Rights / License: Research Collection In Copyright - Non … · · 2017-07-15Diss. ETH No. 15555...
Transcript of Rights / License: Research Collection In Copyright - Non … · · 2017-07-15Diss. ETH No. 15555...
Research Collection
Doctoral Thesis
Mechanisms of hydrodenitrogenation of amines over sulfidedNiMo, CoMo, and Mo supported on Al₂O₃
Author(s): Zhao, Yonggang
Publication Date: 2004
Permanent Link: https://doi.org/10.3929/ethz-a-004779056
Rights / License: In Copyright - Non-Commercial Use Permitted
This page was generated automatically upon download from the ETH Zurich Research Collection. For moreinformation please consult the Terms of use.
ETH Library
Diss. ETH No. 15555
Mechanisms of Hydrodenitrogenation of Amines Over
Sulfided NiMo, CoMo, and Mo Supported on Al2O3
A dissertation submitted to the
SWISS FEDERAL INSTITUTE OF TECHNOLOGY ZURICH
for the degree of
DOCTOR OF CHEMICAL ENGINEERING
Presented by
Yonggang Zhao
Bachelor and Master of Chemical Engineering, Fushun Petroleum Institute Born on 21 September 1974
in Jiangsu, China
Accepted on the recommendation of
Prof. Dr. Roel Prins, examiner Prof. Dr. Michèle Breysse, co-examiner
Zürich, 2004
I
CONTENTS
Abstract................................................................................................................................ V
Zusammenfassung ........................................................................................................... VII
1 Introduction.......................................................................................................................... 1
1.1 Hydrodenitrogenation .................................................................................................. 1
1.2 Structure of the catalysts.............................................................................................. 3
1.3 The role of H2, H2/H2S and active sites ....................................................................... 8
1.4 Reaction mechanism.................................................................................................. 12
1.4.1 Aromatic C-N bond breaking............................................................................ 12
1.4.2 C-N bond cleavage of aliphatic amines ............................................................ 15
1.4.3 C-N bond cleavage of hetero cyclic amines ..................................................... 18
1.5 Experimental.............................................................................................................. 21
1.5.1 Unit ................................................................................................................... 21
1.5.2 Product analysis ................................................................................................ 24
1.5.3 Catalysts............................................................................................................ 24
1.5.4 Weight time....................................................................................................... 25
1.5.5 Reactants and reaction conditions used in every chapter ................................. 25
1.6 References.................................................................................................................. 28
2 On the role of β-hydrogen atoms in the hydrodenitrogenation of 2-methylpyridine
and 2-methylpiperidine ..................................................................................................... 35
2.1 Abstract...................................................................................................................... 35
2.2 Introduction................................................................................................................ 35
2.3 Results........................................................................................................................ 37
2.3.1 HDN of 2-methylpyridine................................................................................. 37
2.3.2 HDN of 2-methylpiperidine.............................................................................. 39
2.3.3 Comparison of piperidine, 2-methylpyridine, and 2,6-dimethylpiperidine ... 40
2.3.4 HDN of 1-aminohexane.................................................................................... 43
II
2.3.5 HDN of 2-aminohexane ....................................................................................45
2.4 Discussion ..................................................................................................................47
2.5 References ..................................................................................................................51
3 Investigation of the mechanism of the hydrodenitrogenation of n-hexylamine over
sulfided NiMo/γ-Al2O3 .......................................................................................................53
3.1 Abstract ......................................................................................................................53
3.2 Introduction ................................................................................................................53
3.3 Results ........................................................................................................................55
3.3.1 HDS of pentanethiol and hydrogenation of hexene ..........................................55
3.3.2 HDN of hexylamine ..........................................................................................58
3.3.3 HDN of dihexylamine .......................................................................................63
3.3.4 HDN of trihexylamine.......................................................................................66
3.4 Discussion ..................................................................................................................69
3.4.1 Hexylamine .......................................................................................................69
3.4.2 Dihexylamine ....................................................................................................71
3.4.3 Trihexylamine ...................................................................................................74
3.4.4 General discussion.............................................................................................75
3.5 Conclusions ................................................................................................................78
3.6 References ..................................................................................................................79
4 Mechanisms of the hydrodenitrogenation of alkylamines with secondary and tertiary
α-carbon atoms over sulfided NiMo/γ-Al2O3 ...................................................................81
4.1 Abstract ......................................................................................................................81
4.2 Introduction ................................................................................................................81
4.3 Results ........................................................................................................................82
4.3.1 2-Pentylamine and 2-pentanethiol.....................................................................82
4.3.2 3-Methyl-2-butylamine and 3-methyl-2-butanethiol ........................................85
4.3.3 3,3-Dimethyl-2-butylamine...............................................................................88
4.3.4 2-Methylcyclohexylamine.................................................................................90
4.3.5 2-Methyl-2-butylamine and 2-methyl-2-butanethiol ........................................91
4.4 Discussion ..................................................................................................................94
III
4.4.1 HDN and HDS mechanism............................................................................... 94
4.4.1.1 Acid-base mechanism........................................................................... 94
4.4.1.2 Metal-like mechanism........................................................................... 96
4.4.2 HDN of amines with secondary α-carbon atoms .............................................. 99
4.4.3 HDN of 2-methyl-2-butylamine and benzylamine ......................................... 102
4.5 Conclusions.............................................................................................................. 103
4.6 References................................................................................................................ 105
5 Mechanisms of HDN of Alkylamines and HDS of alkanethiol on NiMo/Al2O3,
CoMo/Al2O3, and Mo/Al2O3 ......................................................................................... 107
5.1 Abstract.................................................................................................................... 107
5.2 Introduction.............................................................................................................. 107
5.3 Results...................................................................................................................... 109
5.3.1 Simultaneous reaction of pentylamine and hexanethiol ................................. 111
5.3.2 Simultaneous reaction of 2-hexylamine and 2-pentanethiol........................... 116
5.3.3 2-Methyl-2-butylamine and 2-methyl-2-butanethiol ...................................... 122
5.4 Discussion................................................................................................................ 124
5.4.1 Pentylamine..................................................................................................... 124
5.4.2 2-Hexylamine.................................................................................................. 126
5.4.3 2-Methyl-2-butylamine................................................................................... 128
5.5 Conclusion ............................................................................................................... 130
5.6 References................................................................................................................ 131
6 Mechanism of the hydrodenitrogenation of adamantylamine and neopentylamine
over sulfided NiMo/γ-Al2O3............................................................................................. 133
6.1 Abstract.................................................................................................................... 133
5.2 Introduction.............................................................................................................. 133
6.3 Results...................................................................................................................... 134
6.3.1 Neopentylamine .............................................................................................. 134
6.3.2 Adamantylamines and adamantanethiol ......................................................... 136
6.4 Discussion................................................................................................................ 138
6.4.1 HDN of neopentylamine ................................................................................. 138
IV
6.4.2 HDN of AdNH2 and HDS of AdSH................................................................140
6.5 Conclusion................................................................................................................143
6.6 References ................................................................................................................144
7 Mechanism of the direct hydrodenitrogenation of naphthylamine over sulfided
NiMo/γ-Al2O3 ....................................................................................................................147
7.1 Abstract ....................................................................................................................147
7.2 Introduction ..............................................................................................................147
7.3 Results ......................................................................................................................149
7.4 Discussion ................................................................................................................154
7.4.1 Direct denitrogenation.....................................................................................154
7.4.2 Direct desulfurization......................................................................................160
7.5 Conclusions ..............................................................................................................162
7.6 References ................................................................................................................162
8 Concluding remarks.........................................................................................................165
8.1 Conclusion................................................................................................................165
8.2 Outlook.....................................................................................................................167
8.3 References ................................................................................................................170
Acknowledgements
Publications
Curriculum Vitae
V
Abstract
The hydrodenitrogenation (HDN) of alkylamines has been studied over sulfided NiMo,
CoMo, and Mo catalysts supported on γ-Al2O3 at reaction conditions in the range of 3-5 MPa,
270-350 °C, and 10-100 kPa H2S. The heterocyclic amines 2-methylpyridine and 2,6-
dimethylpyridine as well as the alkylamines of n-hexylamine, 2-pentylamine, and 2-methyl-2-
butylamine were chosen as HDN models. The corresponding alkanethiols were studied as
well.
In the HDN of 2-methylpiperidine, the products 2-methylpyridine, 3,4,5,6-
tetrahydropyridine, and 2-hexylamine as well as hydrocarbons were formed. Only a trace of
hexylamine was observed. As the reactivity of 2-hexylamine was much higher than that of
hexylamine at the same reaction conditions, the ring opening of 2-methylpiperidine occurred
preferentially between the nitrogen atom and the methylene group, instead of between the
nitrogen atom and the carbon atom bearing the methyl group. It demonstrates that β-hydrogen
atoms are not involved in the HDN of 2-methylpiperidine. This was confirmed with the much
higher HDN conversion of piperidine than 2-methylpiperidine and 2,6-dimethylpiperidine.
The HDN of hexylamine, dihexylamine, and trihexylamine was studied between 300 and
340 °C, 3 and 5 MPa total pressure, 5 and 20 kPa amine pressure, and 10 and 150 kPa H2S
pressure over a sulfided Ni-Mo/γ-Al2O3 catalyst. The conversion of hexylamine and
dihexylamine decreased slightly with H2S pressure, but that of trihexylamine increased
substantially. The conversion increased with the H2 pressure and decreased with increasing
partial pressure of the hexylamines. In the HDN of n-hexylamine, a substantial amount of
hexenes and hexanethiol was formed by elimination and nucleophilic substitution. Two
methods were used to distinguish between elimination and nucleophilic substitution. One is to
test the initial product selectivities at short weight time. The initial alkene selectivities were
low and accounted for only a minor part of the n-alkylamine conversion. Furthermore, the
simultaneous reactions of hexylamine and pentanethiol show that the hexenes/hexane ratio in
the HDN of the hexylamine was almost equal to the pentenes/pentane ratio in the
hydrodesulfurization (HDS) of pentanethiol. Therefore, it was concluded that the majority of
the hexene in the HDN of hexylamine originates from hexanethiol formed by nucleophilic
substitution of hexylamine with H2S.
VI
The HDN of alkylamines with secondary and tertiary α-carbon atoms and benzylamine
as well as the HDS of the corresponding alkanethiols was studied over sulfided NiMo/Al2O3
CoMo/Al2O3 and Mo/Al2O3 catalysts. The similar alkenes/alkane ratios in the HDN of the
alkylamines and HDS of the corresponding alkanethiols confirm that nucleophilic substitution
is the predominant reaction in the HDN of the amine group attached to secondary carbon
atoms. 2-Methyl-2-butylamine and benzylamine reacted much faster than the amines with
secondary α-carbon atoms. In the HDN of 2-methyl-2-butylamine and benzylamine,
hydrocarbons were formed. Only a trace of the relevant thiol was formed and this amount did
not increase much with increasing H2S pressure. The much higher
methylbutenes/methylbutane ratios in the HDN of 2-methyl-2-butylamine as in the HDS of 2-
methyl-2-butanethiol demonstrate that 2-methyl-2-butylamine and the activated benzylamine
react by means of an E1 mechanism.
The HDN of 1-adamantylamine, 2-adamantylamine and neopentylamine and the HDS of
1-adamantanethiol were studied. The adamantanethiols and neopentanethiol were formed as
the primary products of the adamantylamines and neopentylamine by substitution of the NH2
group with H2S. Adamantane and neopentane were secondary products, demonstrating that
hydrogenolysis can hardly take place. Furthermore, elimination cannot take place and a
classic SN2 substitution is not possible for the adamantylamines either. It is proposed that the
NH2-SH substitution in adamantylamine takes place by adsorption of the amine group at the
metal sulfide surface and by shifting the adamantyl group to a neighbouring sulfur atom.
1-Naphthylamine was studied to test the direct hydrogenolysis in hydrodenitrogenation.
Tetralin, naphthalene, 1,2-dihydronaphthalene and 5,6,7,8-tetrahydro-1-naphthylamine were
formed in the HDN of naphthylamine. The reactions of the intermediates 1,2,3,4-tetrahydro-
1-naphthylamine, 1,2-dihydronaphthalene and 5,6,7,8-tetrahydro-1-naphthylamine were
studied as well. 1-Naphthylamine reacts to 1,2,3,4-tetrahydro-1-naphthylamine by
hydrogenation, and then it reacts by NH3 elimination to 1,2-dihydronaphthalene. 1,2-
dihydronaphthalene subsequently reacts to tetralin as well as naphthalene. This direct
denitrogenation of naphthylamine to naphthalene could take place by hydrogenation of 1-
naphthylamine to 1,2-dihydro-1-naphthylamine, followed by NH3 elimination or followed by
a Bucherer-type NH2-SH exchange, dehydrogenation and C-S bond hydrogenolysis.
VII
Zusammenfassung
Die Hydrodenitrogenierung (HDN) von Alkylaminen wurde über sulfidierten, auf γ-
Al2O3 geträgerten NiMo, CoMo, und Mo Katalysatoren unter Reaktionsbedingungen von 3-5
MPa, 270-350 °C und 10-100 kPa H2S untersucht. Als Modellverbindungen für die HDN
wurden sowohl die heterocyclischen Amine 2-Methylpyridin und 2,6-Dimethylpyridin als
auch die Alkylamine n-Hexylamin, 2-Pentylamin, und 2-Methyl-2-butylamin gewählt.
Ebenfalls untersucht wurden die entsprechenden Alkylthiole.
Während der HDN von 2-Methylpiperidin wurden die Produkte 2-Methylpyridin, 3,4,5,6-
Tetrahydropyridin und 2-Hexylamin sowie Kohlenwasserstoffe gebildet. Hexylamin ist nur in
Spuren beobachtet worden. Da die Reaktivität von 2-Hexylamin unter denselben
Reaktionsbedingungen viel höher war als die von Hexylamin, fand die Ringöffnung von 2-
Methylpiperidin bevorzugt zwischen dem Stickstoffatom und der Methylengruppe statt,
anstatt zwischen dem Stickstoffatom und dem methylsubstituierten Kohlenstoff. Dies zeigt,
dass β-Wasserstoffatome in der HDN von 2-Methylpiperidin nicht involviert sind. Der viel
höhere HDN Umsatz von Piperidin verglichen mit dem Umsatz von 2-Methylpiperidin
bestätigt dies.
Die HDN von Hexylamin, Dihexylamin und Trihexylamin wurde über sulfidiertem Ni-
Mo/γ-Al2O3 Katalysator zwischen 300-340 °C, 3-5 MPa Totaldruck, 5-20 kPa
Aminpartialdruck und 10-150 kPa Schwefelwasserstoffpartialdruck untersucht. Während der
Umsatz von Hexylamin und Dihexylamin mit zunehmendem H2S Druck leicht abnahm,
nahm derjenige von Trihexylamin deutlich zu. Der Umsatz nahm mit zunehmendem H2 Druck
zu und mit zunehmendem Partialdruck der Hexylamine ab. Während der HDN von n-
Hexylamin bildete sich durch Eliminierung und nucleophile Substitution eine erhebliche
Menge an Hexenen und Hexanthiol. Für die Unterscheidung zwischen Eliminierung und
nucleophiler Substitution wurden zwei Methoden angewendet. Die eine besteht darin, die
anfänglichen Produktselektivitäten nach kurzer Kontaktzeit zu untersuchen. Die anfänglichen
Alkenselektivitäten waren klein und trugen nur wenig zum Umsatz von n-Alkylamin bei.
Weiter zeigte die gleichzeitige Reaktion von Hexylamin und Pentanthiol, dass das
Hexene/Hexan Verhältnis in der HDN von Hexylamin beinahe gleich war wie das Verhältnis
Pentene/Pentan in der Hydrodesulfurisierung (HDS) von Pentanthiol. Daraus wurde
VIII
geschlossen, dass die Mehrheit des in der HDN von Hexylamin auftretenden Hexens von
Hexanthiol stammen muss, welches durch nucleophile Substitution von Hexylamin mit H2S
entsteht.
Sowohl die HDN von Alkylaminen mit sekundären und tertiären α-Kohlenstoffatomen
und von Benzylamin, als auch die HDS der entsprechenden Alkanthiolen, wurde über
sulfidierten NiMo/Al2O3, CoMo/Al2O3 und Mo/Al2O3 Katalysatoren untersucht. Die
ähnlichen Alkene/Alkan Verhältnisse während der HDN der Alkylamine und während der
HDS der entsprechenden Thiole bestätigen, dass nucleophile Substitution der bevorzugte
Reaktionsmechanismus der HDN der Aminogruppe am sekundären Kohlenstoffatom ist. 2-
Methyl-2-butylamin und Benzylamin reagierten viel schneller als die Amine mit sekundären
α-Kohlenstoffatomen. Während der HDN von 2-Methyl-2-butylamin und Benzylamin wurde
die Bildung von Kohlenwasserstoffen beobachtet. Das relevante Thiol wurde nur in Spuren
gebildet, und dessen Menge nahm mit steigendem H2S Druck nur wenig zu. Das viel höhere
Methylbutene/methylbutan Verhältnis während der HDN von 2-Methyl-2-butylamin als
während der HDS von 2-Methyl-2-butanthiol zeigt, dass 2-Methyl-2-butylamin und das
aktivierte Benzylamin nach einem E1 Mechanismus reagieren.
Die HDN von 1-Adamantylamin, 2-Adamantylamin und Neopentylamin und die HDS
von 1-Adamantanthiol wurden ebenfalls untersucht. Durch Substitution der NH2 Gruppe mit
H2S wurden als primäre Produkte die Adamantanthiole und Neopentanthiol gebildet.
Adamantan und Neopentan waren sekundäre Reaktionsprodukte, was zeigt, dass
Hydrogenolyse praktisch nicht stattfinden kann. Im weiteren ist für die Adamantylamine
weder eine Elimination noch eine klassische SN2 Substitution möglich. Es ist vorgeschlagen,
dass die NH2-SH Substitution an Adamantylamin über Adsorption der Aminogruppe auf der
Metallsulfidoberfläche und anschliessende Verschiebung der Adamantylgruppe zu einem
benachbahrten Schwefelatom abläuft.
Um die direkte Hydrogenolyse in der HDN zu testen, wurde 1-Naphthylamin untersucht.
Während der HDN von Naphthylamin wurden Tetralin, Naphthalin, 1,2-Dihydronaphthalin
und 5,6,7,8-Tetrahydro-1-naphthylamin gebildet. Die Reaktionen der Zwischenprodukte
1,2,3,4-Tetrahydro-1-naphthylamin, 1,2-Dihydronaphthalin und 5,6,7,8-Tetrahydro-1-
naphthylamin wurden ebenfalls untersucht. 1-Naphthylamin reagiert zuerst über
Hydrogenolyse zu 1,2,3,4-Tetrahydro-1-naphthylamin und dann über NH3 Eliminierung zu
1,2-Dihydronaphthalin. 1,2-Dihydronaphthalin reagiert dann zu Tetralin und Naphthalin.
IX
Diese direkte Entstickung von Naphthylamin zu Naphthalin könnte über Hydrierung von 1-
Naphthylamin zu 1,2-Dihydro-1-naphthylamin und anschliessende NH3 Eliminierung oder
über anschliessenden Austausch von NH2-SH nach Bucherer, Hydrierung und Hydrogenolyse
der C-S Bindung ablaufen.
Introduction Chapter 1 1
1. Introduction
1.1 Hydrodenitrogenation
In an oil refinery, hydrotreating is a very important process to diminish the contents of
sulfur, nitrogen, and metals in the oil fraction so that fewer air-polluting emission of sulfur
and nitrogen oxides are formed when these oil fractions are burned in cars and trucks.
Furthermore, most catalysts used in a refinery for the processing of oil fractions cannot
tolerate sulfur and metals. The main reactions in hydrotreating are hydrodesulfurization
(HDS), hydrodenitrogenation (HDN), hydrodeoxygenation (HDO) and hydrodemetalization
(HDM). It has long been recognized that HDN is more difficult than HDS. People did not pay
much attention to HDN in the old days as only comparatively small quantities of nitrogen
compounds present in conventional petroleum feedstocks. However, this situation has
changed because of the growing need to process heavy and low-quality feed stocks, which are
rich in highly refractory nitrogen compounds. Furthermore, environmental legislation requires
a deep reduction of the sulfur content of gasoline and diesel fuel to 50 ppm by the year 2005.
Nitrogen-containing compounds inhibit the hydrodesulfurization (HDS) of sulfur-containing
compounds through competitive adsorption. At the low level of sulfur that must be reached in
deep HDS, however, nitrogen compounds will be harmful. Therefore, it is important to
understand HDN. In recent years, there has been considerable interest in the development of
more effective HDN catalysts [1], as witnessed by the rapidly expanding literature in this area
[2-10].
Industrial hydrotreating catalysts contain molybdenum and cobalt or nickel, supported
on γ-Al2O3. Since oil fractions always contain sulfur, a metal or metal oxide that would be
introduced as the catalyst would quickly become sulfided by the H2S that is produced during
hydrotreating. In practice, one therefore sulfides supported metal oxides under controlled
conditions before starting the hydrotreating process. When supported alone on alumina, Mo
sulfide has a much higher activity for the removal of S, N and O atoms than Co and Ni
sulfide. Therefore, molybdenum sulfide is traditionally considered to be the catalyst. Sulfided
Co-Mo/Al2O3 and Ni-Mo/Al2O3 catalysts, on the other hand, have substantially higher
catalytic activity than Mo/Al2O3. As a consequence, cobalt and nickel are referred to as the
Introduction Chapter 1 2
promoters of the Mo activity [4,7,10,11]. Cobalt is used mainly as a promoter for sulfided
Mo/Al2O3 in HDS, while nickel is the choice for HDN. In addition to molybdenum and cobalt
or nickel, hydrotreating catalysts often contain additives such as phosphorus, boron, fluorine
or chlorine, which may influence the catalytic as well as the mechanical properties of the
catalyst [12-15]. The chemical properties of tungsten are similar to those of Mo; tungsten is,
however, more expensive and its industrial use is, therefore, limited. Especially when
hydrogenation is carried out under severe conditions, such as in hydrocracking, sulfided Ni-
W/Al2O3 catalysts have advantages over sulfided Ni-Mo/Al2O3 catalysts. The concentration of
the metals is usually 1-4 wt% for Co and Ni, 8-16 wt% for Mo and 12-25 wt% for W. The
NiW catalysts have the highest activity for aromatic hydrogenation at low hydrogen sulfide
partial pressure of the three catalyst combinations mentioned above [16].
Hydrotreating catalysts originated in the 1920s when German researchers developed
unsupported metal sulfide catalysts to liquefy coal. However, it was not until the 1970s that
the structures of these catalysts and the mechanisms of their catalytic action began to be
understood. It was established that under catalytic reaction conditions, most of the
molybdenum in industrial hydrotreating catalysts is present as small MoS2 particles in the
pores of the γ-Al2O3. It was not until the 1980s that the location of the cobalt and the nickel
promoter ions in the hydrotreating catalysts was more or less determined. The role of
phosphate and fluorine additives is still under investigation. Supports other than γ-Al2O3 like
amorphous silica-alumina, are also used in commercial units and their functions are topics of
academic and industrial research.
Typical hydrotreating catalysts need to be sulfided before achieving their active state.
The sulfiding is traditionally done during the start-up phase by exposing the catalyst to the
sulfur-containing feed or by adding H2S to the hydrogen. The sulfiding procedures have
significant influence on the catalyst activity and stability. After a period of use, the catalysts
will lose activity. There are several reasons for this. One is sintering of the active phase; the
second is decomposition of the active phase; the third is the covering of the active sites by
reactants and products; the fourth is coking, and the fifth is the deposition of metal sulfides.
The fourth and the fifth are the most important causes of activity loss. Regeneration of the
used catalysts is possible, however. It is important that the burn-off is carefully controlled to
avoid catalyst overheating, which could irreversibly change the active phase of the catalyst.
Introduction Chapter 1 3
1.2 Structure of the catalysts
The preparation of hydrotreating catalysts can be done by a two-step pore volume
impregnation procedure or by co-impregnation. In the sequential pore volume impregnation,
γ-Al2O3 is impregnated with an aqueous solution of (NH4)6Mo7O24, followed by drying and
calcination (heating in air). In a second step, the resulting material is impregnated with an
aqueous solution of Co(NO3)2 or Ni(NO3)2 and dried and calcined. Alternatively, and
preferentially used in the industry, all inorganic materials are co-impregnated, and the
resulting catalyst precursor is dried and calcined. Several studies indicated that there are
interactions between Mo and Ni or Co in the oxidic state. Thus, it is known that the order of
impregnation and calcination - first Mo and then Co or Ni or vice versa - plays an important
role in the activity of the final sulfided catalyst. Catalysts in which the support is impregnated
first with a solution containing Mo invariably have a higher activity. It has been suggested
that the nickel or cobalt cations interact with the polymolybdate phase by forming a metal
heteropolymolybdate [17,18]. Several publications deal with this project: For instance, the
infrared absorption bands of NO adsorbed on Co-Mo/Al2O3 are shifted from those of NO on
Co/Al2O3 [19], and Raman bands indicating polymeric molybdenum oxide species decrease in
intensity with increasing the cobalt loading in an oxidic Co-Mo/Al2O3 catalyst [20]. These
results suggest that nickel or cobalt cations interact especially with the most polymeric
molybdenum oxide species to form species in which nickel or cobalt and molybdenum
interact. In this way the promoter cations stay at the surface and close to the molybdenum
cations and are positioned to form the active Ni-Mo-S structure during sulfidation.
Furthermore, the promoter ions interact to a lesser extent with the support and thus can be
used more efficiently after sulfidation.
After the impregnation, drying, and calcination steps, the oxidic hydrotreating catalyst is
formed. The real hydrotreating catalysts will be sulfided in a mixture of H2 with one or more
sulfur-containing compounds such as H2S, CS2, dimethyl disulfide, thiophene, and even
elemental sulfur. The oil fraction to be treated can be used for the sulfidation as well. The
properties of the final sulfidic catalyst depend to a great extent on the calcination and
sulfidation conditions. High-temperature calcination induces a strong interaction between
molybdenum and nickel or cobalt cations and the alumina support. Consequently, it is
difficult to transform the oxidic species into sulfides. Mössbauer spectroscopy of Co-
Introduction Chapter 1 4
Mo/Al2O3 catalysts showed that at increasingly high calcination temperatures, increasingly
more of Co2+ ions are incorporated into the bulk of the alumina, forming a spinel [21]. The
higher the calcination temperature, the higher the sulfidation temperature needed to bring
these cations back to the surface to provide a high catalytic activity for hydrotreating. When
the sulfidation temperature is too high, the metal sulfides sinter or do not form the
catalytically active Co-Mo-S phase. Optimum calcination and sulfidation temperatures are in
the range 673-773 K for Al2O3-supported catalysts [22].
The sulfidation mechanism was investigated by temperature-programmed sulfidation, in
which the oxidic catalyst was heated in a flow of H2S and H2, and the consumption of H2S
and H2 and the evolution of H2O were measured continuously [23]. It was found that H2S is
taken up and H2O is formed, even at room temperature, indicating a sulfur-oxygen exchange
reaction. This conclusion was confirmed by Cattaneo and Prins [24] with quick extended X-
ray absorption fine structure (QEXAFS) studies (Figure 1.1, phase 2), which also
demonstrated that the Mo (VI) species containing both oxygen and sulfur transform into
intermediate MoS3-type species at temperatures between 520 and 570 K (Figure 1.1, phase 3).
At higher temperatures, the MoS3 is reduced to MoS2 (Figure 1.1, Phase 4) with concomitant
H2 consumption and H2S evolution. [23].
During sulfidation as well as during actual hydrotreating, the conditions are highly
reducing with H2S always present; thermodynamics predict that molybdenum should be in the
MoS2 form. Nevertheless, XPS has shown that complete sulfidation of Mo, and especially of
W, is difficult to attain. Apparently, some of the Mo(VI) ions interact so strongly with the
Al2O3 support that they can only be sulfided above 500°C. Mo and W can be sulfided more
easily, however, in the presence of Ni or Co ions. EXAFS studies of Mo K-edge absorption
spectra demonstrated that, in well-sulfided Mo/Al2O3 catalysts, the Mo-S and Mo-Mo
distances are the same as in MoS2 [25,26], the only difference being that, in the catalyst, each
Mo ion is surrounded, on average, by fewer than six Mo atoms, as in the case of pure MoS2.
EXAFS is a bulk technique by means of which the environment of surface Mo ions as well as
Mo ions in the interior of the MoS2 particles (Fig. 1.2) is determined. As a consequence, the
co-ordination number lower than six indicates that the proportion of surface Mo ions is
substantial and that the MoS2 particles on the surface of the support typically contain about 60
Mo atoms [26,27] as shown in Figure 1.2.
Introduction Chapter 1 5
0 1 2 3 4 5 6
fresh
328
363
398
428
468
498
533
568
603
638
673
673
0
3
6
9
R [Å]
|FT[χ(
k) •
k3 ]|
O Mo
S
Mo
Sulfidation T [K]
Phase 1: Oxidic state
Phase 2: Coexistence of O and S
Phase 3: Intermediate Mo-Mo 2.5 Å
Phase 4: Formation of MoS2
Fig. 1.1. Quick Extended X-ray Absorption Fine Structure of a Mo catalyst.
Fig. 1.2. MoS2 particles, where the small balls are Mo and the big ones are sulfur.
MoS2 has a layer lattice and the sulfur-sulfur interaction between successive MoS2
sandwiches (Fig. 1.3) is weak. Nowadays people pay a lot of attention to the edges of MoS2
[28-30]. Byskov et al. used a model consisting of a periodic single layer of a S-Mo-S slab
with two molybdenum atoms in cross-section (2-Mo model) [28,29]. This model exposes the
Mo-edge at one side and the S-edge at the opposite side (Fig. 1.4). They added two sulfur
atoms per molybdenum atom on the Mo-edge of the stoichiometric model, and thus, the two
edges are fully covered by sulfur atoms. Raybaud et al. used a model consisting of two layers
Introduction Chapter 1 6
of S-Mo-S sheets as shown in Figure 1.5 [30]. The Mo24S48 unit is periodically repeated along
the directions parallel to the edge surfaces, while perpendicular to the edge surface a vacuum
layer of 12.8 Å separates the neighboring units.
Fig. 1.3. Surface structure of the bulk MoS2 lattice for the (100) plane.
Fig. 1.4. Stoichiometric MoS2 including two rows of MoS2 units (A) top view; (B) side
view.
SMo S
Mo
(A) (B)
Fig. 1.5. Perspective view of optimised MoS2 edge surfaces: (A) with bare Mo-edge and fully
saturated S-edge; (B) with fully sulfided Mo- and S-edge
Introduction Chapter 1 7
The incorporation of cobalt or nickel into the MoS2 structure can significantly increase
catalyst activity for hydrotreating reactions. Nickel atoms may be present in three forms after
sulfidation: as Ni3S2 crystallites on the support, as nickel atoms adsorbed on the edges of
MoS2 crystallites (the so-called Ni─Mo─S phase), and as nickel cations at octahedral or
tetrahedral sites in the γ-Al2O3 lattice (Fig. 1.6). Depending on the relative concentrations of
nickel or cobalt and molybdenum and on the pretreatment conditions, a sulfided catalyst may
contain a relatively large amount of either Ni3S2 or Co9S8 or the Ni─Mo─S (or Co─Mo─S)
phase.
Ni NiAl2O3
Ni3S2SMoNi
Fig. 1.6. Three forms of nickel present in a sulfided Ni-Mo/Al2O3 catalyst: as active sites on
the MoS2 edges (the so-called Ni─Mo─S phase), as segregated Ni3S2, and as Ni2+
ions in the support lattice.
It is well accepted that promoter atoms are located at MoS2 edges, but the exact location
of the promoter atoms relative to molybdenum and sulfur, and the mechanism of their effect is
still under debate. Using the 2-Mo model, Byskov et al. [29] found that the configuration with
Mo atoms substituted by cobalt atoms at the S-edge (Fig 1.7A) is more stable than the
structure with Mo atoms substituted by cobalt atoms at the Mo-edge (Fig 1.7B). However,
Raybaud et al. [31] concluded that the substitution of Mo atoms by Co atom on the Mo-edge
(Fig 1.7C) is energetically preferred on the fully sulfided S-edge (Fig 1.7D).
Introduction Chapter 1 8
Co
S Mo
Co
(A) (B)
(C) (D)
Co
Fig. 1.7. Co-Mo-S models with cobalt atoms at different locations: (A) two Mo atoms are
substituted by Co atoms at the fully sulfided S-edge; (B) two Mo atoms are
substituted by Co atoms at the fully sulfided Mo-edge; (C) three Mo atoms are
substituted by Co atoms at the bare Mo-edge; (D) three Mo atoms are substituted by
Co atoms at the fully sulfided S-edge.
1.3 The role of H2, H2/H2S and Active sites
A review on the chemistry of catalytic hydrotreating was published by Topsøe et al.
[32]. Several recent reviews deal specifically with the individual reactions, such as HDS [33-
35], HDN [36-37], HDO [38], and HYD [39]. These studies predominantly focused on
reactants such as hetero ring model compounds contaning sulfur, nitrogen and oxygen. Real
feedstocks have also been evaluated. The first attempt to review the information on hydrogen
activation was made in 1988 by Moyes [40]. The activity of the commercial hydrotreating
catalysts results from the ability of Mo(W)S2 to adsorb and activate hydrogen. The presence
of active surface hydrogen is critical for hydroprocessing reactions to proceed at desirable
Introduction Chapter 1 9
rates. In addition, active surface hydrogen extends the catalyst life by slowing down
deactivation.
The adsorption and desorption of hydrogen on MoS2 was reported by Badger et al. [41].
The reaction of H2 with Mo(W)S2 involves several processes occurring in parallel. It results in
a gradual removal of sulfur as H2S leading to a decrease in the stoichiometric S/Mo(W) ratio.
At the same time, H2 is adsorbed on the metal sulfide phase from where it can be transferred
to take part in various reactions. This is indicated by the study on the hydrogen adsorption by
MoS2 prepared by different methods from ammonium tetrathiomolybdate (ATTM) published
by Kalthod and Weller [42]. Polz et al. [43] proposed a heterolytic splitting of H2 on Mo-S
pairs giving a hydride and SH group,
H2 + □- S2- = □ – H + - SH-
as well as homolytic H2 dissociation with the aid of the S2-2 species giving two SH groups.
H2 + S22- = 2 SH-
The H2S dissociation can occur on the same vacancy as follows
H2S + □- S2- = □ – SH + - SH-
The study of Barbour and Campbell [44] supports the involvement of the reactions.
Theoretical and spectroscopic studies [45-56] provide some evidence for the occurrence of
several forms of Mo-SH and Mo-H groups. The results of Maternova [57,58] obtained by
AgNO3 titration and of Li et al. [59] obtained by TPR and TPD support the presence of at
least two kinds of SH groups, one processing reversible hydrogen and the other irreversible
hydrogen. It was suggested that the former is a highly labile hydrogen that can migrate from
one sulfur ion to another. Such hydrogen can be desorbed by purging in N2 in 723 K. Some
experimental evidence for the presence of entities other than SH, e.g., Mo-H was presented by
Jalowiecki et al. [60]. These authors confirmed the presence of another form of reactive
hydrogen (H*) on MoS2 previously reduced in H2 between 373 and 973 K.
Introduction Chapter 1 10
Sundberg et al. [45] reported that promoters increased the amount of adsorbed hydrogen.
However, increasing the H2 pressure by a factor of 20 resulted only in a threehold increase in
the hydrogen retention. The bond strength between hydrogen and Co(Ni) promoted Mo(W)S2
was greater than that for the unpromoted sulfides. In this regard, the direct involvement of
Co(Ni) in hydrogen bonding cannot be ruled out. Beneficial effects of promoters on the
hydrogen activation by Mo(W)S2 may be attributed to the increased rate of hydrogen
activation although this issue has not yet been studied experimentally.
Several studies indicated the importance of the H2S/H2 ratio and temperature on the
distribution of vacancies and SH groups [61-67]. This was further advanced, most notably by
Kasztelan et al. [30, 68-70] who recognized the presence of several types of sulfur and metal
ions, each of them playing a specific role during hydrogen activation. For example, the role of
corner and edge S ions differs significantly from that of S ions in the basal plane. It is
generally accepted that the latter cannot dissociate molecular hydrogen; however, they are
capable of storing active hydrogen after dissociation of the molecular hydrogen occurring on
other sites.
It was commonly assumed that the catalytically active sites in a hydrotreating catalyst
are the Mo cations at the surface of the MoS2 crystallites with at least one sulfur vacancy so
that the reacting molecule can chemically bind to the Mo cation [4,7,31]. Since sulfur anions
in the basal planes of MoS2 are much more difficult to remove than anions at edges and
corners, exposed Mo ions are predominantly present at edges and corners. Catalysis therefore
occurs at MoS2 edges and corners rather than on the basal plane, as verified by a surface
science study. A MoS2 single crystal, with a high basal plane to edge surface area ratio, had a
low HDS activity. Its activity increased after the sulfur atoms were sputtered from the basal
plane and after exposure of the Mo ions [71].
The HDS and HDN activities of a MoS2/Al2O3 catalyst increase substantially upon
addition of Co or Ni. Several explanations for the promoter function of Co and Ni have been
proposed [4,7,31]. The most famous model is that in which the promotion effect is ascribed to
cobalt present in the Co-Mo-S phase, with cobalt ions located at the MoS2 surface; a
significant contribution of separate Co9S8 was excluded [72]. This so-called Co-Mo-S model
(or Ni-Mo-S model for Ni-Mo catalysts) is currently the most widely accepted model. The
Co-Mo-S model itself does not indicate whether the catalytic activity comes from Mo
promoted by the presence of Co or from the Co sites themselves. Both cobalt and nickel
Introduction Chapter 1 11
sulfide, supported on carbon, have a higher HDS activity than MoS2/C [73]. Therefore it has
been suggested that the cobalt in the Co-Mo-S phase and the nickel in the Ni-Mo-S phase
might be the catalysts and not the promoters. In the past, the idea that Co and Ni might be the
catalyst in sulfided Co-Mo and Ni-Mo systems was rejected, because sulfided Co/Al2O3 and
Ni/Al2O3 catalysts have a very low HDS activity. However, during the usual catalyst pre-
treatment of Co/Al2O3 or Ni/Al2O3 catalysts and in the absence of Mo, cobalt and nickel ions
interact strongly with Al2O3. Therefore, during subsequent sulfidation, the metal ions are not
sulfided at all and do not contribute to the HDS activity. Alternatively, severe sulfidation
brings the metal ions back to the surface but lowers their dispersion and activity.
Carbon-supported cobalt and nickel sulfide catalysts, when carefully prepared, are
indeed highly active. The activity of a sulfided Co-Mo/C catalyst, based on the number of Co
atoms, compared much better with the number of estimated surface Co atoms in a sulfided
Co/C catalyst than with the number of estimated edge Mo atoms in a MoS2/C catalyst [74].
The observation that the hydrogenation pattern of Co-Mo and Ni-Mo catalysts resembles that
of sulfided Co respectively Ni catalysts and is different from that of supported MoS2 is further
evidence that Co and Ni are the catalytic sites rather than Mo. IR [75] and Mo EXAFS
investigations [26,76] showed that Mo is fully coordinated and is not accessible to substrate
molecules in Co-Mo and Ni-Mo catalysts. The Mo cannot, therefore, be catalytically active.
An EXAFS study by Startsev et al. of the adsorption of selenophene (the Se analogue of
thiophene) on a sulfided Ni-Mo/Al2O3 catalyst seemed to confirm this conclusion. They
claimed that selenophene adsorption changed the Ni but not the Mo EXAFS spectrum,
indicating that selenophene coordinated to Ni and not to Mo [77]. This would prove that Mo
is not accessible to selenophene. Medici et al. observed that the Ni EXAFS measured by
Startsev et al. cannot be simulated by selenophene adsorbed on Ni [78]. Leliveld et al. have
repeated the selenophene EXAFS studies of Startsev et al. [79]. They showed that after
reaction of selenophene and hydrogen with a Co-Mo catalyst at 200°C Se was exclusively
coordinated to the Co atoms. At 400°C, on the other hand, the Se atoms were found in bridge
positions between Co and Mo. The authors interpreted this as proof for two different vacancy
sites in which the Se atom can adsorb. One site has a S vacancy associated with Co, while the
other site has a vacancy between Co and Mo. In subsequent work, the authors admitted,
however, that an alternative explanation is possible [80]. It may be that at higher temperature
redistribution takes place of the Se atoms that are originally bonded to Co in a terminal
Introduction Chapter 1 12
position and S atoms in bridging positions between Co and Mo. Because of the stronger M-Se
bonds, this seems a very likely possibility. Nevertheless, it seems clear that the first reaction
takes place on Co and not on Mo.
Recent discussions on the catalytic sites have concentrated on a combined action of Ni
(or Co) and Mo [29-31,81-82]. In HDS, a sulfur-containing molecule is supposed to adsorb on
a site with a sulfur vacancy and react to a hydrocarbon molecule and a sulfur atom. This
sulfur atom occupies the vacancy and must be removed by hydrogenation before the catalytic
cycle can start all over again. It has been pointed out that a sulfur atom between a Ni (Co) and
Mo atom is less strongly bonded than a sulfur atom between two Mo atoms. Therefore it can
be more easily removed. This would explain the promoter action of Ni and Co on Mo in HDS.
If HDN were to occur analogously, a nitrogen atom should be taken up by the metal sulfide
catalyst particles and later be removed by hydrogenation. This seems less likely than the
equivalent sulfur uptake and removal in HDS, and suggests that in HDN different sites are
used as in HDS.
1.4 Reaction Mechanism
Environmental legislation requires a deep reduction of the sulfur content of gasoline and
diesel fuel to 50 ppm by the year 2005. Nitrogen-containing molecules inhibit the
hydrodesulfurization (HDS) of sulfur-containing compounds through competitive adsorption.
In the past, this was not a severe problem, as the amount of these nitrogen-containing
compounds in petroleum is much lower than the amount of sulfur-containg molecules.
However, a low level of sulfur has to be reached in deep HDS. In that case, nitrogen
compounds will be harmful. Therefore, it is very important to understand the mechanism of
the removal of nitrogen. A deep understanding of these mechanisms will be essential to
develop the necessary catalyst.
1.4.1 Aromatic C-N bond breaking
Introduction Chapter 1 13
C-N bonds in aromatic rings are much stronger than those in aliphatic rings.
Consequently, C-N bonds in rings, as in pyridine and pyrrole, can be broken only after
hydrogenation of the ring to give piperidine and pyrrolidine, respectively. Direct cleavage of a
C-X bond external to an aromatic ring, as in C6H5-X, to give a benzene and HX is the rule for
X = Cl and SH. For X = OH and NH2, direct cleavage occurs only to a limited extent and only
at high H2/H2S ratios and high temperatures (above 400 °C). The strength of the C-O bond in
phenol and that of the C-N bond in aniline are increased through conjugation with the
aromatic ring. Under normal hydrotreating conditions, the C-N bond can therefore be broken
only when it is aliphatic. Hydrogenation of the N-containing heterocycle or of the aromatic
ring to which the amine group is attached is necessary in order to obtain a substantial degree
of nitrogen removal. This difference is a consequence of the lower energies of the C6H5-Cl
and C6H5-SH bonds relative to the C6H5-OH and C6H5-NH2 bonds.
How aniline reacts to benzene is not clear. This reaction is often referred to as
hydrogenolysis, because it behaves as a concerted reaction in which the C-N bond is split and
the fragments are simultaneously hydrogenated. Such hydrogenolysis reactions are well
known for hydrocarbons over metal catalysts, and are believed to occur on an ensemble of
metal atoms. The hydrocarbon molecule adsorbs on the metal surface, and neighbouring
carbon atoms bind to neighbouring metal atoms. In this way, the C-C bond is weakened and
broken, and H atoms on nearby metal atoms are used to hydrogenate the fragments. It is
difficult to believe that such a hydrogenolysis reaction can occur on the surface of metal
sulfides. First, the distance between the metal atoms in a metal sulfide is much longer than in
a metal, thus making it unlikely that there is a geometrical fit for the four-center M-(C)-M-(N)
reaction intermediate. Second, if hydrogenolysis of C-N bonds occurs, then hydrogenolysis of
C-C bonds should also occur, because these bonds are not much stronger than the aliphatic C-
N bonds. Over sulfided catalysts, in the presence of H2S, only very little C-C bond breaking
occurs, however.
Thus, how is benzene formed directly from aniline and, in the same way, how is
naphthalene formed directly from naphthylamine? Although a definite answer has not been
given yet, some hypothesis can be suggested. One explanation for the direct denitrogenation
would be to assume that aniline is hydrogenated to tetrahydroaniline, which undergoes
elimination to cyclohexadiene. Cyclohexadiene then quickly reacts to cyclohexene or
Introduction Chapter 1 14
benzene. Since tetrahydroaniline is not flat, the elimination of ammonia is possible in the anti
conformation.
Another explanation is the partial hydrogenation of aniline to dihydroaniline, followed
by elimination of ammonia. At first glance, this explanation seems flawed because, owing to
the planar structure of the cyclohexadience molecule, the NH2 group on the C1 atom and the H
atom on the neighbouring C2 atom are in the eclipsed conformation. This would mean that the
subsequent elimination (e.g. of 1,2-dihydro-aniline to benzene and ammonia) must occur in
the anti-periplanar rather than in the syn-antiplanar conformation. A closer look at 1,2-
dihydro-aniline suggests, however, that the elimination will not occur by an E2 mechanism,
but by an E1 mechanism. The cyclohexadienyl carbocation resulting from scission of the C-N
bond will be strongly stabilized by conjugation with butadiene fragment, as shown in Figure
1.8.
NH2 NH2 NH3
H2 H+ -H+
Fig. 1.8 C-N hydrogenolysis via partial hydrogenation, protonation and syn-elimination.
The third possibility is hydrogenolysis on a single Mo or Ni (Co) atom on the metal
sulfide surface. The drawback of the single-atom hydrogenolysis mechanism is that no
examples for single-atom hydrogenolysis over metal surfaces have been provided. The fourth
explanation is that aniline might have reacted to thiophenol through enol-keto tautomerism
and NH2-SH exchange by addition of H2S and elimination of NH3 analogous to reactions
described in [83-85]. The resulting thiophenol could then have reacted to benzene [6] as
shown in Figure 1.9.
NH2+NH2
HH
+NH3SH +SH SH
+H+ +H2S -NH3 -H+
Fig. 1.9 NH2-SH exchange by addition of H2S and elimination of NH3.
Introduction Chapter 1 15
To determine whether the hydrogenolysis of an aryl C-N bond is real or apparent, we
studied the HDN of 1-naphthylamine and the reactions of the possible intermediates 1,2,3,4-
tetrahydro-1-naphthylamine, 1,2-dihydronaphthalene and 5,6,7,8-tetrahydro-1-naphthylamine
in chapter 7.
1.4.2 C-N bond cleavage of aliphatic amines
The removal of nitrogen is easier from aliphatic amines than from aromatic amines. It is
generally accepted that the first step in the HDN of nitrogen-containing aromatic molecules is
the hydrogenation of the heterocyclic ring. Only after the breaking of the aromaticity can C-N
bond cleavage in the resulting saturated molecules take place. Several mechanisms of the C-N
bond scission and nitrogen removal have been proposed [5,86-88]. Nelson and Levy [86]
were the first to suggest Hofmann-type elimination and nucleophilic substitution as
mechanisms for C-N bond scission of aliphatic nitrogen-containing molecules. The initial step
in C-N bond scission is the addition of a proton to a nitrogen lone pair with the formation of a
quarternary ammonium compound, which provides a better leaving group than the amine
group. C-N bond scission can then occur via elimination of a β−hydrogen atom with the
formation of an alkene (Fig. 1.10) or via nucleophilic substitution of the amine group at the
α−carbon atom by a sulfhydryl group to form an alkanethiol (Fig. 1. 11).
H C N HH
HC C C NH C NC ++
BH
+ B+
Fig. 1.10. Hofmann elimination mechanism.
C N+H2S
HS- + C N S C
H
+ NH
S C
H
+ H2 H2S H C+
+
Fig. 1.11. Nucleophilic substitution of the amine group by an SH group, followed by
hydrogenolysis of the C-S bond.
Introduction Chapter 1 16
Laine proposed a different type of HDN mechanism where the C-N bond scission in
saturated heterocycles requires metal atoms or ions rather than acidic sites. Piperidine is
activated by a metal site and the intermediate formed is attacked by H2S to form a C-S bond
and leads to the C-N bond scission (Fig. 1.12).
NH
MHN
MH-
H2SHN
M H-SH2
+
MH2NH
SHH2N SMH2
H2N SH H2 H2N
H2
-NH3
+ +
++
-H2S/fast
Fig. 1.12. Reaction pathways for piperidine HDN
An alternative explanation of C-N bond breaking in molecules that lack β-H atoms is
direct hydrogenolysis of the alkylamine to a hydrocarbon and NH3 (Fig. 1.13).
C N + H2 C H HN+
Fig. 1.13. Direct hydrogenolysis of the alkylamine.
Proof of real hydrogenolysis of an aliphatic C-N bond, that is a concerted reaction in
which a C-N bond is broken and C-H and N-H bonds are formed, has not yet been presented
for metal sulfide catalysts. Therefore, the term hydrogenolysis should not be used to describe
C-N bond breaking until hydrogenolysis has been proven mechanistically.
Another alternative mechanism for the HDN of alkylamines would be
dehydrogenation to an imine (Fig. 1.14) followed by the addition of H2S, the removal of NH3,
as well as hydrogenation and hydrogenolysis.
Introduction Chapter 1 17
C
H
NH
-H2C N
H2SC
SH
NH
C S HN+
C SH2
C
H
SHH2
H2S + C
H
H
Fig. 1.14 Mechanism of hydrogenolysis of an alkylamine by means of an imine.
Concurrent with the β-hydrogen elimination and nucleophilic substitution reactions, a
disproportionation reaction can occur between two alkylamine molecules, and this
complicates the study of the HDN reaction mechanism. This disproportionation results in the
formation of a dialkylamine and ammonia in case of an alkylamine (Fig. 1.15) and in the
formation of a trialkylamine and alkylamine in case of a dialkylamine; it takes place even on
alumina at low hydrogen pressure [89,90]. Substantial amounts of the disproportionation
products N-pentylpiperidine, dicyclohexylamine, and dipentylamine were observed in the
HDN of piperidine [90], cyclohexylamine [91], and pentylamine [6,92] respectively. An
imine mechanism can explain the formation of dialkylamine as well (Fig. 1.16). For example,
secondary ethylimine will be formed in HDN of ethylamine. Then the imine will react with an
ethylamine to form an intermediate, followed by elimination and hydrogenation to form
secondary ethylamine.
C
+ C N+
H
HH
N:H H
C
C N+
H
HH
NH H
C
C :N
H
HH
N+
H H
+
Fig. 1.15. Disproportionation mechanism.
NH2
NH N NHNH2
NH
Fig. 1.16. Imine mechanism to form a secondary amine in the HDN of ethylamine.
Introduction Chapter 1 18
Several studies [92-98] have dealt with the HDN of aliphatic amines over different
catalysts. Portefaix et al. showed in HDN studies at 2 MPa over sulfided NiMo/Al2O3 that an
increase in the rate of formation of denitrogenation products occurred when going from
neopentylamine to amylamine and tert-amylamine; this increase corresponds to the increase in
the number of β−hydrogen atoms in the neopentylamine, amylamine and tert-amylamine. This
was taken as proof that aliphatic C-N bond cleavage takes place by Hofmann elimination.
Vivier et al. were the first to prove that C-N bond cleavage by nucleophilic substitution can
take place in the HDN of amines [99]. They observed that benzylamine and α,α-
diphenylmethylamine, which do not have β−hydrogen atoms and thus cannot react by
elimination, react fast to toluene and diphenylmethane respectively. Benzylamine and α,α-
diphenylmethylamine react most probably by nucleophilic substitution of the amine group by
an SH group followed by rapid hydrogenolysis of the intermediate thiol.
Cattenot et al. showed that both elimination and nucleophilic substitution play a role in
the C-N bond scission of pentylamines on unsupported transition-metal sulfides at
atmospheric pressure [92]. The ratio of the two mechanisms depended on the type of metal
sulfide catalyst and the type of amine. Over MoS2, n-pentylamine reacted by nucleophilic
substitution with H2S to pentanethiol as well as with another n-pentylamine molecule to
dipentylamine. Pentenes were observed as secondary products and supposed to be formed by
elimination from dipentylamine. These findings suggest that the molecular structure is one of
the most important factors in HDN and that different molecules may undergo nitrogen
removal by different mechanisms.
A detailed investigation of the HDN of alkylamines will be carried out over sulfided
NiMo/Al2O3. The aim of the work was to determine which of the different mechanisms
(elimination, substitution, and disproportionation) plays the major role in the HDN of
aliphatic amines. In Chapter 3 we will present our results of the HDN of the linear n-
alkylamines hexylamine, dihexylamine, and trihexylamine, while in chapter 5 we will present
our results of the HDN of alkylamines with the amine group attached to secondary and
tertiary carbon atoms.
1.4.3 C-N bond cleavage of hetero cyclic amines
Introduction Chapter 1 19
Heterogeneous compounds like pyridine, quinoline, and acridine are the main nitrogen-
containing compounds in oil. Pyridine is often considered to be the simplest heterocyclic
nitrogen compound and is often used as a model compound for comparing the HDN activity
of catalysts and studying the HDN mechanism [100-104]. The supposed reaction scheme is
shown in Figure 1.16.
N N NH2 NH2
NH3 C5H10 C5H12
H
-
Figure 1.16 HDN network of pyridine over sulfided Ni-Mo/Al2O3 catalysts
The HDN of pyridine proceeds via the hydrogenation of the pyridine ring to piperidine,
followed by denitrogenation [6,105-108]. The hydrogenation of pyridine to piperidine is
inferred to take place on sulfur-deficient sites of the metal sulfide surface since a negative
effect of H2S was observed on this step. This reaction is favored by hydrogen as well
[6,107,109]. The second step, ring-opening of piperidine and the following nitrogen-removal
reaction, might occur by elimination and lead to alkene intermediates. This reaction is favored
by H2S [6,110,112]. At high H2S partial pressure, 2-methyl-thiacyclopentane and
thiacyclohexene were found as products [110,111]. The high H2S concentration induced a
nucleophilic substitution of the amino group of 5-aminopentene-1 by the SH group, and the
resulting 5-thiopentene-1 reacted intramolecularly to give a five- and six-membered
thiacyclo-alkane. The fact that these two molecules were observed in the HDN of pyridine,
piperidine, and 5-aminopentene-1, but not in that of 1-pentylamine, strongly supports the
mechanism shown in Figure 1.17. That ring opening and removal of nitrogen are promoted by
H2S indicates that these reactions take place on relatively sulfur-rich sites on the metal sulfide
surface.
Introduction Chapter 1 20
+
N NH2
SSH
N NH2
S
Fig. 1.17. Possible HDN mechanism of pyridine via SH nucleophilic substitution
Thermodynamics may play an important role in the HDN of aromatic heterocycles. In
general, at higher temperatures the equilibrium between ring hydrogenation and
dehydrogenation shifts to the dehydrogenation side leading to a decrease in the equilibrium
concentration of the hydrogenated compound, and thus a decrease in nitrogen removal rate.
Satterfield and Cocchetto [101] observed that under such conditions, the overall rate of
pyridine HDN showed a maximum with increasing temperature. Therefore, the rate-
determining step in the HDN of pyridine is changing with reaction conditions. At a
temperature of 573 K and pressure of 3 MPa, the rate of hydrogenation of pyridine is about
the same as that of the ring-opening of piperidine over NiMo, CoMo, and NiW catalysts
[6,62,109,112]. H2S may have a promoting as well as a poisoning effect on the HDN of
pyridine under these conditions. At a low H2S/H2 ratio, pyridine hydrogenation to piperidine
is not inhibited, but ring opening of piperidine is occurring slowly, thus leading to a low
overall HDN conversion to C5 hydrocarbons. Under such conditions, the effect of adding
nickel to molybdenum is not very important for the total HDN to give C5 products [112]. This
observation has led some authors to conclude that nickel does not promote molybdenum in
HDN as it does in HDS [113,114]. However, at higher H2S/H2 ratio, hydrogenation of
pyridine to piperidine is retarded by H2S, but ring opening is accelerated. Thus, even at a
lower pyridine conversion, the yield of hydrocarbons increases [109]. At an even higher
H2S/H2 ratio, the negative effect of H2S is responsible for low conversions to piperidine and
low HDN conversions. At lower hydrogen pressures and higher reaction temperatures,
pyridine hydrogenation becomes rate determining, and H2S becomes increasingly toxic. At
higher hydrogen pressures and lower temperatures, H2S acts as a promoter because the ring-
Introduction Chapter 1 21
opening reaction becomes rate determining. Under such conditions, nickel clearly promotes
the HDN of pyridine.
During the HDN of pyridine and piperidine, a higher molecular weight product, N-
pentylpiperidine is formed readily, especially when the partial pressure of intermediates
(piperidine and pentylamine) is high. It is formed by the disproportionation of piperidine and
pentylamine [6, 100, 114]. Even though N-pentylpiperidine can be denitrogenated at high
conversion of the reactant, this makes the HDN netwok of pyridine quite complicated to
study. It is well known that a nucleophilic attack is hindered by substitution on the α carbon
atom [115]. Therefore, 2-methylpyridine could be the right molecule to study the HDN
mechanism.
Portefaix et al. observed that the HDN reaction of 2,6-dimethylpiperidine was faster
than that of piperidine [93]. Their result suggests that the presence of a methyl group leads to
faster ring opening. Thereofore, it confirms that β−hydrogen has a very important influence
on the C-N bond cleavage.
Cerny and Trka performed their investigations in an autoclave at 15.5 MPa and 250 °C.
They concluded that the 2-methylpiperidine ring opens preferentially on the side that does not
contain the methyl group and that the HDN reactions of more substituted pyridine derivatives
are slower [116], namely, ring opening of 2-methylpiperidine by C-N bond cleavage occurred
preferentially on the CH2-N side and not on the CH(CH3)-N side. In the other words,
β−hydrogen does not show positive influence on the C-N bond cleavage. Further study will be
shown in Chapter 2.
1.5 Experimental
1.5.1 Unit
The hydrodenitrogenation reactions were performed in a continuous flow fixed-bed
reactor, where the solid catalyst is situated in the middle of a tubular reactor heated with an
oven. A simplified scheme (Fig. 1.18) shows the main parts of the unit.
Introduction Chapter 1 22
Fig. 1.19. Simplified scheme of the unit.
The reactants (amine and sulfur compounds) were solved in cyclohexane, which was
used as a solvent, and in heptane, which is used as reference for the chromatograph. The feed
mixture is pumped with a syringe pump (Isco, Modell 500D). H2S was fed to maintain the
properties of the sulfided catalyst constant. H2 was fed to keep the high pressure. The gas
flows of hydrogen (H2) and of the hydrogen-hydrogensulfide mixture (90% H2/10% H2S)
were controlled by two mass flow controllers (Brooks, Series 5850E). Gases and liquids are
mixed before being heated, in order to avoid plugs in the liquid inlet. The liquid reactants
were fed co-current to the gases at the top of the reactor inlet. Gases and liquids are then
heated at 240°C in the pre-heater to achieve a homogeneous gas phase. Then, the gas feed
flows to the reactor, where the temperature was monitored and controlled inside the catalytic
bed. Above the catalytic bed, 8 g SiC was added to the reactor to achieve plug-flow
conditions and the desired temperature. The active catalyst (in general 0.05 g) was diluted
with 8 g SiC to obtain good heat transfer. The fixed-bed reactor was held in place using glass
wool and a metal support. A more detailed description of the reactor is given in the
dissertation of M. Flechsenhar [117]. After the reaction, the product flows to a 6-port valve
heated at 300°C and at the same pressure as the reactor, to maintain all products in the gas
phase (Fig. 1.19).
Introduction Chapter 1 23
612
3
4 5
Reactor
Condenser inlet
He inlet
He outlet
Column
Injector
Sample
Fig. 1.19. Sampling system with the 6-port-valve.
Helium, which was the carrier gas for the column, was continuously flowing in the
valve (inlet port 4, outlet port 3). Reaction products were continuously flowing through the
sample loop (50 µl, port 5-2) and then to the condenser. When the sample was taken, the
valve turned 60°, to connect the helium inlet to the sample loop, so that the contents of the
sample loop was purged in the injector. After 12 seconds the valve returned to the original
position. In this period the reactor flow was led directly to the condenser. This sampling
system shows no mass balance problem if the operation temperature is above 240°C
(depending on the boiling point of the reactant used). The reaction products were injected
from the valve into the gas chromatograph equipped with a 30 m PTA-5 fused-silica capillary
column for quantitative analysis. Detection was made with a flame ionisation detector (FID)
as well as with a pulsed flame photometric detector (PFPD), which is especially sensitive to
nitrogen and sulfur-containing compounds. The detection method is discussed more
extensively in the dissertation of F. Rota [118]. After the sampling valve, the products were
condensed in the condenser (40 °C). The heavy products were collected as a fluid in the
condenser and the light gases (H2, H2S, and NH3) were purged outside the back pressure
regulator. The mixed gas was purified in a NaOH water solution. The total pressure was
maintained constant using a back pressure regulator. A safety system was integrated with the
Introduction Chapter 1 24
unit. In case of over- and under-pressure (± 0.3 MPa) and in case of over- and under-heating
(± 20 °C) the inlet gases of H2 and H2/H2S mixture and liquid were interrupted.
1.5.2 Product analysis
The product analysis was performed online with a Varian gas chromatograph (Modell
3800) equipped with a flame ionization detector (FID) and a pulsed flame photometric
detector (PFPD). The FID detector was used to determinate the concentration of the carbon-
containing species. The signal of the FID is proportional to the number of carbon atoms
present in the molecule. All products were calibrated in the presence of heptane as an internal
standard to calculate the amount of each species. The response factors (Rf) of the compounds
were determined by injecting a known amount together with a known amount of heptane.
Then, with the relationship ni/nHeptane=Rf·Ai/AHeptane, it was possible to calculate the number of
moles of the compound i (ni) during the reaction.
1.5.3 Catalysts
The Ni(Co)Mo/γ-Al2O3 (Condea, BET surface area 210 m2/g, total pore volume 0.44
cm3/g) catalyst used in this work contained 8 wt% Mo and 3 wt% Ni or Co and was prepared
by successive pore-volume impregnation of γ-Al2O3 (CONDEA, pore volume: 0.5 cm3·g-1,
specific area: 230 m2·g-1) with an aqueous solution of (NH4)6Mo7O24·4H2O (Aldrich) and then
with an aqueous solution of Ni(Co)(NO3)2·6H2O (Aldrich). The catalysts were dried in air at
ambient temperature for 4 hours and then dried in an oven at 120°C for 15 hours after each
impregnation step. Then, they were finally calcined at 500°C for 4 hours. The catalysts were
crushed and sieved to 230 mesh (0.067 mm). The two catalysts were prepared in large amount
(20 g) so that all catalyst used in the catalytic experiments were from the same batch.
A sample of 0.050 g of catalyst was diluted with 8 g SiC to achieve plug-flow
conditions in the continuous flow fixed bed reactor. The oxidized form of the catalyst was
sulfided in situ with a mixture of 10 % H2S in H2 (25 ml/min) at 370°C and 1.0 MPa for 4
hours. The oxidized form of the catalyst is assumed to be mostly NiO and MoO3, whose
activities are practically zero. During sulfidation they transform to NiS, MoS2, and H2O.
Introduction Chapter 1 25
1.5.4 Weight time
Weight time was defined as τ = wcat / nfeed, where wcat denotes the catalyst weight and
nfeed the total molar flow fed to the reactor. The weight time (τ) was changed by varying the
liquid and gaseous reactant flow rates, while their relative ratio was kept constant. In the
calculation of the weight time equations (1), (2) and the weight time definition are combined
to equation (3), and an ideal gas mixture behavior is assumed.
total total
reactant reactant
n pn p
= (1)
totaltotal feed reactant
reactant
( % ) pn F wtp
ρ•
= ⋅ ⋅ ⋅ (2)
cat cat cat
feed,totaltotal
feed reactantreactant
( % )
w wmoln pF wt timep
τ
ρ•
= = =
⋅ ⋅ ⋅
g (3)
ρfeed denotes the density of the liquid mixture, which contains 70% cyclohexane. Therefore
the density was always around 0.72 g/ml. •
F is the flow rate (ml/min) pumped with the
syringe pump. The unit of weight time is g·min/mol (1 g.min/mol = 0.68 · 10-3 g.h/l).
1.5.5 Reactants and reaction conditions used in every chapter
The NiMo/γ-Al2O3 catalyst was used in Chapter 2. After sulfidation of the catalysts, the
pressure was increased to 5.0 MPa, and the liquid reactant was fed to the reactor by means of
a high-pressure syringe pump (ISCO 500D). Blank experiments with and without SiC were
carried out at 573 and 623 K. The composition of the gas-phase feed in most experiments
consisted of 5 kPa amine reactant, 140 kPa decane (as solvent for the amine), 20 kPa heptane
(as reference for GC analysis), 20 kPa H2S, and 4.8 MPa H2 (unless indicated otherwise).
Mass spectrometry and NMR spectroscopy were used to identify the reaction products. The
MS analysis was performed with an Agilent 6890 gas chromatograph equipped with a HP-
5MS capillary column (crosslinked 5% PH ME siloxane, 30 m × 0.25 mm × 0.25 µm) and
Introduction Chapter 1 26
with an Agilent 5973 mass selective detector. The temperature of the injector was 270 °C, the
initial temperature of the column oven was 80 °C, and heating to 300 °C started after 2 min at
20 °C /min. 1H and 13C NMR spectra of isolated compounds were recorded on a Bruker DPX-
300 instrument at 300 and 75 MHz, respectively, at room temperature using CDCl3 as a
solvent.
The NiMo/γ-Al2O3 catalyst was used in Chapter 3 as well. After sulfidation, the pressure
was increased to 3 or 5 MPa, and the liquid reactant was fed to the reactor by means of a high-
pressure syringe pump (ISCO 500D). Cyclohexane, decane, and octane were used as solvents
and heptane as an internal standard for GC analysis. The hydrogen pressure was varied from
2.8 to 4.8 MPa. Three types of molecules were used as reactants to study the HDS, HDN, and
hydrogenation reactions simultaneously. We used pentanethiol and hexanethiol as thiols,
hexylamine and cyclohexylamine as amines, and 1-hexene and cyclohexene as alkenes. The
choice of the alkanethiol, alkylamine, and alkene in a simultaneous HDS, HDN, and
hydrogenation experiment was primarily made so as to obtain separate peaks in the GC
analysis. All the chemicals were purchased as commercial standards from Aldrich and Fluka.
The partial pressure of the alkanethiols and alkenes was kept at 5 kPa, while the partial
pressure of the amines was 5, 10, or 20 kPa. The H2S pressure was varied between 10 and 150
kPa and the experiments were carried out at 300, 320, and 340 °C. When changing the partial
pressure of the reactant, the solvent flow was adapted to keep the partial pressure of hydrogen
constant. The product selectivity (S) was defined as the number of molecules converted to a
certain product (nP) divided by the number of converted reactant molecules (nR), both
multiplied by their number of carbon atoms, CnP and CnR respectively: S = (nP*CnP) /
(nR*CnR). With this definition, the mass balance of the carbon atoms is preserved. For
instance, in the reaction the selectivity of DHA
is 66.7% and the selectivity of hexanethiol is 33.3%. The original feed was usually re-entered
after the set of HDN experiments to check whether the activity of the catalyst had remained
constant. Then the whole reactor set-up was cleaned for another series of experiments. The
gases used were hydrogen (PanGas 4.0) and a mixture of 10% H2S in H2 (Linde).
6 13 3 2 6 13 2 6 13( ) ( )C H N H S C H NH C H SH+ → + ,
In Chapter 4, the same catalyst as in Chapter 2 and 3 was used. The total pressure in the
HDN and HDS experiments was 3 MPa and the partial pressure of the alkylamines was 5 kPa.
The experiments were carried out at 270, 300, and 340 °C with a partial pressure of H2S of 10
or 100 kPa. When changing the partial pressure of the reactant, the solvent flow was adapted
Introduction Chapter 1 27
to keep the partial pressure of hydrogen constant. The accuracy in the measured conversion
was 2% (relative). 2-Pentylamine (Lancaster), 2-pentanethiol (Aldrich), 3-methyl-2-
butylamine (Aldrich, 98%), 3-methyl-2-butanethiol (ABCR, 100%), 3,3-dimethyl-2-
butylamine (ABCR, 100%), 2-methylcyclohexylamine (Fluka), 2-methyl-2-butylamine
(Aldrich, 98%), 2-methyl-2-butanethiol (TCI, Japan, 99%), benzylamine (Fluka, purum), α-
toluenethiol (Fluka, purum), and cyclohexane (Fluka, puriss.) were all used as purchased in
the HDN tests. 2-Methyl-3-pentylamine was synthesized from 2-methylpentanone-3 by
reaction with hydroxylamine and reduction of the resulting oxime with LiAlH4 in ether. The
product was purified by destillation.
In chapter 5, HDN and HDS experiments were carried out over NiMo/Al2O3,
CoMo/Al2O3 and Mo/Al2O3 catalysts. The total pressure was 3 MPa and the partial pressure
of the alkylamines was 5 kPa. The HDS of alkanethiols was studied at 5 kPa in the
simultaneous reaction with 5 kPa of the relevant alkylamine. The experiments were carried
out at 270, 300, 320 and 370 °C with a partial pressure of H2S of 10 or 100 kPa. When
changing the partial pressure of H2S, the hydrogen pressure was adapted to keep the total
pressure constant. Pentylamine (Aldrich), hexanethiol (Fluka), 2-hexylamine (Lancaster), 2-
pentanethiol (Aldrich), 2-methyl-2-butylamine (Aldrich, 98%), 2-methyl-2-butanethiol
(ABCR), and cyclohexane (Fluka, puriss.) were all used as purchased.
In chapter 6, a NiMo/Al2O3 catalyst was used in the HDN and HDS reactions. The total
pressure was 3 MPa and the partial pressure of neopentylamine was 5 kPa in all experiments.
The partial pressures of the adamantylamines and 1-adamantanethiol were kept at 1 kPa to
avoid condensation. The experiments were carried out at 300 and 340 °C with a partial
pressure of H2S of 10 and 100 kPa.1-Adamantylamine (Aldrich, pract.), 2-adamantylamine
(Fluka, purum), 1-adamantanethiol (Apin Chemicals Limited), adamantane (Fluka, purum),
neopentylamine (TCI, purum) and cyclohexane (Fluka, puriss.) were all used as received. For
the purpose of solubility, 1-adamantylamine, 2-adamantylamine, 1-adamantanethiol and
neopentylamine were dissolved in cyclohexane.
In chapter 7, the NiMo/Al2O3 and CoMo/Al2O3 catalysts were in HDN reactions. For the
purpose of solubility, 1-naphthylamine (Aldrich, 98%), 1,2,3,4-tetrahydro-1-naphthylamine
(Acros, 98%), 1,2-dihydronaphthalene (Fluka, 98%), 5,6,7,8-tetrahydro-1-naphthylamine
(Aldrich, 99%), tetralin (ABCR) and naphthalene (ABCR) were dissolved in benzene or
toluene (Fluka, puriss. p.a.). The catalytic experiments were performed in a stirred 17-ml
Introduction Chapter 1 28
autoclave as well as in a microflow reactor. The NiMo/Al2O3 and CoMo/Al2O3 catalysts that
were used in the autoclave were sulfided in a glass flow reactor by heating at 5 °C/min and
then sulfiding at 400 °C for 4 h in a mixture of 10% H2S in H2. Thereafter, nitrogen was
passed through the reactor at the same temperature for 0.5 h and subsequently the catalysts
were cooled to room temperature. The reactor was opened to air at room temperature. An
amount of 5 to 10 mg of catalyst was transferred to the autoclave and resulfided for 1 h in a
mixture of 10% H2S in H2 at 400 °C (heating rate 5 °C/min) and 0.35 MPa. After
resulfidation, the catalyst was cooled to room temperature, the autoclave was opened and the
reaction mixture (0.6 or 1.0 ml) was added quickly. After closing the autoclave, hydrogen was
added up to 0.6 MPa at room temperature and the temperature was increased to the reaction
temperature (between 300 and 350 °C). Liquid samples of 0.05 to 0.1 ml were collected at
different times and analyzed off-line by gas chromatography (Shimadzu GC-14 B), using an
HP1 (cross-linked methyl siloxane) or a DB-5ms (5%-phenyl methylpolysiloxane) column
and a flame ionization detector. The experiments in the microflow reactor were carried out
with the NiMo/Al2O3 catalyst only. A sample of 0.02 g NiMo/Al2O3 diluted with 8 g SiC was
first dried for 2 h at 400 °C and then sulfided for 4 h in situ with a mixture of 10% H2S in H2
at 1 MPa. After sulfidation, the pressure was increased to reaction pressure and the solution of
the reactand in toluene was fed to the reactor with a high-pressure syringe pump.
1.6 References [1] M.J. Girgis, B.C. Gates, Ind. Eng. Chem. Res 30 (1991) 2021.
[2] J.K. Minderhoud, J.A.R. van Veen, Fuel Proc. Tech. 35 (1993) 87.
[3] R.A. Sanchez-Delgado, Organometallic Modeling of the HDS and HDN Reactions,
Kluwer Academic, 2002.
[4] T.C. Ho, Catal. Rev. Sci. Eng. 30 (1988) 117.
[5] R.M. Laine, Catal. Rev. Sci. Eng. 25 (1983) 459.
[6] H. Schulz, M. Schon, N.M. Rahman, Stud. Surf. Sci. Catal. 27 (1986) 201.
[7] R. Prins R., V.H.J. de Beer, G.A. Somorjau, Catal. Rev. Sci. Eng. 31 (1989) 1.
[8] A.N. Startsev, Catal. Rev. Sci. Eng. 37 (1995) 353.
Introduction Chapter 1 29
[9] B. Delmon, G.F. Froment, Catal. Rev. Sci. Eng. 38 (1996) 69.
[10] R. Prins, Handbook of Heterogeous Catalysis 4, Ed. Ertl G., Knözinger H., Weitkamp
J., Wiley-VCH Wienheim (1997) 1908.
[11] H. Knözinger, in Proc. 9th Int. Congress Catal. (Eds. M.J. Phillips, M. Ternan), The
Chemical Institute of Canada, Ottawa, 5 (1988) 20.
[12] M. Lewandowski, Z. Sarbak, Fuel 79 (2000) 487.
[13] M. Jian, R. Prins, J. Catal. 179 (1998) 18.
[14] M.Y. Sun, D. Nicosia, R. Prins, Catal. Today 86 (2003) 173.
[15] F. Gioia, F. Murena, J. Hazardous Materials 57 (1998) 177.
[16] A. Stanislaus, B.H. Cooper, Catal. Rev-Sci. Eng. 36/1 (1994) 75.
[17] M. Adachi, C. Contescu, J.A. Schwarz, J. Catal. 162 (1996) 66.
[18] S. Kasztelan, J. Grimblot, J.P. Bonnelle, J. Phys. Chem. 91 (1987) 1503.
[19] N. Topsøe, H. Topsøe, J. Catal. 75 (1982) 354.
[20] X. Gao, Q. Xin, Catal. Lett. 18 (1993) 409.
[21] C. Wivel, B.S. Clausen, R. Candia, S. Morup, H. Topsøe, J. Catal. 87 (1984) 497.
[22] R. Prada Silvy, P. Grange, B. Delmon, Stud. Surf. Sci. Catal. 52 (1990) 233.
[23] P. Arnoldy, J.A.M. Van den Heijkant, G.D. de Bok, J.A. Moulijn, J. Catal. 92 (1985)
35.
[24] R. Cattaneo, T. Weber, T. Shido, R. Prins, J. Catal. 191 (2000) 225.
[25] B.S. Clausen, H. Topsøe, R. Candia, J. Villadsen, B. Lengeler, J. Als-Nielsen, F.
Christensen, J. Phys. Chem. 85 (1981) 3868.
[26] S.M.A.M Bouwens, R. Prins, V.H.J. de Beer, D.C. Koningsberger, J. Phys. Chem. 94
(1990) 3711.
[27] T. Shido, R. Prins, J. Phys. Chem. B 102 (1998) 8426.
[28] L.S. Byskov, B. Hammer, J.K. Norskov, B.S. Clausen, H. Topsøe, Catal. Lett. 47
(1997) 177.
[29] L.S. Byskov, J.K. Norskøv, B.S. Clausen, H. Topsoe, J. Catal. 187 (1999) 109.
[30] P. Raybaud, J. Hafner, G. Kresse, S. Kasztelan, H. Toulhoat, J. Catal. 189 (2000) 129.
[31] P. Raybaud, J. Hafner, G. Kresse, S. Kasztelan, H. Toulboat, J. Catal. 190 (2000) 128.
[32] H. Topsøe, B.C.Clausen, F.E. Massoth, Hydrotreating Catalysis. In Catalysis, Science
and Technology (Eds. J. Anderson, M. Boudart, Springer Berlin, 1996 Vol. 11.
Introduction Chapter 1 30
[33] T. Kabe, A. Ishikawa, W. Qian, Hydrodesulfurization and Hydrodenitrogenation,
Wiley–VCH 1999.
[34] F.E. Massoth, Adv. Catal. 27 (1978) 265.
[35] R.R. Chianelli, Catal. Rev. Sci. Eng. 26 (1984) 361.
[36] S. Ren, Z. Wang, Y. Hu, Ranliao Huaxue Xuebo 15 (1987) 255.
[37] G. Perot, Catal. Today 10 (1991) 447.
[38] E. Furimsky, F.E. Massoth, Catal. Today 52 (1999) 381.
[39] L. Qu, R. Prins, J. Catal. 207 (2002) 286.
[40] R.B. Moyes, in Hydrogen Effects in Catalysis (Eds. Z. Paal, P.G. Menon) Marcel
Dekker, 1988, 583.
[41] E.M. Badger, R.H. Griffith, W.B.S. Newling, Proc. Soc. (Lond.) 1949, A 197, 184.
[42] D.G. Kalthod, S.W. Weller, J. Catal. 95 (1985) 455.
[43] J. Polz, H. Zeilinger, B. Muller, H. Knözinger, J. Catal. 120 (1989) 22.
[44] J. Barbour, K.C. Campbell, Chem. Commun. (1982) 1371.
[45] P. Sundberg, R.B. Moyes, J. Tomkinson, Bull. Soc. Chim. Belg. 100 (1991) 967.
[46] P. Ratnasamy, J. Fripiat, Trans. Faraday Soc. 66 (1970) 2897.
[47] N. Topsøe, H. Topsøe, J. Catal. 139 (1993) 641.
[48] S. Vasudevan, J.M. Thomas, C.J. Wright, C. Sampson, J. Chem. Soc., Chem.
Commun. 1982, 418.
[49] C.J. Wright, D. Fraser, C. Sampson, R.B. Moyes, P.B.J. Wells, J. Chem. Soc. Faraday
Trans. 176 (1980) 1585.
[50] C. Vogdt, T. Butz, A. Lerf, H. Knözinger, Polyhedron 5 (1986) 95.
[51] C.J. Wright, D. Fraser, R.B. Moyes, P.B. Wells, Appl. Catal. 1 (1981) 49.
[52] C. Sampson, J.M. Thomas, S. Vasudevan, C.J. Wright, Bull. Soc. Chim. Belg. 90
(1981) 1215.
[53] P.N. Jones, E. Knözinger, W. Langel, R.B. Moyes, Tomkinson, Surf. Sci. Catal.
207(1988) 159.
[54] M. Lacroix, C. Dumonteil, M. Breysse, Am. Chem. Soc. Div. Petr. Chem. Prepr. 1998
35.
[55] C.G. Wiegenstein, K.H. Schultz, Surf. Sci. 396 (1998) 284.
[56] P.C.H. Mitchel, D.A. Green, E. Payen, A.C. Evans, J. Chem. Soc. Faraday Trans. 91
(1995) 4467.
Introduction Chapter 1 31
[57] J. Maternova, Appl. Catal. 3 (1982) 3.
[58] J. Maternova, Appl. Catal. 6 (1983) 61.
[59] X.S. Li, Q. Xin, X.X. Guo, P. Grange, B. Delmon, J. Catal. 137 (1992) 385.
[60] L. Jalowiecki, A. Aboulaz, S. Kasztelan, J. Grimblot, J.P. Bonelle, J. Catal. 120 (1989)
108.
[61] L. Qu, R. Prins, Appl. Catal. A 250 (2003) 105.
[62] C. Moreau, C. Aubert, R. Durand, N. Zimita, P. Geneste, Catal. Today 4 (1988) 117.
[63] L. Vivier, S. Kasztelan, G. Perot, Bull. Soc. Chim. Belg. 100 (1991) 801.
[64] L. Vivier, P. d’Araujo, S. Kasztelan, G. Perot, Bull. Soc. Chim. Belg. 100 (1991) 806.
[65] M.S. Rana, B.N. Srinivas, S.K. Maity, G. Murali Dhar, T.S.R. Prasado Rao, J. Catal.
195 (2000) 31.
[66] M.L. Occelli, R. Chianelli, Hydrotreating Technology for Pollution Control, Marcel
Dekker 1996.
[67] H. Topsøe, B.S. Clausen, N.Y. Topsøe, P. Zeuthen, Stud. Surf. Sci. Catal. 53 (1990)
77.
[68] S. Kasztelan, Langmuir 6 (1990) 590.
[69] A. Wambeke, L. Jalowiecki, S. Kasztelan, J. Grimblot, J.P. Bonnelle, J. Catal. 109
(1988) 320.
[70] S. Kasztelan, L. Jalowiecki, A. Wambeke, J. Grimblot, J.P. Bonnelle, Bull. Soc. Chim.
Belg. 96 (1987) 1003.
[71] M.H. Farias, A.J. Gellman, G.A. Somorjai, R.R. Chianelli, K.S. Liang, Surf. Sci. 140
(1984) 181.
[72] C. Wivel, R. Candia, B.S. Clausen, S. Mørup, H. Topsøe, J. Catal. 68 (1981) 453.
[73] J.C. Duchet, E.M. van Oers, V.H.J. de Beer, R. Prins, J. Catal. 80 (1983) 386.
[74] J.P.R Vissers, V.H.J. de Beer, R. Prins, J. Chem. Soc. Faraday Trans. 83 (1987) 2145.
[75] N. Topsøe, H. Topsøe, J. Catal. 84 (1983) 386.
[76] S.P.A. Louwers, R. Prins, J. Catal. 133 (1992) 94.
[77] A.N. Startsev, S.A. Shkuropat, V.V. Kriventsov, D.I. Kochubey,
K.I. Zamaraev, Mendeleev Commun. (1991) 6.
[78] L. Medici, M. Harada, R. Prins, Mendeleev Commun. (1996) 121.
[79] R.G. Leliveld, A.J. van Dillen, J.W. Geus, D.C. Koningsberger, J. Phys. Chem. B 101
(1998) 11160.
Introduction Chapter 1 32
[80] R.G. Leliveld, A.J. van Dillen, J.W. Geus, D.C. Koningsberger, J. Catal. 175 (1998)
108.
[81] J.K. Nørskov, B.S. Clausen, H. Topsøe, Catal. Lett. 13 (1992) 1.
[82] H. Toulhoat, P. Raybaud, S. Kasztelan, G. Kresse, J. Hafner, Catal. Today 50 (1999)
629.
[83] C.D. Chang, W.H. Lang, US Pat. 4380669 (1980), to Mobil Comp.; C.D. Chang, P.D.
Perkins, Eur. Pat. 0082613 (1983), to Mobil Comp.
[84] T. Stamm, H.W. Kouwenhoven, D. Seebach, R. Prins, J. Catal. 155, (1995) 268.
[85] A. Tschumper, R. Prins, Appl. Catal. 172 (1998) 285; 174 (1998) 129.
[86] N. Nelson, R.B. Levy, J. Catal. 58 (1979) 485.
[87] M. Zdrazil, J. Catal. 141 (1993) 316.
[88] S. Rajagopal, R. Miranda, J. Catal. 141 (1993) 318.
[89] M. Fikry Ebeid, J. Pasek, Coll. Czech. Chem. Commun. 35 (1970) 2166.
[90] P. Hogen, J. Pasek, Coll. Czech. Chem. Commun. 39 (1974) 3696.
[91] S. Eijsbouts, C. Sudhakar, V.H.J. de Beer, R. Prins, J. Catal. 127 (1991) 605.
[92] M. Cattenot, J.L. Portefaix, J. Afonso, M. Breysse, M. Lacroix, G. Perot, J. Catal. 173
(1998) 366.
[93] J.L. Portefaix, M. Cattenot, M. Guerriche, J. Thivolle-Cazat, M. Breysse, Catal. Today
10 (1991) 473.
[94] J.L. Portefaix, M. Cattenot, M. Guerriche, M. Breysse, Catal. Lett. 9 (1991) 127.
[95] M. Egorova, Y. Zhao, P. Kukula, R. Prins, J. Catal. 206 (2002) 263; and chapter 2
[96] M. Breysse, J. Afonso, M. Lacroix, J.L. Portefaix, M. Vrinat, Bull. Soc. Chim. Belg.
100 (1991) 923.
[97] J.H. Lee, C.E. Hamrin, B.H. Davis, Appl. Catal. A 111 (1994) 11.
[98] P. Clark, X. Wang, P. Deck, S.T. Oyama, J. Catal. 210 (2002) 116.
[99] L. Vivier, L. Dominguez, G. Perot, S. Kasztelan, J. Mol. Catal. 67 (1991) 267.
[100] M. J. Ledoux, A. Bouassida, R. Benazouz, Appl. Catal. 9 (1984) 41.
[101] C. N. Satterfield, J. F. Cocchetto, AIChE J. 21 (1975) 1107.
[102] A. Benitez, J. Ramirez, A. Vazquez, D. Acosta, A. Lopez Agudo, J. Catal. 133 (1995)
103.
[103] H. S. Joo, J. A. Guin, Fuel Process. Techn. 49 (1996) 137.
[104] J. Cinibulk, Z. Vit, Appl. Catal. A 204 (2000) 107.
Introduction Chapter 1 33
[105] P. Betancourt, A. Rives, C. E. Scott, R. Hubaut, Catal. Today 57 (2000) 201.
[106] C. N. Satterfield, M. Modell, J.A. Wilkens, Ind. Eng. Chem. Proc. Des. Dev. 19
(1980) 154.
[107] R. T. Hanlon, Energy Fuels 1 (1987) 424.
[108] G. C. Hadjiloizou, J. B. Butt, J. S. Dranoff, Ind. Eng. Chem. Res. 31 (1992) 2503.
[109] M. Jian, J.L. Rico Cerda, R. Prins, Bull. Soc. Chim. Belg. 104 (1995) 225.
[110] M. Cerny, Coll. Czech. Chem. Commun. 47 (1982) 928.
[111] M. Cerny, A. Trka, Coll. Czech. Chem. Commun. 48 (1983) 1749.
[112] Z. Vit, M. Zdrazil, J. Catal. 119 (1989) 1.
[113] M. J. Ledoux, G. Agostini, R. Benazouz, O. Michaux, Bull. Soc. Chim. Belg. 93
(1984) 635.
[114] N.Y. Topsøe, H. Topsøe, F.E. Massoth, J. Catal. 119 (1989) 252.
[115] J. March, Advanced Organic Chemistry, 3rd ed., Chap. 10. Wiley, New York, 1985.
[116] M. Cerny, Coll. Czech. Chem. Commun. 44 (1979) 85.
[117] M. Flechsenhar, Dissertation 12471, ETH Zürich, 1997.
[118] F. Rota, Dissertation 14085, ETH Zurich, 2001.
HDN of 2-Methylpiperidine Chapter 2 35
2. On the role of β-hydrogen atoms in the hydrodenitrogenation of 2-
methylpyridine and 2-methylpiperidine
2.1 Abstract
The hydrodenitrogenation (HDN) of 2-methylpyridine and its intermediate products 2-
methylpiperidine, 1-aminohexane, and 2-aminohexane was studied. The presence of most
intermediates could be explained by a combination of pyridine ring hydrogenation, piperidine
ring opening by elimination, and nitrogen removal by elimination, as well as by nucleophilic
substitution of the amino group by a sulfhydryl group, followed by elimination of H2S or
hydrogenolysis of the C–S bond. Aminoalkenes, which are expected to be the primary
products of the ring opening of alkylpiperidine, were not observed, probably because of fast
hydrogenation to the corresponding amines. The ring opening of 2-methylpiperidine occurred
preferentially between the nitrogen atom and the methylene group, rather than between the
nitrogen atom and the carbon atom bearing the methyl group. This was confirmed by
comparative HDN experiments of piperidine, 2-methylpiperidine, and 2,6-dimethylpiperidine.
Although the methyl groups offer extra β hydrogen atoms, these primary hydrogen atoms are
not used for elimination. Instead, the methyl groups hinder the adsorption leading to the
elimination of the β hydrogen atoms on the side of the molecule bearing the methyl group.
2.2 Introduction
Heterocyclic compounds like pyridine, quinoline, and acridine are the main nitrogen-
containing compounds in oil. They are removed by hydrodenitrogenation (HDN) in a
hydrotreating process in which gasoline or gas oil is treated with hydrogen over a metal
sulfide catalyst like nickel promoted molybdenum sulfide (Ni–MoS2) supported on alumina
[1]. Several groups have studied the HDN of pyridine [2–7], because, as the smallest nitrogen-
containing heterocyclic molecule, pyridine was believed to be the simplest model molecule to
HDN of 2-Methylpiperidine Chapter 2 36
study HDN. Although the network of reactions taking place in the HDN of pyridine is now
well understood, the study of the kinetics of the HDN of pyridine proved to be extremely
difficult. The reason for this difficulty is the occurrence of a side reaction of piperidine, the
first intermediate in the HDN of pyridine. Two piperidine molecules disproportionate to N-
pentylpiperidine and ammonia [1–5]. Opening of the piperidine ring and removal of ammonia
can take place from piperidine as well as from N-pentylpiperidine. Consequently the network
of the HDN becomes very complicated and a trustworthy kinetic analysis of the separate
reactions is almost impossible.
The disproportionation of piperidine to N-pentylpiperidine takes place by nucleophilic
substitution at the carbon atom in the α position to the nitrogen atom in the piperidine ring
(Fig. 2.1) [3, 8]. It is well known that a nucleophilic attack is hindered by substitution on the α
carbon atom [9]. Substitution of a hydrogen atom by a methyl group on the α carbon atom
might therefore hinder the disproportionation so much, that it is strongly suppressed and that
it hardly interferes with the other reactions taking place during the HDN of pyridine and
piperidine. Therefore we decided to study the HDN of 2-methylpyridine and 2-
methylpiperidine.
NN
+H H H
NH H
NH N
H HN
NH H
N N + NH3
-H
+ H2
Fig. 2.1. Mechanism of the disproportionation of piperidine.
2-Methylpyridine and 2-methylpiperidine were studied before by Cerny and Trka [10,
11] and Ren et al. [12]. Ren et al. studied the Langmuir–Hinshelwood–Hougen–Watson
kinetics of the HDN of 2-methylpyridine in a continuous-flow reactor at 4.9 MPa and 240–
280 °C [12]. They observed 2-methylpiperidine as primary product and hexane and
cyclohexane as final products. No intermediates between 2-methylpiperidine and hexane were
reported. Cerny and Trka performed their investigations in an autoclave at 15.5 MPa and 250
°C. Because of the high H2 pressure, low temperature, and absence of H2S in their
HDN of 2-Methylpiperidine Chapter 2 37
experiments, mainly ring hydrogenation and only a small amount of products due to nitrogen
removal were observed. They concluded that the 2-methylpiperidine ring opens preferentially
on the side that does not contain the methyl group and that the HDN reactions of more
substituted pyridine derivatives are slower [10]. This is in disagreement with the results of
Portefaix et al., who observed that the HDN reaction of 2,6-dimethylpiperidine was faster
than that of piperidine [13]. Their result suggests that the presence of a methyl group leads to
faster ring opening. Portefaix et al. performed their HDN work at the much lower H2 pressure
of 2 MPa and relatively high H2S pressure of 33.3 kPa; this may explain the different results.
Further study is clearly called for.
Another reason for studying the HDN of 2-methylpiperidine is the presence of three
additional hydrogen atoms on the methyl carbon atom in β position relative to the nitrogen
atom. HDN occurs (partly) via Hofmann elimination in which, on the one hand, the bond
between the α carbon atom and the nitrogen atom is broken and, on the other hand, the bond
between a hydrogen atom and the β carbon atom is broken. Portefaix et al. compared the
HDNs of piperidine, 3,5-dimethylpiperidine, and 2,6-dimethylpiperidine and concluded that
Hofmann elimination is quicker when more β hydrogen atoms are present [13]. This
implicates elimination of a β H atom from the methyl groups in 2,6-dimethylpiperidine as an
important step in the HDN of this molecule. Portefaix et al. reported only the conversion of
the reactant and nothing about the resulting products. Therefore, it seemed of interest to
investigate if the elimination reaction of 2-methylpiperidine takes place by removal of a
hydrogen atom from the methyl group and leads preferentially to 1-aminohexane.
2.3 Results
2.3.1. HDN of 2-Methylpyridine
The results of the HDN of 2-methylpyridine at 340 °C, 4.8 MPa H2, and 20 kPa H2S are
shown in Figure 2.2. No products with mass higher than that of the reactant (such as
condensation products) were observed and the mass balance was always better than 95%. The
product selectivities show (Fig. 2.2b) that 2-methylpiperidine is the only primary product, as
expected, since the HDN of heterocyclic N-containing aromatic molecules can occur only
HDN of 2-Methylpiperidine Chapter 2 38
after ring hydrogenation [1, 3, 14, 15]. The maximum yield of 34% 2-methylpiperidine and its
selectivity against 2-methylpyridine conversion indicate that the ratio of the effective rate
constants of formation and further reaction of 2-methylpiperidine is about 0.8 [16].
Fig. 2.2. Relative concentrations (a) and selectivities (b) of the products of the HDN of 2-
methylpyridine as a function of weight time.
2-Hexene (cis and trans), 1-hexene, and hexane were observed as the main secondary
products (Fig. 2.2b). These products are actually expected to be tertiary products, because
HDN of aliphatic amines is generally considered to occur by Hofmann elimination or by
nucleophilic substitution of the NH2 group by an SH group followed by elimination or
hydrogenolysis [1, 14]. In either case, the nitrogen atom of 2-methylpiperidine is removed in
two steps. The first step is a ring opening by C–N bond breaking and the second step is the
removal of the nitrogen atom in the form of ammonia by breaking the other C–N bond. Of the
products that are possible after the first C–N bond breakage, only traces of 1-aminohexane
and 2-aminohexane were observed. The reason is that their rates of further reaction are much
higher than their rates of formation, as discussed in Sections 4 and 5. These amines have a
high basicity and thus larger equilibrium adsorption constants than 2-methylpyridine. Even at
low concentration they may therefore have an important (inhibiting) influence on the HDN
kinetics [15]. For that reason, the HDN of 2-methylpiperidine, the primary product of the
HDN of 2-methylpyridine, and of 1-aminohexane and 2-aminohexane, the expected
secondary (or tertiary, see below) products, were studied in detail as well.
HDN of 2-Methylpiperidine Chapter 2 39
2.3.2. HDN of 2-Methylpiperidine
The HDN of 2-methylpiperidine was carried out at 340 °C, 4.8 MPa H2, and 20 kPa
H2S. Figure 2.3 shows that at least four compounds have nonzero selectivity at zero
conversion of 2-methylpiperidine and thus might be considered primary products. Three of
these products were identified by their GC retention times and mass spectra as 1-
aminohexane, 2-aminohexane, and 2-methylpyridine. The yield of 2-aminohexane was much
higher than that of 1-aminohexane. This confirms the results of Cerny [10], although they
were obtained under quite different conditions, and suggests that the bond between the N
atom and the methylene group is more easily broken than that between the N atom and the
CH(CH3) group. This is also perfectly in line with the results of Cattenot et al. [8] and Vivier
et al. [17], indicating that the amino group bonded to a methylene group cleaves very easily
by nucleophilic substitution (SN2).
GC–MS showed that the fourth compound had a molecular weight of 97, but no
commercially available compound could be found that had the same retention time and a
matching mass spectrum. Therefore, the product of the HDN reaction was collected and a
fraction that contained the basic nitrogen-containing molecules was separated from a
hydrocarbon fraction. Since pulsed flame photometric detection had shown that the fourth
compound contains a nitrogen atom, it was extracted from the HDN product with an aqueous
HCl solution. Neutralization of this aqueous extract and subsequent extraction with
chloroform gave a chloroform solution of all primary products as well as the remaining 2-
methylpiperidine. After evaporation of the chloroform, the mixture of nitrogen-containing
compounds was separated by column chromatography using silicagel and a 50 : 50 : 1
solution of CH3OH: CHCl3 :NH4OH (25% aqua solution of NH3) as a mobile phase. The
fraction containing the fourth unknown product was evaporated and the raw material obtained
with a purity of 90% was analyzed by NMR spectroscopy. The 1H NMR spectrum of the
fourth compound in CDCl3 showed peaks at δ 3.46–3.52 (m, 2H, CH2N), 2.13 (t of t, 3 J =6.5
Hz, 5 J =1.8 Hz, 2H, CH2C=), 1.91 (t, 5 J =1.8 Hz, 3H, CH3), 1.62–1.71 (m, 2H, CH2CH2C=),
and 1.51–1.59 ppm (m, 2H, CH2CH2N), while its 13C NMR spectrum in CDCl3 showed peaks
at δ 168.45 (s, CN), 48.98 (t, CN), 30.26 (t, CH2C=), 27.33 (q, CH3), 21.57 (t, CH2CH2N),
and 19.52 ppm (t, CH2CH2C==). The NMR spectra together with the mass spectrum obtained
(MS (EI, 70 eV) m/z 97 (M+, 63), 96 (11), 69 (61), 68 (21), 56 (26), 55 (17), 54 (15), 42 (100),
HDN of 2-Methylpiperidine Chapter 2 40
41 (65), 39 (33), 28 (41), 27 (26)) enabled us to identify the fourth primary product as 2-
methyl-3,4,5,6-tetrahydropyridine. Both NMR and mass spectra were in accord with the
spectra assigned to this molecule in the literature [18, 19].
Fig. 2.3 Relative concentration (a) and selectivities (b) of the products of the HDN of 2-
methylpiperidine as a function of weight time.
HDN experiments with 2-methylpiperidine under conditions other than 340 °C and 5
MPa suggested that 2-methyl-3,4,5,6-tetrahydropyridine is formed by a catalytic as well as a
thermal reaction. Experiments in the empty steel reactor and in the reactor filled with SiC
only, without catalyst, showed that the 2-methyl-3,4,5,6-tetrahydropyridine yield increased by
increasing the temperature from 300 to 350 °C, as expected for a simple dehydrogenation
reaction. Over the NiMo/Al2O3 catalyst diluted with SiC, however, the yield decreased
substantially from 300 to 350 °C. The 2-methyl-3,4,5,6-tetrahydropyridine yield was also
lower over the catalyst than over the SiC or in the empty reactor. This suggests that in the
presence of the catalyst, there occurs not only dehydrogenation but also reactions to other
products, which lower the yield of 2-methyl-3,4,5,6-tetrahydropyridine.As expected, in the
presence of the catalyst, decreasing the H2 pressure from 5 to 3 MPa raised the 2-methyl-
3,4,5,6-tetrahydropyridine yield.
2.3.3. Comparison of Piperidine, 2-Methylpiperidine, and 2,6-Dimethylpiperidine
As indicated in the Introduction, Portefaix et al. reported that the amount of HDN
product was larger for 2,6-dimethylpiperidine than for piperidine under the following reaction
conditions: 275 °C, 2 MPa H2, and 33.3 kPa H2S [13]. They related this to the presence of
HDN of 2-Methylpiperidine Chapter 2 41
more β H atoms in 2,6-dimethylpiperidine, which would facilitate Hofmann elimination.
These results seem to contradict those of Cerny [10] and our results for 2-methylpiperidine
described in the previous section, which suggested that the ring opening occurs preferentially
between the nitrogen atom and the methylene group and not between the nitrogen atom and
the carbon atom bearing the methyl group. We therefore decided to repeat the measurements
of Portefaix et al. under their conditions.
From the results presented in Figure 2.4 it is clear that the conversion of piperidine is
very slow and hardly reaches 2% at a weight time of 10 g.min/mol, whereas the conversion of
2-methylpiperidine is almost 20% and that of cis-2,6-dimethylpiperidine is more than 50% at
the same weight time. The conversions at a weight time of 2.4 g.min/mol are in good
agreement with those of Portefaix et al. [13]. Analyzing the resulting products, however, we
found that for piperidine the main product was not that of HDN but pyridine. For 2-
methylpiperidine the main products were 2-methylpyridine (17%) and 2-methyl-3,4,5,6-
tetrahydropyridine (64%), and for cis-2,6-dimethylpiperidine the main products were 2,6-
dimethylpyridine (2%), 2,6-dimethyl-3,4,5,6-tetrahydropyridine (66%), and trans-2,6-
dimethylpiperidine (32%), with the selectivities in parentheses.
Fig. 2.4 Total conversion in the HDN of piperidine, 2-methylpiperidine, and cis-2,6-
dimethylpiperidine as a function of weight time at 275 °C, 2 MPa, and 33.3 kPa
H2S.
The observed high selectivities to fully dehydrogenated pyridine molecules at low
conversions are not in contradiction with thermodynamics, which indicates that the
pyridine/piperidine ratio cannot be higher than 0.01 at 275 °C and 2 MPa H2 [20]. The
tetrahydropyridine/piperidine ratio can be much higher, however. Portefaix et al. apparently
HDN of 2-Methylpiperidine Chapter 2 42
underestimated the latter ratio and the isomerization of cis-2,6-dimethylpiperidine to trans-
2,6-dimethylpiperidine, when assuming, without any product analysis, that most of the
piperidine-type molecules would convert to HDN products. At high weight time and high
conversion, thermodynamic controls and HDN products will indeed dominate. At low
conversion, however, kinetics may dominate the product distribution and it is in this regime
that mechanistic results should be obtained.
All the compounds mentioned above are products of dehydrogenation and
isomerization, and not of HDN or C–N bond cleavage. The selectivities for ring opening and
HDN were calculated from the sum of the observed amines and saturated and unsaturated
hydrocarbons and amounted to 9% for piperidine, 5% for 2-methylpiperidine, and 0.7% for
cis-2,6-dimethylpiperidine at 2.4 g.min/mol. These selectivities are small; dehydrogenation
and isomerization (for cis-2,6-dimethylpiperidine) dominate at the low H2 pressure of 1.8
MPa. The yield (selectivity times conversion) of these ring opening and HDN products was
indeed higher for cis-2,6-dimethylpiperidine than for piperidine, as reported by Portefaix et al.
[13].
We also studied these three piperidine molecules under conditions in which elimination
is the dominating reaction, so that a fair comparison of the HDN rates of the three molecules
could be made. At 340 °C, 5 MPa, and 20 kPa H2S the 2-methylpiperidine conversion was
20% lower than that of piperidine, while the conversion of cis-2,6-dimethylpiperidine was
higher than that of piperidine below a weight time of τ = 5.5 g.min/mol and lower above this
value. The reason for the high initial reaction rate of cis-2,6-dimethylpiperidine is fast
isomerization of cis to trans-2,6-dimethylpiperidine. For τ >5.5 g.min/mol, the equilibrium
between cis- and trans-2,6-dimethylpiperidine is established, and other, slower reactions
determine the reaction rate of both isomers of 2,6-dimethylpiperidine. Even at 340 °C, 5 MPa,
and 20 kPa H2S, conversion to products other than obtained by HDN is not negligible for
these three molecules. For piperidine the total selectivity for dehydrogenation to pyridine and
disproportionation to N-pentylpiperidine was always below 10%. For 2-methylpiperidine and
cis-2,6-dimethylpiperidine the selectivities to the dehydrogenation products (substituted
pyridine and tetrahydropyridine) were 22 and 24% respectively, at the lowest weight time
measured (1.4 g.min/mol). Taking into account only the products of hydrodenitrogenation, we
found that piperidine undergoes HDN 30% faster than 2-methylpiperidine and 50% faster
than 2,6-dimethylpiperidine (Fig. 2.5).
HDN of 2-Methylpiperidine Chapter 2 43
Fig. 2.5 HDN conversions in the HDN of piperidine, 2-methylpiperidine, and cis-2,6-
dimethylpiperidine as a function of weight time at 340 °C, 5 MPa, and 20 kPa H2S.
2.3.4. HDN of 1-Aminohexane
The HDN of 1-aminohexane becomes fast above 300 °C (Fig. 2.6) and 2-hexene and
hexane are the main products. A plot of the product selectivities versus weight time (Fig. 2.7)
shows that 1-hexene and trans- and cis-2-hexene are primary products. According to the
Hofmann elimination mechanism only 1-hexene can be formed from 1-aminohexane.
However, the isomerization of 1-hexene to 2-hexene is so fast above 300 °C that it is difficult
to distinguish if 2-hexene is a primary or secondary product. Addition of 1-pentene to the feed
indeed showed that the isomerization to cis- and trans-2-pentene was fast. This means that the
ratio of 1-hexene to 2-hexene above 300 °C is determined mainly by thermodynamics and
hardly by the kinetics of the formation of these alkenes. Consequently, the ratio of 1-hexene
to 2-hexene cannot be used to distinguish between the two ways of C–N bond breakage in 2-
methylpiperidine either. At 260 °C the conversion of 1-aminohexane is less than 5%, even at
high weight time (20 g.min/mol), versus 50% at 300 °C. Comparison of the selectivity plots at
300 °C (Fig. 2.7) and 260 °C (Fig. 2.8) confirms that trans- and cis-2-hexene are secondary
products, because the selectivities decrease at decreasing temperature and conversion.
HDN of 2-Methylpiperidine Chapter 2 44
Fig. 2.6. Conversion and relative product concentration in the HDN of hexylamine between
280 and 340 °C and at τ = 5 g.min/mol
Fig. 2.7. Product selectivities of the HDN of hexylamine at 300 °C.
Fig. 2. 8 Product selectivities in the HDN of 1-aminohexane at 260 °C
HDN of 2-Methylpiperidine Chapter 2 45
At the higher H2S pressure of 80 kPa, the selectivity to hexane was higher than at 16
kPa. HDN activity was hardly influenced by this change in H2S partial pressure (at a constant
H2 pressure of 3.8 MPa), but selectivity did change. Not only was hexane selectivity higher,
but 2-hexene selectivity was substantially lower and 1-hexene selectivity higher at 80 kPa
H2S. Apparently, isomerization of 1-hexene to cis- and trans-2-hexene requires vacancies at
the metal sulfide surface. The higher selectivity toward hexane formation indicates that
nucleophilic attack of H2S on 1-aminohexane must have led to hexanethiol, which very
quickly reacted to hexane by hydrogenolysis and 1-hexene by elimination [1].
2.3.5. HDN of 2-Aminohexane
The HDN of 2-aminohexane was complicated by the formation of di-2-hexylamine, a
disproportionation product of the reaction of two 2-aminohexane molecules (Fig. 2.1). As
expected for this molecule with two chiral atoms (2-aminohexane itself has one chiral atom),
the gas chromatogram showed two peaks of equal intensity, equal mass spectra, and only a
small difference in retention time. One peak belongs to the (R,R)- and (S,S)-isomers, the other
to the meso (R,S)-isomer.
Experiments between 220 and 350 °C showed that not only di-2-hexylamine but also 1-
hexene, and cis- and trans-2-hexene behave as a primary product (Fig. 2.9). Hofmann
elimination explains why 1-hexene as well as 2-hexene is formed. The activation energy for
elimination is higher than that of nucleophilic substitution because the hexene selectivity
increased with temperature. The selectivity of di-2-hexylamine is very high at low
temperature. At the lowest temperature studied (220 °C), it was higher than 90% when
extrapolated to zero 2-aminohexane conversion. In this case, the only other product was
hexane. The formation of hexane is explained by nucleophilic substitution of the NH2 group
by an SH group, followed by hydrogenolysis of the C–S bond.
Increasing the H2S pressure from 16 to 80 kPa, at the same H2 pressure of 3.8 MPa, led
to faster conversion of 2-aminohexane to hydrocarbons, while production of the
disproportionation product di-2-hexylamine decreased (Fig. 2.10). Whereas at 16 kPa H2S it
reached a maximum yield of 25%, at 80 kPa H2S the maximum yield was only 10%. At the
higher H2S partial pressure, a new intermediate was observed. It behaved as a primary product
and was analyzed to be 2-hexanethiol. This intermediate is formed by an SN2 reaction
HDN of 2-Methylpiperidine Chapter 2 46
between 2-aminohexane and H2S. At higher H2S partial pressure, it will be formed faster and
will hydrogenolyze less rapidly to hexane because of fewer vacancies on the metal sulfide
surface. Therefore, it is easier to observe 2-hexanethiol at higher H2S pressure.
Fig. 2.9. Relative concentrations (a) and selectivities (b) of the products of the HDN of 2-
hexylamine as a function of weight time at 20 kPa H2S.
Fig. 2.10. Product selectivities of the HDN of 2-hexylamine as a function of weight time at 80
kPa H2S.
HDN of 2-Methylpiperidine Chapter 2 47
2.4 Discussion
Combining the results of the HDN of 2-methylpyridine, 2-methylpiperidine, 1-
aminohexane, and 2-aminohexane, we arrive at the reaction scheme presented in Fig. 2.11.
For all intermediates, except two, direct relationships between parent and daughter molecules
could be established by measuring the product selectivities as a function of weight time and
extrapolating to zero weight time. Thus, 2-methylpiperidine proved to be the primary product
of 2-methylpyridine, while 2-methylpyridine as well as 2-methyl-3,4,5,6-tetrahydropyridine
behaved as primary products of 2-methylpiperidine.
The other two apparent primary products in the HDN of 2-methylpiperidine, 1-
aminohexane and 2-aminohexane, should actually be secondary rather than primary products.
If opening of the piperidine ring would occur by Hofmann elimination, it would lead to 5-
amino-1-hexene when the C–N bond with the methylene group is broken, and to 6-amino-1-
hexene and 6-amino-2-hexene when the C–N bond with the CH(CH3) group is broken (Fig.
2.11). These products were not detected in the HDN of 2-methylpiperidine. The equivalent of
5-amino-1-hexene has never been observed in the HDN of pyridine either [6]. The reason is
most probably that these aminoalkenes adsorb strongly on the catalyst surface because of the
presence of a nitrogen atom in the molecule and are very quickly hydrogenated to the
corresponding saturated amines before they desorb from the catalytic site. Alternatively, if
opening of the pyridine ring would occur by nucleophilic attack by H2S, then 5-
aminohexanethiol, 6-aminohexanethiol, and 6-amino-2-hexanethiol would be primary
products. Thiols react very quickly by elimination to alkenes and by hydrogenolysis to
alkanes. In the first and most important case, aminoalkenes should be formed; in the latter
case, amines. Again, because of strong adsorption and fast hydrogenation, the aminoalkenes
have not been detected. As a result, only 1-aminohexane and 2-aminohexane occur in the
product, their selectivities do not go to zero at low conversion, and they behave as (quasi)
primary products in the HDN of 2-methylpiperidine.
HDN of 2-Methylpiperidine Chapter 2 48
NH
N
N
NH2
H2N
NH2
H2N
Fig. 2.11. Scheme of the reaction network of the HDN of 2-methylpiperidine and 2-
methylpyridine.
2-Methylpyridine and 2-methyl-3,4,5,6-tetrahydropyridine both behaved as primary
products in the HDN of 2-methylpiperidine. One might expect 2-methyl-3,4,5,6-
tetrahydropyridine to be the dehydrogenation intermediate between 2-methylpiperidine and 2-
methylpyridine, in which case it is surprising that 2-methylpyridine behaves as a primary
product too. If the rate of dehydrogenation of 2-methyl-3,4,5,6-tetrahydropyridine to 2-
methylpyridine is of the same order of magnitude as its rate of desorption from the catalytic
site, both molecules might have nonzero selectivities at zero 2-methylpiperidine conversion.
Another explanation could be that 2-methyl-3,4,5,6-tetrahydropyridine is (partially) produced
by a thermal dehydrogenation reaction, while 2-methylpyridine is directly, without desorption
of intermediates, produced by a catalytic reaction. We have not studied this question any
further, because it is only a side effect in our study of the HDN of 2-methylpyridine and 2-
methylpiperidine.
An investigation of the HDN of 1-aminohexane and 2-aminohexane is important not
only to gain a better understanding of the kinetics of the HDN of 2-methylpyridine and 2-
methylpiperidine, but also to understand how the ring opening of the piperidine ring takes
place. Because of the methyl group in α position to the nitrogen atom, C–N bond breakage in
2-methylpiperidine can take place in two ways: between the nitrogen atom and the carbon
atom of the methylene group, or between the nitrogen atom and the carbon atom carrying the
methyl group. The latter possibility should prevail if, as suggested by Portefaix et al. [13, 21]
and Cattenot et al. [8], the number of β H atoms determines the course of the Hofmann
elimination reaction. Unfortunately, the ratio of the concentrations of 1-aminohexane and 2-
aminohexane cannot be used as a direct measure of the ratio of the N–CH2 and N–CH(CH3)
bond breaks. The reason is that the concentrations of these amines depend not only on their
HDN of 2-Methylpiperidine Chapter 2 49
rates of formation, but also on their rates of reaction to hexenes and hexane. Thus, the very
small amount of 1-aminohexane that is produced in the HDN of 2-methylpiperidine (Fig.
2.3b) can be due to either slow breaking of the N–CH(CH3) bond, or rapid disappearance of
1-aminohexane by HDN, or both. For that reason, it was necessary to investigate the HDNs of
1-aminohexane and 2-aminohexane separately.
Comparison of the conversions of 2-aminohexane (Fig. 2.9a) and 1-aminohexane (not
shown) showed that the reactivity of 2-aminohexane is higher than that of 1-aminohexane.
Despite a higher reactivity, much more 2-aminohexane than 1-aminohexane was detected in
the HDN of 2-methylpiperidine (Fig. 2.3b). This proves that the first C–N bond break in 2-
methylpiperidine occurs predominantly between the nitrogen atom and the carbon atom of the
methylene group. If only the number of β H atoms plays a role, as suggested by Portefaix et
al. [13], then 2.5 times more 1-aminohexane than 2-aminohexane should have been formed.
Actually, 3 to 4 times more 2-aminohexane was formed! It is clear that the number of β H
atoms is not the most important factor in Hofmann elimination. The same conclusion was
reached in the HDN of 2-methylcyclohexylamine, in which the type of β H atom proved to be
the most important factor [22]. Thus, the β H atom at the tertiary carbon atom was removed
much faster than the β H atom at the secondary carbon atom, leading to more 1-
methylcyclohexene than 3-methylcyclohexene. Analogously, the results of the HDN of 2-
aminohexane described in Section 2.3.5 demonstrated that 3 to 4 times more 2-hexene was
produced than 1-hexene, although there are 1.5 times fewer H atoms on the CH2 group in β
position to the nitrogen atom than on the CH3 group in 2-aminohexane. It is clear that the ease
with which the C–H bond breaks plays an important role in the elimination. A hydrogen atom
on a tertiary carbon atom is more easily abstracted by a base than β H atoms on secondary or
primary carbon atoms. This is the basis of the Zaytzev rule, which states that elimination
preferentially leads to more substituted alkenes [9].
The fact that much more 2-aminohexane than 1-aminohexane is formed in the HDN of
2-methylpiperidine further indicates that the methyl group actually has a negative rather than
a positive influence on the elimination. All β H atoms on the two carbon atoms in β position
to the nitrogen atom belong to methylene groups. Thus, they should have the same tendency
to be eliminated. If the methyl group played no role in elimination, neither positive nor
negative, then, on the basis of the number and type (methylene) of H atoms, equal amounts of
2-aminohexane and 1-aminohexane should have been formed. The fact that the rate of
HDN of 2-Methylpiperidine Chapter 2 50
breakage of the N–CH(CH3) bond is lower than that of the N–CH2 bond indicates that the
methyl group hinders the adsorption of 2-methylpiperidine in a conformation in which the
nitrogen atom and the β H atom of the methylene group next to the CH(CH3) group approach
the metal sulfide surface. Such a steric hindrance does not exist for the adsorption of the other
side of the 2-methylpiperidine molecule on the metal sulfide surface. Our results are in good
agreement with the rule that nucleophilic substitution is favored at low temperature, while
elimination is favored at high temperature. The H2S pressure may also steer the reaction in
different directions. At low H2S pressure, nucleophilic substitution is dominated by the
reaction of an amine reactant with another amine molecule, leading to disproportionation
products such as di-2-hexylamine (Fig. 2.9). In this sense, the metal sulfide surface that is
depleted of sulfur behaves similarly to a metal surface, on which disproportionation of amines
is important as well [23]. At high H2S pressure, H2S becomes the dominant nucleophile that
reacts with the amine, transforming the amine into a thiol molecule that reacts relatively
quickly to an alkene by elimination and to an alkane by hydrogenolysis. Hydrogenolysis
requires sulfur vacancies at the metal sulfide surface. Consequently, an increase in H2S
pressure has a positive effect on hexane formation than lower H2S pressures, because more
thiol is formed. At higher H2S pressures, however, hardly any vacancies are available
anymore and the thiol can undergo only elimination to an alkene.
From the higher rate of HDN conversion of cis-2,6-dimethylpiperidine than of
piperidine, Portefaix et al. concluded that the rate of elimination of ammonia from an amine is
larger when more β Hatoms are present [13]. Our analysis of all the products of the HDN
reactions of piperidine, 2-methylpiperidine, and cis-2,6-dimethylpiperidine shows that this
conclusion is not correct. Indeed, the rate of disappearance of cis-2,6-dimethylpiperidine is
higher than that of piperidine (Fig. 4) at 275 °C, 2 MPa H2, and 33.3 kPa H2S. However, the
majority of the product at 2 MPa H2 is not formed by elimination, but rather by
dehydrogenation and isomerization. On the other hand, at 340 °C and 5 MPa,
dehydrogenation is much less important and the main products were formed by ring opening
and HDN. Under such conditions, the rate of elimination decreases from piperidine to 2-
methylpiperidine to cis-2,6-dimethylpiperidine. This is then in agreement with our
observation that the 2-methylpiperidine ring is preferentially opened between the N atom and
the methylene group. Thus, it is clear that, contrary to the proposal of Portefaix et al. [13], the
addition of a methyl group in α position to the nitrogen atom in piperidine does not increase
HDN of 2-Methylpiperidine Chapter 2 51
the HDN rate. On the contrary, the methyl group constitutes a strong steric hindrance for the
right adsorption conformation of the nitrogen atom and the β H atom.
2.5 References
[1] R. Prins, Adv. Catal. 46 (2001) 399.
[2] J. Sonnemans, W.J. Neyens, P. Mars, J. Catal. 34 (1974) 230.
[3] H. Schulz, M. Schon, N.M. Rahman, Stud. Surf. Sci. Catal. 27 (1986) 201.
[4] R.T. Hanlon, Energy Fuels 1 (1987) 424.
[5] G.C. Hadjiloizou, J.B. Butt, J.S. Dranoff, Ind. Eng. Chem. Res. 31 (1992) 2503.
[6] M. Jian, R. Prins, Catal. Lett. 35 (1995) 193.
[7] R. Pille, G. Froment, Stud. Surf. Sci. Catal. 106 (1997) 403.
[8] M. Cattenot, J.L. Portefaix, A. Afonso, M. Breysse, M. Lacroix, G. Perot, J. Catal.
173 (1998) 366.
[9] J. March, “Advanced Organic Chemistry,” 3rd ed., Chap. 10. Wiley, New York, 1985.
[10] M. Cerny, Coll. Czech. Chem. Commun. 44 (1979) 85.
[11] M. Cerny, A. Trka, Czech. Chem. Commun. 48 (1983) 3413.
[12] S. Ren, Z. Wang, Y. Hu, Ranliao Huaxue Xuebo 15 (1987) 255.
[13] J.L. Portefaix, M. Cattenot, M. Gerriche, J. Thivolle-Cazat, M. Breysse, Catal. Today
10 (1991) 473.
[14] N. Nelson, R.B. Levy, J. Catal. 58 (1979) 485.
[15] G. Perot, Catal. Today 10 (1991) 447.
[16] O. Levenspiel, Chemical Reaction Engineering, 3rd ed., Chap. 8. Wiley, New York,
1999.
[17] L. Vivier, V. Dominguez, G. Perot, S. Kasztelan, J. Mol. Catal. 67 (1991) 267.
[18] G. Asensio, M.E. Gonzales-Nunez, C.B. Bernardini, R. Mello, W. Adam, J. Am.
Chem. Soc. 105 (1983) 6877.
[19] D.H. Hua, W.M. Shou, S. Narasimha Bharathi, T. Katsuhira, A.A. Bravo, J. Org.
Chem. 55 (1990) 3682.
[20] J.F. Cocchetto, C.N. Satterfield, Ind. Eng. Chem. Process Des. Dev. 15 (1976) 272.
HDN of 2-Methylpiperidine Chapter 2 52
[21] J.F. Portefaix, M. Cattenot, M. Guerriche, M. Breysse, Catal. Lett. 9 (1991) 127.
[22] F. Rota, V. Ranade, R. Prins, J. Catal. 201 (2001) 389.
[23] G. Meitzner, W.J. Mykytka, J.H. Sinfelt, J. Catal. 98 (1986) 513.
HDN of n-Hexylamines Chapter 3 53
3. Investigation of the mechanism of the hydrodenitrogenation of n-
hexylamines over sulfided NiMo/γ-Al2O3
3.1 Abstract
The hydrodenitrogenation (HDN) of n-hexylamine, dihexylamine, and trihexylamine
was studied between 300 and 340 °C, 3 and 5 MPa total pressure, 5 and 20 kPa amine
pressure, and 10 and 150 kPa H2S pressure over a sulfided Ni-Mo/γ-Al2O3 catalyst. The
conversion increased with the H2 pressure and decreased with increasing partial pressure of
the hexylamines. The conversion of hexylamine and dihexylamine decreased slightly with
H2S pressure, but that of trihexylamine increased substantially. The contributions of
elimination and nucleophilic substitution to the HDN were determined by the initial product
selectivities at short weight time. The initial alkene selectivities were low and accounted for
only a minor part of the n-alkylamine conversion. Since the hexene/hexane branching ratio in
the HDN of the alkylamines was almost equal to that in the hydrodesulfurization of
pentanethiol in the presence of an alkylamine, it was concluded that the majority of hexene in
the HDN of the hexylamines originates from hexanethiol. Nucleophilic substitution of the
hexylamines with H2S to give an alkanethiol is the predominant HDN reaction of all three n-
hexylamines.
3.2 Introduction
Reactions such as hydrodenitrogenation (HDN) and hydrodesulfurization (HDS) take
place in the hydrotreating of oil fractions, one of the most important catalytic processes in the
petroleum industry. A considerable number of studies have led to a better understanding of
the mechanisms involved in these reactions. It is generally accepted [1-9] that the first step in
the HDN of nitrogen-containing aromatic molecules is the hydrogenation of the heterocyclic
ring. Only after the breaking of the aromaticity can C-N bond cleavage in the resulting
HDN of n-Hexylamines Chapter 3 54
saturated molecules take place. Several mechanisms of the C-N bond scission and nitrogen
removal have been proposed [1,10-12]. Nelson and Levy [1] were the first to suggest
Hofmann-type elimination and nucleophilic substitution as mechanisms for C-N bond scission
of aliphatic nitrogen-containing molecules. The initial step in C-N bond scission is the
addition of a proton to a nitrogen lone pair with the formation of a quarternary ammonium
compound, which provides a better leaving group than the amine group. C-N bond scission
can then occur via elimination of a β−hydrogen atom with the formation of an alkene or via
nucleophilic substitution of the amine group at the α−carbon atom by a sulfhydryl group to
form an alkanethiol.
Several studies [6,13-18] have dealt with the HDN of aliphatic amines over different
catalysts. Portefaix et al. showed in HDN studies at 2 MPa over sulfided NiMo/Al2O3 that an
increase in the number of β−hydrogen atoms in pentylamines and piperidines increased the
conversion of these molecules [6,13]. This was taken as proof that aliphatic C-N bond
cleavage takes place by Hofmann elimination. However, they did not measure the reaction
products and ascribed the total conversion of the amines to the rate of HDN. We showed that
a substantial part of the conversion went to dehydrogenated molecules rather than HDN
products [14]. If only the HDN products were taken into account, then the introduction of a
methyl group onto the α−carbon atom of piperidine and, thus an increase in the number of
β−hydrogen atoms, actually decreased the rate of nitrogen removal. Furthermore, ring
opening of 2-methylpiperidine by C-N bond cleavage occurred preferentially on the CH2-N
side and not on the CH(CH3)-N side. This and the observation of thiol intermediates in the
HDN of 2-methylpiperidine [14] and methylcyclohexylamine [19] suggests that nucleophilic
substitution is even important in the HDN of aliphatic amines that contain β-hydrogen atoms.
Vivier et al. were the first to prove that C-N bond cleavage by nucleophilic substitution
can take place in the HDN of amines [7]. They observed that benzylamine and α,α-
diphenylmethylamine, which do not have β−hydrogen atoms and thus cannot react by
elimination, react fast to toluene and diphenylmethane respectively. Benzylamine and α,α-
diphenylmethylamine react most probably by nucleophilic substitution of the amine group by
an SH group followed by rapid hydrogenolysis of the intermediate thiol. The C-N bond
cleavage in these molecules may be of the SN1 and not SN2 type, because of the stabilizing
influence of the phenyl groups on the intermediate carbenium ion that results from removal of
the amine group. Cattenot et al. showed that both elimination and nucleophilic substitution
HDN of n-Hexylamines Chapter 3 55
play a role in the C-N bond scission of pentylamines on unsupported transition-metal sulfides
at atmospheric pressure [15]. The ratio of the two mechanisms depended on the type of metal
sulfide catalyst and the type of amine. Over MoS2, n-pentylamine reacted by nucleophilic
substitution with H2S to pentanethiol as well as with another n-pentylamine molecule to
dipentylamine. Pentenes were observed as secondary products and supposed to be formed by
elimination from dipentylamine. These findings suggest that the molecular structure is one of
the most important factors in HDN and that different molecules may undergo nitrogen
removal by different mechanisms.
Concurrent with the β-hydrogen elimination and nucleophilic substitution reactions, a
disproportionation reaction can occur between two alkylamine molecules, and this
complicates the study of the HDN reaction mechanism. This disproportionation results in the
formation of a dialkylamine and ammonia in case of an alkylamine and in the formation of a
trialkylamine and alkylamine in case of a dialkylamine; it takes place even on alumina at low
hydrogen pressure [20,21]. Substantial amounts of the disproportionation products N-
pentylpiperidine, dicyclohexylamine, and dipentylamine were observed in the HDN of
piperidine [22], cyclohexylamine [23], and pentylamine [15,24] respectively.
We decided to carry out a detailed investigation of the HDN of alkylamines over
sulfided NiMo/Al2O3. The aim of our work was to determine which of the different
mechanisms (elimination, substitution, and disproportionation) plays the major role in the
HDN of aliphatic amines. To elucidate the mechanism it is necessary to determine the
primary HDN products and to compare the influence of different reaction conditions on
product distribution. Accordingly, we carried out experiments at low weight times and under
different reaction conditions to check the formation of primary products and, thus, to prove
the mechanism of the HDN of aliphatic amines. In this work we will present our results of the
HDN of the linear n-alkylamines hexylamine, dihexylamine, and trihexylamine, while in
subsequent work we will publish our results of the HDN of alkylamines with the amine group
attached to secondary and tertiary carbon atoms.
3.3. Results
3.3.1. HDS of pentanethiol and hydrogenation of hexene
HDN of n-Hexylamines Chapter 3 56
To compare the relative rates of the HDS of alkanethiols, the HDN of alkylamines, and
the hydrogenation of alkenes, we performed the simultaneous conversion of octanethiol,
hexylamine, and 1-pentene at 300 °C and 3 MPa in the presence of 10 kPa H2S. It is clear
from the results presented in Figure 3.1 that an alkanethiol reacts very much faster than an
alkylamine and an alkene.
0 2 4 6 8 100
20
40
60
80
100
HA
1-pentene
octanethiolC
onve
rsio
n, %
Weight time, g.min/mol
Fig. 3.1. Conversions of octanethiol and hexylamine, and yield of pentane in the simultaneous
HDS of 5 kPa octanethiol, HDN of 5 kPa hexylamine (HA), and hydrogenation of
5 kPa 1-pentene at 300 °C, 3 MPa, and 10 kPa H2S.
Figure 3.2 shows the HDS conversion of pentanethiol at 300 °C and 3 MPa in the presence of
different pressures of hexylamine and H2S; it was equally fast as that of octanethiol (Fig. 3.1).
0 2 4 6 8 100
20
40
60
80
100
10 kPa HA, 50 kPa H2S
20 kPa HA, 50 kPa H2S
5 kPa HA, 50 kPa H2S
5 kPa HA, 10 kPa H2S
Pen
tane
thio
l con
vers
ion,
%
Weight time, g.min/mol
Fig. 3.2. Conversion of 5 kPa pentanethiol at 300 0C and 3 MPa in the presence of 5, 10 and
20 kPa hexylamine (HA), and 10 and 50 kPa H2S.
The pentanethiol conversion decreased strongly with increasing H2S pressure from 10 to 50
kPa and less strongly with increasing hexylamine pressure from 5 to 20 kPa. The non-zero
selectivities at short weight time (Fig. 3.3) demonstrate that 1-pentene and pentane are
HDN of n-Hexylamines Chapter 3 57
primary products, while the low initial selectivity of 2-pentene indicates that 2-pentene is a
secondary product. The pentene selectivity decreased with increasing hexylamine pressure as
well as with increasing H2S partial pressure, while the opposite dependence was observed for
the complementary pentane. As a consequence, the molar pentenes/pentane ratio (pentenes
stands for the sum of all the pentenes) decreased with increasing partial pressure of
hexylamine and H2S (Fig. 3.4).
0 1 2 3 4 50
10
20
30
40
50
cis-2-C5=
trans-2-C5=
1-C5=
C5
Sele
ctiv
ity, %
Weight time, g.min/mol
Fig. 3.3. Product selectivities to pentane, 1-pentene, trans-2-pentene, and cis-2-pentene in the
HDS of 5 kPa pentanethiol at 300 °C and 3 MPa in the presence of 20 kPa
hexylamine, 5 kPa cyclohexene, and 10 kPa H2S.
0 2 4 6 8 100.0
0.2
0.4
0.6
0.8
1.0
Hex
enes
/hex
ane
Weight time, g.min/mol0 2 4 6 8 10
0.0
0.3
0.6
0.9
1.2
1.5
Pen
tene
s/pe
ntan
e
Weight time, g.min/mol
Fig. 3.4. Pentenes/pentane ratio and hexenes/hexane ratio in the simultaneous HDS of 5 kPa
pentanethiol and HDN of 5 or 20 kPa hexylamine at 300 °C and 3 MPa, and 10 or 50
kPa H2S.
5 kPa HA and 10 kPa H2S 20 kPa HA and 10 kPa H2S
20 kPa HA and 50 kPa H2S
The conversion of 1-hexene to hexane was high in the absence of alkylamine; at 300 °C
in the presence of 10 kPa H2S it was already 40% at τ = 0.8 g.min/mol (Fig. 3.5).
HDN of n-Hexylamines Chapter 3 58
0 3 6 9 120
20
40
60
80
100
20 kPa CHA, 300°C
5 kPa CHA, 300°C
20 kPa CHA, 340°C
300°C
C6 Y
ield
, %
Weight time, g.min/mol
Fig. 3.5 Conversion of 5 kPa 1-hexene at 300 and 340 °C and 3 MPa in the presence or
absence of 5 or 20 kPa cyclohexylamine (CHA), 5 kPa pentanethiol (PT), and 10
kPa H2S.
300 °C 5 kPa PT, 20 kPa CHA, and 340 °C
5 kPa PT, 5 kPa CHA, and 300 °C 5 kPa PT, 20 kPa CHA, and 300 °C
In the presence of 5 kPa cyclohexylamine and 10 kPa H2S, however, it was only 8% at τ = 0.8
g.min/mol. Like at 300 °C (Fig. 1), the alkene conversion at 340 °C in the presence of an
alkylamine (Fig. 3.5) was much lower than the conversion of pentanethiol (98%). The
conversion of 1-hexene to hexane decreased substantially with increasing hexylamine and
H2S partial pressure, while in the presence of alkylamine the influence of H2S was only
moderate. This indicates that the alkylamine is more strongly adsorbed than H2S.
3.3.2. HDN of Hexylamine
The conversion of 5 kPa hexylamine at 300 °C and 5 MPa in the presence of 50 kPa of
H2S was 6% at low weight time (0.9 g.min/mol) and reached 37% at high weight time (8.7
g.min/mol). It almost doubled when the reaction temperature was increased from 300 to 320
°C (Fig. 3.6). Figure 3.7 shows the corresponding product distributions at 300 °C (Fig. 3.7A)
and 320 °C (Fig. 3.7B). The main product of the HDN of hexylamine was hexane, which
behaved as a primary product because the selectivity extrapolates to a non-zero value at τ = 0.
The selectivity of the hexenes (the sum of 1-hexene, 2-hexene, and 3-hexene) decreased with
decreasing contact time at 300 °C, suggesting that hexene may be a secondary product or a
secondary as well as a primary product. At 340 °C, hexene behaved as a primary product. The
HDN of n-Hexylamines Chapter 3 59
selectivity of 1-hexanethiol increased with decreasing weight time, showing that it is a
primary product.
0 2 4 6 8 100
20
40
60
80
20 kPa HA, 300 °C, 3 MPa
5 kPa HA, 300 °C, 3 MPa
5 kPa HA, 300 °C, 5 MPa
5 kPa HA, 320 °C, 5 MPa
Con
vers
ion,
%
Weight time, g.min/mol
Fig. 3.6. Influence of the partial pressure of hexylamine (HA) on its conversion as a function
of weight time at 300 and 320 °C, 3 and 5 MPa, and 50 kPa H2S.
5 kPa HA at 320 °C and 5 MPa 5 kPa HA at 300 °C and 5 MPa
5 kPa HA at 300 °C and 3 MPa 20 kPa HA at 300 °C and 3 MPa
0 2 4 6 8 100
20
40
60
80
C6-SH1-C6
=
C6=
C6
A
Sele
ctiv
ity, %
Weight time, g.min/mol
0 2 4 6 8 100
20
40
60
80
C6-SH1-C6
=
C6=
C6
B
Sele
ctiv
ity, %
Weight time, g.min/mol
Fig. 3.7. Product selectivities in the HDN of 5 kPa hexylamine at 5 MPa, 50 kPa H2S, and 300 0C (A), and 320 0C (B).
hexane, hexenes, 1-hexene, hexanethiol
As 1-hexene is too easily hydrogenated at 5 MPa total pressure, it does not give us much
information about the HDN mechanism. Therefore, the total pressure was decreased from 5 to
3 MPa to obtain less severe hydrogenation conditions. This change resulted in a decrease in
the hexylamine conversion that was about proportional to the change in H2 pressure (Figs. 3.6
and 3.8). Figure 3.9 shows that, at 50 kPa H2S and short contact time (1.4 g.min/mol), the
HDN of n-Hexylamines Chapter 3 60
main product was hexanethiol (45% selectivity), followed by hexane (39% selectivity) and
hexenes (12%). At 3 MPa, dihexylamine was observed as well (4% at short contact time). The
selectivities of 2-hexene and 3-hexene were zero. This shows that initially no isomerization of
1-hexene takes place. At τ = 14.2 g.min/mol, the hexanethiol selectivity decreased to 6% and
the hexane selectivity increased to 62%. Isomerization had become important and the sum of
the selectivities of 1-hexene, 2-hexene, and 3-hexene reached 30%. The hexanethiol
selectivity was much higher at 3 than at 5 MPa (cf. Figs. 3.7A and 3.9). Some dihexylimine
was observed as well. It was also observed in a blank experiment carried out without the
catalyst in the empty Inconel 718 reactor. In the product mixture of the HDN of hexylamine,
the yield of dihexylimine reached only 0.4% at τ = 1.4 g.min/mol, while during the HDN of
dihexylamine and trihexylamine it amounted to 1.2% at τ = 1.0 g.min/mol. These small
amounts of dihexylimine were ignored.
0 2 4 6 8 100
20
40
60
80
50 kPa H2S100 kPa H2S
150 kPa H2S
300 °C, 3 MPa10 kPa H2S
320 °C, 5 MPa
50 kPa H2S
Con
vers
ion,
%
Weight time, g.min/mol
Fig. 3.8. Influence of H2S on the conversion of 5 kPa hexylamine at
320 °C, 5 MPa, and 50 kPa H2S 320 °C, 5 MPa, and 150 kPa H2S
300 °C, 3 MPa, and 10 kPa H2S 300 °C, 3 MPa, and 50 kPa H2S
300 °C, 3 MPa, and 100 kPa H2S
At 3 MPa and 300 °C, the hexane selectivity extrapolated to a non-zero value at τ = 0
and the hexene selectivity to a zero or very low value (Fig. 3.9), indicating that hexane
behaves like a primary and hexene like a secondary product. This ratio decreased with
increasing hexylamine and H2S partial pressure. The hexenes/hexane ratio, obtained from the
HDN of hexylamine, was smaller than the pentenes/pentane ratio, obtained from the HDS of
pentanethiol (Fig. 3.4), at small weight time, but approached the pentenes/pentane ratio at
high weight time.
HDN of n-Hexylamines Chapter 3 61
0 4 8 12 160
10
20
30
40
50
Hex
enes
, %
w e ight tim e, g .m in/m ol0 4 8 1 2 1 6
0
2 0
4 0
6 0
8 0
Hex
ane,
%
W e ig h t tim e , g .m in /m o l
0 4 8 1 2 1 60
2 0
4 0
6 0
Hex
anet
hiol
, %
W e ig h t t im e , g .m in /m o l0 4 8 12 16
0
5
10
15
20
25
Dih
exyl
amin
e, %
W e ig h t tim e , g .m in /m o l
Fig. 3.9. Product selectivities in the HDN of 5 kPa hexylamine at 300 °C, 3 MPa, and
10 kPa H2S ( ), 50 kPa H2S ( ), and 100 kPa H2S ( ).
The conversion of hexylamine decreased by about a factor of three at τ = 5.8 g.min/mol
when increasing the hexylamine partial pressure from 5 to 20 kPa while keeping the other
parameters constant at 3 MPa, 300 °C, and 50 kPa H2S (Fig. 3.6). The selectivities of
hexanethiol, hexane, and hexene did not change much with the change of the hexylamine
partial pressure; only the initial selectivity of dihexylamine increased from 4 to 7%. At 300
°C, the hexylamine conversion decreased slightly when the H2S pressure was increased from
10 to 50 and 100 kPa (Fig. 3.8). At 320 °C and 5 MPa, a stronger decrease was obtained. With
increasing H2S partial pressure, the selectivity of the hexenes decreased, the hexane
selectivity was about constant, the hexanethiol selectivity increased strongly, and the
dihexylamine selectivity decreased strongly (Fig. 3.9). The weight time dependencies of the
products show that dihexylamine and hexanethiol are primary products, that 1-hexene
behaves like a secondary product, and that hexane might be formed as a primary as well as a
secondary product.
To obtain a greater amount of product, and thus increase the accuracy of the selectivity
measurements at short weight time, the reaction temperature was increased to 340 °C. Figure
10 shows the conversion of hexylamine at 340 °C and 3 MPa in the presence of 10 or 50 kPa
H2S and 5 kPa pentanethiol. Because of the fast HDS reaction, this means that the actual H2S
HDN of n-Hexylamines Chapter 3 62
pressure amounted to 15 or 55 kPa. The hexylamine conversion decreased strongly when its
partial pressure was increased from 5 to 20 kPa and the dihexylamine selectivity increased
(Fig. 3.10). Their non-zero initial selectivities show that dihexylamine and hexanethiol are
primary products, while hexane and hexene behave like primary products. With increasing
H2S pressure, the dihexylamine and hexene selectivities decreased, while the hexanethiol
selectivity increased sharply. The selectivity of hexane did not change much with H2S
pressure.
0 2 4 6 8 100
20
40
60
80
100
20 kPa HA, 50 kPa H2S
20 kPa HA, 10 kPa H2S
10 kPa HA, 50 kPa H2S
10 kPa HA, 10 kPa H2S5 kPa HA, 50 kPa H2S
HA
con
vers
ion,
%
Weight time, g.min/mol
0 2 4 6 8 100
20
40
60
80
100
Hex
ane,
%
Weight time, g.min/mol
0 2 4 6 8 100
10
20
30
40
50
Hex
enes
, %
Weight time, g.min/mol
0 2 4 6 8 100
5
10
15
Hex
anet
hiol
, %
Weight time, g.min/mol
0 2 4 6 8 100
5
10
15
20
Dih
exyl
amin
e, %
Weight time, g.min/mol
Fig. 3.10. Conversion and product selectivities in the HDN of 5, 10, and 20 kPa hexylamine
(HA) at 340 0C and 3 MPa, in the presence of 5 kPa pentanethiol, and 10 and 50
kPa H2S.
5 kPa HA and 10 kPa H2S 20 kPa HA and 10 kPa H2S
20 kPa HA and 50 kPa H2S
The alkenes/alkane ratio in the HDN of hexylamine and HDS of pentanethiol increased
when increasing the temperature from 300 to 340 °C (cf. Figs. 3.4 and 3.11). The
HDN of n-Hexylamines Chapter 3 63
hexenes/hexane ratio was equal to the pentenes/pentane ratio resulting from the HDS of
pentanethiol at two different H2S and hexylamine partial pressures (Fig. 3.11). Both ratios
decreased with increasing weight time, due to the hydrogenation of the alkene to the alkane.
The decrease was less steep at higher hexylamine as well as H2S partial pressure because of
the slower hydrogenation of the alkene under those conditions.
0 2 4 6 8 100.0
0.5
1.0
1.5
2.0
2.5
10 kPa HA, 50 kPa H2S
Alk
enes
/alk
ane
Weight time, g.min/mol0 2 4 6 8 10
0.0
0.5
1.0
1.5
2.0
2.5
10 kPa HA, 10 kPa H2SAlk
enes
/alk
ane
Weight time, g.min/mol
0 2 4 6 8 100.0
0.5
1.0
1.5
2.0
2.5
20 kPa HA, 10 kPa H2S
Alk
enes
/alk
ane
Weight time, g.min/mol
0 2 4 6 8 100.0
0.5
1.0
1.5
2.0
2.5
20 kPa HA, 50 kPa H2S
Alk
enes
/alk
ane
Weight time, g.min/mol
Fig. 3.11. Pentenes/pentane ( ) and hexenes/hexane ( ) ratios in the simultaneous HDS of 5
kPa pentanethiol and HDN of 10 and 20 kPa hexylamine (HA) at 340 0C and 3
MPa, and 10 and 50 kPa H2S.
3.3.3. HDN of Dihexylamine
The conversion of dihexylamine (Fig. 3.12) was much faster than that of hexylamine
(Fig. 3.6). At 320 °C, almost complete conversion was reached already at short contact time.
The conversion of dihexylamine hardly changed when the total pressure was increased from 3
to 5 MPa (not shown). Figure 3.13 shows the product selectivities in the HDN of 5 kPa
dihexylamine at 300 °C, 3 MPa, and 50 kPa H2S. It is apparent from the selectivities at low
HDN of n-Hexylamines Chapter 3 64
weight time that hexylamine, hexanethiol, and trihexylamine behave like primary products
and that hexene probably behaves like a secondary product. The hexane selectivity decreased
with decreasing weight time, but it is not clear whether it extrapolates to zero. Therefore we
cannot say whether hexane is a secondary or a primary as well as secondary product. The
selectivities did not change significantly when the total pressure was increased from 3 to 5
MPa. The main difference was observed at low weight time, where the hexanethiol selectivity
was lower and the hexane selectivity higher at 5 MPa. The selectivity patterns at short weight
time were the same at both pressures, as were the conclusions about primary and secondary
products. When the dihexylamine partial pressure was increased from 5 to 20 kPa, while the
other reaction conditions remained constant at 300 °C, 3 MPa, and 50 kPa H2S, the
conversion decreased substantially (Fig. 3.12). At the same time, the THA selectivity
increased strongly, the thiol selectivity remained the same, but the hexylamine selectivity
decreased at shorter weight time (cf. Figs. 3.13 and 3.14A). Trihexylamine, hexanethiol, and
hexylamine are primary products.
0 2 4 6 80
20
40
60
80
100
20 kPa DHA, 300 °C, 3 MPa
300 °C, 3 MPa320 °C, 5 MPa
Con
vers
ion,
%
Weight time, g.min/mol
Fig. 3.12. Conversion of 5 or 20 kPa dihexylamine (DHA) at 300 or 320 °C, 3 or 5 MPa, and
50 kPa H2S.
5 kPa DHA at 320 °C and 5 MPa 5 kPa DHA at 300 °C and 3 MPa
20 kPa DHA at 300 °C and 3 MPa
While the conversion of dihexylamine decreased only slightly when increasing the H2S
pressure from 10 to 50 kPa (not shown), it influenced the product distribution substantially.
With increasing partial pressure of H2S, the hexanethiol selectivity increased strongly, the
trihexylamine selectivity decreased sharply, the hexane and hexenes selectivities decreased,
and the hexylamine selectivity remained the same (Fig. 3.13). The much higher rate of
HDN of n-Hexylamines Chapter 3 65
trihexylamine formation shows that disproportionation is favored by a low partial pressure of
H2S. At the high pressure of 20 kPa dihexylamine and low pressure of 10 kPa H2S, the
trihexylamine selectivity even reached 61% at short contact time, while the hexylamine
selectivity was 26% and the sum of the thiol, hexene, and hexane selectivities was about 15%
(Fig. 3.14B). Under these conditions, dihexylamine reacts predominantly by
disproportionation. Extrapolation to zero weight time gave a hexylamine selectivity of 25%;
thus, hexylamine is a primary product.
0 1 2 3 4 50
6
12
18
24
Hex
enes
, %
Weight time, g.min/mol
0 1 2 3 4 50
3
6
9
12
1-H
exen
e, %
W e ight tim e , g .m in /m o l
0 1 2 3 4 50
5
10
15
20
25
Hex
ane,
%
W eight tim e, g .m in /m ol0 1 2 3 4 5
0
15
30
45
60
Hex
ylam
ine,
%
W eight tim e, g.m in/m ol
0 1 2 3 4 50
4
8
12
16
Trih
exyl
amin
e, %
W e igh t tim e , g .m in /m o l0 1 2 3 4 5
0
10
20
30
40
Hex
anet
hiol
, %
W eight tim e, g .m in/m ol
Fig. 3.13. Product selectivities in the HDN of 5 kPa dihexylamine at 300 °C, 3 MPa,
and 10 kPa H2S ( ), 50 kPa H2S ( ), and 100 kPa H2S ( ).
HDN of n-Hexylamines Chapter 3 66
0 1 2 3 4 50
20
40
60
80 B
C6-SH1-C6
=C6
=
C6
C6-NH2
(C6)3N
Sele
ctiv
ity, %
Weight time, g.min/mol0 2 4 6
0
15
30
45
60 A
(C6)3N
C6=
1-C6=
C6-SH
C6
C6-NH2
Sele
ctiv
ity, %
Weight time, g.min/mol
Fig. 3.14. Product selectivities in the HDN of 20 kPa dihexylamine at 300 °C, 3 MPa, and (A)
50 and (B) 10 kPa H2S.
hexylamine, hexane, hexenes, hexanethiol, 1-hexene, trihexylamine.
3.3.4 HDN of Trihexylamine
The HDN conversion of trihexylamine at 300 °C and 5 MPa was very high; already at
short weight time (0.9 g.min/mol) it reached 82% (not shown). The main products at τ = 0.9
g.min/mol were dihexylamine (43%), hexane (26%), and hexylamine (16%), while the 1-
hexanethiol selectivity was 7% (Fig. 3.15). At 320 °C, the conversion of trihexylamine was
even 92% at τ = 0.9 g.min/mol, and at this temperature dihexylamine (28%), hexane (35%),
and hexylamine (23%) were the main products. The time dependency of the products show
that dihexylamine and 1-hexanethiol are primary products, and that hexane, hexylamine, 2-
hexene, and 3-hexene are secondary products, while 1-hexene behaves like a primary product
(not shown).
0 2 4 6 8 100
15
30
45
60
C6-SH 1-C6=
C6=
C6-NH2
(C6)2NH
C6
Secl
ectiv
ity, %
Weight time, g.min/mol
Fig. 3.15. Product selectivities in the HDN of 5 kPa trihexylamine at 300 °C, 5 MPa, and 50
kPa H2S.
dihexylamine, hexane, hexylamine, hexene, 1-hexene, hexanethiol.
HDN of n-Hexylamines Chapter 3 67
The conversion of 5 kPa trihexylamine at τ = 0.9 g.min/mol and in the presence of 50
kPa H2S decreased to 70% when the total pressure was decreased from 5 to 3 MPa (Fig. 3.16).
Increasing the partial pressure of trihexylamine from 5 to 10 kPa at 3 MPa decreased the
conversion even further to 50% (Fig. 3.16). Lowering the pressure from 5 to 3 MPa increased
the selectivities of the primary products hexanethiol and dihexylamine and decreased those of
the secondary products hexylamine and hexane (cf. Figs. 3.15 and 3.17). Figure 3.17 shows
the product distributions in the HDN of 5 kPa trihexylamine at 300 °C, 3 MPa, and 50 kPa
H2S. At short contact time, the main products were dihexylamine (63%) and 1-hexanethiol
(20%), both being primary products. The initial selectivity of 1-hexene was less than 3%.
Under these conditions, hexane and 1-hexene are clearly secondary products. At high weight
time, the dihexylamine selectivity decreased substantially to less than 5%. The hexanethiol
selectivity also decreased; only the hexylamine and hexane selectivities increased to 32 and
44% respectively. The product selectivities at 5 and 10 kPa trihexylamine were the same at
short contact time but changed less fast with weight time at 10 kPa than at 5 kPa
trihexylamine (Fig. 3.17).
0 1 2 3 4 50
20
40
60
80
100
10 kPa THA, 50 kPa H2S
5 kPa THA, 10 kPa H2S
5 kPa THA, 50 kPa H2S
Con
vers
ion,
%
Weight time, g.min/mol
Fig. 3.16. Conversion in the HDN of 5 or 10 kPa trihexylamine (THA) at 300 °C and 3 MPa,
in the presence of 10 or 50 kPa H2S.
5 kPa THA, 10 kPa H2S 5 kPa THA, 50 kPa H2S
10 kPa THA, 50 kPa H2S
Whereas the conversion of hexylamine decreased slightly and that of dihexylamine
decreased even less when the H2S pressure was increased from 10 to 50 kPa, the conversion
of trihexylamine increased (Fig. 3.16). At short contact time, the selectivities of dihexylamine
HDN of n-Hexylamines Chapter 3 68
and hexylamine were the same at both H2S pressures, but the hexanethiol selectivity was
much higher and the selectivities of hexane and 1-hexene lower at 50 kPa H2S pressure (Fig.
3.17).
0 1 2 3 4 5 60
5
1 0
1 5
2 0
2 5
Hex
enes
, %
W e ig h t tim e , g .m in /m o l 0 1 2 3 4 5 60
3
6
9
12
1-H
exen
e, %
Weight time, g.min/mol
0 1 2 3 4 5 60
10
20
30
40
Hex
ylam
ine,
%
W eight tim e, g.m in/mol0 1 2 3 4 5 6
0
10
20
30
40
50
Hex
ane,
%
W e igh t tim e , g .m in /m o l
0 1 2 3 4 5 60
5
10
15
20
25
Hex
anet
hiol
, %
Weight time, g.min/mol0 1 2 3 4 5 6
0
20
40
60
80
Dih
exyl
amin
e, %
W e igh t tim e , g .m in /m o l
Fig. 3.17. Product selectivities in the HDN of 5 or 10 kPa trihexylamine (THA) at 300 °C, 3
MPa, and 10 and 50 kPa H2S.
5 kPa THA, 10 kPa H2S 5 kPa THA, 50 kPa H2S
10 kPa THA, 50 kPa H2S
HDN of n-Hexylamines Chapter 3 69
3.4 Discussion
To determine which mechanism is responsible for the HDN reaction of an n-alkylamine
one can measure the product selectivities as a function of weight time (τ) and determine
whether these extrapolate to a non-zero or zero value at τ = 0. If a product such as 1-hexene
has a zero selectivity at zero weight time, then it cannot be a primary product and elimination
cannot play a role. If its selectivity is non-zero at τ = 0, then 1-hexene might be a primary
product, and elimination might be important. A zero selectivity at zero weight time does not
automatically mean that a product is a secondary product, however, nor does a non-zero initial
selectivity automatically mean that a product is primary. For instance, hexylamine may react
slowly by substitution to hexanethiol, which then reacts fast to hexane and 1-hexene. The
selectivities of hexane and 1-hexene may then extrapolate to a non-zero initial selectivity and
these molecules may then appear to be primary products, although they are secondary. At the
same time, the fast consecutive reaction of the real primary product hexanethiol will decrease
its initial selectivity and the contribution of the substitution mechanism will be
underestimated. With such potential pitfalls in mind, we will analyze the initial selectivities
observed in the HDN of the hexylamines and try to determine the responsible mechanism(s).
Since elimination is generally considered to be the main HDN mechanism [1,6,13,18], we
will pay particular attention to the initial selectivity of hexene. Because isomerization of 1-
hexene to 2- and 3-hexene is fast, we will use the initial selectivity of the sum of all hexenes,
rather than that of 1-hexene, as a measure of the contribution of elimination.
3.4.1. Hexylamine
Three reactions are possible in the HDN of n-hexylamine: elimination to 1-hexene,
substitution to hexanethiol, and disproportionation to dihexylamine:
6 13 2 6 12 3C H NH C H NH→ + (1)
6 13 2 2 6 13 3C H NH H S C H SH NH+ → + (2)
6 13 2 6 13 2 32 ( )C H NH C H NH NH→ + (3)
HDN of n-Hexylamines Chapter 3 70
At 300 °C, the hexanethiol selectivity increased and the disproportionation selectivity to
dihexylamine decreased with increasing H2S pressure (Fig. 3.9), because the higher H2S
pressure favors the substitution of the NH2 group of hexylamine by H2S over the substitution
by another hexylamine molecule. The hexene selectivity was lower at high H2S pressure and
decreased strongly with decreasing weight time for all H2S pressures. Unfortunately,
measurements below τ = 0.8 g.min/mol were not possible because the gas flow rate could not
be increased further and using less than 50 mg catalyst led to channelling and to a decreased
conversion and thus lower accuracy of the measurement. Therefore, the values to which the
hexene selectivities extrapolate at τ = 0 could not be determined with high precision.
Nevertheless, the steep decrease of the hexene selectivity towards τ = 0 indicates that the
hexene selectivity is in any case smaller than 5% at 50 and 100 kPa H2S and smaller than 20%
at 10 kPa H2S (Fig. 3.9).
6 13 6 12 2C H SH C H H S→ + (4)
Under our conditions, the equilibrium of the decomposition of hexanethiol (reaction 4)
lies to the right [25]. This means that the hexene formed in the HDN of hexylamine will
hardly react with H2S. Also a disappearance of hexene by hydrogenation plays a minor role,
as the hydrogenation of 1-pentene showed. It was strongly inhibited by hexylamine and the
yield of pentane was only 5% at τ = 0.8 g.min/mol. This means that the observed hexene
selectivities in the HDN of hexylamine are truly representative for the discussed HDN
mechanisms.
The values of 5% for the hexene selectivity at 50 and 100 kPa H2S and of 20% at 10 kPa
H2S at short weight time demonstrate that elimination is not the major mechanism in the HDN
of hexylamine. These values are even upper limits to the contribution of elimination to the
HDN of hexylamine (Eq. 1). The reason is that 1-hexene can not only be formed by direct
elimination of hexylamine (Eq. 1), but also by substitution of hexylamine followed by
elimination of the resulting thiol to 1-hexene (Eqs. 2 and 4). The pentenes/pentane ratio from
the HDS of pentanethiol and the hexenes/hexane ratio from the HDN of hexylamine,
measured simultaneously in the same reaction mixture, are about the same (Fig. 3.4). This
similarity of the alkene/alkane branching ratio shows that substitution is the predominant
route for nitrogen removal from hexylamine and that the contribution of elimination is even
HDN of n-Hexylamines Chapter 3 71
less than 5% at 50-100 kPa H2S and 20% at 10 kPa H2S. This also explains why, in the HDN
of hexylamine, the selectivity of hexene is higher at low H2S pressure. Namely, the HDS of
pentanethiol demonstrated that the pentanethiol conversion increased strongly with decreasing
H2S pressure (Fig. 2.2).
At 340 °C, hexene and hexane behaved as primary products (Fig. 2.10), but the
hexenes/hexane ratio was very similar to the pentenes/pentane ratio measured in the
simultaneous HDS of pentanethiol (Fig. 2.11). Both ratios were not only very similar, but also
reacted in the same way on changes in the H2S and hexylamine partial pressures. This shows
that also at 340 °C the main reaction of hexylamine is substitution by H2S to form
hexanethiol. At the higher temperature of 340 °C, the subsequent decomposition of
hexanethiol becomes so fast, that hexene and hexane appear to be primary products.
Hexane is a product in the HDN of all three hexylamines. It is supposed to be formed
from hexanethiol
(5) 6 13 2 6 14 2C H SH H C H H S+ → +
The mechanism of this reaction is unclear. It might be a real hydrogenolysis reaction as
on a metal surface or as reported in the homogeneous reaction of aliphatic and aromatic thiols
with the Cp’2Mo2Co2S3(CO)4 cluster (Cp’ stands for pentamethylcyclopentadienyl) [26].
Another possibility would be that the alkene formed by elimination from the alkanethiol (Eq.
4) is hydrogenated before desorbing from the catalyst surface. Both mechanisms explain the
observed lower hexene/hexane ratio at higher H2 pressure (total pressure), lower temperature,
and higher H2S and hexylamine pressure. The increase of the hexane selectivity with
increasing τ (Figs. 3.9 and 3.10) is due to increased hexanethiol decomposition (Eq. 5) as well
as hydrogenation of hexene. These two factors oppose each other in the production of hexene,
and explain the maximum in the hexenes/hexane ratio at 300 °C (Fig.3.4).
3.4.2 Dihexylamine
Like hexylamine, dihexylamine can react in three ways: by elimination to hexylamine
and 1-hexene, by substitution to hexylamine and 1-hexanethiol, and by disproportionation to
hexylamine and trihexylamine:
HDN of n-Hexylamines Chapter 3 72
6 13 2 6 12 6 13 2( )C H NH C H C H NH→ + (6)
6 13 2 2 6 13 6 13 2( )C H NH H S C H SH C H NH+ → + (7)
6 13 2 6 13 3 6 13 22( ) ( )C H NH C H N C H NH→ + (8)
Furthermore, because of the much higher reactivity of trihexylamine than of hexylamine
and dihexylamine, trihexylamine can react back to dihexylamine and 1-hexene or 1-
hexanethiol (Eqs. 9 and 10).
6 13 3 6 12 6 13 2( ) ( )C H N C H C H NH→ + (9)
6 13 3 2 6 13 6 13 2( ) ( )C H N H S C H SH C H NH+ → + (10)
At 320 °C and 5 MPa we observed only a trace amount of trihexylamine, which means
that, under these conditions, either disproportionation hardly took place or that trihexylamine
reacted away faster than it was formed. The product selectivity suggests that 1-hexene is a
primary product, which would mean that 1-hexene is formed by elimination of dihexylamine.
At 300 °C and 5 MPa (not shown) and 3 MPa, however, 1-hexene behaves more like a
secondary product. The maximum hexene selectivities at τ = 0 are 3% at 50 and 100 kPa H2S
and 8% at 10 kPa H2S (Fig. 3.13). Because of the stoichiometry of Eq. 6 this means that the
maximum relative contribution of elimination to the HDN of dihexylamine are 6 and 16%
respectively. Like in the case of hexylamine, these percentages overestimate the contribution
of elimination substantially, because the hexene/hexane ratio in the HDN of dihexylamine
was not much different from that of the HDS of hexanethiol. This means that reaction 7
followed by reaction 4 is the main pathway for the formation of hexene.
Trihexylamine behaved as a primary product, meaning that dihexylamine quickly forms
trihexylamine by disproportionation. At 320 °C less trihexylamine was observed as at 300 °C,
probably because trihexylamine quickly reacts to hexanethiol and dihexylamine by
substitution. This would also explain why the conversion of dihexylamine and trihexylamine
never reached 100%, neither at 320 °C nor at high weight time. This can be explained by re-
formation of these molecules by a disproportionation reaction of two molecules of
HDN of n-Hexylamines Chapter 3 73
hexylamine to dihexylamine, or two molecules of dihexylamine to trihexylamine, or of one
molecule of dihexylamine and one molecule of hexylamine to trihexylamine.
The trihexylamine selectivity in the HDN of dihexylamine increased with increasing
partial pressure of dihexylamine, while the hexanethiol selectivity stayed almost the same,
and the hexylamine selectivity became much lower (cf. Figs. 3.13 and 3.14). At first glance
this seems strange; if the thiol selectivity stays the same, one would expect that the
hexylamine selectivity is also the same (Eq. 7). We can only explain the increased
trihexylamine selectivity and decreased hexylamine selectivity by assuming that the
hexylamine obtained from the decomposition of dihexylamine reacts with dihexylamine to
form trihexylamine. Although the trihexylamine selectivity increased and trihexylamine has a
much higher reactivity, fewer active centers are available for trihexylamine to decompose at
high partial pressure of dihexylamine. This also explains why the conversion of dihexylamine
decreased at higher partial pressure of dihexylamine. We therefore suggest that at 300 °C, 3
MPa, and 50 kPa H2S dihexylamine reacts by substitution by H2S to hexylamine and 1-
hexanethiol and by disproportionation to trihexylamine and hexylamine (Scheme 3.1). 1-
Hexene and hexane are subsequently formed from 1-hexanethiol. An analysis of the mass
balance supports the conclusion that mainly substitution and disproportionation and hardly
any elimination occurred in the HDN of dihexylamine.
C6 N
C6
C6
C6HN C6 C6 NH2 C6 SH+
+ C6
C6
C6=
NH2
Scheme 3.1. HDN network of dihexylamine.
Cattenot et al. studied the HDN of n-pentylamine over unsupported MoS2 at 275 °C and
atmospheric pressure [15]. They observed dipentylamine and pentanethiol as primary
HDN of n-Hexylamines Chapter 3 74
products and pentenes as secondary products. From the initially larger amount of
dipentylamine, they concluded that the majority of the pentenes formed from dipentylamine
and only a fraction from pentanethiol. Under our conditions (300-320 °C, 3 MPa,
NiMo/Al2O3), however, dihexylamine mainly reacted by substitution with another amine
molecule and with H2S, and hexanethiol was the main source of hexene.
3.4.3 Trihexylamine
Trihexylamine can only react by elimination to dihexylamine and 1-hexene and by
nucleophilic substitution by H2S to dihexylamine and 1-hexanethiol (Eqs. 9 and 10). As
shown in Section 3.3.3, dihexylamine will react further to hexylamine, hexanethiol, hexane,
and 1-hexene. At higher temperature, the scission of the C-N bond of dihexylamine becomes
faster, and thus the selectivity of dihexylamine in the HDN of the trihexylamine is much
lower at 320 than at 300°C and substantially more hexylamine and hexane are formed at 320
°C. At lower total pressure, the C-N cleavage rate of trihexylamine is lower and the selectivity
of dihexylamine increases. High selectivities to dihexylamine and hexanethiol and a low
selectivity to hexenes were obtained at short weight time (Fig. 3.17). The dihexylamine
selectivity extrapolates to 67% at τ = 0 (Fig. 3.17), as is expected for both elimination and
nucleophilic substitution (Eqs. 9 and 10). Note that the selectivity is defined so as to preserve
the carbon mass balance, which is the hexyl balance in this case. The hexene selectivity
extrapolates to 2% at τ = 0 at 50 kPa H2S and to 8% at 10 kPa H2S (Fig. 3.17). With a
maximum selectivity of 33.3% for hexene (Eq. 9), this means that the relative contribution of
elimination to the HDN of trihexylamine is 6 and 24% respectively. Like in the case of n-
hexylamine and dihexylamine, these percentages overestimate the contribution of elimination
substantially, because the hexene/hexane ratio in the HDN of trihexylamine was not much
different from that of the HDS of hexanethiol. This means that reaction 10 followed by
reaction 4 is the main pathway for the formation of hexene. The reason that the hexene
selectivity was higher at the lower H2S pressure of 10 kPa, is due to the higher reactivity of
alkanethiol to alkene under these conditions. Assuming that the reactivity of hexanethiol is
similar to that of pentanethiol, we expect less hexanethiol and more 1-hexene in the HDN of
trihexylamine at lower H2S pressure. This means that elimination of 1-hexene from
trihexylamine to form dihexylamine plays a minor role at all H2S pressures. The qualitative
HDN of n-Hexylamines Chapter 3 75
conclusion about the dominant role of nucleophilic substitution was substantiated by the mass
balance of the amounts of the products from the HDN of trihexylamine (Scheme 3.2).
C6 N
C6
C6
C6 NH
C6
+
+
C6=
C6C6
C6
C6
NH2 SH
SH
Scheme 3.2. HDN network of trihexylamine.
Substitution explains the higher conversion of trihexylamine at higher H2S pressure and
the higher selectivity to hexanethiol and lower selectivity to hexane and hexene. At a higher
H2S pressure there will be more SH groups at the surface and trihexylamine will react faster
with SH to dihexylamine and hexanethiol. At the same time, the hexanethiol decomposition
will be more inhibited because there are fewer vacancies on which hexanethiol can adsorb;
this causes an increase in the hexanethiol selectivity and a decrease in the hexane plus hexene
selectivity. The increased number of SH groups at the catalyst surface not only increases the
formation of dihexylamine from trihexylamine, but also the further (substitution) reaction of
dihexylamine to hexylamine and hexanethiol. As a consequence, the H2S pressure does not
have a significant influence on the dihexylamine selectivity at short weight time.
3.4.4. General Discussion
The conversion of the hexylamines increased with increasing basicity of the amine in
the gas phase in the order hexylamine < dihexylamine < trihexylamine. The conversion of all
three hexylamines increased with increasing total pressure (higher H2 pressure). Since the
primary reactions of the hexylamines (be it elimination, nucleophilic substitution, or
disproportionation) are chemically independent of hydrogen, the positive influence of
hydrogen must be due to a secondary effect. Most probably it is caused by the number of
HDN of n-Hexylamines Chapter 3 76
vacancies on the catalyst surface, which is generated by the H2S/H2 ratio and, at constant H2S
pressure, increases with the H2 pressure. On the other hand, when increasing the H2S pressure
at constant total pressure, the conversion of the hexylamines did not show a strong decrease.
While it decreased slightly for hexylamine and dihexylamine, it even increased with the H2S
pressure for trihexylamine. This may be due to two competing factors. On the one hand, H2S
decreases the number of vacancies on the catalyst surface and, thus, the reaction rate. On the
other hand, more sulfur on the catalyst surface increases the nucleophilic substitution by H2S,
which was demonstrated to be the main HDN reaction. These two factors are almost equal for
hexylamine and dihexylamine, but the second factor is more important in the case of
trihexylamine. This might be due to a strong adsorption of trihexylamine and even the
replacement of H2S from the catalyst surface.
An additional explanation for the increased activity of trihexylamine and slightly
decreased activities of hexylamine and dihexylamine at increasing H2S pressure could be that
only the latter two molecules can undergo disproportionation. At low H2S pressure,
disproportionation determines the conversion of hexylamine and dihexylamine but not that of
trihexylamine. At increasing H2S pressure, nucleophilic substitution increases in importance
for all three molecules, while the contribution of disproportionation to the conversion
decreases. This would explain why the conversion of trihexylamine continuously increased
with increasing H2S pressure, while the contributions of disproportionation and nucleophilic
substitution are influenced in opposite ways by H2S.
Our results clearly demonstrate that nucleophilic substitution is the dominant HDN
mechanism for unbranched alkylamines, dialkylamines, and trialkylamines. For the first two
molecules the substitution can occur by an amine as well as by H2S, while for the
trialkylamine only substitution by H2S is possible. Portefaix et al. showed that more β-H
atoms led to a higher HDN conversion of piperidines [6,13]. We showed, however, that the
higher conversion of 2-methylpiperidine was not due to HDN but to dehydrogenation to 2-
methylpyridine [14]. In fact, 2-methylpiperidine preferentially underwent ring opening to 2-
aminohexane by C-N bond cleavage of the α carbon atom that did not carry the methyl group.
Thus, the extra three β-H atoms on this methyl group did not increase the HDN conversion.
To prove that this has nothing to do with the type of H atoms (on primary or secondary carbon
atoms), we performed an HDN experiment with 2-ethylpiperidine. Also for this molecule,
with two extra β H atoms on a secondary carbon atom, ring opening occurred mainly on the
HDN of n-Hexylamines Chapter 3 77
less sterically hindered side of the piperidine ring. We take this as evidence that, also for
piperidine-like molecules, HDN occurs by nucleophilic substitution rather than by
elimination, as is also the case for dihexylamine, which is a secondary amine like piperidine.
A remaining question is how the nucleophilic substitution of alkylamines takes place at
the catalyst surface. Hydroxyl and amine groups are very poor leaving groups in nucleophilic
substitution [27,28]. Protonation of the amine group or complexing with a Lewis acid group
gives a better leaving group and this might be the main role of the catalyst surface. The
nucleophile that attacks the α-carbon atom can either be an alkylamine, or an SH- or S2- group
at the catalyst surface. Laine suggested that a metal-assisted nucleophilic substitution might
occur via a metal alkyl or alkylidene intermediate [10]. Another mechanism could be a
sequence of dehydrogenation, addition, elimination, and hydrogenation reactions, which is an
established method to replace a hydroxyl or amine group. For instance, the conversion of an
amine into an alcohol takes place via
32 2 2-NH-H H O H
2 2 2 2R-CH -NH R-CH=NH R-CHOH-NH R-CH=O R-CH -OH⎯⎯→ ⎯⎯⎯→ ⎯⎯⎯→ ⎯⎯→
Similarly, an alkylamine can be transformed into an alkanethiol. In that case, one
obtains a reaction sequence that has the same overall stoichiometry as the nucleophilic
substitution reaction of an alkylamine and H2S, giving an alkanethiol and NH3. All four
reactions in the sequence are known to take place on metals. For instance, copper is the metal
of choice in the reaction sequence from alcohol to amine [29]. If such reactions take place on
copper, they may also take place on metal sulfides. The dehydrogenation of amines to nitriles
has been reported by Portefaix et al. in the HDN of pentylamine at low pressure (0.1 MPa)
[15]. In the present work we never observed nitriles, but did observe imines in low
concentration. This may be due to the higher H2 pressure used in our work than in that of
Portefaix et al. Further work has to clarify which mechanism is responsible for the HDN of
alkylamines. For the moment, we can only conclude that, in the HDN of linear alkylamines,
elimination plays a minor role and that a reaction with the stoichiometry of nucleophilic
substitution can explain all observations.
HDN of n-Hexylamines Chapter 3 78
3.5 Conclusions
Our results show that the removal of the nitrogen atom from alkylamines occurs mainly
by a nucleophilic substitution of the alkylamine to an alkanethiol, which subsequently reacts
to an alkene or alkane and H2S. This makes sense from the point of view of organic
chemistry. The aliphatic C-N bond is strong and the amine group is thus a bad leaving group.
Also, an SH- group is too weak a base to remove the hydrogen atom from the β−carbon atom.
As a consequence, Hofmann elimination of an alkylamine to an alkene and ammonia is an
unlikely reaction. Nucleophilic substitution, on the other hand, may very well occur with an
SH- group, because it is a strong nucleophile [15,27]. Even the NHR group can act as a
nucleophile. At the same time, the α-carbon atom in a n-alkylamine is easily accessible for the
nucleophile. Thus, nucleophilic substitution of the NH2 group of an amine by an SH group,
leading to an alkanethiol, as well as by an NHR group (leading to disproportionation)
occurred readily. The much higher reaction rates of dihexylamine and trihexylamine than the
hexylamine might, on the one hand, be caused by their higher number of carbon atoms and,
thus, higher adsorption constants and, on the other hand, by their higher basicities. The
nucleophilic substitution is aided by protonation of the nitrogen atom by a Brønsted acid or by
interaction of the nitrogen atom with a Lewis acid in order to create a better leaving group
[27]. The Ni atom at the catalyst surface may act as the Lewis acid site.
The reaction sequence dehydrogenation-H2S addition-NH3 elimination-hydrogenation
may explain the HDN of hexylamines equally well. This reaction scheme has the same overall
stoichiometry as nucleophilic substitution and therefore reacts in the same way on
temperature, pressure, and other parameters. Also the metal-assisted nucleophilic substitution
proposed by Laine [10] explains our results.
The reason that elimination has long been held to be the dominant nitrogen removal
reaction is due to the fact that relatively large amounts of alkenes are observed in HDN at
longer weight time. Our results show that, at least for n-alkylamines, these large amounts of
alkenes are caused by the elimination reaction of alkanethiols and not by the elimination of
alkylamines. It is the fast reaction of alkanethiols that obscures the true origin of the alkenes.
Only when measuring at short reaction time can one identify the true origin of the alkene.
HDN of n-Hexylamines Chapter 3 79
3.6 References
[1] N. Nelson, R.B. Levy, J. Catal. 58 (1979) 485.
[2] C.N. Satterfield, M. Modell, J.A. Wilkens, Ind. Eng. Chem. Proc. Des. Dev. 19 (1980)
154.
[3] R. Ramachandran, F.E. Massoth, Chem. Eng. Commun. 18 (1982) 239.
[4] R.T. Hanlon, Energy & Fuels 1 (1987) 424.
[5] M.J. Girgis, B.C. Gates, Ind. Eng. Chem. Res. 30 (1991) 2021.
[6] J.L. Portefaix, M. Cattenot, M. Guerriche, J. Thivolle-Cazat, M. Breysse, Catal. Today
10 (1991) 473.
[7] L. Vivier, V. Dominguez, G. Perot, S. Kasztelan, J. Mol. Catal. 67 (1991) 267.
[8] U.S. Ozkan, S. Ni, L. Zhang, E. Moctezuma, Energy & Fuels 8 (1994) 249.
[9] R. Prins, Adv. Catal. 46 (2001) 399.
[10] R.M. Laine, Catal. Rev.-Sci. Eng. 25 (1983) 459.
[11] M. Zdrazil, J. Catal. 141 (1993) 316.
[12] S. Rajagopal, R. Miranda, J. Catal. 141 (1993) 318.
[13] J.L. Portefaix, M. Cattenot, M. Guerriche, M. Breysse, Catal. Lett. 9 (1991) 127.
[14] M. Egorova, Y. Zhao, P. Kulula, R. Prins, J. Catal. 206 (2002) 263; and chapter 2.
[15] M. Cattenot, J.L. Portefaix, J. Afonso, M. Breysse, M. Lacroix, G. Perot, J. Catal. 173
(1998) 366.
[16] M. Breysse, J. Afonso, M. Lacroix, J.L. Portefaix, M. Vrinat, Bull. Soc. Chim. Belg.
100 (1991) 923.
[17] J.H. Lee, C.E. Hamrin, B.H. Davis, Appl. Catal. A 111 (1994) 11.
[18] P. Clark, X. Wang, P. Deck, S. T. Oyama, J. Catal. 210 (2002) 116.
[19] F. Rota, R. Prins, J. Catal. 202 (2001) 195.
[20] M. Fikry Ebeid, J. Pasek, Coll. Czech. Chem. Commun. 35 (1970) 2166.
[21] P. Hogen, J. Pasek, Coll. Czech. Chem. Commun. 39 (1974) 3696.
[22] R. Pille, G. Froment, Stud. Surf. Sci. Catal. 106 (1997) 403.
[23] S. Eijsbouts, C. Sudhakar, V.H.J. de Beer, R. Prins, J. Catal. 127 (1991) 605.
[24] H. Schulz, M. Schon, N.M. Rahman, Stud. Surf. Sci. Catal. 27 (1986) 201.
[25] J.G. Speight, in “The Desulfurization of Heavy Oils and Residua”. Marcel Dekker
Journals, 2000.
HDN of n-Hexylamines Chapter 3 80
[26] M.D. Curtis, S.H. Druker, J. Am. Chem. Soc. 119 (1997) 1027.
[27] R. Bruckner, “Advanced Organic Chemistry”. Academic Press, 2002.
[28] J. Clayden, N. Greeves, S. Warren, P. Wothers, “Organic Chemistry”. Oxford Uni.
Press, 2001.
[29] T. Mallat, A. Baiker, in “ Handbook of Heterogeneous Catalysis” (G. Ertl, H.
Knözinger, J. Weitkamp, Eds.). Vol 5, P. 2334, Wiley-VCH.
HDN of Alkylamines Chapter 4 81
4. Mechanisms of the hydrodenitrogenation of alkylamines with secondary
and tertiary α-carbon atoms on sulfided NiMo/Al2O3
4.1 Abstract
The hydrodenitrogenation (HDN) of alkylamines with secondary and tertiary α-carbon
atoms (2-pentylamine, 3-methyl-2-butylamine, 3,3-dimethyl-2-butylamine, 2-
methylcyclohexylamine, 2-methyl-2-butylamine) and benzylamine as well as the
hydrodesulfurization (HDS) of corresponding alkanethiols were studied over sulfided
NiMo/Al2O3. Alkanethiols and dialkylamines were primary products in the HDN of the
amines with secondary α-carbon atoms, formed by substitution of the amine group by H2S or
an alkylamine. Alkanes and alkenes were secondary products, formed from elimination and
hydrogenolysis of the alkanethiols, as confirmed by the similar alkenes/alkane ratios in the
HDN of the alkylamines and HDS of the corresponding alkanethiols. 2-Methyl-2-butylamine
and benzylamine reacted much faster than the amines with secondary α-carbon atoms.
Methylbutenes and methylbutane were the primary products of 2-methyl-2-butylamine, and
toluene was the primary product of benzylamine. This and the different
methylbutenes/methylbutane ratios in the HDN of 2-methyl-2-butylamine and HDS of 2-
methyl-2-butanethiol indicate that 2-methyl-2-butylamine, with a tertiary α-carbon atom, and
the activated benzylamine react by means of an E1 mechanism.
4.2 Introduction
Nitrogen atoms are removed from hetero-aromatic compounds by hydrogenation of the
aromatic ring, which contains the nitrogen atom, and breaking of the resulting aliphatic C-N
bonds to form a hydrocarbon molecule and ammonia [1-5]. Hydrogenation is not required for
the removal of a sulfur atom from an aromatic ring that contains a sulfur atom, as in
(di)benzothiophene, because the relatively weak C-S bond can be broken by hydrogenolysis
HDN of Alkylamines Chapter 4 82
[6,7]. The presence of a large amount of alkenes and a minor amount of alkanes in the
reaction product of HDN suggests that aliphatic C-N bond breaking occurs mainly by
elimination of NH3 [1,3]. Nucleophilic substitution of the alkylamine by H2S, followed by C-
S bond hydrogenolysis, explains the presence of alkanes [1,4]. Cattenot et al. showed,
however, that in the HDN of the linear n-pentylamine over unsupported MoS2 at 275 °C and
atmospheric pressure the formation of pentenes was negligible at short weight time and that
the major product was dipentylamine [8]. At higher weight times pentenes and pentanethiol
were observed and the production of the pentenes was ascribed mainly to the elimination of
pentylamine from the dipentylamine and, in part, to the elimination of H2S from pentanethiol.
We showed that, over a sulfided NiMo/Al2O3 catalyst at 300 to 340 °C and elevated pressure
(3 MPa), n-hexylamine, di-n-hexylamine, and tri-n-hexylamine react predominantly by
nucleophilic substitution of the amines by H2S and not by elimination of NH3 [9]. The
resulting hexanethiol reacts very fast to hexenes as well as to hexane. The very low selectivity
of the hexenes at low weight time demonstrates that the hexenes are secondary not primary
products in the HDN of n-hexylamines. As a consequence, the hexene/hexane ratio in the
HDN product mixture is determined by the HDS reaction, as demonstrated by the similar
alkenes/alkane ratios in the simultaneous HDN of hexylamine and HDS of pentanethiol [9].
The nucleophilic substitution by a good nucleophile such as SH- is aided by the
accessibility of the α-carbon atom in linear alkylamines. The accessibility of the α-carbon
atom decreases with substitution and, at the same time, the number of β-hydrogen atoms
increases. Thus, amines with secondary and tertiary α-carbon atoms may react differently as
linear alkylamines [8]. Therefore, we report the results of the organic part of our mechanistic
investigation of the removal of ammonia from such alkylamines in this work. In future work
we will report on the inorganic aspects, the catalytic sites and the mode(s) of adsorption.
4.3 Results
4.3.1. 2-Pentylamine and 2-pentanethiol
The conversion of 2-pentylamine at 300 °C in the presence of 10 kPa H2S was 12% at τ =
0.9 g.min/mol and 53% at 8.9 g.min/mol (Fig. 4.1). The conversion increased slightly when
HDN of Alkylamines Chapter 4 83
the H2S pressure was increased from 10 to 100 kPa but increased considerably with an
increase in temperature from 300 to 340 °C. The products were 2-pentanethiol, di-(2-
pentyl)amine, 1-pentene, 2-pentene, and pentane (Fig. 4.2). The selectivity of 2-pentanethiol
increased with decreasing weight time, showing that 2-pentanethiol is a primary product (of
the nucleophilic substitution of 2-pentylamine with H2S). Di-(2-pentyl)amine, the
disproportionation product of two molecules of 2-pentylamine, behaved as a primary product
as well. The selectivity of 2-pentanethiol increased and that of di-(2-pentyl)amine decreased
with increasing H2S pressure.
0 2 4 6 8 100
20
40
60
80
100
3M2B
A C
over
sion
, %
Weight time, g.min/mol0 2 4 6 8 10
0
20
40
60
80
100
2-PA
Con
vers
ion,
%
Weight time, g.min/mol
Fig. 4.1 Conversions of 2-pentylamine (2-PA) and 3-methyl-2-butylamine (3M2BA) at 300
°C ( and ) and 340 °C ( and ), and 10 kPa (drawn line) and 100 kPa H2S
(dashed line).
The selectivity of the pentenes (i.e. the sum of the two pentene isomers) as a function of
weight time was different at 300 than at 340 °C. At 300 °C the selectivity extrapolates to zero
with decreasing weight time, which indicates that pentenes are a secondary product. On the
other hand, at 340 °C the pentenes behaved like a primary product because their selectivity
extrapolates to a non-zero value at τ = 0. At both temperatures, the selectivity of pentane
decreased with decreasing weight time to a value close to zero.
The conversion of 2-pentanethiol in the presence of 10 kPa H2S and 5 kPa hexylamine
was much larger than that of the corresponding 2-pentylamine. Thus, at τ = 0.9 g.min/mol the
conversion of the thiol was 83% at 300 °C and 95% at 340 °C (Fig. 4.3), while that of the
amine was 12% at 300 °C and 42% at 340 °C (Fig. 4.1). The conversion of 2-pentanethiol
decreased strongly when the H2S pressure was increased from 10 to 100 kPa. The main
products at 300 °C and 10 kPa H2S were 1-pentene (21%), 2-pentene (52%), and pentane
HDN of Alkylamines Chapter 4 84
(27%) at τ = 0.9 g.min/mol (not shown). All three molecules were primary products. The
selectivity of 1-pentene decreased and that of pentane increased with weight time, while the
selectivities of cis- and trans-2-pentene were constant. This difference between the pentenes
must be due to the easier hydrogenation of the terminal 1-pentene. The pentenes/pentane ratio,
obtained from the HDS of 2-pentanethiol, was similar to that obtained in the HDN of 2-
pentylamine (Fig. 4.4).
0 2 4 6 8 100
20
40
60
80
100
Pent
ane,
%
Weight time, g.min/mol
0 2 4 6 8 100
20
40
60
80
Pent
enes
, %
Weight time, g.min/mol
0 2 4 6 8 100
5
10
15
20
25
Pent
anet
hiol
, %
Weight time, g.min/mol
10
0 2 4 6 8 100
20
40
60
80
0
Di-2
-PA
, %
Weight time, g.min/mol
Fig. 4.2 Product selectivities in the HDN of 2-pentylamine (2-PA) at 300 °C ( and ) and
340 °C ( and ), and 10 kPa (drawn line) and 100 kPa H2S (dashed line).
0 2 4 6 8 100
20
40
60
80
100
2-PT
Con
vers
ion,
%
Weight time, g.min/mol
0 2 4 6 8 100
20
40
60
80
100
3M2B
T co
nver
sion
, %
Weight time, g.min/mol
Fig. 4.3 Conversions in the HDS of 2-pentanethiol (2-PT) and 3-methyl-2-butanethiol
(3M2BT) in the presence of hexylamine at 300 °C ( and ) and 340 °C ( and ),
and 10 kPa (drawn line) and 100 kPa H2S (dashed line).
HDN of Alkylamines Chapter 4 85
0 2 4 6 8 100
1
2
3
42-PA
Pent
enes
/pen
tane
Weight time, g.min/mol
0 2 4 6 8 100
1
2
3
42-PT
Pent
enes
/pen
tane
Weight time, g.min/mol
Fig. 4.4 Pentenes/pentane ratio in the HDN of 2-pentylamine (2-PA) and HDS of 2-
pentanethiol (2-PT) in the presence of hexylamine at 300 °C ( and ) and 340
°C ( ), and 10 kPa (drawn line) and 100 kPa H2S (dashed line).
4.3.2. 3-Methyl-2-butylamine and 3-methyl-2-butanethiol
The conversion of 3-methyl-2-butylamine was slightly lower than that of 2-pentylamine
(Fig. 4.1). It increased weakly with increasing H2S pressure and increased substantially with
increasing temperature. The main products were 3-methyl-2-butanethiol, di-(3-methyl-2-
butyl)amine, 2-methyl-2-butene, 2-methyl-1-butene, 3-methyl-1-butene and 2-methylbutane.
The selectivities of 3-methyl-2-butanethiol and di-(3-methyl-2-butyl)amine increased with
decreasing weight time (Fig. 4.5). This shows that they are primary products of the
nucleophilic substitution of 3-methyl-2-butylamine with H2S and with another molecule of 3-
methyl-2-butylamine, respectively. As expected for a molecule with two chiral atoms (3-
methyl-2-butylamine has one chiral atom), the gas chromatogram of di-(3-methyl-2-
butyl)amine showed two peaks of equal intensity and only a small difference in retention
time; the corresponding products had equal mass spectra. One GC peak is due to the (R,R)
and (S,S) isomers, the other to the meso (R,S) isomer. The selectivity of di-(3-methyl-2-
butyl)amine decreased with increasing H2S pressure, while the reverse was true for the
selectivity of 3-methyl-2-butanethiol.
HDN of Alkylamines Chapter 4 86
0 2 4 6 8 100
20
40
60
80
100
2-M
ethy
lbut
ane,
%
Weight time, g.min/mol0 2 4 6 8 10
0
10
20
30
40
502-
Met
hyl-2
-but
ene,
%
Weight time, g.min/mol
0 2 4 6 8 100
2
4
6
8
10
3-M
ethy
l-1-b
uten
e, %
Weight time, g.min/mol
0 2 4 6 8 100
5
10
15
20
2-M
ethy
l-1-b
uten
e, %
Weight time, g.min/mol
0 2 4 6 8 100
10
20
30
40
3-M
ethy
l-2-b
utan
ethi
ol, %
Weight time, g.min/mol0 2 4 6 8 10
0
10
20
30
40
50
Di-a
min
e, %
Weight time, g.min/mol
Fig. 4.5 Product selectivities in the HDN of 3-methyl-2-butylamine at 300 °C ( and ) and
340 °C ( and ), and 10 kPa (drawn line) and 100 kPa H2S (dashed line).
As in the HDN of its isomer 2-pentylamine (Section 4.3.1), the selectivities of the alkene
products in the HDN of 3-methyl-2-butylamine differed at 300 and 340 °C. At 300 °C the
methylbutene selectivities were low at τ = 0 and increased with weight time, while at 340 °C
and τ = 0 they were higher and decreased with weight time. At both temperatures, all three
methylbutenes behaved as primary products (Fig. 4.5) as the selectivities extrapolated to non-
zero with decreasing weight time to zero.
HDN of Alkylamines Chapter 4 87
The conversion of 3-methyl-2-butanethiol in the presence of 10 kPa H2S and 5 kPa
hexylamine was much larger than that of the corresponding 3-methyl-2-butylamine: The
conversion of the thiol was 74% at τ = 0.9 g.min/mol at 300 °C and 100% at 340 °C (Fig.
4.3), while that of the corresponding amine was 7% at 300 °C and 29% at 340 °C (Fig. 4.1).
The main products at 300 °C and 10 kPa H2S were 3-methyl-1-butene (17%), 2-methyl-2-
butene (49%), and 2-methylbutane (29%) at τ = 0.9 g.min/mol, while the selectivity of 2-
methyl-1-butene was 5% (not shown). The methylbutenes/methylbutane ratio obtained from
the HDS of 3-methyl-2-butanethiol in the presence of hexylamine was slightly lower than that
obtained from the HDN of 3-methyl-2-butylamine at low weight time but was similar at high
weight time (Fig. 4.6).
0 2 4 6 8 100
1
2
3
4 3M2BA
C5= /C
5
Weight time, g.min/mol
0 2 4 6 8 100
1
2
3
4
C5= /C
5
3M2BT
Weight time, g.min/mol
Fig. 4.6 Methylbutenes/methylbutane ratio in the HDN of 3-methyl-2-butylamine (3M2BA)
and HDS of 3-methyl-2-butanethiol (3M2BT) in the presence of hexylamine at 300
°C ( and ) and 340 °C ( and ), and 10 kPa (drawn line) and 100 kPa H2S
(dashed line).
To determine whether a methyl-shift rearrangement occurred during the HDN of 3-
methyl-2-butylamine, the HDN of 2-methyl-3-pentylamine was measured. The selectivity of
the rearrangement products 3-methyl-2-pentene and 3-methylpentane was below 1%. The
main products (di-(2-methyl-3-pentyl)amine, 2-methyl-1-pentene, 2-methyl-2-pentene, 2-
methyl-3-pentene, 2-methylpentane, and 2-methyl-3-pentanethiol) corresponded to those in
the HDN of 3-methyl-2-butylamine and were present in similar amounts (Fig. 4.5). In
addition, a constant selectivity of 2% was observed for 2-methyl-4-pentene.
HDN of Alkylamines Chapter 4 88
4.3.3. 3,3-Dimethyl-2-butylamine
The conversion of 3,3-dimethyl-2-butylamine was even lower than that of 3-methyl-2-
butylamine: only 2% at 300 °C and 10 kPa H2S at τ = 0.9 g.min/mol but 24% at τ = 8.9
g.min/mol. The conversion increased strongly with increasing temperature and increased
slightly with increasing H2S pressure (Fig. 4.7). 3,3-Dimethyl-2-butanethiol formed and
behaved as a primary product, because its selectivity increased with decreasing weight time
(Fig. 4.8). We did not observe a disproportionation product, probably because of the
combined steric hindrance of the tertiary butyl and methyl groups attached to the same carbon
atom as the NH2 group.
In addition to the normal products (3,3-dimethyl-1-butene and 2,2-dimethylbutane),
rearranged products (2,3-dimethylbutane, 2,3-dimethyl-2-butene, and 2,3-dimethyl-1-butene)
also formed. The sum of the selectivities of these rearranged products was 43% at 300 °C and
63% at 340 °C and 10 kPa H2S at τ = 0.9 g.min/mol. As in the case of 2-pentylamine and 3-
methyl-2-butylamine, the selectivity of 3,3-dimethyl-1-butene in the HDN of 3,3-dimethyl-2-
butylamine as a function of weight time was different at 300 °C than at 340 °C. At 300 °C, the
3,3-dimethyl-1-butene selectivity went through a maximum, while at 340 °C it decreased
continuously with increasing weight time.
0 2 4 6 8 100
20
40
60
80
100
3,3D
M2B
A C
onve
rsio
n, %
Weight time, g.min/mol
0 2 4 6 8 100
20
40
60
80
100
2-M
CH
A C
onve
rsio
n, %
Weight time, g.min/mol
Fig. 4.7 Conversions of 3,3-dimethyl-2-butylamine (3,3DMBA) and 2-
methylcyclohexylamine (2-MCHA) at 300 °C ( and ) and 340 °C
( and ), and 10 kPa (drawn line) and 100 kPa H2S (dashed line).
HDN of Alkylamines Chapter 4 89
Fig. 4.8 Product selectivities in the HDN of 3,3-dimethyl-2-butylamine at 300 °C ( and )
and 340 °C ( and ), and 10 kPa (drawn line) and 100 kPa H2S (dashed line).
0 2 4 6 8 100
10
20
30
40
50
2,2-
Dim
ethy
lbut
ane,
%
Weight time, g.min/mol0 2 4 6 8 10
0
10
20
30
402,
3-D
imet
ht
e,%
enyl
-2-b
u
Weight time, g.min/mol
0 2 4 6 8 100
10
20
30
2,3-
Dim
ethy
lbut
ane,
%
Weight time, g.min/mol100 2 4 6 8
0
5
10
15
20
3,3-
dim
ethy
l--b
uten
e,%
Weight time, g.min/mol
1
0 2 4 6 8 100
10
20
30
2,3-
Dim
ethy
l-1-lb
uten
e,%
Weight time, g.min/mol
0 2 4 6 8 100
10
20
30
40
SH
3,3-
Dim
ethy
l-2-b
utan
ethi
ol, %
Weight time, g.min/mol
HDN of Alkylamines Chapter 4 90
4.3.4. 2-Methylcyclohexylamine
2-Methylcyclohexylamine has the same molecular structure as 3-methyl-2-butylamine,
with the amine group attached to a secondary α-carbon atom and a methyl group on the
neighboring β-carbon atom. The conversion of 2-methylcyclohexylamine was lower than that
of 3-methyl-2-butylamine, both at 300 °C and 340 °C, and H2S had a positive influence on the
conversion of both amines (cf. Figs. 4.1 and 4.7). The selectivities of the methylcyclohexenes,
methylcyclohexane, 2-methylcyclohexanethiol, and di-(2-methylcyclohexyl)amine products
(Fig. 4.9) were similar to those of the respective alkenes, alkane, alkanethiol, and
dialkylamine products in the HDN of 3-methyl-2-butylamine respectively (Fig. 4.5).
0 2 4 6 8 100
20
40
60
80
MC
H, %
Weight time, g.min/mol0 2 4 6 8 10
0
20
40
60
80
MC
HE,
%
Weight time, g.min/mol
0 2 4 6 8 100
10
20
30
40
50
MC
HTH
IOL,
%
Weight time, g.min/mol
0 2 4 6 8 100
10
20
30
40
50
Di-2
-am
ine,
%
Weight time, g.min/mol
Fig. 4.9 Product selectivities of methylcyclohexenes (MCHE), methylcyclohexane (MCH) 2-
methylcyclohexanethiol (MCHTHIOL) and di(2-methylcyclohexyl)amine (Di-2-
amine) in the HDN of 2-methylcyclohexylamine at 300 °C ( and ) and 340 °C
( and ), and 10 kPa (drawn line) and 100 kPa H2S (dashed line).
HDN of Alkylamines Chapter 4 91
4.3.5. 2-Methyl-2-butylamine and 2-methyl-2-butanethiol
The HDN of 2-methyl-2-butylamine occurred fast; the conversion was already 7% at 270
°C and 10 kPa H2S at low weight time (0.9 g.min/mol) and reached 90% at τ = 8.9 g.min/mol
(Fig. 4.10). The main products were 2-methyl-2-butene, 2-methyl-1-butene, and 2-
methylbutane (Fig. 4.11). The non-zero selectivities of these products at τ = 0 showed that
they behaved as primary products. The selectivity of 2-methyl-1-butene decreased and that of
2-methylbutane increased with increasing weight time due to isomerization and
hydrogenation. A separate experiment of 2-methyl-1-butene in the presence of hexylamine at
270 °C, 3 MPa and 10 kPa H2S showed 30% conversion to 2-methyl-2-butene and 10%
conversion to 2-methylbutane at τ = 10 g.min/mol. The selectivity of 3-methyl-1-butene, the
isomerization product of 2-methyl-2-butene, was less than 1% over the whole range of weight
times. A small amount of 2-methyl-2-butanethiol was observed that increased with increasing
H2S pressure. Its selectivity was 1.5% at τ = 0.9 g.min/mol and 100 kPa H2S. Increasing the
H2S pressure from 10 to 100 kPa hardly influenced the conversion and product selectivities of
2-methyl-2-butylamine.
0 2 4 6 8 100
20
40
60
80
100
BA C
onve
rsio
n, %
Weight time, g.min/mol0 2 4 6 8 10
0
20
40
60
80
100
2M2B
A C
onve
rsio
n, %
Weight time, g.min/mol
Fig. 4.10 Conversions in the HDN of 2-methyl-2-butylamine (2M2BA) and benzylamine
(BA) at 270 °C ( and ), and 10 kPa (drawn line) and 100 kPa H2S (dashed line).
HDN of Alkylamines Chapter 4 92
0 2 4 6 8 100
20
40
60
80
2-M
ethy
l-2-b
uten
e, %
Weight time, g.min/mol
0 2 4 6 8 100
10
20
30
40
50
2-M
ethy
l-1-b
uten
e, %
Weight time, g.min/mol
0 2 4 6 8 100
10
20
30
2-M
ethy
lbut
ane,
%
Weight time, g.min/mol
0 2 4 6 8 100
1
2
3
4
5
SH
2M2B
T, %
Weight time, g.min/mol
Fig. 4.11 Product selectivities in the HDN of 2-methyl-2-butylamine at 270 °C ( and ), and
10 kPa (drawn line) and 100 kPa H2S (dashed line).
The conversion of 2-methyl-2-butanethiol at 270 °C in the presence of 10 kPa H2S and 5
kPa hexylamine was much larger than that of the equivalent amine: It was already 66% at τ =
0.9 g.min/mol (Fig. 4.12), while the conversion of the corresponding amine was only 20%
(Fig. 4.10). The conversion of 2-methyl-2-butanethiol decreased with increasing H2S
pressure. The main products at τ = 0.9 g.min/mol were 2-methyl-1-butene (35%), 2-methyl-2-
butene (39%), and methylbutane (25%) (not shown). The methylbutenes/methylbutane ratio,
obtained in the HDS of 2-methyl-2-butanethiol, was 3 at 270 °C and 10 kPa at τ = 0.9
g.min/mol, which is very different from the value of 16.5 obtained in the HDN of 2-methyl-2-
butylamine under the same conditions (Fig. 4.13).
HDN of Alkylamines Chapter 4 93
0 2 4 6 8 100
20
40
60
80
100
2M2B
T C
onve
rsio
n, %
Weight time, g.min/mol
Fig. 4.12 Conversion in the HDS of 2-methyl-2-butanenethiol (2M2BT) in the presence of
hexylamine at 270 °C ( and ), and 10 kPa (drawn line) and 100 kPa H2S
(dashed line).
0 2 4 6 8 100
6
12
18
2M2BT
2M2BA
C5= /C
5
Weight time, g.min/mol
Fig. 4.13 Methylbutenes/methylbutane ratio in the HDN of 2-methyl-2-butylamine (2M2BA)
and HDS of 2-methyl-2-butanethiol (2M2BT) in the presence of hexylamine at 270
°C ( and ), and 10 kPa (drawn line) and 100 kPa H2S (dashed line).
The conversion of benzylamine was lower than that of 2-methyl-2-butylamine at 270 °C
(32% for benzylamine and 52% for 2-methyl-2-butylamine at τ = 3.4 g.min/mol) and was not
influenced by the H2S pressure (Fig. 4.10). The main product at 10 kPa H2S was toluene (>
HDN of Alkylamines Chapter 4 94
99%); only traces of α-toluenethiol (benzyl mercaptan), methylcyclohexene, and
methylcyclohexane formed. At 100 kPa H2S, the α-toluenethiol selectivity was 3% at low
weight time and even smaller at high weight time. The conversion of α-toluenethiol was
100% under all conditions at 270 °C, even at the lowest weight time, and toluene was the only
product.
4.4 Discussion
4.4.1 HDN and HDS mechanisms
As shown above, the products of the HDN of alkylamines are dialkylamines, alkanethiols,
alkenes, and alkanes, while alkenes and alkanes are the products of the HDS of alkanethiols.
Dialkylamines and alkanethiols are formed by substitution of the NH2 group in alkylamines
by an alkylamine or H2S respectively. The alkenes can be formed by elimination of NH3 from
the alkylamines or of H2S from the alkanethiols. The alkanes are formed by hydrogenolysis,
which is the rupture of the C-N or C-S bond and simultaneous hydrogenation. Dialkylamines,
alkanethiols, and alkenes can, in principle, form by acid-base catalysis as well as metal-like
catalysis, while alkanes can only form by metal-catalyzed hydrogenolysis.
4.4.1.1 Acid-base mechanisms
In acid-base catalyzed elimination [1,3] and nucleophilic substitution [1,4] reactions of
alkylamines, the amine group reacts first with a proton or a Lewis acid in order to create a
better leaving group (In the subsequent schemes, only the reaction with a proton is indicated).
Then a concerted bimolecular reaction takes place in the E2 elimination as well as SN2
nucleophilic substitution (Scheme 4.1), in which a base or nucleophile reacts with the
protonated amine and ammonia is split off. In the E2 mechanism, the base subtracts a
hydrogen atom from the β-carbon atom of the alkylamine. That is why this elimination can
HDN of Alkylamines Chapter 4 95
only occur when β-H atoms are present and why it was proposed that the HDN of alkylamines
occurs faster when a greater number of β-H atoms is present [3]. In the SN2 reaction, the base
attacks the α-carbon atom of the alkylamine. In classic nucleophilic substitution this attack
takes place from the backside of the molecule, at the side opposite to the leaving amine group,
with inversion of the configuration at the α-carbon atom.
B + R
H2C
CH2
NH3+ E2
BH+ + R
HC
CH2+ NH3
HS- + R1 C NH3+
R2
R3
SN2SH C R1
R2
R3
+ NH3
Scheme 4.1 E2 and SN2 reactions of an alkylamine to an alkene and alkanethiol.
In the E1 elimination and SN1 nucleophilic substitution mechanisms, ammonia splits off
from the protonated amine in the rate-limiting step and forms a carbenium ion (Scheme 4.2).
The carbenium ion quickly undergoes proton abstraction to form an alkene (E1), or it reacts
with a nucleophile, such as H2S or another alkylamine, to form an alkanethiol or
dialkylamine, respectively (SN1). E1 and SN1 mechanisms are likely to occur only if relatively
stable carbenium ions (such as a benzyl, allyl, or tertiary trialkyl carbenium ion) can be
formed. Linear alkylamines will not react by E1 and SN1 mechanisms, because they would
lead to an unstable primary carbenium ion. Secondary alkyl carbenium ions are more stable
than primary ions and may form from alkylamines with a secondary α-carbon atom. They will
only form, however, if sufficiently strong acid sites are available and this is not the case on
the surface of metal sulfides. On the other hand, HDN reactions are carried out at least 300
°C. Such temperatures, which are much higher than normally used for organic reactions,
might increase the participation of carbenium ions [11].
The E1 and E2 reactions of an alkylamine lead to an alkene and the SN1 and SN2 reactions
to an alkanethiol or dialkylamine. The dialkylamine can undergo further nucleophilic
substitution to give an alkanethiol. As this and other studies [9,12-15] have shown,
alkanethiols react fast to alkenes and alkanes. Acid-base catalyzed elimination explains the
HDN of Alkylamines Chapter 4 96
formation of an alkene from a thiol by an E1 or E2 reaction. In both cases, the reaction rate
and, thus, the conversion might be positively influenced by the H2S pressure. In the E1
reaction protonation of the SH group takes place before the rate-determining breaking of the
C-S bond. As a consequence, the E1 reaction may be aided by H2S. In the E2 reaction, a
higher H2S pressure increases the concentration of the S2- or SH- base at the catalyst surface.
However, in the HDS reactions that we studied, H2S had a strong negative influence on the
conversion of all n-alkanethiols, be it primary, secondary, or tertiary alkanethiols (Table 4.1).
This must be due to the fact that H2S adsorbs rather strongly on the catalyst surface and
inhibits the adsorption of the alkanethiol.
H2C C NH3+
R2
R3
H2C C+
R2
R3
+ NH3R1 R1
H
R2R1
R3
H2C C SH
R2
R3
R1
H2SSN1E1
+H++ H+
Scheme 4.2 E1 and SN1 reactions of an alkylamine to an alkene and alkanethiol.
4.4.1.2 Metal-like mechanisms
Substitution and elimination of alkylamines can be catalyzed not only by acids and bases,
but also by metals. Nucleophilic substitution can take place by a series of metal-catalyzed
reactions: dehydrogenation of an amine to an imine, addition of H2S, elimination of NH3, and
hydrogenation of the resulting thioaldehyde to a thiol (Eq. 1) [16]. Analogous reaction
schemes for homogeneous catalysts have been proposed by Laine [17]:
R1R2C=NHR1R2CH-NH2 R1R2C(SH)-NH2
R1R2C(SH)-NH2 R1R2C=S R1R2CH-SH (1)
HDN of Alkylamines Chapter 4 97
Furthermore, the elimination of NH3 from an amine to form an alkene can be metal-
catalyzed [2,5]. In a first step, C-N bond hydrogenolysis would take place. This is an easy
reaction on metal catalysts [18]; on supported platinum, for example, it is already fast at
around 150 °C [19]. The resulting alkyl fragment can lose a β−hydrogen atom and form an
alkene, or it can add a hydrogen atom to form an alkane. Both reactions are well-known in
Fischer-Tropsch catalysis by metals. The fast and reversible addition of the H atom to the
alkene and rupture from the alkyl fragment explains the double bond shift in the alkene. C-S
bond breaking (hydrogenolysis) has also been demonstrated for the homogeneous
Cp*2Mo2Co2S3(CO)4 cluster catalyst (Cp* stands for pentamethylcyclopentadienyl) [20].
2-Pentanethiol and di-(2-pentyl)amine were primary products and pentane and pentenes
were secondary products in the HDN of 2-pentylamine at 300 °C (cf. Fig. 4.2). The alkenes
and alkane are formed from the corresponding alkanethiol, that is formed from the alkylamine
2 2RNH + H S RSH + NH→ 3
2
(2)
and dialkylamine by substitution of the amine group by H2S
2RNHR + H S RSH + RNH→ (3)
The pentenes/pentane ratio in the HDN of 2-pentylamine was very similar to that obtained in
the HDS of 2-pentanethiol in the presence of hexylamine (Fig. 4.4). This demonstrates that
the branching ratio between alkenes and alkane is determined by the thiol and that the thiol is
an intermediate between alkylamine and hydrocarbons.
At 340 °C, the pentenes behaved like a primary product (Fig. 4.2) due to the fast
formation of pentenes from 2-pentanethiol (Fig. 4.3). The decrease in the selectivity of the
alkenes and increase in the selectivity of the alkane with weight time at 340 °C (Fig. 4.2) is
due to hydrogenation of the alkenes. At this elevated temperature the hydrogenation of the
alkenes is relatively fast because of the high conversion (and thus decreased inhibition) of the
alkylamine. The HDN of 2-pentylamine is, thus, similar to that of n-hexylamine [9]. This
shows that amines with primary and secondary α-carbon atoms do not undergo elimination to
alkenes or hydrogenolysis to an alkane.
HDN of Alkylamines Chapter 4 98
Table 4.1 Conversions and effect of H2S on conversion at time τ, and product selectivities at τ = 1 g.min/mol. Reactant T, τ, conv. select. °C g.min/ % effect H2S Cn
=/Cn RNHR, RSH, mol % %
2M2BA 270 3 30 (+) 16.5 0 0
300 1 46 (+) 16 0 0
2M2BT 270 1 65 -- 3.0
BA 270 3 25 0 0 0 0
TT 270 1 100 0
2PA 300 3 27 (+) 2.9 80 2
340 3 75 + 35 1
2PT 300 1 83 -- 2.9
340 1 95
3M2BA 300 3 18 + 3.5 45 9
340 3 70 + 3.5 10 3
3M2BT 300 1 75 -- 2.5
340 1 100 -- 3.2
2MCHA 300 3 10 (+) 1.0 38 4
340 1 55 (+) 3.5 10 1
3,3DM2BA 300 3 5 + 1.5 0 10
340 3 35 + 4.1 0 2
HA [9] 300 3 8 - 0.9 22 10
320 3 24 (-) 1.3 20 7
HT 300 1 90 -- 2.0
320 1 100 2.0
DHA [9] 300 1 55 0.8 14* 7
THA [9] 300 1 55 + 0.8 58 7
HDN of Alkylamines Chapter 4 99
4.4.2 HDN of amines with secondary α-carbon atoms
The reactions of 3-methyl-2-butylamine, 2-methylcyclohexylamine, and 3,3-dimethyl-2-
butylamine showed many similarities (cf. Figs. 4.5, 4.8, 4.9). The conversions vary within a
factor of three in the order 3-methyl-2-butylamine > 2-methylcyclohexylamine > 3,3-
dimethyl-2-butylamine (Table 4.1). Alkanethiol and dialkylamine are primary products (non-
zero selectivities at τ = 0) and alkenes and alkane are primary as well as secondary products
(increasing selectivity with τ). The initial alkene selectivities at 300 °C and 100 kPa H2S were
about 22% for 3-methyl-2-butylamine, 8% for 2-methylcyclohexylamine, and 34% for 3,3-
dimethyl-2-butylamine. The non-zero selectivities of the alkenes at τ = 0 are due either to
direct elimination of NH3 from the respective alkylamine or to a relatively slow reaction of
the alkylamine to the corresponding alkanethiol followed by a fast reaction of this thiol to the
alkenes. The latter seems more feasible, because it is hardly likely that 2-pentylamine does
not undergo elimination, but that the introduction of a methyl group on the neighboring β-
carbon atom induces such a change in the mechanism. The higher initial selectivities at 340
than at 300 °C are, thus, due to an even faster formation of alkenes from the alkanethiol and
the decrease with weight time at 340 °C to a relatively fast hydrogenation of the alkenes. This
agrees with the fact that the HDS rates of the alkanethiols decrease less than the conversions
of the corresponding alkylamines as a result of methyl substitution (Table 4.1). Thus, the HDS
of the intermediate is relatively faster for the methyl-substituted thiol than for the
unsubstituted thiol and the final secondary products tend to behave like primary products.
Furthermore, the alkenes/alkane ratio, obtained in the HDN of 3-methyl-2-butylamine, is
about the same as that obtained in the HDS of the corresponding 3-methyl-2-butanethiol (Fig.
4.6) and similar to the alkenes/alkane ratio obtained in the HDN of 2-pentylamine (Fig. 4.4).
These ratios of the alkylamines and alkanethiols with a secondary α-C atom are much lower
than the ratio obtained for 2-methyl-2-butylamine, an alkylamine with a tertiary α-C atom (cf.
Section 4.4.3). In Section 4.4.3 we will show that the latter alkylamine reacts by direct
elimination of ammonia. The alkenes/alkane ratios thus confirm that the alkylamines with a
secondary α-C atom do not react by elimination of ammonia.
The amines with secondary α-carbon atoms reacted faster than the primary n-hexylamine
[9] but much slower than 2-methyl-2-butylamine with a tertiary α-carbon atom and the
activated benzylamine (section 4.3.5). 2-Pentylamine and 3,3-dimethyl-2-butylamine, both
HDN of Alkylamines Chapter 4 100
with secondary α-carbon atoms, reacted four times faster at 300 °C than n-hexylamine and
neopentylamine [21], both with primary α-carbon atoms, respectively (Table 4.1). This
indicates that a pure SN2 mechanism cannot be responsible for the reaction of 2-pentylamine
and 3,3-dimethyl-2-butylamine. In a pure SN2 mechanism the extra methyl group on the α-
carbon atom would hinder the approach of a nucleophile, and the rate of reaction of the
alkylamine with a secondary α-carbon atom would be lower than that of the corresponding
alkylamine with a primary α-carbon atom [22]. The higher rates of the amines with secondary
α-carbon atoms are probably due to a weakening of the C-N bond because of the higher ionic
character of the C-N bond for a secondary carbon atom. The limit would be the dissociation of
the amine group with the formation of a secondary carbenium ion. This extreme situation did
not occur, however, for the amines in this study. In that case, an SN1 or E1 reaction would
have taken place and H2S would not have influenced the reaction rate.
The Wagner-Meerwein-type rearrangement of the carbon skeleton in the reaction of 3,3-
dimethyl-2-butylamine is ascribed to a nucleophilic substitution to 3,3-dimethyl-2-
butanethiol, followed (partly) by γ elimination aided by the neighboring group effect of the
methyl groups on the β-carbon atom (Scheme 4.3), analogous to reactions of branched
alcohols [23]. As to be expected, the rearrangement in the reaction of 2-methyl-3-pentylamine
was much less pronounced (< 1%).
NH2
SH
SH
+H2S
-NH3
-H2S
-H2S
H
Scheme 4.3 Reaction network of 3,3-dimethyl-2-butylamine.
HDN of Alkylamines Chapter 4 101
The HDN of cyclohexylamine and 2-methylcyclohexylamine and the HDS of the
corresponding thiols was described recently [15,24]. These compounds also have secondary
α-carbon atoms. The removal of ammonia from cyclohexylamine and 2-
methylcyclohexylamine was mainly (60-70%) ascribed to an E2 elimination reaction to
cyclohexene and methylcyclohexenes respectively, because cis-2-methylcyclohexylamine (c-
MCHA) reacted much faster to 1-methylcyclohexene than trans-2-methylcyclohexylamine (t-
MCHA) [24]. Elimination is assumed to occur with the β-H atom in an anti-periplanar
configuration relative to the axial NH2 group. As the conformations in Scheme 4.4 show, the
elimination of c-MCHA should then be faster than that of t-MCHA because of the presence of
the hydrogen atom on the tertiary carbon atom in the anti position to the amine group [24]. In
t-MCHA the methyl group occupies this anti position and cannot be eliminated. It was
overlooked, however, that nucleophilic substitution of the amine group by an SH group may
also explain the faster reaction of c-MCHA. As in the anti-periplanar E2 elimination, SN2
substitution can only occur when the leaving group is in the axial position (Scheme 4.4). In
that case, the methyl group in the anti position in t-MCHA hinders the approach of an SH
nucleophile from the backside of the carbon atom bearing the amine group and, thus, the SN2
reaction should also be faster for c-MCHA than for t-MCHA. Thus, our conclusion, that 2-
methylcyclohexylamine reacts by substitution and not by elimination, does not contradict the
experimental results described in refs. 15 and 24.
E2
SN2
c-MCHA t-MCHA
NH2
CH3
H
NH2
CH3
NH2
CH3
NH2
H
CH3
B
SHSH
Scheme 4.4 E2 and SN2 reactions of cis- and trans 2-methylcyclohexylamine.
HDN of Alkylamines Chapter 4 102
4.4.3 HDN of 2-methyl-2-butylamine and benzylamine
2-Methyl-2-butylamine and benzylamine reacted much faster than the other amines
studied. Whereas 2-pentylamine, 3-methyl-2-butylamine, 3,3-dimethyl-2-butylamine, and 2-
methylcyclohexylamine (all with secondary α-carbon atoms) did not show appreciable
conversion below 300 °C, 2-methyl-2-butylamine, with a tertiary α-carbon atom, already
reached a conversion of 30% at τ = 3 g.min/mol at 270 °C. The product distribution was also
different. For the most part, alkenes (2-methyl-1-butene and 2-methyl-2-butene) and an alkane
(methylbutane), but no di-(2-methyl-2-butyl)amine, and only a trace of 2-methyl-2-
butanethiol were observed. This behavior indicates that 2-methyl-2-butylamine, with a tertiary
α-carbon atom, reacts by a different mechanism than the amines with secondary α-carbon
atoms. Furthermore, because H2S does not influence the reaction rate, the most likely
mechanisms are E1 elimination and SN1 nucleophilic substitution.
If 2-methyl-2-butylamine were to react by a classic organic E1 or SN1 mechanism, then it
would be protonated and would react by ammonia removal to the tertiary isopentyl carbenium
ion (Scheme 4.2). In the E1 mechanism, this ion would then react further to 2-methyl-1-
butene and 2-methyl-2-butene by proton removal and the formation of methylbutane would be
unaccounted for. However, on the metal sulfide surface, the 2-methyl-2-butylamine adsorbs
with the nitrogen lone pair on an Mo or Ni atom. After C-N bond breaking, the isopentyl
carbenium ion will either move to a neighboring Mo or Ni atom or to a sulfur atom. If the
carbenium ion binds to a metal atom, an electron transfer reaction may take place with the
formation of the isopentyl radical. As in Fischer-Tropsch chemistry on a metal surface [2],
this alkyl radical may react to an alkene by removal of a hydrogen atom, or it may add a
hydrogen atom and become an alkane. If the carbenium ion binds to a sulfur atom, then
adsorbed 2-methyl-2-butanethiol forms and the mechanism changes to the SN1 type. 2-
Methyl-2-butanethiol can react to 2-methylbutenes as well as to methylbutane.
The methylbutenes/methylbutane ratio of the products of the HDN of 2-methyl-2-
butylamine was about five times larger than that obtained in the HDS of the corresponding 2-
methyl-2-butanethiol (Fig. 4.13). This demonstrates that the methylbutenes/methylbutane
ratio in the HDN is not determined by the thiol and that 2-methyl-2-butylamine reacts by an
E1 rather than an SN1 mechanism. In agreement with this conclusion, only 0.3% thiol was
observed in the HDN of 2-methyl-2-butylamine in the presence of 10 kPa H2S at τ = 1
HDN of Alkylamines Chapter 4 103
g.min/mol; even in the presence of 100 kPa H2S, the initial selectivity of the thiol was only
1.5% (Fig. 4.11). Figure 4.12 demonstrates that 2-methyl-2-butanethiol reacts rather slowly in
the presence of 100 kPa H2S. Thus, if this thiol had been an intermediate in the HDN of 2-
methyl-2-butylamine, then a quantity larger than 1.5% would have been observed.
Benzylamine cannot react by elimination, because it has no β-hydrogen atoms. The high
reaction rate and lack of an effect of the H2S pressure suggest that benzylamine does not react
by an SN2 but by an SN1 reaction. Protonation of the amine group and the removal of
ammonia would lead to the relatively stable benzyl carbenium ion. As indicated above for the
isopentyl carbenium ion, the benzyl carbenium ion will move to a neighboring Mo or Ni atom
or to a sulfur atom. If it binds to a metal atom, an electron transfer reaction may take place
with the formation of the benzyl radical, which can be hydrogenated to toluene. If the
carbenium ion binds to a sulfur atom, adsorbed α-toluenethiol forms and may react to toluene.
4.5 Conclusions
Our former [9] and present results show that alkylamines with the NH2 group attached to a
primary or secondary carbon atom react by substitution of the NH2 group by an SH or amine
group to form an alkanethiol or a dialkylamine. After subsequent substitution by H2S the
dialkylamine also reacts to an alkanethiol. The alkanethiol finally reacts to an alkene or alkane
and H2S. Only an alkylamine with the NH2 group attached to a tertiary or activated carbon
atom reacts directly to an alkene or alkane. The C-N bonds of alkylamines with primary and
secondary α-carbon atoms are too strong to be easily broken. For such alkylamines
elimination is, therefore, too difficult and they react by other mechanisms. The stabilization of
the tertiary or benzyl carbenium cation is necessary to weaken the C-N enough for elimination
to take place.
The proposal by Portefaix et al. [3], that alkylamines react by elimination and that the
number of β-H atoms determines their HDN rate, is thus incorrect. The fact that 2-methyl-2-
butylamine reacts much faster than n-pentylamine has nothing to do with the four times larger
number of β-H atoms but has everything to do with the fact that the NH2 group of the former
amine is attached to a tertiary α-C atom and the NH2 group of the latter amine to a primary α-
HDN of Alkylamines Chapter 4 104
C atom. Even when elimination occurs, as for 2-methyl-2-butylamine, the selectivity for 2-
methyl-2-butene is higher than that for 2-methyl-1-butene, although there are three times
more β-H atoms on the terminal methyl groups than on the internal methylene group. We
checked that this is not due to a fast isomerization of 2-methyl-1-butene to 2-methyl-2-butene.
The higher selectivity for 2-methyl-2-butene is due to the fact that in an E1 mechanism the
leaving group is gone before the proton. The product is thus determined by thermodynamic
factors and Zaitsev’s rule applies: the double bond goes preferentially to the most highly
substituted carbon atom. We conclude therefore that the number of β-H atoms, as proposed by
Portefaix et al., does not determine the HDN rate of the alkylamines. In alkylamines with
primary and secondary α-C atoms the number of β-H atoms is of no importance, because
substitution rather than E2 elimination takes place. In alkylamines with a tertiary α−C atom
the reverse occurs: the hydrogen atom is preferentially removed from the β-C atom with the
lowest number of β-H atoms!
Even though elimination of the NH2 group from an alkylamine does not take place when it
is attached to a primary or secondary carbon atom, removal of nitrogen does take place for
such alkylamines. The substitution of the NH2 group by H2S leads to an alkanethiol and
ammonia and, thus, to total denitrogenation. Substitution by an amine leads to a dialkylamine
and ammonia and to 50% nitrogen removal. The rest of the nitrogen is removed in the
subsequent substitution of the dialkylamine by H2S to an alkanethiol and the original
alkylamine. High partial pressures of H2S and alkylamine increase the rate of transformation
of alkylamine to alkanethiol and, thus, of denitrogenation. At the same time, however, the rate
of sulfur removal from the alkanethiol decreases.
While we have pinpointed the types of reactions that alkylamines undergo, we have not
answered the question as to how these reactions are catalyzed by the supported metal sulfide.
Our future work will address the question of how the substitution of the NH2 group by H2S
takes place on the surface of nickel- and cobalt-promoted and unpromoted MoS2.
Acknowledgement.
The authors thank T. Schmid for synthesizing 2-methyl-3-pentylamine and Dr. P. Kukula for
discussions.
HDN of Alkylamines Chapter 4 105
4.6 References
[1] N. Nelson, R.B. Levy, J. Catal. 58 (1979) 485.
[2] H. Schulz, M. Schon, N.M. Rahman, Stud. Surf. Sci. Catal. 27 (1986) 201.
[3] J.L. Portefaix, M. Cattenot, M. Guerriche, J. Thivolle-Cazat, M. Breysse, Catal. Today
10 (1991) 473.
[4] L. Vivier, V. Dominguez, G. Perot, S. Kasztelan, J. Mol. Catal. 67 (1991) 267.
[5] R. Prins, Adv. Catal. 46 (2001) 399.
[6] M. Houalla, N.K. Nag, A.V. Sapre, D.H. Broderick, B.C. Gates, AIChE J. 24 (1978)
1015.
[7] J. Mijoin, G. Pérot, F. Bataille, J.L. Lemberton, M. Breysse, S. Kasztelan, Catal. Lett.
71 (2001) 139.
[8] M. Cattenot, J.L. Portefaix, J. Afonso, M. Breysse, M. Lacroix, G. Perot, J. Catal. 173
(1998) 366.
[9] Y. Zhao, P. Kukula, R. Prins, J. Catal. 221 (2004) 441; and chapter 3.
[10] M. Egorova, Y. Zhao, P. Kulula, R. Prins, J. Catal. 206 (2002) 263.
[11] H. Knözinger, A. Scheglila, J. Catal. 17 (1970) 252.
[12] P. Desikan, C.H. Amberg, Can. J. Chem. 42 (1964) 843.
[13] D.L. Sullivan, J.G. Ekerdt, J. Catal. 178 (1998) 226.
[14] M.J. Girgis, B.C. Gates, Ind. Eng. Chem. Res. 30 (1991) 2021.
[15] F. Rota, R. Prins, J. Catal. 202 (2001) 195.
[16] R. Prins, in “ Handbook of Heterogeneous Catalysis” (G. Ertl, H. Knözinger, J.
Weitkamp, Eds.). Vol 4, p. 1916, Wiley-VCH.
[17] R.M. Laine, Catal. Rev.-Sci. Eng. 25 (1983) 459.
[18] G. Meitzner, W.J. Mykytka and J.H. Sinfelt, J. Catal. 98 (1986) 513.
[19] Triyono, R. Kramer, Appl. Catal. A 100 (1993) 145.
[20] M.D. Curtis, S.H. Druker, J. Am. Chem. Soc. 119 (1997) 1027.
[21] Y. Zhao, J. Czyzniewska, R. Prins, Catal. Lett. 88 (2003) 155.
[22] J. Clayden, N. Greeves, S. Warren, P. Wothers, ‘’Organic Chemistry’’. Oxford Univ.
Press, 2001.
[23] H. Pines, J. Manassen, Adv. Catal. 16 (1966) 49.
[24] F. Rota, V.S. Ranade, R. Prins, J. Catal. 200 (2001) 389.
HDN over Hydrotreating Catalysts Chapter 5 107
5. Mechanisms of HDN of Alkylamines and HDS of alkanethiol on NiMo/Al2O3, CoMo/Al2O3, and Mo/Al2O3
5.1 Abstract
The simultaneous hydrodenitrogenation (HDN) of alkylamines and hydrodesulfurization
(HDS) of alkanethiols, with the NH2 and SH groups attached to primary, secondary, and
tertiary carbon atoms were studied at 270 and 270-320 °C and 3 MPa over sulfided
NiMo/Al2O3, CoMo/Al2O3, and Mo/Al2O3 catalysts. Pentylamine and 2-hexylamine reacted
by substitution with H2S to alkanethiols and with another amine molecule to dialkylamines.
Alkenes and alkanes were not formed directly from pentylamine and 2-hexylamine, but
indirectly by elimination and hydrogenolysis of the alkanethiol intermediates, as confirmed by
their secondary behaviour and the similar alkenes/alkane ratios in the simultaneous reactions
of amines and thiols. Only 2-methyl-2-butylamine, with the NH2 group attached to a tertiary
carbon atom, produced alkenes as primary products by E1 elimination. NiMo/Al2O3 and
CoMo/Al2O3 have a higher activity for the HDS of alkanethiols than Mo/Al2O3 and H2S has a
negative influence. This shows that the thiols react on vacancies on the catalyst surface
(Lewis acid sites). Mo/Al2O3 is the best HDN catalyst and H2S has a positive influence for the
HDN of amines with the NH2 group attached to a secondary and tertiary carbon atom. This
indicates that the HDN of alkylamines occurs on Brønsted acid sites.
5.2 Introduction
Environmental legislation in the coming years requires a further reduction of the sulfur
content of gasoline and diesel fuel to 10 ppm. Deep hydrodesulfurization (HDS) technology
has to be implemented to reach this low level of sulfur. Nitrogen-containing compounds are
harmful in deep HDS, as they inhibit the HDS of sulfur-containing compounds through
HDN over Hydrotreating Catalysts Chapter 5 108
competitive adsorption [1-3]. Therefore, it is important to know how nitrogen-containing
molecules are removed by hydrotreating catalysts. Nelson and Levy were the first to suggest
mechanisms for the hydrodenitrogenation (HDN) of alkylamines and they suggested that
nucleophilic substitution and Hofmann β-H elimination are responsible for HDN [4]. For each
of these mechanisms evidence has been published [5-8] and it has been concluded that the
mechanism depends on the alkylamine as well as on the catalyst [7]. We found that over
sulfided NiMo/Al2O3 the substitution mechanism is predominant in the HDN of alkylamines
with the NH2 group attached to a primary or secondary carbon atom [9,10]. Hofmann β-H
elimination hardly takes place in these alkylamines. Nevertheless, a large amount of alkenes is
formed in the HDN of an alkylamine by fast decomposition of the corresponding alkanethiol,
which is formed by the substitution of the alkylamine with H2S. On the other hand, the HDN
of tertiary amines does occur via an E1 mechanism [10].
Cattenot et al. concluded that the HDN mechanism not only depends on the alkylamine
but also on the catalyst [7]. Different metal sulfides may have different acidities and therefore
different catalytic properties. According to their results of unsupported metal sulfides, the
acidity decreases in the order NbS3 > MoS2 > RuS2 > Rh2S3. The HDN reaction of
alkylamines might help to discriminate between the acid properties of NiMo/Al2O3,
CoMo/Al2O3, and Mo/Al2O3. In the HDN of an alkylamine, the products are a dialkylamine,
an alkanethiol, alkenes, and an alkane. They can, in principle, be formed by acid-base
catalysis (Scheme 5.1). The amine group reacts with a proton or a Lewis acid and, at the same
time, a nucleophile attacks the α-carbon atom. The ammonia is split off, while a substitution
product is formed. That acid-base chemistry can take place at metal sulfide surfaces was
confirmed by FTIR spectroscopy of pyridine adsorbed at high temperature [11]. Maugé and
coworkers found evidence that H2S adsorption leads to a substantial increase in the Brønsted
acidity of the sulfide phase [12-14].
Substitution may not only be catalyzed by acidic sites, but also by metallic sites [15].
Several groups have discussed the electronic properties of the ideal reactive MoS2 surface and
pointed out that this surface has electron acceptor property and that it is metallic [16-22]. As
shown in Scheme 5.2, a sequence of dehydrogenation of amine to imine, H2S addition,
ammonia elimination, and hydrogenation of the thioaldehyde transforms an alkylamine into
an alkanethiol.
HDN over Hydrotreating Catalysts Chapter 5 109
For these reasons, we compared the HDN mechanisms over the three catalysts in the
simultaneous reactions of pentylamine and hexanethiol, the simultaneous reactions of 2-
hexylamine and 2-pentanethiol, the HDN of 2-methyl-2-butylamine, and the HDS of 2-
methyl-2-butanethiol in the presence of hexylamine over the three metal sulfide catalysts.
H2S + R1 C NH2
R2
H
SN2C SH
R2
H
+ NH3R1
Scheme 5.1 Direct substitution of an alkylamine to an alkanethiol by acid-base catalysis.
C N
H
H
H
C N
SH
H
HN
H-H2 H2S
S + NH3
S C SH
H
C H
H
+ H2SH2 H2
Scheme 5.2 Indirect substitution of an alkylamine to an alkanethiol by metal catalysis.
5.3 Results
As we showed for sulfided NiMo/Al2O3 [9,10] and as we will show here for sulfided
Mo/Al2O3 and CoMo/Al2O3, the HDN of alkylamines is not a simple reaction but is the sum
of two parallel reactions, each of which consists of several consecutive reactions. To
disentangle this complex network of reactions, we will use two methods. One method is to
concentrate on selectivities instead of yields. The advantage of selectivities over yield is that
they allow, in principle, to more easily distinguish between primary and secondary products.
For instance, in a series of consecutive reactions from reactant A to product D
1 2 3k k kA B C⎯⎯→ ⎯⎯→ ⎯⎯→D
the yield of the primary product B as well as of the secondary product C initially increase with
reaction time (Fig. 5.1 A), but this behaviour would also be observed for parallel reactions of
A to B and A to C. Only if one can measure at very short time, will one be able to distinguish
HDN over Hydrotreating Catalysts Chapter 5 110
between consecutive and parallel reactions and prove the secondary character of C. By
plotting the same experimental results as selectivities, however, the selectivity of B decreases
and that of C increases with time initially for a consecutive reaction (Fig. 5.1B), while both
selectivities increase for parallel reactions of A to B and C. Even though the product yields
are initially low, and thus the uncertainties in the selectivities high, it is still easy to
distinguish between an increase or a decrease of the selectivity with time. Therefore, we will
use selectivities to unravel the reaction mechanisms.
0 1 2 3 4 50
20
40
60
80
100B
0 1 2 3 4 50
50
100A
D
B
A
Yiel
d, %
k*t
D
C
B
Sele
ctiv
ity, %
k*t
C
Figure 5.1 Product selectivities and yields in a consecutive reaction from reactant A to
product D (with equal rate constants for each reaction step).
The second method that we use is the comparison of the alkene/alkane ratio in the HDN of
an alkylamine with the ratio in the HDS of a similar alkanethiol. The idea is that if an
alkylamine would react to alkenes and an alkane without going through an intermediate
alkanethiol, the alkene/alkane ratio should not be the same as that in the HDS of the
alkanethiol. If, on the other hand, the alkylamine first reacts to an alkanethiol, which then
reacts to a mixture of alkenes and alkane, the alkenes/alkane branching ratio will be
determined by the alkanethiol. Thus the branching ratio will be the same in HDN and in HDS.
The ratio of elimination to hydrogenolysis of an alkanethiol, and thus the resulting
alkenes/alkane ratio, is sensitive to the coverage on the catalyst surface and thus to the
presence and partial pressures of alkylamine and alkanethiol. Therefore, the branching ratio in
HDS was always determined in the presence of an amount of alkylamine equivalent to that
used in HDN.
HDN over Hydrotreating Catalysts Chapter 5 111
5.3.1. Simultaneous reaction of pentylamine and hexanethiol
The conversion of 5 kPa pentylamine in the simultaneous HDN and HDS at 320 °C in the
presence of 5 kPa hexanethiol and 10 kPa H2S was 12% at τ = 1.6 g.min/mol and 60% at 8.9
g.min/mol over NiMo/Al2O3 (Fig. 5.2); the conversion over CoMo/Al2O3 was similar. The
conversion over both catalysts decreased slightly when the H2S pressure was increased from
10 to 100 kPa. Over Mo/Al2O3, the conversion of pentylamine was much higher than over
NiMo/Al2O3 and CoMo/Al2O3 and it decreased substantially with increasing H2S pressure
from 10 to 100 kPa H2S.
Figure 5.2 Conversion of pentylamine in the presence of hexanethiol at 10 ( ) and
100 ( ) kPa H2S, and 320 °C over sulfided NiMo/Al2O3, CoMo/Al2O3
and Mo/Al2O3 catalysts.
0 2 4 6 8 100
20
40
60
80
100Mo/Al2O3
C5-N
H2 C
onve
rsio
n, %
Weight time, g.min/mol
40
60
0 2 4 6 8 100
20
NiMo/Al2O3
C5-N
H2 C
onve
rsio
n, %
Weight time, g.min/mol 0 2 4 6 8 100
20
40
60CoMo/Al2O3
C5-N
H2 C
onve
rsio
n, %
Weight time, g.min/mol
HDN over Hydrotreating Catalysts Chapter 5 112
The products over the three catalysts were pentanethiol, dipentylamine, pentenes (i.e. the
sum of the two pentene isomers), and pentane (Figs. 5.3-5.5). The selectivities to
dipentylamine and pentanethiol increased in the order NiMo<CoMo<Mo, while the reverse
order was observed for pentane and the pentenes. Dipentylamine behaved as a primary
product over all three catalysts. It was present in much larger quantity over Mo/Al2O3 than
over NiMo/Al2O3 and CoMo/Al2O3. The selectivity of pentanethiol over the NiMo/Al2O3 and
CoMo/Al2O3 catalysts increased with decreasing weight time, suggesting that pentanethiol is
a primary product as well. The Mo/Al2O3 catalyst behaved differently at 100 kPa H2S (Fig.
5.5).
0 2 4 6 8 100
20
40
60
Pent
ane,
%
Weight time, g.min/mol0 2 4 6 8 10
0
10
20
30
40
50Pe
nten
es, %
Weight time, g.min/mol
0 2 4 6 8 10
0
5
10
15
20
25
Pent
anet
hiol
, %
Weight time, g.min/mol0 2 4 6 8 10
0
5
10
15
20
25
DPA
, %
Weight time, g.min/mol
Figure 5.3 Product selectivities in the HDN of pentylamine in the presence of
hexanethiol at 10 ( ) and 100 ( ) kPa H2S, and 320 °C over sulfided
NiMo/Al2O3 (DPA = dipentylamine).
HDN over Hydrotreating Catalysts Chapter 5 113
50
Figure 5.4 Product selectivities in the HDN of pentylamine in the presence of
hexanethiol at 10 ( ) and 100 ( ) kPa H2S, and 320 °C over sulfided
CoMo/Al2O3 (DPA = dipentylamine).
When decreasing the weight time, the selectivity of pentanethiol first increased, then reached
a maximum, and decreased at shorter weight time. This result was checked several times, by
repeating the experiments. The selectivity of pentanethiol strongly increased and that of
dipentylamine strongly decreased with increasing H2S pressure over all three catalysts.
The selectivities of the pentenes and pentane as a function of weight time were different
over the three catalysts (Figs. 5.3-5.5). The pentenes and pentane behaved as primary
products over NiMo/Al2O3 and CoMo/Al2O3, as the selectivities extrapolate to a non-zero
value with decreasing weight time. Since the selectivities initially increase with weight time,
the pentenes and pentane are secondary or even tertiary products as well. Over Mo/Al2O3,
0 2 4 6 8 100
10
0 2 4 6 8 100
20
40
60
Pent
enes
, %
Weight time, g.min/mol
40
30
20
Pent
ane,
%
Weight time, g.min/mol
40
0 2 4 6 8 100
10
0 2 4 6 8 100
10
20
30
40
DPA
, %
Weight time, g.min/mol
30
20
Pent
anet
hiol
, %
Weight time, g.min/mol
HDN over Hydrotreating Catalysts Chapter 5 114
however, the selectivities extrapolate to zero with decreasing weight time, which indicates
that over this catalyst the hexenes and hexane are secondary or tertiary products only.
0 2 4 6 8 100
20
40
60
80
Pent
ane,
%
Weight time, g.min/mol0 2 4 6 8 10
0
10
20
30
Pent
enes
, %Weight time, g.min/mol
0 2 4 6 8 100
20
40
60
Pen
tane
thio
l, %
Weight time, g.min/mol0 2 4 6 8 10
0
20
40
60
80
100
DPA
, %
Weight time, g.min/mol
Figure 5.5 Product selectivities in the HDN of pentylamine in the presence of
hexanethiol at 10 ( ) and 100 ( ) kPa H2S, and 320 °C over Mo/Al2O3
(DPA = dipentylamine).
The conversion of 5 kPa hexanethiol in the presence of 10 kPa H2S and 5 kPa pentylamine
(simultaneous HDS and HDN) was much higher than that of pentylamine. At τ = 1.6
g.min/mol it was already 100% at 320 °C (Table 5.1), while that of pentylamine was 12% at
320 °C over NiMo/Al2O3 and CoMo/Al2O3 (Fig. 5.2). The conversion of hexanethiol
decreased when the H2S pressure was increased from 10 to 100 kPa, most strongly over
Mo/Al2O3 (Table 5.1). Nevertheless, for all three catalysts the conversion of the thiol was
complete at 320 °C at short τ. This means that the presence of H2S during the HDN was
actually 5 kPa higher than the indicated 10 or 100 kPa due to H2S in the feed. The
HDN over Hydrotreating Catalysts Chapter 5 115
hexenes/hexane ratio, obtained in the HDS of hexanethiol, was similar to the
pentanes/pentane ratio obtained in the HDN of pentylamine over all three catalysts (Fig. 5.6).
2.0
Figure 5.6 Hexenes/hexane ratio in the HDS of hexanethiol (HT) and pentenes/pentane
ratio in the HDN of pentylamine (PA) at 10 ( ) and 100 ( ) kPa H2S,
and 320 °C over sulfided NiMo/Al2O3, CoMo/Al2O3, and Mo/Al2O3
(DPA = dipentylamine).
0 2 4 6 8 100.0
0.5
1.0
1.5
HT, NiMo/Al2O3
C6= /C
6
0 2 4 6 8 100.0
0.5
1.0
1.5
2.0
Weight time, g.min/mol
PA, NiMo/Al2O3
C5= /C
5
0 2 4 6 8
Weight time, g.min/mol
10
2.0
1.5
1.0
0.5
0.0
HT, CoMo/Al2O3
C6= /C
6
Weight time, g.min/mol0 2 4 6 8 10
0.0
0.5
1.0
1.5
2.0PA, CoMo/Al2O3
C5= /C
5
Weight time, g.min/mol
2.0
0 2 4 6 8 100.0
0.5
1.0
1.5
HT, Mo/Al2O3
C6
6
Weight time, g.min/mol0 2 4 6 8 10
0.0
0.5
1.0
1.5
2.0PA, Mo/Al2O3
C5= /C
5
Weight time, g.min/mol
= /C
HDN over Hydrotreating Catalysts Chapter 5 116
TABLE 5.1 Conversions (%) of hexanethiol, 2-pentanethiol, and 2-methyl-2-butanethiol in the presence
of pentylamine, 2-hexylamine, and hexylamine respectively at 270 and 320 °C, 10 and 100
kPa H2S at τ = 1.6 g.min/mol over NiMo/Al2O3, CoMo/Al2O3, and Mo/Al2O3
Hexanethiol 2-pentanethiol 2-methyl-2-butanethiol
320 °C 320 °C 270 °C
H2S (kPa) 10 100 10 100 10 100
NiMo 100 90 100 89 76 32
CoMo 100 79 99 83 77 26
Mo 88 24 98 59 65 22
The conversion of pentylamine over sulfided Zn/Al2O3 and Cd/Al2O3 was lower than 4%
at 320 °C, 10 kPa H2S and τ = 8.9 g.min/mol.
5.3.2. Simultaneous reaction of 2-hexylamine and 2-pentanethiol
The conversion of 2-hexylamine was higher than that of pentylamine over NiMo and
CoMo and increased weakly with increasing H2S pressure (cf. Figs. 5.2 and 5.7). The main
products were 2-hexanethiol, di-(2-hexyl)amine, hexenes (i.e. the sum of the three hexene
isomers), and hexane. Like for pentylamine, the selectivities for disproportionation and
substitution increased in the order NiMo<CoMo<Mo, while those for hexane and hexenes
decreased in the same order (Figs. 5.8-5.10). The selectivities of 2-hexanethiol and di-(2-
hexyl)amine increased with decreasing weight time (Figs. 5.8-5.10), which shows that they
are primary products. The selectivity of di-(2-hexyl)amine decreased with increasing H2S
pressure, while the reverse was true for the selectivity of 2-hexanethiol.
The selectivities of the hexenes and hexane as a function of weight time were different
over the three catalysts (Figs. 5.8-5.10). Hexenes and hexane behave as primary products
over NiMo/Al2O3 and CoMo/Al2O3, as the selectivities extrapolate to non-zero values with
decreasing weight time, as well as secondary products because the selectivities initially
increase with increasing weight time. Over Mo/Al2O3, however, the selectivities extrapolate
HDN over Hydrotreating Catalysts Chapter 5 117
to zero with decreasing weight time, which indicates that hexenes and hexane are secondary
or tertiary products only.
0 2 4 6 8 100
20
40
60
80
100NiMo/Al2O3
2-H
exyl
amin
e C
onve
rsio
n, %
Weight time, g.min/mol
0 2 4 6 8 100
20
40
60
80
100CoMo/Al2O3
2-H
exyl
amin
e C
onve
rsio
n, %
Weight time, g.min/mol
0 2 4 6 8 10
0
20
40
60
80
100Mo/Al2O3
2-C
6-NH
2 Con
vers
ion,
%
Weight time, g.min/mol
Figure 5.7 Conversion of 2-hexylamine in the presence of 2-pentanethiol at 10 ( )
and 100 ( ) kPa H2S, and 320 °C over sulfided NiMo/Al2O3,
CoMo/Al2O3 and Mo/Al2O3
HDN over Hydrotreating Catalysts Chapter 5 118
0 2 4 6 8 100
20
40
60
80
0 2 4 6 8 100
20
40
60
Hex
enes
, %
Weight time, g.min/mol
ne, %
Hex
a
Weight time, g.min/mol
0 2 4 6 80
10
20
30
10
Figure 5.8 Product selectivities in the HDN of 2-hexylamine in the presence of
2-pentanethiol at 10 ( ) and 100 ( ) kPa H2S, and 320 °C over sulfided
NiMo/Al2O3
2-H
exa
io
Weight time, g.min/mol0 2 4 6 8 10
0
10
20
30
40
Di-2
-hex
ylam
ine,
%
Weight time, g.min/mol
l, %
neth
HDN over Hydrotreating Catalysts Chapter 5 119
0 2 4 6 8 10
0
20
40
60
80
Hex
ane,
%
Weight time, g.min/mol0 2 4 6 8 10
0
20
40
60
Hex
enes
, %Weight time, g.min/mol
0 2 4 6 8 100
10
20
30
2-H
exan
ethi
ol, %
Weight time, g.min/mol0 2 4 6 8 10
0
10
20
30
40
50
Di-2
-hex
ylam
ine,
%
Weight time, g.min/mol
Figure 5.9 Product selectivities in the HDN of 2-hexylamine in the presence of
2-pentanethiol at 10 ( ) and 100 ( ) kPa H2S, and 320 °C over sulfided
CoMo/Al2O3
HDN over Hydrotreating Catalysts Chapter 5 120
0 2 4 6 8 100
20
40
60
80
Hex
ane,
%
Weight time, g.min/mol
0 2 4 6 8 100
10
20
30
40
Hex
enes
, %
Weight time, g.min/mol
0 2 4 6 8 10
0
10
20
30
40
2-H
exan
ethi
ol, %
Weight time, g.min/mol0 2 4 6 8 10
0
20
40
60
80
100D
i-2-h
exyl
amin
e, %
Weight time, g.min/mol
Figure 5.10 Product selectivities in the HDN of 2-hexylamine in the presence of
2-pentanethiol at 10 ( ) and 100 ( ) kPa H2S, and 320 °C over sulfided
Mo/Al2O3
The conversion of 2-pentanethiol in the simultaneous reaction with 2-hexylamine in the
presence of 10 kPa H2S was much larger than that of the corresponding amine. It was almost
100% at τ = 1.6 g.min/mol at 320 °C (Table 5.1), while that of the corresponding amine was
30% (Fig. 5.7). The conversion of hexanethiol decreased when the H2S pressure was
increased from 10 to 100 kPa, most strongly over Mo/Al2O3. The pentenes/pentane ratio
obtained from the HDS of 2-pentanethiol was the same as the hexenes/hexane ratio obtained
from the HDN of 2-hexylamine over NiMo/Al2O3 and CoMo/Al2O3 (Fig. 5.11). The ratio
originating from 2-pentanethiol was slightly lower than the one from 2-hexylamine over
Mo/Al2O3.
HDN over Hydrotreating Catalysts Chapter 5 121
0 2 4 6 8 100.0
0.5
1.0
1.5
2.02-PT, NiMo/Al2O3
C5= /C
5
Weight time, g.min/mol0 2 4 6 8 10
0.0
0.5
1.0
1.5
2.02-HA, NiMo/Al2O3
C6= /C
6
Weight time, g.min/mol
0 2 4 6 8 10
0.0
0.5
1.0
1.5
2.02-PT, CoMo/Al2O3
C5= /C
5
Weight time, g.min/mol
0 2 4 6 8 100.0
0.5
1.0
1.5
2.02-HA, CoMo/Al2O3
C6= /C
6
Weight time, g.min/mol
0 2 4 6 8 10
0.0
0.5
1.0
1.5
2.02-PT, Mo/Al2O3
C5= /C
5
Weight time, g.min/mol
0 2 4 6 8 100.0
0.5
1.0
1.5
2.02-HA, Mo/Al2O3
C6= /C
6
Weight time, g.min/mol
Figure 5.11 Pentenes/pentane ratio in the HDS of 2-pentanethiol (2-PT) and
hexenes/hexane ratio in the HDN 2-hexylamine (2-HA) at 10 ( ) and 100
( ) kPa H2S, and 320 °C over sulfided NiMo/Al2O3, CoMo/Al2O3, and
Mo/Al2O3
HDN over Hydrotreating Catalysts Chapter 5 122
5.3.3. 2-Methyl-2-butylamine and 2-methyl-2-butanethiol
The HDN of 2-methyl-2-butylamine occurred fast; the conversion was already 17% at 270
°C and 10 kPa H2S at low weight time (1.6 g.min/mol) and reached 75% at τ = 8.9 g.min/mol
over NiMo/Al2O3 (Fig. 5.12). The conversions over CoMo/Al2O3 and Mo/Al2O3 were even
higher than over NiMo/Al2O3. In all cases, the main products were 2-methyl-2-butene, 2-
methyl-1-butene, and 2-methylbutane (not shown), and the methylbutenes were primary
products and methylbutane a secondary product.
0 2 4 6 8 100
20
40
60
80
100CoMo/Al2O3
2M2B
A C
onve
rsio
n, %
Weight time, g.min/mol0 2 4 6 8 10
0
20
40
60
80
100NiMo/Al2O3
Figure 5.12 Conversion of 2-methyl-2-butylamine (2M2BA) at 270 °C, 10 ( ) and
100 ( ) kPa H2S over sulfided NiMo/Al2O3, CoMo/Al2O3, and Mo/Al2O3
2M2B
A C
onve
rsio
n,
Weight time, g.min/mol
%
0 2 4 6 8 100
20
40
60
80
100Mo/Al2O3
2M2B
Aon
ve
Weight time, g.min/mol
rsio
n, %
C
HDN over Hydrotreating Catalysts Chapter 5 123
The conversion of 2-methyl-2-butanethiol at 270 °C in the presence of 10 kPa H2S and 5
kPa hexylamine was much larger than that of the equivalent amine. It was already 76% at τ =
1.6 g.min/mol over NiMo/Al2O3 (Table 5.1), while the conversion of the corresponding amine
was only 17% (Fig. 5.12). The conversion of 2-methyl-2-butanethiol strongly decreased with
increasing H2S pressure. The methylbutenes/methylbutane ratio, obtained in the HDS of 2-
methyl-2-butanethiol in the presence of hexylamine, was very different from the one obtained
in the HDN of 2-methyl-2-butylamine under the same conditions over all three catalysts (Fig.
5.13).
The conversion of 2-methyl-2-butylamine over sulfided Cd/Al2O3 at 10 kPa H2S and 270
°C was only 6% at τ = 8.9 g.min/mol. It increased to 25 % at 300 °C and 100 % at 320 °C.
20
Figure 5.13 Methylbutenes/methylbutane ratio in the HDN of 2-methyl-2-
butylamine (2M2BA) and HDS of 2-methyl-2-butanethiol (2M2BT) in
the presence of hexylamine at 270 °C, 10 ( ) and 100 ( ) kPa H2S over
sulfided NiMo/Al2O3, CoMo/Al2O3 and Mo/Al2O3
0 2 4 6 8 100
5
15
10
NiMo/Al2O3
2M2BT
2M2BA
C5= /C
5
Weight time, g.min/mol
40
30
20
10
0 2 4 6 8 100
Mo/Al2O3
2M2BT
2M2BA
C5= /C
5
Weight time, g.min/mol
0 2 4 6 8 100
5
10
15
20
2M2BT
2M2BA
CoMo/Al2O3
C5= /C
5
Weight time, g.min/mol
HDN over Hydrotreating Catalysts Chapter 5 124
5.4. Discussion
5.4.1. Pentylamine
The conversion of pentylamine at 320 °C was similar over NiMo/Al2O3 and CoMo/Al2O3
(Fig. 5.2), but lower than over Mo/Al2O3. It decreased more stongly over Mo/Al2O3 with
increasing H2S pressure from 10 to 100 kPa H2S. At first glance, these results seem to be in
contradiction with literature, which reports that NiMo/Al2O3 is a better HDN catalyst than
CoMo/Al2O3 and much better than Mo/Al2O3 [23-25]. However, these publications are
mainly concerned with the HDN of aromatic N-containing molecules for which
hydrogenation and not the final N-removal step is rate determining. Indeed, the position of the
maximum in the selectivity of the pentenes at lower τ for NiMo (Fig. 5.3) than for CoMo (Fig.
5.4) than for Mo (Fig. 5.5) indicates that hydrogenation is fastest for the NiMo catalyst and
slowest for the Mo catalyst. Also, the product selectivities in the HDN of pentylamine should
be considered in detail. The selectivity of dipentylamine, formed by the disproportionation
reaction of pentylamine, was 20% over NiMo/Al2O3 (Fig. 5.3), 40% over CoMo/Al2O3 (Fig.
5.4), and 80% over Mo/Al2O3 (Fig. 5.5) at τ = 1.6 g.min/mol, 320 °C, and 10 kPa H2S. Thus,
the higher conversion over Mo/Al2O3 is mainly due to the higher disproportionation of
pentylamine to dipentylamine. In section 5.4.3, we will suggest an explanation for the higher
activity of Mo/Al2O3.
Hexanethiol converted very fast at 320 °C (Table 5.1). The conversion was already 100%
at τ = 1.6 g.min/mol over both NiMo/Al2O3 and CoMo/Al2O3 at 10 kPa H2S. At 100 kPa H2S,
NiMo/Al2O3 performs slightly better than CoMo/Al2O3 in the HDS of hexanethiol, but it is
fair to say that both catalysts are about equally good in the HDN of alkylamines as well as in
the HDS of alkanethiols. The HDS of hexanethiol is definitely slower over Mo/Al2O3,
however, and decreased strongly with increasing H2S pressure from 10 to 100 kPa. We
ascribe the lower H2S conversion over Mo/Al2O3 to the much lower number of sulfur
vacancies on MoS2 than on Co or Ni-promoted MoS2 [19,20] and the decreased conversion
with increasing H2S pressure to the filling of these vacancies by H2S.
Pentanethiol was a primary product in the HDN of pentylamine over all three catalysts, as
the selectivity extrapolated to a non-zero value at time zero (Figs. 5.3-5.5). At 100 kPa H2S,
the selectivity of pentanethiol increased with decreasing weight time over NiMo/Al2O3,
HDN over Hydrotreating Catalysts Chapter 5 125
seemed to reach a maximum at short weight time over CoMo/Al2O3 (Fig. 5.4), and clearly
showed a maximum with decreasing weight time over Mo/Al2O3 (Fig. 5.5). This was checked
by repeating the experiments several times. Pentanethiol can be formed in two ways. One is
substitution of pentylamine with H2S
C5H11NH2 + H2S C5H11SH + NH3 (1)
and the other is substitution of dipentylamine with H2S
C5H11NH2 + C5H11NH2 C5H11NHC5H11 + NH3 (2)
C5H11NHC5H11 + H2S C5H11SH + C5H11NH2 (3)
Pentanethiol is a primary product in Eq. 1 and a secondary product in Eqs. 2 and 3. This
explains that the selectivity of pentanethiol is non-zero at τ = 0 (Eq. 1) and then increases with
increasing weight time at 100 kPa H2S (Eqs. 2 and 3). At higher weight time, the pentanethiol
selectivity decreases because of the reaction of the thiol to an alkene by elimination and to an
alkane by hydrogenolysis. The initial increase of the pentanethiol selectivity with increasing
weight time at 100 kPa H2S proves that dipentylamine, formed by disproportionation of
pentylamine, reacts fast with H2S to form pentanethiol and pentylamine over Mo/Al2O3. Over
NiMo/Al2O3, dipentylamine reacts indeed faster with H2S to form pentylamine and
pentanethiol than pentylamine reacts with H2S to pentanethiol [9]. For NiMo/Al2O3, the
decomposition of the formed pentanethiol is very fast, while for Mo/Al2O3 it is slower (Table
5.1). This explains that for NiMo/Al2O3 no maximum in the thiol selectivity is observed
(above τ = 1.6 g.min/mol, the lowest weight time that could be obtained), while for Mo/Al2O3
a clear maximum is observed. For CoMo/Al2O3, with an intermediate rate of thiol
decomposition (Table 5.1), there is an indication of a maximum (Fig. 5.4).
Pentene and pentane behaved as primary as well as secondary products over NiMo/Al2O3
and CoMo/Al2O3 (Figs. 5.3,5.4), while they behaved only as secondary products over
Mo/Al2O3. However, the pentenes/pentane ratio in the simultaneous reaction of pentylamine
and hexanethiol was very similar to the hexenes/hexane ratio originating from hexanethiol
over all three catalysts (Fig. 5.6). This demonstrates that the alkenes and alkane are
determined by the thiol, which is an intermediate between alkylamine and hydrocarbons.
HDN over Hydrotreating Catalysts Chapter 5 126
Thus, substitution is the predominant reaction in the HDN of pentylamine over all three
catalysts. That the pentenes and pentane behave as primary products is due to the very fast
decomposition of the intermediate pentanethiol (Table 5.1). Only when this decomposition is
slowed down, as over Mo/Al2O3 (Table 5.1), the formation of the pentenes and pentane is
clearly not primary. The fact that their selectivities seem to extrapolate to zero for τ > 0 (Fig.
5.5), suggests that the pentenes and pentane might even mainly be tertiary products (cf. curve
D in Fig. 5.1B with the pentenes and pentane selectivities in Fig. 5.5, both at low τ). They
would then be formed from dipentylamine (Scheme 5.3).
RNH2
H 2S
-NH 3
RSH
R= + H2S
RH + H2SH2
RNH2-NH
3 RNRH
H2S -RNH2
Scheme 5.3 Reaction network for the removal of nitrogen from alkylamines with the amine
group attached to a primary or secondary carbon atom.
5.4.2. 2-Hexylamine
The reaction rate of 2-hexylamine was faster than that of pentylamine over NiMo/Al2O3
and CoMo/Al2O3, because of the weaker C-N bond of the alkylamine with the NH2 group
attached to a secondary carbon atom [10]. The conversion of 2-hexylamine increases slightly
with increasing H2S pressure, while that of pentylamine decreases with increasing H2S
pressure. The influence of H2S is at least twofold. On the one hand, it has a positive influence
as a reaction partner in the nucleophilic substitution of the alkylamine to an alkanethiol. On
the other hand, it adsorbs on the catalyst surface and hinders reaction, as shown by the
negative influence on the disproportionation of the alkylamines to dialkylamines and on the
reaction of alkanethiols (Table 5.1). The much stronger effect of H2S on the Mo/Al2O3
HDN over Hydrotreating Catalysts Chapter 5 127
catalyst (Fig. 5.2), which shows the highest selectivity for dipentylamine (Fig. 5.5), suggests
that the inhibition of the disproportionation explains the negative influence of H2S on the
HDN of pentylamine. For 2-hexylamine also the increased acidity of the catalyst surface by
H2S adsorption and dissociation [12-14,26] may play a role. The ionic character of the C-N
bond for secondary carbon atoms may be increased by an increased proton concentration at
the catalyst surface.
Like for pentylamine, the selectivity to the dialkylamine (di-(2-hexyl)amine in this case) is
much higher over Mo/Al2O3 than over NiMo/Al2O3 and CoMo/Al2O3 (Figs. 5.8-5.10). With
increasing weight time, the selectivity of di-(2-hexyl)amine decreased much stronger over the
Ni(Co) promoted catalysts than over Mo/Al2O3. It shows that disproportionation easily takes
place over Mo/Al2O3, but that the dialkylamine disproportionation product reacts slower over
Mo/Al2O3 than over NiMo/Al2O3 and CoMo/Al2O3. The selectivity to hexane, the final
product in the HDN of 2-hexylamine, always increased with increasing weight time over all
three catalysts. On the other hand, the selectivity of the hexenes first increased with increasing
weight time and later decreased over NiMo/Al2O3 and CoMo/Al2O3. This must be due to the
high HDN conversion and thus decreased inhibition of the hydrogenation of the hexenes by
the amines at higher weight time. The hexenes/hexane ratio in the simultaneous reaction of 2-
hexylamine and 2-pentanethiol was very similar to the pentenes/pentane ratio originating from
2-pentanethiol over all three catalysts (Fig. 5.11). This demonstrates that the hexenes are
mainly produced from 2-hexanethiol, formed by substitution of H2S in the HDN of 2-
hexylamine over all three catalysts. The substitution of 2-hexylamine with H2S can be a direct
substitution (Eq. 1) or an indirect substitution after the disproportionation reaction of 2-
hexylamine (Eqs. 2 and 3). The indication of a maximum in the 2-hexanethiol selectivity at
short τ for Mo/Al2O3 (Fig. 5.10) suggests that at least for this catalyst both routes play a role.
Because of the secondary carbon atom in 2-hexylamine and 2-hexanethiol, these molecules
react faster than pentylamine and pentanethiol respectively. As a consequence, maxima due to
formation of intermediates shift to shorter τ for reactants with the secondary carbon atoms.
This explains why only an indication of a maximum is observed for 2-hexanethiol (Fig. 5.10)
and why the maxima in the alkenes are at shorter τ in the HDN of 2-hexylamine than in the
HDN of pentylamine (cf. Figs. 5.8-5.10 with Figs. 5.3-5.5).
HDN over Hydrotreating Catalysts Chapter 5 128
5.4.3 2-Methyl-2-butylamine
Pentylamine and 2-hexylamine did not show an appreciable conversion below 300 °C, but
the conversion of 2-methyl-2-butylamine was very high already at 270 °C. Hydrocarbons and
only a trace of thiol (not shown) were formed in the HDN of 2-methyl-2-butylamine, in
accordance with our former results for NiMo/Al2O3 [10]. Methylbutenes and methylbutane
were primary products and the methylbutenes/methylbutane branching ratio in the HDN of 2-
methyl-2-butylamine was about five times higher than that obtained in the HDS of 2-methyl-
2-butanethiol in the presence of hexylamine over NiMo/Al2O3, twenty times larger over
CoMo/Al2O3, and thirty times larger over Mo/Al2O3 at τ = 1.6 g.min/mol at 270 °C and 10
kPa H2S (Fig. 5.13). The branching ratio in the HDS of the thiol with the SH group attached
to a tertiary carbon atom (Fig. 5.13) is similar to that of thiols with the SH group attached to a
primary (Fig. 5.6) or secondary carbon atom (Fig. 5.11). The branching ratio in the HDN of
the amine group attached to a tertiary carbon atom is totally different, however. Therefore, 2-
methyl-2-butanethiol cannot be the intermediate in the HDN of 2-methyl-2-butylamine.
Also the very small amount of thiol formed shows that the amine with the NH2 group
attached to a tertiary carbon atom does not, or hardly, react by substitution with H2S to the
corresponding thiol. In that case a much higher concentration of 2-methyl-2-butanethiol
should be observed. In fact, the conversion of 2-methyl-2-butanethiol in the presence of an
amine is lower at 270 °C than that of hexanethiol and 2-pentanethiol at 320 °C (Table 5.1).
Nevertheless, the thiol selectivities in the HDN of pentylamine (Figs. 5.3-5.5) and 2-
hexylamine (Figs. 5.8-5.10) were high and that of 2-methyl-2-butylamine was very low. 2-
Methyl-2-butylamine thus neither reacts by substitution with H2S to a thiol, nor by
substitution with another amine molecule to a dialkylamine, since this disproportionation
product was not observed over any of the catalysts. This may be due to steric hindrance at the
α-carbon atom. Because the main primary products in the HDN of 2-methyl-2-butylamine are
alkenes, we conclude that this HDN mechanism takes place by elimination. The weak
influence of H2S and the tertiary carbon atom suggest that the elimination is of the E1 type
over all three catalysts, as discussed in detail previously for NiMo/Al2O3 [10].
2-Methyl-2-butylamine reacted fastest over Mo/Al2O3 and slowest over NiMo/Al2O3 (Fig.
5.12). This HDN reaction is most probably catalysed by acid sites. These may be Lewis acid
sites, consisting of a sulfur vacancy on a molybdenum, cobalt and nickel atom, or Brønsted
HDN over Hydrotreating Catalysts Chapter 5 129
acid sites constituted of H atoms on the sulfur atoms (protons of SH- groups). The activity
order would mean that the acidity of sulfided Mo/Al2O3 is higher than that of CoMo/Al2O3
than of NiMo/Al2O3. In an ionic model, this might be ascribed to the higher charge of Mo4+
than on Co2+ or Ni2+ (Lewis sites) or to the weaker bonding of H+ to S2- which is bonded to
Mo4+ than to Co2+ or Ni2+ (Brønsted site).
The conversion of 2-methyl-2-butylamine increased slightly with increasing H2S pressure
from 10 to 100 kPa. This cannot be explained by competition of H2S with the amine for the
adsorption sites on the catalyst surface. Since nucleophilic substitution hardly occurred for 2-
methyl-2-butylamine, this cannot explain the positive influence of H2S either. On the other
hand, H2S introduces protons at the catalyst surface and thus increases the acidity of the
catalyst. While it is impossible to turn a primary carbon atom into a carbenium ion, the
tertiary carbon atom of 2-methyl-2-butylamine can easily form a carbenium ion after
protonation by H2S and splitting off of NH3. With increasing H2S pressure, the conversion of
2-methyl-2-butylamine thus increases over all three catalysts. This explanation would mean
that Brønsted and not Lewis acid sites are responsible for the elimination of NH3 from the
amine with an NH2 group attached to a tertiary carbon atom. It explains why Mo/Al2O3, the
catalyst with the lowest number of sulfur vacancies, has the highest activity.
Mo/Al2O3 not only has the highest activity in the HDN of 2-methyl-2-butylamine, but also
in that of pentylamine (Fig. 2). The latter reactivity is based on two reactions, however, the
disproportionation to dipentylamine and the substitution to pentanethiol. Since the majority
product is dipentylamine, this suggests that also the disproportionation is acid catalysed and
that therefore the order of the three catalysts is the same for the HDN of 2-methyl-2-
butylamine and of pentylamine.
The conclusion that MoS2 has a stronger acidity than Co-MoS2 and Ni-MoS2 catalysts is
in accordance with DFT calculations of the heat of adsorption of NH3 on Lewis and Brønsted
acid sites on the surface of promoted and non-promoted M-MoS2 systems [20]. Travert et al.
found that the heat of adsorption on metal sites that were not fully coordinated by sulfur
atoms (CUS, Lewis sites) was stronger for Mo than for Co and Ni and decreased with
increasing sulfur coordination. The heat of adsorption on Brønsted SH groups was stronger
for SH groups attached to Mo atoms, as is to be expected since a stronger metal-sulfur bond
will induce a higher protonic character of the SH group. Investigation of Petit et al. showed
HDN over Hydrotreating Catalysts Chapter 5 130
that H2S adsorption leads to an increase in the number of Brønsted acid sites and a decrease in
the number of Lewis acid sites on sulfided Mo/Al2O3 and CoMo/Al2O3 [12].
5.5. Conclusion
The results in the HDN of alkylamines over CoMo/Al2O3 and Mo/Al2O3 are in agreement
with our former work over NiMo/Al2O3 [9,10]. The alkylamine with the amine group attached
to a primary or secondary carbon atoms reacts by substitution of the NH2 group by SH or an
amine to form an alkanethiol or a dialkylamine respectively over all three catalysts. The
alkanethiol forms an alkene and an alkane. Only tertiary alkylamines react directly to
hydrocarbons by an E1 mechanism over all three catalysts. An E2 mechanism hardly takes
place in the HDN of alkylamines over our catalysts. The fact that Mo/Al2O3, the catalyst with
the lowest number of sulfur vacancies, has the highest HDN activity and that H2S has a
positive influence on the HDN of the alkylamines with NH2 group attached to a tertiary
carbon atom suggests that alkylamines react on Brønsted acid sites. The HDS of alkanethiols
on the other hand need vacancies because Mo/Al2O3 has the lowest HDS activity and H2S has
a strong negative influence on all these alkanethiols, with the SH group attached to a tertiary
carbon atom.
The very low conversions of pentylamine over sulfided Zn/Al2O3 and Cd/Al2O3
demonstrate that just acid-base chemistry at a metal sulfide surface is not enough for the HDN
of alkylamines.
Acknowledgment
We thank Dr. M. Breysse for stimulating discussions.
HDN over Hydrotreating Catalysts Chapter 5 131
References
[1] T. Kabe, A. Ishihara, W. Qian, Hydrodesulfurization and Hydrodenitrogenation:
Chemistry and Engineering, Wiley-VCH, 1999.
[2] T.C. Ho, J. Catal. 219 (2003) 442.
[3] M. Egorova, R. Prins, J. Catal. 221 (2004) 11.
[4] N. Nelson, R.B. Levy, J. Catal. 58 (1979) 485.
[5] J.L. Portefaix, M. Cattenot, M. Guerriche, J. Thivolle-Cazat, M. Breysse, Catal. Today
10 (1991) 473.
[6] L. Vivier, V. Dominguez, G. Perot, S. Kasztelan, J. Mol. Catal. 67 (1991) 267.
[7] M. Cattenot, J.L. Portefaix, J. Afonso, M. Breysse, M. Lacroix, G. Perot, J. Catal. 173
(1998) 366.
[8] P. Clark, X. Wang, P. Deck, S.T. Oyama, J. Catal. 210 (2002) 116.
[9] Y. Zhao, P. Kukula, R. Prins, J. Catal. 221 (2004) 441; chapter 3.
[10] Y. Zhao, R. Prins, J. Catal. 222 (2004) 532; chapter 4.
[11] N.Y. Topsøe, H. Topsøe, J. Catal. 139 (1993) 641.
[12] C. Petit, F. Maugé, J.C. Lavalley, Stud. Surf. Sci. Catal. 106 (1997) 157.
[13] G. Berhault, M. Lacroix, M. Breysse, F. Maugé, J.C. Lavalley, H. Nie, L. Qu, J. Catal.
178 (1998) 555.
[14] A. Travert, F. Maugé, Stud. Surf. Sci. Catal. 127 (1999) 269.
[15] R. Prins, in: G. Ertl, H. Knözinger, J. Weitkamp (Eds.), Handbook of Heterogeneous
Catalysis, vol. 4, Wiley-VCH, New York, 1997, p. 1916.
[16] P. Raybaud, J. Hafner, G. Kresse, H. Toulhoat, Surf. Sci. 407 (1998) 237.
[17] L.S. Byskov, J.K. Nørskov, B.S. Clausen, and H. Topsøe, J. Catal. 187 (1999) 109.
[18] P. Raybaud, J. Hafner, G. Kresse, S. Kasztelan, H. Toulhoat, J. Catal. 189 (2000) 129.
[19] S. Cristol, J.F. Paul, E. Payen, D. Bougead, S. Clémendot, F. Hutschka, J. Phys.
Chem. B 104 (2000) 11220.
[20] A. Travert, H. Nakamura, R.A. van Santen, S. Cristol, J.F. Paul, E. Payen, J. Am.
Chem. Soc. 124 (2002) 7084
[21] V. Alexiev, R. Prins, T. Weber, Phys. Chem. Chem. Phys. 2 (2002) 1815.
[22] T. Todorova, V. Alexiev, R. Prins, T. Weber, Phys. Chem. Chem. Phys. 6 (2004)
3023.
HDN over Hydrotreating Catalysts Chapter 5 132
[23] M. Egorova, Y. Zhao, P. Kukula, R. Prins, J. Catal. 206 (2002) 263; chapter 2
[24] R. Prins, V.H.J. de Beer, G.A. Somorjai, Catal. Rev. Sci. Eng. 31 (1989) 1.
[25] H. Topsøe, B.S. Clausen, F.E. Massoth, Hydrotreating Catalysis in Catalysis, Science
and Technology, (J. Anderson, M. Boudart, Eds.) Springer Berlin, 11 (1996 )
[26] R.J. Girgis, B.C. Gates, Ind. Eng. Chem. Res. 30 (1991) 2021.
[27] J. Quartararo, S. Mignard, S. Kasztelan, J. Catal. 192 (2000) 307.
[28] J. Clayden, N. Greeves, S. Warren, P. Wothers, Organic Chemistry, Oxford Unv.
Press 2001 p. 483.
[29] M. Breysse, E. Furimsky, S. Kasztelan, M. Lacroix, G. Perot, Catal. Rev. 44 (2002)
651.
HDN of Adamantylamine Chapter 6 133
6 Mechanism of the hydrodenitrogenation of adamantylamine and
neopentylamine on sulfided NiMo/Al2O3
6.1 Abstract
The hydrodenitrogenation of 1-adamantylamine, 2-adamantylamine and neopentylamine
and the hydrodesulfurization of 1-adamantanethiol were studied over sulfided NiMo/Al2O3.
None of these amines can react by ammonia elimination and a classic SN2 substitution is not
possible for the adamantylamines either. The adamantanethiols and neopentanethiol were
the primary products of the adamantylamines and neopentylamine by substitution of the
NH2 group with H2S. These alkanethiols reacted by C-S hydrogenolysis to adamantane and
neopentane. It is proposed that the NH2-SH substitution in adamantylamine takes place by
adsorption of the amine group at the metal sulfide surface and migration of the adamantyl
group to a neighbouring sulfur atom. The fact that neopentane and adamantane were
secondary products demonstrates that hydrogenolysis of the aliphatic C-N bond of these
amines does not take place over sulfided NiMo/Al2O3 at 3 MPa and 270-340 °C.
6.2 Introduction
The cleavage of a carbon-sulfur (C-S) or a carbon-nitrogen (C-N) bond is the final step in
hydrodesulfurization (HDS) or hydrodenitrogenation (HDN) reactions. C-S bond breaking
may occur through elimination and homolytic C-S bond scission (hydrogenolysis), as
demonstrated in the HDS of alkanethiols [1-3] and aromatic thiols like (di)benzothiophene
[4,5]. Hydrogenolysis has also been observed in the homogeneous reaction of aliphatic and
aromatic thiols to hydrocarbons with the Cp’2Mo2Co2S3(CO)4 cluster [6].
HDN of Adamantylamine Chapter 6 134
The C-N bond in aromatic molecules like pyridine can only be broken after hydrogenation
of the aromatic heterocycle [7-9]. Even the removal of ammonia from aniline and alkylaniline
takes mainly place after hydrogenation of the (alkyl)aniline to (alkyl)cyclohexylamine and
less than 10% (alkyl)benzene was formed [8-10]. This hydrogenolysis of the aryl C-N bond is
only apparent, however [11]. In reality the breaking of the aryl C-N bond takes place by
partial hydrogenation of the arylamine. In the case of aniline it occurs by hydrogenation to
1,2-dihydroaniline and elimination of the NH2 group to form benzene. This NH2 elimination
is of the E1 type, because of the conjugation between the empty 2pz orbital on the carbon
atom and the butadiene structure in the resulting cyclohexadienyl C6H7+ carbocation [11].
Aliphatic C-N bond breaking has been claimed to occur exclusively by elimination [7,12]
and nucleophilic substitution by H2S followed by C-S bond hydrogenolysis [7,13]. Although
the C-N bond in aliphatic amines is weaker than in arylamines, alkanes were not observed as
primary products in the reaction of n-alkylamines [3]. The absence of hydrogenolysis might
be due to the fact that n-alkylamines can react faster by other reactions. To test that
possibility, we studied the removal of ammonia from 1-aminoadamantane, 2-
aminoadamantane and neopentylamine, which hardly have other possibilities to react than by
hydrogenolysis of the aliphatic C-N bond. 1-Adamantylamine has the amino group at a
carbon bridgehead and cannot, therefore, react via elimination or via SN2 substitution.
Furthermore, an SN1 reaction does not seem logical [14]; thus, 1-adamantylamine might react
via hydrogenolysis. 2-Adamantylamine cannot react by elimination either, and because SN1
and SN2 substitution do not seem likely, hydrogenolysis would be a good possibility to react.
The same holds for neopentylamine: The lack of a β-hydrogen atom makes elimination
impossible, the heavy substitution at the β-carbon atom makes SN2 substitution very difficult,
and SN1 substitution is unlikely at a primary carbon atom [14,15].
6.3 Results
6.3.1. Neopentylamine
The conversion of neopentylamine was very low at 300 °C and 3 MPa, about 4% at τ = 9
g.min/mol. It increased strongly when increasing the temperature to 340 °C and decreased
HDN of Adamantylamine Chapter 6 135
slightly when increasing the H2S pressure from 10 to 100 kPa (Fig. 6.1). Neopentylamine
reacted to 2,2-dimethylpropanethiol (neopentanethiol), di-neopentylamine and di-
neopentylimine, which behaved as primary products, and 2,2-dimethylpropane (neopentane),
which behaved as a secondary product (Fig. 6.2).
0 2 4 6 8 100
10
20
30
40
300 oC
340 oC, 100 kPa H2S
340 oCC
onve
rsio
n, %
Weight time, g.min/ mol
Fig. 6.1 Conversion of neopentylamine at 3 MPa and 300 or 340 °C, and 10 or 100 kPa H2S.
0 2 4 6 8 100
20
40
60
80
100
2,2-
dim
ethy
lpro
pane
, %
Weight time, g.min/ mol0 2 4 6 8 10
0
10
20
30
40
50
2,2-
dim
ethy
lpro
pane
thio
l, %
Weight time, g.min/ mol
0 2 4 6 8 100
10
20
30
40
50
di-im
ine,
%
Weight time, g.min/ mol0 2 4 6 8 10
0
10
20
30
40
di-a
min
e, %
Weight time, g.min/ mol
Fig. 6.2 Product selectivities in the HDN of neopentylamine at 3 MPa and 300 ( ) or 340 °C
( and ), and 10 or 100 kPa H2S (open symbols).
HDN of Adamantylamine Chapter 6 136
6.3.2. Adamantylamines and adamantanethiol
The conversion of 1-adamantylamine (1-AdNH2) at 300 °C and 3 MPa in the presence of
10 kPa H2S was lower than that of 2-adamantylamine (2-AdNH2). Increasing the H2S pressure
from 10 to 100 kPa had a positive effect on the conversion of 1-AdNH2 and a very strong
positive effect on that of 2-AdNH2 (Fig. 6.3). Only two products were observed in the HDN
of 1-AdNH2, adamantane and 1-adamantanethiol (1-AdSH) (Fig. 6.4); no products of
isomerization, like 2-AdNH2 and 2-AdSH, were observed. The sum of the adamantane and 1-
adamantanethiol selectivities thus was always 100%. The same holds true for adamantane and
2-AdSH in the HDN of 2-AdNH2 at 100 kPa H2S. Therefore only the adamantanethiol
selectivities of 1-AdSH at 10 and 100 kPa H2S and of 2-AdSH at 100 kPa H2S are presented
in Figure 6.4. They show that the adamantanethiols are the primary products of the HDN of
the adamantylamines, which reacted further to adamantane. At 10 kPa H2S, 2-AdNH2 not only
reacted to 2-AdSH and adamantane, but also to di(2-adamantylamine) and di(2-
adamantylimine) (Fig. 6.5).
0 2 4 6 8 10 120
20
40
60
80
100
2-AdNH2, 10 kPa H2S
1-AdNH2, 10 kPa H2S
1-AdNH2, 100 kPa H2S
2-AdNH2, 100 kPa H2S
Con
vers
ion,
%
Weight time, g.min/mol
Fig. 6.3 Conversion of 1-aminoadamantane (1-AdNH2) and 2-aminoadamantane (2-AdNH2)
at 300 °C, 3 MPa and 10 and 100 kPa H2S.
HDN of Adamantylamine Chapter 6 137
1-AdSH reacted much faster than 1-AdNH2 and 2-AdNH2 in the presence of 10 kPa H2S,
but in the presence of 100 kPa H2S it reacted just a bit faster than 1-AdNH2 (Fig. 6.6). In both
cases, adamantane was the only product. In both cases, hexylamine was added to the reactants
to simulate the presence of an alkylamine during the HDS of the alkanethiol.
0 2 4 6 8 100
20
40
60
80
100
1-AdSH, 10 kPa H2S
2-AdSH, 100 kPa H2S
1-AdSH, 100 kPa H2S
Sel
ectiv
ity, %
Weight time, g.min/mol
Fig. 6.4 Selectivity of 1-adamantanethiol (1-AdSH) in the HDN of 1-aminoadamantane at 300
°C, 3 MPa and 10 and 100 kPa H2S and selectivity of 2-adamantanethiol (2-AdSH)
in the HDN of 2-aminoadamantane at 300 °C and 100 kPa H2S.
0 2 4 6 8 100
20
40
60
80
di-aminedi-imine
2-AdSH
Ad
Sele
ctiv
ity, %
Weight time, g.min/mol
Fig. 6.5 Product selectivities in the HDN of 2-adamantylamine at 300 °C, 3 MPa and 10 kPa
H2S.
HDN of Adamantylamine Chapter 6 138
0 2 4 6 8 100
20
40
60
80
100
100 kPa H2S
10 kPa H2S
AdSH
con
vers
ion
%
Weight time, g.min/mol
Fig. 6.6 Conversion of 1 kPa 1-adamantanethiol in the presence of 5 kPa hexylamine at 300
°C, 3 MPa and 10 and 100 kPa H2S.
6.4 Discussion
6.4.1 HDN of neopentylamine
Neopentylamine has no β-H atoms and Hofmann elimination can therefore not take place
[12]. In an SN1 reaction a primary carbenium ion has to be formed. This is highly unlikely, as
demonstrated by the fact that no rearrangement products such as methylbutane or
methylbutene were observed in the HDN of neopentylamine. Thus, neopentylamine can only
react by SN2 substitution or hydrogenolysis.
Portefaix et al. already reported that neopentylamine hardly reacted at 270 °C [12]. Our
results show that even at 300 °C neopentylamine reacted very slowly and an appreciable
conversion (16% at τ = 3 g.min/mol) was only obtained at 340 °C (Fig. 6.1). The main
products were neopentanethiol (2,2-dimethylpropanethiol), dineopentylamine and
dineopentylimine. The selectivity of neopentane decreased strongly with decreasing weight
HDN of Adamantylamine Chapter 6 139
time and was zero at τ = 0 at 300 °C. At 340 °C the selectivity extrapolated to zero at τ = 0 as
well (Fig. 6.2). This demonstrates that the direct hydrogenolysis reaction of neopentylamine
to neopentane does not occur at 300 and 340 °C.
The behaviour of the neopentanethiol selectivity as a function of weight time at 300 °C
suggests that this molecule is formed by a primary as well as by a secondary reaction, because
the selectivity is non zero at τ = 0 and increases with τ. The primary reaction may be
2 2RNH H S RSH NH+ → + 3 (1)
and the secondary reaction
2 2RNHR H S RSH RNH+ → + (2)
The secondary reaction would be in agreement with the high initial selectivity of
dineopentylamine and dineopentylimine at 300 °C and with the about 20 times higher
reactivity of dihexylamine than hexylamine [3]. At 340 °C the rate of reaction (1) increases
and this would explain why at 340 °C and 100 kPa H2S neopentanethiol seems to be formed
only as a primary product (cf. the continuously decreasing selectivity with τ in Fig. 6.2).
The low reactivity of neopentylamine is easy to understand. Hofmann elimination is
impossible, SN1 substitution is unlikely and hydrogenolysis does not take place. The
occurrence of neopentanethiol, dineopentylamine and dineopentylimine as primary products
indicates that nucleophilic substitution of neopentylamine by H2S and by neopentylamine are
responsible for the formation of neopentanethiol and dineopentylamine respectively. The
secondary nature of neopentane suggests that the majority of neopentane is formed from
neopentanethiol by C-S hydrogenolysis. SN2 substitution of neopentylamine is hindered by
the tertiary butyl group at the α-carbon atom [14] and neopentylamine therefore reacts much
slower than the non-branched n-hexylamine [3]. The formation of dineopentylimine may be
explained by dehydrogenation of neopentylamine to neopentylimine, followed by addition of
neopentylamine and elimination of NH3 (Scheme 6.1). These reactions are well known in the
metal-catalyzed hydrogenation of nitriles [16] and amination of alcohols [17].
HDN of Adamantylamine Chapter 6 140
R CH2 NH2
H
N
HR R CH2 NH2
R CH2 NH C NH2
H
R
N C
RH2CR
H
+ NH3
Scheme 6.1 Formation of dialkylimine from an alkylamine.
The fact that C-N bond hydrogenolysis did not occur demonstrates how difficult
hydrogenolysis of an aliphatic C-N bond is over a sulfided NiMo/Al2O3 catalyst. C-N bond
hydrogenolysis on metal catalysts is an easy reaction [18], for instance on supported platinum
it is already fast around 150 °C [19]. Primary amines react much faster on Pt than secondary
amines, who react again about an order of magnitude faster than tertiary amines.
Hydrogenolysis of the C-N bond takes place after adsorption of both neighbouring N and C
atoms on the catalyst surface. Substitution on the α-C atom makes adsorption of this α-C
atom difficult, thus decreasing the rate of hydrogenolysis. This reactivity pattern on Pt is
completely opposite to that observed for the alkylamines on sulfided NiMo/Al2O3,
demonstrating that HDN on metal sulfides occurs by a different mechanism than on metals.
6.4.2 HDN of AdNH2 and HDS of AdSH
The sulfided NiMo/Al2O3 catalyst catalyzes the removal of the NH2 group from 1- and 2-
adamantylamine and of the SH group from 1-adamantanethiol. A Hofmann elimination is
impossible for these molecules because a bridgehead double bond as in adamantene is highly
strained and will therefore not form easily, as expressed in Bredt's rule [14]. Adamantene has
been synthesized by irradiation of diazo compounds, but its lifetime is very short [20]. 1-
Adamantyl compounds cannot react by a standard SN2 reaction either, because the
nucleophile cannot approach the reaction center through the adamantane cage, from the
HDN of Adamantylamine Chapter 6 141
backside, as required for a classic organic SN2 reaction [14,15]. Since adamantane was a
secondary product, hydrogenolysis cannot explain the HDN of the adamantylamines either.
Also an SN1 reaction is unlikely, since the H2S pressure has a strong positive influence on the
conversion of both adamantylamines. This leaves a non-classic SN2 reaction as the only
mechanism to explain the HDN of 1-adamantylamine. One could envisage that the NH2-SH
substitution occurs by alkyl migration on the metal sulfide surface. In that case, the amine
would adsorb at the vacancy on a Mo or Ni atom and then the alkyl group would migrate to a
sulfur atom on the neighbouring metal atom (Scheme 6.2).
Mo
S
S S
S
Mo
S
S
S N
H
H HR
Mo
S
S S
S
Mo
S
S
S N
H
H HR
Mo
S
S S
S
Mo
S
S
S N
H HHR
Scheme 6.2 NH2-SH substitution by alkyl migration on the metal sulfide surface.
2-Adamantylamine reacted faster than 1-adamantylamine, leading to a high conversion
already at small weight time (Fig. 6.3). As a consequence, 2-adamantanethiol reacted faster to
adamantane (Fig. 6.4), because of less inhibition by the amine. At 340 °C 2-adamantylamine
reacted to 2-adamantanethiol and adamantane only, and 2-adamantanethiol was a primary and
adamantane a secondary product. At 300 °C the selectivity pattern of 2-adamantylamine was
more like that of neopentylamine at 340 °C. Diadamantylamine and diadamantylimine were
primary products and adamantane was a secondary product. The selectivity of 2-
adamantanethiol was unequal to zero at τ = 0 and initially increased with weight time. Thus,
2-adamantanethiol is a primary product at 300 °C, formed directly from a nucleophilic
substitution of the adamantylamine by H2S (Eq. 1), as well as a secondary product formed
from di(2-adamantylamine) and di(2-adamantylimine) (Eq. 2). Figure 6.5 shows that, at 300
°C, the latter two molecules are initially formed faster than the thiol. At 340 °C this situation
is reversed, or the di(2-adamantylamine) and di(2-adamantylimine) react very fast with H2S to
2-adamantanethiol.
While 1-adamantylamine could, in principle, react with H2S by a classic SN1 reaction but
not by an SN2 reaction [21,22], 2-adamantylamine and di(2-adamantylamine) can do neither
HDN of Adamantylamine Chapter 6 142
reaction. An SN1 reaction would involve the secondary 2-adamantyl carbenium ion. This
mechanism is ruled out, because the H2S pressure has a strong positive influence on the
conversion of 2-adamantylamine (Fig. 6.3). An SN2 reaction would mean that the SH- or H2S
nucleophile has to react from the inside of a cyclohexane ring on the adamantane surface
(Scheme 6.3) [23]. This reaction is equivalent to that of SH- or H2S with cyclohexylamine
with the amine group in equatorial position, which is known to be impossible; a substituent
can only react when in axial position [14]. We therefore conclude that also the reaction of 2-
adamantylamine to 2-adamantanethiol occurs most likely by a non-classic substitution. In
addition to the possibility suggested above for the reaction of 1-adamantylamine, alkyl
migration on the metal sulfide surface (Scheme 6.2), 2-adamantylamine can undergo
substitution of the amine group by a sequence of dehydrogenation, addition, elimination and
hydrogenation reactions (Scheme 6.4).
HH
H
H2N H
GSH
HH
H
NH2
H
GSH
Scheme 6.3 SN2 reaction of 2-adamantylamine.
C NH H
C NHS H
C SH H
C N
C S
-H2 H2S
-RNH2
H2
Scheme 6.4 NH2-SH substitution by dehydrogenation of an amine to an imine, addition of
H2S, elimination of alkylamine, and hydrogenation of the thioketone.
HDN of Adamantylamine Chapter 6 143
Unlike 1-adamantylamine, 2-adamantylamine can adsorb with its C-N bond parallel to the
metal sulfide surface and become dehydrogenated to 2-adamantylimine. Addition of H2S and
elimination of NH3 leads to 2-thioadamantanone and further hydrogenation gives 2-
adamantanethiol. Similarly, addition of 2-adamantylamine to 2-adamantylimine and
elimination of NH3 gives di(2-adamantylimine) (cf. Scheme 1). After addition of H2S and
elimination of 2-adamantylamine one obtains 2-thioadamantanone. This sequence of
dehydrogenation, addition, elimination and hydrogenation reactions is very well known in
metal catalysis and used in the synthesis of amines from alcohols [17]. The high initial
selectivities of di(2-adamantylimine) and di(2-adamantylamine) suggest that they may indeed
be intermediates in the substitution reaction of 2-adamantylamine by H2S.
The hydrogenolysis of the C-S bond may occur as observed by Curtis and Druker for the
reactions of alkyl and aryl thiols in the homogeneous reaction with a Mo2Co2 cluster [6].
After adsorption of the neopentane- or adamantanethiol with the sulfur atom on a Mo or Ni
atom at the metal sulfide surface, the neopentyl and adamantyl groups may move to a
neighbouring metal atom. The alkyl intermediate can then be hydrogenated to an alkane.
Alternatively, alkanethiol and SH groups adsorbed on neighboring metal atoms may react to
an alkane and two adsorbed sulfur atoms.
Only 2-adamantanethiol, and no 1-adamantanethiol or 1-adamantylamine, was observed in
the reaction of 2-adamantylamine. Neither were 2-adamantanethiol and 2-adamantylamine
observed in the reaction of 1-adamantylamine. That no isomerization of 1-adamantyl to 2-
adamantyl compounds, nor the reverse reaction, was observed is due to the fact that the
isomerization of the 1-adamantyl to the 2-adamantyl carbenium ion and back is extremely
difficult. The reason is that the 2pz orbitals on the C1 and C2 carbon atoms must be parallel in
the transition state. Because of the rigid cage structure of adamantane, this is impossible [24].
6.5 Conclusion
Neopentylamine reacts to dineopentylamine, dineopentylimine, neopentanethiol and
neopentane at the high temperature of 340 °C. The formation of dineopentylimine can be
explained by dehydrogenation of neopentylamine to neopentylimine. The NH2 group from
HDN of Adamantylamine Chapter 6 144
another neopentylamine molecule can add to the neopentylimine to form an intermediate.
After elimination of ammonia, dineopentylimine will be formed, which after hydrogenation
leads to dineopentylamine. Substitution of neopentylamine with H2S leads to neopentanethiol.
This thiol could also be formed by dehydrogenation of a neopentylamine to an imine, addition
of H2S, elimination of ammonia, and hydrogenation of the thioketone. Neopentane was a
secondary product in the HDN of neopentylamine. This demonstrates that neopentane was
formed by decomposition of neopentanethiol formed by substitution of neopentylamine with
H2S. The direct hydrogenolysis of neopentylamine to neopentane hardly took place.
1-Adamantylamine reacts only to 1-adamanethiol and adamantane, while 2-
adamantylamine reacts to di-2-adamantylamine, di-2-adamantylimine, 2-adamanethiol and
adamantane. No isomerization took place between 1-adamantyl and 2-adamantyl compounds.
The very low reactivity of 1-adamantylamine at 300 °C shows that a carbenium ion cannot be
formed. Therefore, SN1 or E1 reactions cannot take place in the HDN of 1-adamantylamine.
H2S cannot attack the α carbon from the back through the adamantane cage, which shows that
a SN2 reaction cannot take place either. An E2 mechanism is unlikely, as no double bond
cannot be formed on a bridge-head carbon atom. Our explanation of the formation of 1-
adamantanethiol in the HDN of 1-adamantylamine is that the alkyl group shifts to a
neighboring sulfur atom.
6.6 References
[1] B.C. Wiegand, C.M. Friend, P. Uvdal, M.E. Napier, Surf. Sci. 355 (1996) 311.
[2] M.J. Girgis, B.C. Gates, Ind. Eng. Chem. Res. 30 (1991) 2021.
[3] Y. Zhao, P. Kukula, R. Prins, J. Catal. 221 (2004) 441; and chapter 3.
[4] M. Houalla, N.K. Nag, A.V. Sapre, D.H. Broderick, B.C. Gates, AIChE J. 24 (1978)
1015.
[5] J. Mijoin, G. Pérot, F. Bataille, J.L. Lemberton, M. Breysse, S. Kasztelan, Catal. Lett.
71 (2001) 139.
HDN of Adamantylamine Chapter 6 145
[6] M.D. Curtis, S.H. Druker, J. Am. Chem. Soc. 119 (1997) 1027.
[7] N. Nelson, R.B. Levy, J. Catal. 58 (1979) 485.
[8] H. Schulz, M. Schon, N.M. Rahman, Stud. Surf. Sci. Catal. 27 (1986) 201.
[9] C. Moreau, C. Aubert, R. Durand, N. Zmimita, P. Geneste, Catal. Today 4 (1988) 117.
[10] M. Jian, F. Kapteijn, R. Prins, J. Catal. 168 (1997) 491.
[11] Y. Zhao, J. Czyzniewska, R. Prins, Catal. Lett. 88 (2003) 155; and chapter 7.
[12] J.L. Portefaix, M. Cattenot, M. Guerriche, J. Thivolle-Cazat, M. Breysse, Catal. Today
10 (1991) 473.
[13] L. Vivier, V. Dominguez, G. Perot, S. Kasztelan, J. Mol. Catal. 67 (1991) 267.
[14] M.B. Smith, J. March, Advanced Organic Chemistry (Wiley, New York, 5th Ed.,
2001).
[15] J. Clayden, N. Greeves, S. Warren, P. Wothers, Organic Chemistry. Oxford Univ.
Press, 2001.
[16] C. De Bellefon, P. Fouilloux, Catal. Rev. Sci. Eng. 36 (1994) 459.
[17] T. Mallat, A. Baiker, in “ Handbook of Heterogeneous Catalysis” (G. Ertl, H.
Knözinger, J. Weitkamp, Eds.). Vol 5, P. 2334, Wiley-VCH.
[18] G. Meitzner, W.J. Mykytka and J.H. Sinfelt, J. Catal. 98 (1986) 513.
[19] Triyono, R. Kramer, Appl. Catal. A 100 (1993) 145.
[20] E.L. Tae, Z. Zhu, M.S. Platz, J. Phys. Chem. A 105 (2001) 3803.
[21] G.A. Olah, Cage Hydrocarbons (Wiley, New York, 1990).
[22] P. von R. Schleyer, R.D. Nicholas, J.Am. Chem.Soc. 83 (1961) 2700.
[23] J.L. Fry, C.J. Lancelot, L.K.M. Lam, J.M. Harris, R.C. Bingham, D.J. Raber, R.E.
Hall, P. von R. Schleyer, J. Am. Chem. Soc. 92 (1970) 2538.
[24] D.M. Brouwer, in ‘Chemistry and Chemical Engineering of Catalytic Processes’ (R.
Prins and G.C.A. Schuit, Eds.), Sythof-Noordhof, Alphen, 1980, p. 137.
HDN of Naphthylamine Chapter 7 147
7. Mechanism of the direct hydrodenitrogenation of naphthylamine on
sulfided NiMo/Al2O3
7.1 Abstract
The hydrodenitrogenation of 1-naphthylamine was studied over a sulfided NiMo/Al2O3
catalyst between 300 and 350 ºC. 1-Naphthylamine reacted to tetralin, naphthalene, 1,2-
dihydronaphthalene and 5,6,7,8-tetrahydro-1-naphthylamine. To elucidate the reaction
mechanism, the reactions of the intermediates 1,2,3,4-tetrahydro-1-naphthylamine, 1,2-
dihydronaphthalene and 5,6,7,8-tetrahydro-1-naphthylamine were studied as well. The results
show that 1-naphthylamine reacts through hydrogenation to 1,2,3,4-tetrahydro-1-
naphthylamine, which reacts by NH3 elimination to 1,2-dihydronaphthalene. The latter
molecule subsequently reacts by hydrogenation to tetralin as well as by dehydrogenation to
naphthalene. In addition, naphthalene is formed by direct denitrogenation from 1-
naphthylamine. This direct denitrogenation may take place by hydrogenation of 1-
naphthylamine to 1,2-dihydro-1-naphthylamine, followed by NH3 elimination or followed by
a Bucherer-type NH2-SH exchange, dehydrogenation and C-S bond hydrogenolysis.
7.2 Introduction
The cleavage of a carbon-nitrogen (C-N) or a carbon-sulfur (C-S) bond is a crucial step in
hydrodenitrogenation (HDN) and hydrodesulfurization (HDS) reactions respectively.
Hydrogenation of the aromatic ring that contains the S atom does not seem to be required for
the removal of the S atom, because thiophenol reacts almost exclusively to benzene under
HDS conditions [1], while about 80% of dibenzothiophene reacts to biphenyl [2]. This
HDN of Naphthylamine Chapter 7 148
suggests that the relatively weak C-S bond can be broken by hydrogenolysis. Hydrogenolysis
is used here in the mechanistic sense, meaning a reaction on the catalyst surface in which a C-
X bond is broken and C-H and H-X bonds are formed before the product molecule leaves the
catalyst surface. Homolytic C-S bond breaking (hydrogenolysis) was demonstrated in the
homogeneous reaction of aliphatic and aryl thiols on sulfur–containing Mo-Co clusters [3].
Breaking of the C-S bond might occur by nucleophilic aromatic substitution by a hydride ion
as well [4]. It has also been suggested, however, that the hydrogenolysis of the C-S bond (also
called direct desulfurization) is only apparent and actually occurs by hydrogenation of a
neighbouring C-C bond, followed by H2S elimination [2,5].
The main reactions involved in the removal of an N atom from aromatic compounds are
hydrogenation of the aromatic ring which contains the nitrogen atom, and breaking of the
resulting aliphatic C-N bonds to a hydrocarbon molecule and ammonia [1,6-9]. Aliphatic C-N
bond breaking occurs either by elimination [6,7], or by nucleophilic substitution by H2S
followed by C-S bond hydrogenolysis [6,8]. The main HDN product of aniline is therefore
cyclohexane, which is formed via cyclohexylamine and cyclohexene [1,8]. Similarly, the
main product in the HDN of quinoline is propylcyclohexane [9].
Direct breaking of the C-N bond in aniline (also called direct denitrogenation) occurs to a
minor degree as well. For instance, in the HDN of o-propylaniline the selectivity to
propylbenzene was 7% over a NiMo/Al2O3 catalyst and 24% over a Mo/Al2O3 catalyst [10].
For fused aromatic amines like naphthylamine and anthracylamine direct C-N bond breakage
is even more important [11]. If this direct C-N bond breakage in an arylamine occurs by real
hydrogenolysis in the mechanistic sense, then one should expect that hydrogenolysis of an
alkylamine is even easier. The reason for this is that the C-N bond in alkylamines is weaker
than the C-N bond in arylamines because of the conjugation of the NH2 group with the
aromatic ring. Nevertheless, alkylamines seem to react exclusively by β hydrogen elimination
and nucleophilic substitution by H2S followed by C-S bond hydrogenolysis [6-9]. This
suggests that hydrogenolysis in arylamines may not be real but apparent, meaning that the
reaction occurs via an indirect, multi-step mechanism.
To determine whether the hydrogenolysis of an aryl C-N bond is real or apparent, we
studied the HDN of 1-naphthylamine and the reactions of the possible intermediates 1,2,3,4-
tetrahydro-1-naphthylamine, 1,2-dihydronaphthalene and 5,6,7,8-tetrahydro-1-naphthylamine.
HDN of Naphthylamine Chapter 7 149
7.3 Results
Both catalysts were active in the HDN of 1-naphthylamine in the autoclave and the
conversion after 1 h at 300 °C was 22% over NiMo/Al2O3 and 33% over CoMo/Al2O3. 1-
Naphthylamine reacted to tetrahydronaphthalene (tetralin), 1,2-dihydronaphthalene,
naphthalene and a small amount of 5,6,7,8-tetrahydro-1-naphthylamine. The selectivity to 1,2-
dihydronaphthalene decreased with reaction time and no 1,2-dihydronaphthalene was
observed after 30 min over CoMo/Al2O3, while over NiMo/Al2O3 the selectivity to 1,2-
dihydronaphthalene decreased more slowly (Fig. 7.1). At short reaction time, the selectivities
to naphthalene and 1,2-dihydronaphthalene were high and increased with decreasing reaction
time, while the reverse was true for tetralin (Fig. 7.2). This indicates that naphthalene and 1,2-
dihydronaphthalene are formed earlier and tetralin later in the reaction network. At the higher
temperature of 350 °C, the conversion was 70% after 1 h and more naphthalene and less
tetralin were formed than at 300 °C. 1,2-Dihydronaphthalene was only produced at short
reaction time. The conversions in the experiments in the microflow reactor over NiMo/Al2O3
were always much higher than in the autoclave; the main reason being the five times higher
H2 pressure. Conversions and product selectivities are presented in Figures 7.3 and 7.4
respectively. At 300 °C and 3 MPa, in the presence of 10 kPa H2S, tetralin, 5,6,7,8-tetrahydro-
1-naphthylamine, naphthalene and 1,2-dihydronaphthalene behaved like primary products
with non-zero selectivities at zero weight time. The selectivity to 1,2-dihydronaphthalene was
very sensitive to the temperature and H2 pressure. It decreased with increasing H2 pressure
and increasing temperature (Table 7.1). At low weight time the selectivity to 5,6,7,8-
tetrahydro-1-naphthylamine was 30%, but it decreased to zero at high conversion (Figure 7.4).
The naphthalene selectivity was constant at about 10%. Since the tetralin selectivity increased
with weight time, the naphthalene to tetralin ratio decreased with weight time (Figure 7.5).
This ratio was more sensitive to temperature than to H2 pressure (Table 7.1).
HDN of Naphthylamine Chapter 7 150
15 30 45 600
20
40
60
80
100
1,2-DHN
5,6,7,8-THNA
naphthalene
tetralinS
elec
tivity
, %
Time, min
Fig 7.1 Selectivities in the HDN of 1-naphthylamine over NiMo/Al2O3 (closed symbols) and
CoMo/Al2O3 (open symbols) catalysts at 300 °C and 0.6 MPa in the autoclave (1,2-
DHA = 1,2-dihydronaphthalene, 5,6,7,8-THNA = 5,6,7,8-tetrahydro-1-
naphthylamine).
3 4 5 6 7 8 90
15
30
45
60
naphthalene
tetralin
1,2-DHN
B
Sele
ctiv
ity, %
Time, min3 4 5 6 7 8 9
0
15
30
45
60
5,6,7,8-THNA
1,2-DHN
naphthalene
tetralin
A
Sele
ctiv
ity, %
Time, min
Fig. 7.2 Selectivities in the HDN of 1-naphthylamine over CoMo/Al2O3 (A) and NiMo/Al2O3
(B) catalysts at 300 °C and 0.6 MPa at short reaction times in the autoclave (1,2-
DHN = 1,2-dihydronaphthalene, 5,6,7,8-THNA = 5,6,7,8-tetrahydro-1-
naphthylamine).
HDN of Naphthylamine Chapter 7 151
0 1 2 3 4 50
20
40
60
80
100
300°C
320°C
350°C
Con
vers
ion,
%
Weight time, g.min/mol
Fig. 7.3 Conversion of 1-naphthylamine over NiMo/Al2O3 in the microflow reactor at 3 MPa,
10 kPa H2S and 300, 320 and 350 °C.
0 1 2 3 4 50
20
40
60
80
100
1,2-DHN
naphthalene
5,6,7,8-THNA
Tetralin
Sel
ectiv
ity, %
Weight time, g.min/mol
Fig. 7.4 Product selectivities in the HDN of 1-naphthylamine over NiMo/Al2O3 at 300 °C, 3
MPa and 10 kPa H2S in the microflow reactor (1,2-DHN = 1,2-dihydronaphthalene,
5,6,7,8-THNA = 5,6,7,8-tetrahydro-1-naphthylamine).
HDN of Naphthylamine Chapter 7 152
0 1 2 3 4 50.0
0.1
0.2
0.3
0.4
300°C
320°C
350°C
naph
thal
ene/
tetr
alin
Weight time, g.min/mol
Fig. 7.5 Naphthalene to tetralin ratio in the HDN of 1-naphthylamine over NiMo/Al2O3 at 3
MPa, 10 kPa H2S and 300, 320 and 350 °C.
Table 7.1 Selectivity to 1,2-dihydronaphthalene (DHN) in the reaction of 1-naphthylamine
(NA), and the naphthalene to tetralin ratio (N/T) in the reactions of NA, 1,2,3,4-tetrahydro-1-
naphthylamine (THAN) and DHN at 10 kPa H2S and τ = 1.07 g.min/mol.
Conditions % DHN N/T
T (°C) P (MPa) NA NA THAN DHN
300 1 13 0.22
300 2 6 0.20
300 3 3 0.20 0.10 0.10
320 3 0 0.26
350 3 0 0.37 0.19 0.22
To compare the reaction rates of potential intermediates in the HDN of 1-naphthylamine,
we also measured the HDN of 1,2,3,4-tetrahydro-1-naphthylamine and 5,6,7,8-tetrahydro-1-
naphthylamine and the hydrogenation of 1,2-dihydronaphthalene, the latter in the presence of
aniline to simulate the inhibiting effect of an arylamine. At 300 °C, 3 MPa and 10 kPa H2S,
HDN of Naphthylamine Chapter 7 153
the conversions of 1,2,3,4-tetrahydro-1-naphthylamine and 1,2-dihydronaphthalene were
already complete at the lowest weight time possible in our microflow reactor (τ = 1.07
g.min/mol). These conversions of 100% were much higher than that of 1-naphthylamine
(20%). The only products were tetralin and naphthalene. Figure 7.6 shows the naphthalene to
tetralin ratio as a function of time for 1,2,3,4-tetrahydro-1-naphthylamine at 300 and 350 °C
and at 10 kPa H2S; almost identical curves were obtained for 1,2-dihydronaphthalene.
Because of the complete conversion of 1,2,3,4-tetrahydro-1-naphthylamine and 1,2-
dihydronaphthalene already at short weight time, the naphthalene to tetralin ratios decreased
with time because of the hydrogenation of naphthalene to tetralin and the fact that initially a
larger amount of naphthalene was produced than corresponding with thermodynamics. The
naphthalene to tetralin ratios, extrapolated to τ = 0 g.min/mol for the reactions of 1,2,3,4-
tetrahydro-1-naphthylamine and 1,2-dihydronaphthalene, were 0.24 and 0.27 respectively at
350 °C and 3 MPa, and the ratio was 0.11 for both reactions at 300 °C and 3 MPa. These
values are much lower than the values of 0.37 (350 °C, 3 MPa) and 0.21 (300 °C, 3 MPa)
obtained in the HDN of 1-naphthylamine itself (Fig. 7.5).
0 1 2 3 4 50.00
0.05
0.10
0.15
0.20
0.25
0.30
300°C
350°C
naph
thal
ene/
tetra
lin
Weight time, g.min/mol
Fig. 7.6 Naphthalene to tetralin ratio in the HDN of 1,2,3,4-tetrahydro-1-naphthylamine over
NiMo/Al2O3 at 3 MPa, 10 kPa H2S and 300 and 350 °C.
HDN of Naphthylamine Chapter 7 154
The conversion of 5,6,7,8-tetrahydro-1-naphthylamine at 300 °C, 3 MPa and 10 kPa H2S
was only 3% at τ = 1.07 g.min/mol. This indicates that its low selectivity in the HDN of 1-
naphthylamine is not due to a fast subsequent reaction, but to a relatively slow rate of
formation.
7.4 Discussion
7.4.1 Direct denitrogenation
Our results confirm that hydrogenolysis of arylamine C-N bonds is possible, since
naphthalene behaved like a primary product on sulfided NiMo/Al2O3 and CoMo/Al2O3
catalysts. The other main products of the reaction were 1,2-dihydronaphthalene, tetralin and
5,6,7,8-tetrahydronaphthylamine; they result from hydrogenation. The ratio of hydrogenolysis
to hydrogenation strongly depended on the reaction temperature: the higher the reaction
temperature, the higher the ratio. Similar behaviour was observed in the HDN of aniline over
a sulfided NiW/Al2O3 catalyst [13].
The experiments at lower pressure in the autoclave as well as in the microflow reactor
demonstrated that both naphthalene and 1,2-dihydronaphthalene behave as primary products,
with selectivities increasing at shorter reaction time. Whereas naphthalene can be envisaged to
be formed directly from 1-naphthylamine (e.g. by hydrogenolysis), the formation of 1,2-
dihydronaphthalene from 1-naphthylamine has to occur through at least one intermediate. A
logic intermediate would be 1,2,3,4-tetrahydro-1-naphthylamine. This intermediate is very
reactive under our conditions, as shown in the separate experiment in which 1,2,3,4-
tetrahydro-1-naphthylamine already completely reacted to tetraline and naphthalene at the
lowest weight time possible in our microflow reactor. 1,2-Dihydronaphthalene, the expected
primary product obtained by NH3 elimination of 1,2,3,4-tetrahydro-1-naphthylamine, was not
observed in this experiment, as it reacts very fast as well. The very high reactivity of 1,2,3,4-
tetrahydro-1-naphthylamine also explains why it was not observed in the HDN of 1-
naphthylamine even after a short reaction time. This is in accordance with results obtained in
HDN of Naphthylamine Chapter 7 155
the HDN of o-methylaniline, for which the first hydrogenation step was also much slower
than the subsequent nitrogen-removal step [14]. In that reaction, the hydrogenated
intermediate o-methylcyclohexylamine was only observed when a large amount of
cyclohexene was added during reaction, to cause the intermediate to leave the catalyst surface.
When an arylamine is hydrogenated at the catalyst surface, the intermediate cyclohexylamine
apparently undergoes ammonia elimination faster than that it desorbs from the surface and
diffuses out of the catalyst pores. Furthermore, in the case of 1,2,3,4-tetrahydro-1-
naphthylamine, the denitrogenation might be even very fast because it can take place by an E1
elimination mechanism. The reason is that the carbocation resulting from
tetrahydronaphthalene (Scheme 7.1) is strongly stabilized by conjugation with the aromatic
ring and by electron donation from the β CH2 group.
Because 1,2-dihydronaphthalene reacts fast but not extremely fast, it could be detected
and seen to behave as a primary product. Higher temperature and H2 pressure increase the rate
of the reaction of 1,2-dihydronaphthalene. This explains why at 3 MPa and 300 or 350 °C this
intermediate was not observed. The microflow experiment showed that 1,2-
dihydronaphthalene reacts to a 9:1 mixture of tetralin and naphthalene at short weight time.
Apparently, hydrogenation as well as dehydrogenation can take place quickly.
+
Scheme 7.1 Carbocation of tetrahydronaphthalene.
The much higher naphthalene to tetralin ratios observed in the HDN of 1-naphthylamine
than in the reaction of 1,2-dihydronaphthalene (Figs. 7.5 and 7.6 and Table 7.1) indicate that
1,2-dihydronaphthalene is not the only source of naphthalene. The different behaviour of
naphthalene and tetralin as a function of reaction time confirms this (Fig. 7.2): naphthalene
behaves as a primary product and tetralin as a secondary product. This means that additional
naphthalene must be formed by a reaction that takes place earlier in the reaction network than
the formation of 1,2-dihydronaphthalene from 1,2,3,4-tetrahydro-1-naphthylamine. 1,2-
HDN of Naphthylamine Chapter 7 156
Dihydro-1-naphthylamine or 1-naphthylamine could be intermediates for the formation of this
naphthalene.
Moreau et al. proposed that the naphthalene that forms in the HDN of 1-naphthylamine
could be partially hydrogenated to tetralin [11]. To check this, we performed a hydrogenation
of 1-methylnaphthalene in the presence of 1-naphthylamine over sulfided NiMo/Al2O3 and
CoMo/Al2O3 at 300 ºC in the autoclave, but we did not observe any products of the
hydrogenation of 1-methylnaphthalene. This shows that the naphthalene-to-tetralin step does
not take place during the HDN of 1-naphthylamine as long as the 1-naphthylamine
concentration is high enough to inhibit the hydrogenation of aromatic molecules. This
corroborates the results obtained in the reaction of ethylbenzene in the presence of o-
propylaniline, in which ethylbenzene hydrogenation was only 1% over a NiMo/Al2O3 catalyst
at 350 °C and at 60% conversion of o-propylaniline [12].
On the basis of our results we propose a mechanism for the HDN of 1-naphthylamine over
sulfided NiMo/Al2O3 and CoMo/Al2O3 catalysts (Scheme 7.2) that differs in some points
from that proposed by Moreau et al. [11]. There are two pathways, the main one being a
multi-step reaction pathway. First, 1-naphthylamine is partially hydrogenated to 1,2,3,4-
tetrahydro-1-naphthylamine. This intermediate eliminates NH3 and the resulting 1,2-
dihydronaphthalene reacts to tetralin and naphthalene.
NH2 NH2 NH2 NH2
SH
4231
Scheme 7.2 Reaction mechanism for the HDN of 1-naphthylamine.
HDN of Naphthylamine Chapter 7 157
In the second pathway, 1-naphthylamine undergoes direct breaking of the C-N bond to
naphthalene. The question is, if this reaction is really taking place, for instance by homolytic
splitting of the C-N bond and fast hydrogenation of the resulting radicals (like the C-S bond
breaking in thiols [3]) or by the substitution of the amine group by a hydride ion [4]. It is also
possible that the formed naphthalene gives the impression that 1-naphthylamine reacts by
hydrogenolysis, but that this is not the case. For instance, the transformation of 1-
naphthylamine to 1-naphthylthiol by NH2-SH exchange [9] and fast reaction of 1-
naphthylthiol to naphthalene by hydrogenolysis of the C-S bond would look like C-N
hydrogenolysis if 1-naphthylthiol were not observed (route 1 in scheme 7.2). The NH2-SH
exchange (Scheme 7.3) would certainly be enhanced in arylamines with fused aromatic rings
like naphthyl- and anthracylamine, because the aromaticity of fused rings decreases with
increasing number of rings. Thus, the 1,2-C-C bond in naphthalene has more double bond
character than a C-C bond in benzene and the enamine character of 1-naphthylamine is
stronger than that of aniline. The higher enamine contribution in turn means that also the
imine character is higher because of the enamine-imine tautomeric equilibrium. This is
analogous to the greater importance of the keto form in the enol-keto tautomeric equilibrium
for naphthol than for phenol [16]. As a result of the greater imine character, the addition of
H2S to 1-naphthylamine will be easier.
NH2 NH
SH
SH NH2
H2S -NH3
S
Scheme 7.3 Reaction from 1-naphthylamine to 1-thionaphthol by enamine-imine
tautomerism, NH2-SH exchange by addition of H2S and elimination of NH3,
and thioenol-thioketo tautomerism.
HDN of Naphthylamine Chapter 7 158
Also the partial hydrogenation of 1-naphthylamine to 1,2-dihydro-1-naphthylamine
followed by the fast elimination of NH3 and the formation of naphthalene, would give the
impression that hydrogenolysis had occurred (route 2 in scheme 7.2). The formation of a
dihydro compound seems feasible and has already been proposed as an explanation for the
apparent hydrogenolysis of dibenzothiophene [5]. At first sight, this explanation seems flawed
because, due to the planar structure of the cyclohexadiene molecule, the NH2 group on the C1
atom and the H atom on the neighbouring C2 atom are in the eclipsed conformation. This
would mean that the subsequent elimination (e.g. of 1,2-dihydro-aniline to benzene and
ammonia) must occur by syn-elimination although elimination tends to occur in the anti-
periplanar rather than in the syn-antiplanar conformation [16]. A closer look at 1,2-dihydro-
aniline suggests, however, that the elimination will not occur by an E2 mechanism, but by an
E1 mechanism. The reason is that the carbon atom that bears the NH2 group is in α position to
the C3-C6 butadiene fragment. As a consequence, the cyclohexadienyl carbocation resulting
from scission of the C-N bond will be strongly stabilized by conjugation with this butadiene
fragment (Scheme 7.4). An E1 elimination mechanism means, however, that the eclipsed
conformation of the NH2 group on the C1 atom and the H atom on the C2 atom in 1,2-
dihydro-aniline is not an obstacle anymore against elimination.
++
Scheme 7.4 Cyclohexadienyl carbocation.
Another explanation for the direct denitrogenation would be to assume that aniline is
hydrogenated to tetrahydroaniline, which undergoes elimination to cyclohexadiene.
Cyclohexadiene then quickly reacts to cyclohexene or benzene. Since tetrahydroaniline is not
flat, the elimination of ammonia is possible in the anti conformation. In our case of 1-
naphthylamine, this means that the naphthalene would be formed via 1,2,3,4-tetrahydro-1-
naphthylamine (route 4, scheme 7.2). This is, however, in contradiction to the naphthalene-to-
tetralin ratio observed in the reaction of 1,2,3,4-tetrahydro-1-naphthylamine, which is two
times lower than that observed in the reaction of 1-naphthylamine. Also the ratio in the
reaction of 1,2-dihydronaphthalene, the primary product of 1,2,3,4-tetrahydro-1-
HDN of Naphthylamine Chapter 7 159
naphthylamine is lower by a factor two. We conclude that a tetrahydro intermediate cannot
explain the direct denitrogenation of 1-naphthylamine to naphthalene.
1,2-Dihydro-1-naphthylamine may not only react to naphthalene by elimination of NH3,
but also by a Bucherer-like NH2-SH exchange of the amino group of 1,2-dihydro-1-
naphthylamine by addition of NH3 and elimination of H2S, via enamine-imine and thioenol-
thioketo tautomeric equilibria (scheme 7.5 and route 3 in scheme 7.2). The resulting 1,2-
dihydro-1-thionaphthol may dehydrogenate to 1-thionaphthol, which quickly undergoes
hydrogenolysis to naphthalene. A dihydro intermediate could thus explain direct
hydrogenation via a Bucherer-type NH2-SH exchange reaction (Scheme 7.4). Alternatively,
NH2-SH exchange could occur directly in the arylamine via the imine form of the enamine-
imine tautomeric equilibrium (Scheme 7.3).
NH2 NH2 NH
SH SH S
HS NH2
H2H2S
-H2
-NH3
Scheme 7.5 Reaction from 1-naphthylamine to 1-thionaphthol by hydrogenation to dihydro-
1-naphthylamine, followed by a Bucherer-like NH2-SH exchange and
dehydrogenation of the resulting dihydro-1-thionaphthol.
Three mechanisms have thus been identified which can explain the apparent direct C-N
bond breaking in 1-naphthylamine to naphthalene. In the first mechanism (Scheme 7.3, route
1 in scheme 7.2), NH2-SH exchange via the imine form of 1-naphthylamine is followed by
hydrogenolysis of the C-S bond of 1-thionaphthol, while in the second mechanism
hydrogenation of 1-naphthylamine to 1,2-dihydro-1-naphthylamine is followed by NH3
elimination (route 2 in scheme 7.2). A third mechanism would be that 1-naphthylamine is
hydrogenated to a dihydro intermediate which undergoes a Bucherer-type NH2-SH exchange
reaction (Scheme 7.5, route 3 in scheme 7.2). Some experimental observations speak in
HDN of Naphthylamine Chapter 7 160
favour of the second and third alternatives. It has been observed that the ratio of direct versus
indirect C-N bond breaking depends on the catalyst. Thus, a change from Al2O3 to silica-
alumina and fluorination of these supports did not change the toluene to methylcyclohexane
product ratio in the HDN of o-toluidine, although they did improve the activity [17]. This
suggests that direct and indirect C-N bond breaking go through a common intermediate,
which could be a dihydro intermediate. That would eliminate route 1 (Scheme 7.2), but would
still leave route 2 (Scheme 7.2) (hydrogenation followed by NH3 elimination) and route 3
(hydrogenation followed by a Bucherer-type NH2-SH exchange, dehydrogenation and C-S
hydrogenolysis) as possibilities to explain the direct C-N bond breaking. The ratio of direct to
indirect C-N bond breaking would then be determined by the ratio of the two reactions that
dihydro-1-naphthylamine can undergo: further hydrogenation to 1,2,3,4-tetrahydro-1-
naphthylamine or elimination of NH3 to naphthalene (cf. scheme 7.2).
A final comment is appropriate about the small amount of 5,6,7,8-tetrahydro-1-
naphthylamine observed in the HDN product of 1-naphthylamine. This product is due to
hydrogenation of the non-substituted aromatic ring. Since the reactivity of 5,6,7,8-tetrahydro-
1-naphthylamine itself is low, its low concentration in the HDN of 1-naphthylamine points to
a slow rate of formation. This may be due to a weaker adsorption of the benzene part than of
the aniline part of the naphthylamine. The behaviour of 1,2,3,4-tetrahydro-1-naphthylamine
and 5,6,7,8-tetrahydro-1-naphthylamine in the HDN of 1-naphthylamine would then be
similar to that of 1,2,3,4-tetrahydroquinoline and 5,6,7,8-tetrahydroquinoline respectively in
the HDN of quinoline [9]. The small amount of 5,6,7,8-tetrahydro-1-naphthylamine produced
and its low reactivity show that 5,6,7,8-tetrahydro-1-naphthylamine plays only a minor role in
the HDN of 1-naphthylamine under our conditions.
7.4.2 Direct desulfurization
Aliphatic and aromatic thiols undergo HDS with high selectivity to alkanes and aromatic
hydrocarbons, respectively. This suggests that in both cases hydrogenolysis of the C-S bond
actually takes place. However, if amines do not react by real hydrogenolysis, then the
question arises as to whether or not the direct C-S bond breaking in thiols takes place by real
or apparent hydrogenolysis. An alternative explanation for the arylthiol would be partial
HDN of Naphthylamine Chapter 7 161
hydrogenation followed by H2S elimination (e.g., thiophenol reacts to 1,2-dihydrothiophenol
and then to benzene and H2S), as suggested above for the reaction of arylamines to aromatic
molecules. This is not possible, however, for alkane thiols. Alkanes are the main product of
the HDS reaction of alkane thiols at lower temperatures; the only possible explanation is
actual hydrogenolysis. Homolytic C-S bond scission (hydrogenolysis) was demonstrated by
Curtis and Drucker in the homogeneous reaction of aliphatic and aromatic thiols with the
Cp*2Mo2Co2S3(CO)4 cluster (Cp* stands for pentamethylcyclopentadienyl) [3]. By
spectroscopic and kinetic measurements they showed that the thiols react as indicated in
scheme 7.6, in which only the bare structure of the complexes is indicated. The µ3-mode of
coordination of the RS thiolate leads to the activation of the C-S bond for homolytic cleavage
by decreasing the C-S bond dissociation energy. The soft character of sulfur apparently
enables a relatively strong interaction with the soft, low-valent Mo atoms in the metal sulfide;
this promotes C-S bond homolysis, as observed in thiolate complexes [18]. The interaction of
the nitrogen atom in an amine with such Mo atoms does, on the other hand, not lead to C-N
bond scission because the µ3-bonded, harder N atom is less strongly bonded as the equivalent
S atom.
MCo
M S
SCo
S
SH
R
MCo
MS
SCo
S
S
H R
MCo
MS
S
Co
S
S-RH
RSHM
Co
M S
S
Co
S
Scheme 7.6 Reaction of alkyl- and arylthiols on homogeneous Cp*2Mo2Co2S3(CO)4 clusters
(Cp* = C5(CH3)5) [3].
HDN of Naphthylamine Chapter 7 162
7.5 Conclusions
HDN of 1-naphthylamine over sulfided NiMo/Al2O3 and CoMo/Al2O3 catalysts leads to
the formation of tetralin and naphthalene. The high selectivity to 1,2-dihydronaphthalene at
low conversion of 1-naphthylamine shows that one of the HDN pathways is partial
hydrogenation of 1-naphthylamine followed by NH3 elimination. Naphthalene forms with
high selectivity, even at low conversion. This apparent hydrogenolysis can be explained by
hydrogenation of 1-naphthylamine to 1,2-dihydro-1-naphthylamine followed by NH3
elimination. Another possible mechanism is hydrogenation of 1-naphthylamine to dihydro-1-
naphthylamine which undergoes a Bucherer-type NH2-SH exchange, followed by
dehydrogenation to 1-thionaphthol and hydrogenolysis to naphthalene.
7.6 References
[1] H. Schulz, M. Schon and N.M. Rahman, Stud. Surf. Sci. Catal. 27 (1986) 201.
[2] M. Houalla, N.K. Nag, A.V. Sapre, D.H. Broderick and B.C. Gates, AIChE J. 24
(1978) 1015.
[3] M.D. Curtis and S.H. Druker, J. Am. Chem. Soc. 119 (1997) 1027.
[4] C. Moreau, J. Joffre, C. Saenz, J.C. Afonso and J.L. Portefaix, J. Mol. Catal. 161
(2000) 141.
[5] J. Mijoin, G. Pérot, F. Bataille, J.L. Lemberton, M. Breysse and S. Kasztelan, Catal.
Lett. 71 (2001) 139.
[6] N. Nelson and R.B. Levy, J. Catal. 58 (1979) 485.
[7] J.L. Portefaix, M. Cattenot, M. Gerriche, J. Thivolle-Cazat and M. Breysse, Catal.
Today 10 (1991) 473.
[8] L. Vivier, V. Dominguez, G. Perot and S. Kasztelan, J. Mol. Catal. 67 (1991) 267.
[9] R. Prins, Adv. Catal. 46 (2001) 399.
[10] M. Jian and R. Prins, Catal. Today 30 (1996) 127.
HDN of Naphthylamine Chapter 7 163
[11] C. Moreau, L. Bekakra, J.L. Olivé and P. Geneste, in: Proc. 9th Int. Congr. on
Catalysis, Vol. 1, eds. M.J. Philips and M. Ternan (Chem. Inst. of Canada, Ottawa,
1988) p. 58.
[12] M. Egorova, Y. Zhao, P. Kukula and R. Prins, J. Catal. 206 (2002) 263; and chapter 2.
[13] P. Geneste, C. Moulinas and J.L. Olivé, J. Catal. 105 (1987) 254.
[14] F. Rota and R. Prins, Topics in Catal. 11/12 (2000) 327.
[15] F. Rota and R. Prins, J. Mol. Catal. 162 (2000) 359.
[16] M.B. Smith and J. March, Advanced Organic Chemistry (Wiley, New York, 5th Ed.,
2001).
[17] L. Qu and R. Prins, J. Catal. 217 (2003) 284.
[18] J.S. Kim, J.H. Reibenspies and M.Y. Darensbourg, J. Am. Chem. Soc. 118 (1996)
4115.
Concluding remarks Chapter 8 165
8. Concluding remarks
8.1 Conclusion
Since the publication of Nelson and Levy, it has been widely accepted that the
nucleophilic substitution and Hofmann β hydrogen elimination are the main
hydrodenitrogenation (HDN) mechanisms [1-5]. Direct hydrogenolysis was discussed as well.
Direct breaking of the C-N bond in aniline occurs to a minor degree. In the HDN of o-
propylaniline the selectivity to propylbenzene was 7% over a NiMo/Al2O3 catalyst and 24%
over a Mo/Al2O3 catalyst [6]. For naphthylamine, direct C-N bond breakage is even more
important [7]. In order to clarify whether direct C-N bond takes place, the HDN of
neopentylamine and naphthylamine was studied [8,9]. In the HDN of neopentylamine,
neopentane was formed as a secondary product, as the selectivity extrapolates to zero with
decreasing weight time. Neopentane was formed by hydrogenolysis of neopentanethiol, which
was formed by substitution of neopentylamine with H2S. Direct hydrogenolysis of
neopentylamine to neopentane can hardly take place. In the HDN of naphthylamine, the
intermediate 1,2-dihydronaphthalene was detected, which could be formed by elimination of
1,2,3,4-tetrahydronaphthylamine. The HDN studies of the intermediates 1,2,3,4-tetrahydro-1-
naphthylamine, 1,2-dihydronaphthalene and 5,6,7,8-tetrahydro-1-naphthylamine show that 1-
naphthylamine reacts through hydrogenation to 1,2,3,4-tetrahydro-1-naphthylamine, which
reacts by NH3 elimination to 1,2-dihydronaphthalene. The latter molecule subsequently reacts
to tetralin and naphthalene. Direct denitrogenation of 1-naphthylamine to naphthalene could
also occur by hydrogenation of 1-naphthylamine to 1,2-dihydro-1-naphthylamine, followed
by NH3 elimination or followed by a Bucherer-type NH2-SH exchange, dehydrogenation and
C-S bond hydrogenolysis. This suggests that hydrogenolysis in arylamines is not true but
apparent, meaning that the reaction occurs via an indirect and multi-step mechanism.
As direct hydrogenolysis of amines does not take place under normal experimental
conditions, elimination and substitution will be the only HDN mechanisms. To determine
Concluding remarks Chapter 8 166
which mechanism is responsible for the HDN reaction of an alkylamine, one can measure the
product selectivities as a function of weight time (τ) and determine whether they extrapolate
to a non-zero or zero value at τ = 0. If a product such as an alkene has a zero selectivity at
zero weight time, then it cannot be a primary product and elimination cannot play a role. If its
selectivity is non-zero at τ = 0, then the alkene might be a primary product, and elimination
might be important. The alkenes/alkane ratios could also be measured to differentiate the
mechanisms [10-12]. The similar ratios of hexenes/hexane originating from hexylamine and
pentenes/pentane originating from pentanethiol, in the simultaneous reactions of hexylamine
and pentanethiol, show that hexylamine reacts with H2S to form hexanethiol. The latter
molecule reacts to hexenes and hexane. It demonstrates that nucleophilic substitution is the
predominant mechanism. The same mechanism was proved to operate in the HDN of 2-
hexylamine [11]. However, the ratio of methylbutenes/methylbutane in the HDN of 2-methyl-
2-butylamine is totally different as that obtained in the HDS of 2-methyl-2-butanethiol in the
presence of hexylamine, which shows that an E1 mechanism operates in the HDN of an
alkylamine with the amine group attached to a tertiary carbon atom [11]. The results from the
simultaneous HDN of pentylamine and HDS of hexanethiol, and the simultaneous HDN of 2-
hexylamine and HDS of 2-pentanethiol, show that nucleophilic substitution is also the
predominant mechanism over sulfided CoMo/Al2O3 and Mo/Al2O3 catalysts. The HDN of 2-
methyl-2-butylamine, as well as the HDS of 2-methyl-2-butanethiol in the presence of
hexylamine, show that, over all three catalysts, E1 is the main HDN mechanism. Furthermore,
the formation of thiols in the HDN of amines was explained by an acid-base mechanism as
well as by a metal-like catalyzed mechanism [11,12].
In the HDN of 1-adamantylamine, 1-adamanethiol and adamantane were formed. 1-
Adamantanethiol was the primary product, which reacts by C-S hydrogenolysis to
adamantane. As 1-adamantylamine can not react by ammonia elimination, nor by a classic
SN2 substitution, it is proposed that the NH2-SH substitution in adamantylamine takes place
by adsorption of the amine group at the metal sulfide surface and migration of the adamantyl
group to a neighbouring sulfur atom.
Concluding remarks Chapter 8 167
8.2 Outlook
It is clear that substitution is the predominant mechanism in the HDN of alkylamines
with the amine group attached to a primary or secondary carbon atom [9-12]. Only an E1
mechanism operates in the HDN of an alkylamine with the amine group attached to a tertiary
carbon atom [11-12] over NiMo/Al2O3, CoMo/Al2O3 and Mo/Al2O3 catalysts.
A remaining question is how the substitution of the alkylamine with H2S or another
alkylamine takes place on the catalyst surface. The formation of a dialkylamine and
alkanethiol in the HDN of an alkylamine with the NH2 group attached to a primary or
secondary carbon atom could be explained by either an acid-catalysis mechanism or a metal-
catalysis mechanism: dehydrogenation of an amine to an imine, addition of H2S or another
alkylamine, elimination of NH3, and hydrogenation of the resulting thioaldehyde to a thiol
[13].
The disproportionation reaction of two dialkylamine molecules to a trialkylamine and an
alkylamine allows to distinguish between acid and metal catalysis. If the surface is metallic,
N-ethylbutylamine will dehydrogenate to form N-butylethylimine. The NH2 group from
another N-ethylbutylamine molecule can add to the N-butylethylimine to form a diaminal
intermediate. After elimination of butylamine, hydrogenation of the resulting molecule leads
to N,N-diethylbutylamine (Scheme 8.1). N-ethylbutylamine can also dehydrogenate to form
N-ethylbutylimine. In that case, N,N-dibutylethylamine will be formed by imine formation,
addition of an amine, elimination, and hydrogenation of the resulting elimination
intermediate. In the HDN of N-methylhexylamine, however, only dihexylmethylamine should
be formed if the surface is metallic. N,N-dimethylhexylamine cannot be generated, as the
elimination of hexylamine cannot take place because there is no β-H atom available (Scheme
8.2). Therefore, if the surface is only metallic, the molar ratio of N,N-
dihexylmethylamine/N,N-dimethylhexylamine in the HDN of N-methylhexylamine should be
much higher than 1, while the ratio of N,N-dibutylethylamine/N,N-diethylbutylamine in the
HDN of N-ethylbutylamine should be about 1. If the surface is acidic, then the
disproportionation occurs by classic substitution and both ratios should be close to 1.
Concluding remarks Chapter 8 168
C2
HN
C4
HN
C4
C2N
C4
HN
C2
C2N
C4
C2N
C4
+ C4
C2
C2N
C4
C2N
C4
C4
C2N
C4
+ C2
C2
HN
C4
NC4
NC2
H2N
H2N
Scheme 8.1 Reaction mechanism in the disproportionation of N-ethylbutylamine.
HN
C6
C1N
C6
C5HN
C1N
C6
C1
C1N
C6+ C6
C1N
C4
C4
C6
C1N
C6
+ C1
C1
HN
C6
C1
HN C6
NC6
NC5
H2N
H2N
Scheme 8.2 Reaction mechanism in the disproportionation of N-methylhexylamine.
The HDN of a chiral alkylamine can be another test to distinguish between acid-base
and metal catalysis over sulfided NiMo/Al2O3. For instance, in the HDN of R-2-butylamine,
only R,S-di(2-butylamine) can be formed by classic nucleophilic substitution, if the catalyst
surface is acidic (Scheme 8.3). If the catalyst surface is metallic, the disproportionation
product of R-2-butylamine should be a 1:1 mixture of R,R- and R,S-di(2-butylamine), which
can be explained by a sequence of reactions of amine-imine-amine addition-ammonia
elimination-dialkylimine hydrogenation (Scheme 8.4).
Concluding remarks Chapter 8 169
NH2
R
NH2
R
-NH3 NH
R
S
Scheme 8.3 Disproportionation of R-2-butylamine by acid-base catalysis.
NH
NH2
R
NH
R
NH2
R-H2
NH2
-NH3N
R
H2 NH
R
R,S
Scheme 8.4 Disproportionation of R-2-butylamine by metal-like catalysis.
Similar methods can be used to investigate the mechanism of the substitution of the
alkylamine by H2S. If this substitution would occur by classic substitution with Walden
inversion, then R-2-butylamine should form S-2-butanethiol (Scheme 8.5). If the substitution
would be indirect, via an imine, then a racemic 2-butanethiol mixture is expected (Scheme
8.6). It is unclear what the stereochemistry of the substitution proposed in Scheme 8.7 will be.
Maybe it leads to retention of the configuration.
Such and similar methods might help to elucidate the path of the HDN reactions at the
catalyst surface.
NH2H2S + NH2H2S NH3+HS
Scheme 8.5 Classic nucleophilic substitution of R-2-butylamine with H2S to S-2-butanethiol.
Concluding remarks Chapter 8 170
NH
SH
NH2
R-H2
NH2-NH3
SH2
SH
R,S
H2S
Scheme 8.6 Substitution of R-2-butylamine with H2S to R(S)-2-butanethiol by means of an
imine.
Mo
S
S S
S
Mo
S
S
S N
H
H HR
Mo
S
S S
S
Mo
S
S
S N
H
H HR
Mo
S
S S
S
Mo
S
S
S N
H HHR
Scheme 8.7 NH2-SH substitution by alkyl migration on the metal sulfide surface.
8.3 References
[1] N. Nelson, R.B. Levy, J. Catal. 58 (1979) 485.
[2] J.L. Portefaix, M. Cattenot, M. Guerriche, J. Thivolle-Cazat, M. Breysse, Catal. Today
10 (1991) 473.
[3] L. Vivier, V. Dominguez, G. Perot, S. Kasztelan, J. Mol. Catal. 67 (1991) 267.
[4] F. Rota, V.S. Ranade, R. Prins, J. Catal. 200 (2001) 389.
[5] P. Clark, X. Wang, P. Deck, S. T. Oyama, J. Catal. 210 (2002) 116.
[6] M. Jian, R. Prins, Catal. Today 30 (1996) 127.
[7] C. Moreau, L. Bekakra, J.L. Olivé and P. Geneste, in: Proc. 9th Int. Congr. on
Catalysis, Vol. 1, eds. M.J. Philips and M. Ternan (Chem. Inst. of Canada, Ottawa,
1988) p. 58.
[8] Y. Zhao, J. Czyzniewska, R. Prins, Catal. Lett. 88 (2003) 155; and chapter 7.
[9] Y. Zhao, J. Czyzniewska, R. Prins, to be published (2004); and chapter 6.
Concluding remarks Chapter 8 171
[10] Y. Zhao, P. Kukula, R. Prins, J. Catal. 221 (2004) 441; and chapter 3.
[11] Y. Zhao, R. Prins, J. Catal. 222 (2004) 532; and chapter 4.
[12] Y. Zhao, R. Prins, to be published 2004; and chapter 5.
[13] R. Prins, in: G. Ertl, H. Knözinger, J. Weitkamp (Eds.), Handbook of Heterogeneous
Catalysis, vol. 4, Wiley-VCH, New York, 1997, p. 1916.
Acknowledgements
I greatly appreciate that many people contributed to my work especially:
Prof. R. Prins for offering me the opportunity to join his group so that I can fulfil my
dream studying as Ph.D. in western countries. I am very proud of his supervision.
Prof. M. Breysse of Université Pierre et Marie Curie for taking on the task of co-
examiner.
Dr. P. Kukula for many discussions and helpful suggestion and for the friendly
collaboration during my thesis. Dr. L.L. Qu for fruitful help with my hydrotreating unit,
wonderful discussion to the HDN work and teaching me how to be a standard Chinese. Dr. M.
Egorova, one of the my best colleagues in Prof. Prins groups, for kind help and endure to my
bad habits in my first year of PhD, and wonderful collaboration during my Ph.D. work. Ms.
A. Röthlisberger, my nice colleague, for the help of English (French tongue), sharing the
pressure and happiness during my Ph.D study. Samantha and Achim my excellent Swiss
friends for the help to European life and the help of skiing. I enjoyed the wonderful time
driving to ULM for Catalysis Lectures with Daniele and Lukas. Thanks Lukas to translate the
abstract of my English dissertation to German. Ms. M. Schoenberg for her kind help of
English and friendly collaboration. Dr. G. Pirngruber for the kind help and suggestions in
training my oral presentation. T. Schmid and L. Adad for the help of organic synthesis. T.
Todorova for the thoughtful discussion of DFT calculation. M. Lüchinger for the help of
AAS. S.F. Yang, B. Bin, N. Weiher, D. Sirbu, A. Leone, J. Kovacovic, M. Kuba, P. Kanti
Roy, Prof. C., Giambattista, E. Bus, J. van Bokhoven, A. Abraham, and all other members of
the group for their advises and their friendship.
I greatly thank my wife, Xueying Jin, for her love, consistent support during my PhD
work.
Publications
1. M. Egorova, Y. Zhao, P. Kukula and R. Prins
“On the role of β-hydrogen atoms in the hydrodenitrogenation of 2-methylpyridine and 2-
methylpiperidine“
J. Catal. 206 (2002) 263.
2. Y. Zhao, J. Czyzniewska, and R. Prins
“Mechanism of the direct hydrodenitrogenation of naphthylamine over sulfided NiMo/γ-
Al2O3“
Catal. Lett. 88 (2003) 155.
3. Y. Zhao, P. Kukula, and R. Prins
“Mechanisms of the hydrodenitrogenation of alkylamines with secondary and tertiary α-
carbon atoms over sulfided NiMo/γ-Al2O3“
J. Catal. 221 (2004) 441.
4. Y. Zhao and R. Prins
“Mechanisms of the hydrodenitrogenation of alkylamines with secondary and tertiary α-
carbon atoms over sulfided NiMo/γ-Al2O3 “
J. Catal. 222 (2004) 532.
5. Y. Zhao, J. Czyzniewska, and R. Prins
“Mechanism of the hydrodenitrogenation of adamantylamine and neopentylamine over
sulfided NiMo/γ-Al2O3“
To be published (2004).
6. Y. Zhao and R. Prins
“Mechanism of the hydrodenitrogenation of alkylamines over NiMo/γ-Al2O3, CoMo/γ-
Al2O3 and Mo/γ-Al2O3“
To be published (2004).
Curriculum Vitae
Name: Yonggang Zhao
Date of birth: 21th September 1974
Place of birth: Huaiyin (Jiangsu, China)
Nationality: Chinese
Education 1982-1987 Hedong Primary school, Huaiyin county, Jiangsu province
1987-1990 Wuji town Middle school, Huaiyin county, Jiangsu province
1990-1993 Huaiyin County High school, Jiangsu province
1993-1997 Fushun Petroleum Institute, Liaoning Province
Bachelor of Chemical Engineering (refinery)
1997-2000 Fushun Petroleum Institute, Liaoning province
Master of Chemical Engineering
Alkylation of long chain alkene with benzene over solid acid catalyst.
2000-2004 PhD Thesis at the ETH Zürich in the group of Prof. Dr. R. Prins:
Hydrodenitrogenation of Amines over Sulfided NiMo/Al2O3,
CoMo/Al2O3, and Mo/Al2O3