two effects after iron deposition. Zone 1
shows a narrowing of ultra-micropores (<7 Å) due to the incorpo- 3.2. Adsorption experiments
ration of iron nanoparticles. Zone 2 (micropores from 7 to 20 Å)
shows that the deposition of particles increased the surface area, The adsorption isotherms were correlated to the Freundlich and
due to the effect mentioned above. Nonetheless, these results Langmuir models. In order to determine which model has a better

Table 1
Physical surface properties of studied materials.

Sample SBET (m2/g) SDFT (m2/g) Vtotal (cm3/g) Vmicro (cm3/g) Vmeso (cm3/g) Smicro (m2/g) Smeso (m2/g)
AC 952 722 0.381 0.303 0.074 676 35
AC-I 995 724 0.405 0.314 0.087 670 41
442 J.A. Arcibar-Orozco, J.R. Rangel-Mendez / Chemical Engineering Journal 230 (2013) 439–446

Table 2 uptake was barely higher than that of quinoline. For example, at a
Elemental content of studied materials. quinoline concentration of 1000 mg/L (7.74 mmol/L), the molar
Sample Elemental content (%) adsorption capacity (quinoline molar weight: 129.16 mg/mmol)
Al Ca Cu Fe K P Mg Na was 0.480 mmol/g, while at the same indoline concentration the
molar indoline adsorption capacity (indoline molar weight:
AC 0.26 0.14 0.002 0.33 0.04 ND 0.019 0.266
AC-I 0.11 0.01 0.004 2.35 0.03 ND 0.003 0.004
119.16 mg/mmol) was 0.443 mmol/g. After modification, the pres-
ence of iron nanoparticles increases the indoline uptake to a great-
er extent. At an indoline equilibrium concentration of 1000 mg/L,
the AC had an adsorption capacity of 58.4 mg/g, whereas the AC-I
(iron modified) removed 30% more. This synergistic effect of iron
addition is observed also for the quinoline adsorption though to
a lesser extent (about 6%), as can be seen in Fig. 3.

3.2.2. Competitive adsorption of indoline–quinoline


Fig. 4 shows the adsorption isotherms of nitrogen compounds
when indoline and quinoline are present at the same mass concen-
tration at the beginning of the experiment. The indoline adsorption
capacity of AC and AC-I decreased to about 27% and 11%, respec-
tively, compared to their corresponding single isotherm (Fig. 3).
This indicates that AC-I has a 32% higher adsorption capacity (at
1000 mg/L) than AC (Fig. 3). The quinoline uptake decreased about
20% for both materials due to competitive adsorption; however,
AC-I had a 9% higher adsorption capacity than AC.

3.2.3. Effect of sulfur compounds on indoline and quinoline uptake


In order to study the effect of sulfur compounds on the denitro-
genation capacity of carbon, several equimassic concentrations of
quinoline and indoline were mixed with a constant sulfur concen-
Fig. 2. Proton-binding curve for released ions from the surface of studied materials. tration of 50 ppm, as mentioned in the experimental section.
Simultaneously, the uptake of sulfur compounds (DBT and DMDBT)
correlation, the variance (R2) parameter was calculated. The Fre-
undlich and Langmuir parameters as well as the R2 and the stan- was monitored and correlated as a function of total nitrogen con-
dard deviations of experiments are provided in Table S1 in centration in the starting solution.
Denitrogenation results in Fig. 5 showed that, when sulfur com-
Supplementary information. In all cases organic molecule adsorp-
tion is best described by the Freundlich model, perhaps because pounds are present in the starting solutions, the iron-modified
material has a higher indoline and quinoline adsorption capacity
of the material’s heterogeneity; hence, the adsorption isotherm re-
ported in Figs. 3–5 was adjusted by this model. For the sake of than raw activated carbon. For indoline, at a concentration of
1000 mg/L, the AC-I adsorption capacity is 53.5 mg/g, while for
comparison, all isotherms are reported in mass concentration of
the two molecules studied; in order to calculate the nitrogen con- AC it is 45 mg/g, representing an increase of about 20%. Moreover,
quinoline was removed in a larger quantity. Such an increase is
centration corresponding to each of these two organic molecules, it
is necessary to consider that quinoline and indoline have 10.8% and greater at low indoline concentrations and became less intense at
high concentrations. For example, at an equilibrium concentration
11.75% N, respectively.
of 500 mg N/L the adsorption capacity of AC-I is 30% higher than
that of AC, whereas at 1000 mg N/L the difference is just 8%.
3.2.1. Single adsorption of indoline and quinoline Fig. 6 summarizes the adsorption capacity of quinoline and
The isotherms of single indoline and quinoline adsorption onto indoline in single and competitive experiments. The adsorption
AC and AC-I are reported in Fig. 3. Raw activated carbon had a sim- capacity of AC for quinoline–indoline in competition was not af-
ilar adsorption capacity for both nitrogen compounds: the indoline fected by the presence of sulfur compounds, indicating that there

Fig. 3. Single adsorption isotherms of indoline and quinoline over studied materials at 25 °C.
J.A. Arcibar-Orozco, J.R. Rangel-Mendez / Chemical Engineering Journal 230 (2013) 439–446 443

Fig. 4. Competitive adsorption isotherms of indoline and quinoline at 25 °C.

Fig. 5. Adsorption isotherms of quinoline and indoline in competition with 50 ppm S.

is not competition for active adsorption sites between nitrogen and nitrogen concentrations (Fig. 7). It was found that desulfurization
sulfur compounds. On the other hand, the decrease in indoline up- strongly depends on nitrogen concentration. As the total nitrogen
take for AC-I during sulfur competition was higher than when concentration (nitrogen from quinoline and indoline) increased
nitrogen compounds were present. These results suggest that in in the model diesel fuel, the desulfurization capacity of both car-
the presence of iron, sulfur compounds affected nitrogen com- bons (AC and AC-I) decreased. Moreover, the iron-doped sample
pound uptake (mainly indoline). This might be attributed to com- had a lower desulfurization capacity than the unmodified carbon.
petition of sulfur and nitrogen compounds for specific iron sites in It has been reported that iron might enhance the desulfurization
carbon, or to hysteric hindrance of sulfur adsorbed molecules onto capacity of carbon [26,27]; however, those studies were conducted
the surface of activated carbon. in the absence of nitrogen compounds. Our results indicate that
On the other hand, the materials under study also showed con- iron is more akin to nitrogen than sulfur compounds. The
siderable removal of sulfur compounds in the presence of different decreased adsorption capacity of both sulfur compounds did not

Fig. 6. Adsorption capacities of indoline and quinoline of AC and AC-I in presence of a nitrogen contaminant and two sulfur compounds.
444 J.A. Arcibar-Orozco, J.R. Rangel-Mendez / Chemical Engineering Journal 230 (2013) 439–446

high nitrogen concentration (>200 mg/L), the DMDBT adsorption


capacity of carbon materials will not be further reduced.
At the maximum nitrogen concentration, the total desulfuriza-
tion and denitrogenation capacities of AC were 5.08 mg S/g and
10.12 mg N/g, respectively, while for AC-I they were 3.33 mg S/g
and 11.5 mg N/g. These results indicate that iron addition in acti-
vated carbons results in a higher denitrogenation capacity and in
a lower desulfurization uptake. Because of this, we encourage the
use of iron nanoparticles for enhancing the denitrogenation perfor-
mance of adsorbent materials. The denitrogenation capacity in the
presence of 50 ppm S of iron modified carbon (0.821 mmol N/g)
was within the range of other carbon tested for diesel denitrogen-
ation [21] and higher than other adsorbents such as activated alu-
mina, aluminum silicate, and resins [28].
It should be mentioned that the results reported herein did not
take into consideration higher nitrogen concentration because the
objective of our work was to simulate usual nitrogen concentra-
tions found in diesel fuel after hydrotreating [4,11].

3.3. Adsorption mechanism

Since both materials, AC and AC-I, have almost the same tex-
tural properties, differences in adsorption capacity should be ex-
plained by chemical interactions between the material surface
and the adsorbates.
In raw activated carbon, electronegative nitrogen may be
bonded with either polar groups on the surface of carbon or with
p–p interaction between quinoline aromatic ring and graphitic lay-
ers in carbon. Also, the surface of carbon contains oxygen groups
like carboxylic and hydroxyl [29] that may promote the formation
of hydrogen bonds with nitrogen [30]. Xiao et al. [4] reported that
quinoline adsorption onto activated carbon is probably occurring
Fig. 7. DBT and DMDBT adsorption capacities of carbon materials at several
nitrogen concentrations.
by means of p–p interactions and hydrogen bonding, rather than
by polar interactions with electronegative nitrogen.
The iron in the activated carbon is mainly present as a hydrox-
ide [31–33] or an oxide [34,35] (collectively referred to as iron oxy-
follow the same trend: the DBT uptake showed semi-linear behav- hydroxides). Therefore, on iron-modified carbon, nitrogen
ior (R2 < 0.9) for both materials, and the slope value (0.063 L*mg*- molecules might be bonded on either the activated carbon or the
mg N1*g1) indicated that each milligram-per-liter of nitrogen iron oxyhydroxide surface. Since the presence of iron increases
increase caused a decrease of 0.063 mg of DBT per gram of mate- the indoline uptake to a higher extent than the quinoline uptake,
rial. At the highest nitrogen concentration (200 mg N/L), the DBT it is believed that indoline is interacting mostly with the iron oxy-
adsorption capacity falls about 50% for the AC and 66% for AC-I, hydroxides, whereas quinoline interacts mainly with the activated
compared with the experiments that did not contain nitrogen carbon sheets.
compounds. Due to their slightly basic chemical nature, nitrogen compounds
On the other hand, the DMDBT adsorption capacity at the high- were attracted to acidic surfaces [36], as is the case of the iron-
est nitrogen concentration was 45% lower for AC and 55% lower for modified sample. It is important to consider that quinoline is a pla-
AC-I than in the absence of nitrogen compounds. Moreover, as the nar molecule due to the sp2 nitrogen configuration, while the sp3
nitrogen concentration increased the DMDBT adsorption capacity configuration of the indoline’s nitrogen allows the rotation of the
was rather more asymptotic than lineal, suggesting that at a very molecule, making polar interactions between the nitrogen and

Fig. 8. (A) p–p interactions among quinoline and graphitic layers in carbon. (B) Nitrogen in indoline is more capable of interacting with iron nanoparticles.
J.A. Arcibar-Orozco, J.R. Rangel-Mendez / Chemical Engineering Journal 230 (2013) 439–446 445

the carbon surface more feasible. Since quinoline approaches the [7] U.T. Turaga, X. Ma, C. Song, Influence of nitrogen compounds on deep
hydrodesulfurization of 4,6-dimethyldibenzothiophene over Al2O3- and
carbon surface in a parallel manner, it is most likely that the pre-
MCM-41-supported Co–Mo sulfide catalysts, Catal. Today 86 (2003) 265–275.
dominant adsorption mechanism is governed by p–p interactions [8] V. Rabarihoela-Rakotovao, F. Diehl, S. Brunet, Deep HDS of diesel fuel:
with carbon aromatic layers (Fig. 8A). On the other hand, since inhibiting effect of nitrogen compounds on the transformation of the
the indoline molecule approaches the adsorbent surface not only refractory 4,6-dimethyldibenzothiophene over a NiMoP/Al2O3 catalyst, Catal.
Lett. 129 (2009) 50–60.
in a parallel manner (Fig. 8B), it is possible that nitrogen may be [9] H. Yang, J. Chen, C. Fairbridge, Y. Briker, Y.J. Zhu, Z. Ring, Inhibition of nitrogen
bonded with specific active sites. From all hydrogen atoms located compounds on the hydrodesulfurization of substituted dibenzothiophenes in
in the indoline molecule, the one next to the nitrogen atom has the light cycle oil, Fuel Process. Technol. 85 (2004) 1415–1429.
[10] G.C.S. Laredo, J.A.H. De los Reyes, J.D. Luis Cano, J.M. Jesús Castillo, Inhibition
higher charge and is more likely to form hydrogen bonds with the effects of nitrogen compounds on the hydrodesulfurization of
surface of carbon or iron oxyhydroxides (Fig. S3 in Supplementary dibenzothiophene, Appl. Catal. A: – Gen. 207 (2001) 103–112.
information). [11] M. Macaud, M. Sévignon, A. Favre-Réguillon, M. Lemaire, E. Schulz, M. Vrinat,
Novel methodology toward deep desulfurization of diesel feed based on the
selective elimination of nitrogen compounds, Ind. Eng. Chem. Res. 43 (2004)
7843–7849.
4. Conclusions [12] A.J. Hernández-Maldonado, R.T. Yang, Denitrogenation of transportation fuels
by zeolites at ambient temperature and pressure, Angew. Chem. Int. Ed. 43
(2004) 2321.
Results indicated that iron-modified material has a higher
[13] H. Zhang, H. Song, Study of adsorptive denitrogenation of diesel fuel over
adsorption capacity for nitrogen compounds than raw activated mesoporous molecular sieves based on breakthrough curves, Ind. Eng. Chem.
carbon. Of the two nitrogen compounds studied, indoline was Res. 51 (2012) 16059–16065.
[14] X. Ma, L. Sun, C. Song, A new approach to deep desulfurization of gasoline,
more akin to iron particles.
diesel fuel and jet fuel by selective adsorption for ultra-clean fuels and for fuel
Equimassic competition of quinoline and indoline is reflected in cell applications, Catal. Today 77 (2002) 107–116.
a decrease of about 30% in the uptake of both activated carbons; [15] M. Seredych, J. Lison, U. Jans, T.J. Bandosz, Textural and chemical factors
however, the adsorption capacity of the iron-modified carbon affecting adsorption capacity of activated carbon in highly efficient
desulfurization of diesel fuel, Carbon 47 (2009) 2491–2500.
was 32% higher than that of AC. Moreover, the iron-modified mate- [16] J.H. Kim, X. Ma, A. Zhou, C. Song, Ultra-deep desulfurization and
rial has a higher capacity for nitrogen than for sulfur compounds. denitrogenation of diesel fuel by selective adsorption over three different
However, when sulfur is present in the starting solution, the adsorbents: a study on adsorptive selectivity and mechanism, Catal. Today 111
(2006) 74–83.
iron-modified carbon has a 30% lower nitrogen adsorption capacity [17] M. Almarri, X. Ma, C. Song, Role of surface oxygen-containing functional
than AC. groups in liquid-phase adsorption of nitrogen compounds on carbon-based
The increase in the uptake of indoline by iron-containing mate- adsorbents, Energ. Fuel. 23 (2009) 3940–3947.
[18] J. Wen, X. Han, H. Lin, Y. Zheng, W. Chu, A critical study on the adsorption
rials could be related to the non-planar shape of this molecule, of heterocyclic sulfur and nitrogen compounds by activated carbon:
which allows direct interaction among nitrogen and oxygen-con- equilibrium, kinetics and thermodynamics, Chem. Eng. J. 164 (2010) 29–
taining surface groups. It is also hypothesized that such interac- 36.
[19] J.A. Arcibar-Orozco, M. Avalos-Borja, J.R. Rangel-Mendez, Effect of phosphate
tions might be governed by hydrogen bonding. on the particle size of ferric oxyhydroxides anchored onto activated
carbon: As(V) removal from water, Environ. Sci. Technol. 46 (2012) 9577–
9583.
Acknowledgments [20] K. Bourikas, J. Vakros, C. Kordulis, A. Lycourghiotis, Potentiometric mass
titrations: experimental and theoretical establishment of a new technique for
determining the point of zero charge (PZC) of metal (Hydr)oxides, J. Phys.
This project was funded by Ciencia Basica (SEP-CB-2008-01- Chem. B 107 (2003) 9441–9451.
105920). J.A. appreciates the scholarship from CONACYT [21] M. Almarri, X. Ma, C. Song, Selective adsorption for removal of nitrogen
(232598). We also thank the LANBAMA National Laboratory, as compounds from liquid hydrocarbon streams over carbon- and alumina-based
adsorbents, Ind. Eng. Chem. Res. 48 (2008) 951–960.
well as to Dulce Partida, Guillermo Vidriales and Maria del Carmen
[22] A.R. Beltramone, S. Crossley, D.E. Resasco, W.E. Alvarez, T.V. Choudhary,
Rocha for their technical support. We thank Jennifer Eckerly for her Inhibition of the hydrogenation and hydrodesulfurization reactions
assistance in the English Language. by nitrogen compounds over NiMo/Al2O3, Catal. Lett. 123 (2008) 181–
185.
[23] P.D. Paulsen, B.C. Moore, F.S. Cannon, Applicability of adsorption equations to
argon, nitrogen and volatile organic compound adsorption onto activated
Appendix A. Supplementary material
carbon, Carbon 37 (1999) 1843–1853.
[24] D. Lozano-Castelló, D. Cazorla-Amorós, A. Linares-Solano, Usefulness of CO2
Supplementary data associated with this article can be found, in adsorption at 273 K for the characterization of porous carbons, Carbon 42
the online version, at http://dx.doi.org/10.1016/j.cej.2013.06. (2004) 1233–1242.
[25] J. Garrido, A. Linares-Solano, J.M. Martin-Martinez, M. Molina-Sabio, F.
102. Rodriguez-Reinoso, R. Torregrosa, Use of nitrogen vs. carbon dioxide in the
characterization of activated carbons, Langmuir 3 (1987) 76–81.
[26] M. Seredych, T.J. Bandosz, Adsorption of dibenzothiophenes on activated
References carbons with copper and iron deposited on their surfaces, Fuel Process.
Technol. 91 (2010) 693–701.
[1] C. Song, X. Ma, New design approaches to ultra-clean diesel fuels by deep [27] J.A. Arcibar-Orozco, J.R. Rangel-Mendez, T.J. Bandosz, Desulfurization of model
desulfurization and deep dearomatization, Appl. Catal. B – Environ. 41 (2003) diesel fuel on activated carbon modified with iron nanoparticles: effect
207–238. of tert-butylbenzene and naphthalene concentrations, Submitted,
[2] U. EPA, Technical Bulletin: Nitrogen oxides (NOx), why and how they are 2013.
controlled, in: O.o.a. Quality (Ed.), Research Triangle Park, NC 27711, [28] G.C. Laredo, P.M. Vega-Merino, F. Trejo-Zárraga, J. Castillo, Denitrogenation of
November, 1999. middle distillates using adsorbent materials towards ULSD production: a
[3] G.W. Mushrush, M.A. Quintana, J.W. Bauserman, H.D. Willauer, Post-refining review, Fuel Process. Technol. 106 (2013) 21–32.
removal of organic nitrogen compounds from diesel fuels to improve [29] H.P. Boehm, Surface oxides on carbon and their analysis: a critical assessment,
environmental quality, J. Environ. Sci. Heal. A, Part A 46 (2011) 176–180. Carbon 40 (2002) 145–149.
[4] J. Xiao, C. Song, X. Ma, Z. Li, Effects of aromatics, diesel additives, nitrogen [30] N. Li, J. Zhu, X. Ma, Q. Zha, C. Song, Tailoring of surface oxygen-containing
compounds, and moisture on adsorptive desulfurization of diesel fuel over functional groups and their effect on adsorptive denitrogenation of liquid
activated carbon, Ind. Eng. Chem. Res. 51 (2012) 3436–3443. hydrocarbons over activated carbon, AIChE J., 2012.
[5] P. Zeuthen, K.G. Knudsen, D.D. Whitehurst, Organic nitrogen compounds in gas [31] C. Nieto-Delgado, J.R. Rangel-Mendez, Anchorage of iron hydro(oxide)
oil blends, their hydrotreated products and the importance to hydrotreatment, nanoparticles onto activated carbon to remove As(V) from water, Water Res.
Catal. Today 65 (2001) 307–314. 46 (2012) 2973–2982.
[6] A. Stanislaus, A. Marafi, M.S. Rana, Recent advances in the science and [32] V. Fierro, G. Muñiz, G. Gonzalez-Sánchez, M.L. Ballinas, A. Celzard, Arsenic
technology of ultra low sulfur diesel (ULSD) production, Catal. Today 153 removal by iron-doped activated carbons prepared by ferric chloride forced
(2010) 1–68. hydrolysis, J. Hazard. Mater. 168 (2009) 430–437.
446 J.A. Arcibar-Orozco, J.R. Rangel-Mendez / Chemical Engineering Journal 230 (2013) 439–446

[33] W.R. Meyer, S.H. Pulcinelli, C.V. Santilli, A.F. Craievich, Formation of colloidal [35] W. Wang, J.Y. Howe, B. Gu, Structure and morphology evolution of hematite
particles of hydrous iron oxide by forced hydrolysis, J. Non-Cryst. Solids 273 (a-Fe2O3) nanoparticles in forced hydrolysis of ferric chloride, J. Phys. Chem. C
(2000) 41–47. 112 (2008) 9203–9208.
[34] D. Daichuan, H. Pinjie, D. Shushan, Preparation of uniform a-Fe2O3 particles by [36] E. Furimsky, F.E. Massoth, Deactivation of hydroprocessing catalysts, Catal.
microwave-induced hydrolysis of ferric salts, Mater. Res. Bull. 30 (1995) 531– Today 52 (1999) 381–495.
535.