Innovative preparation of catalysts by aerosol route for the Fischer – Tropsch synthesis Maria Joana Figueiredo Rodrigues Thesis to obtain the Master of Science Degree in Chemical Engineering Supervisors: Dr. Alexandra Chaumonnot (IFPEN) Dr. Antoine Fécant (IFPEN) Prof. Carlos Henriques (IST) Examination Committee Chairperson: Prof. José Madeira Lopes (IST) Supervisor: Prof. Carlos Manuel Faria de Barros Henriques (IST) Members of the committee: Prof. Maria Filipa Gomes Ribeiro (IST) October 2015
85
Embed
Innovative preparation of catalysts by aerosol route for ... · Usually Fischer – Tropsch (FT) catalysts are obtained through the deposition of cobalt active phase onto the support.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Innovative preparation of catalysts by aerosol route for the
Fischer – Tropsch synthesis
Maria Joana Figueiredo Rodrigues
Thesis to obtain the Master of Science Degree in
Chemical Engineering
Supervisors: Dr. Alexandra Chaumonnot (IFPEN)
Dr. Antoine Fécant (IFPEN)
Prof. Carlos Henriques (IST)
Examination Committee
Chairperson: Prof. José Madeira Lopes (IST)
Supervisor: Prof. Carlos Manuel Faria de Barros Henriques (IST)
Members of the committee: Prof. Maria Filipa Gomes Ribeiro (IST)
October 2015
This page was intentionally left blank
“If we knew what we were doing it would not be called research, would it?”
Albert Einstein
This page was intentionally left blank
v
Resumo
Em termos gerais, os catalisadores usados no processo de Fischer – Tropsch (FT) são obtidos
através da deposição do cobalto (fase ativa) no suporte. Este óxido é habitualmente obtido pelo spray
– drying de uma solução que contém precursores inorgânicos. O principal objetivo deste trabalho é não
só a síntese de catalisadores para o processo de FT através da técnica de spray – drying, mas também
a sua caracterização.
Foram sintetizados sólidos suportados quer em sílica quer em alumina, onde se verificou uma
grande dispersão do cobalto, o que originou uma forte interação molecular entre o precursor de cobalto
e o precursor inorgânico molecular, confirmando-se assim a existência quer de silicatos de cobalto ou
aluminatos de cobalto. De modo a diminuir esta interação, foram modificados os vários parâmetros da
preparação, como por exemplo, a alteração do pH das soluções de atomização, de modo a induzir
atração ou repulsão electroestática entre o precursor de cobalto e o precursor de sílica ou alumina na
solução inicial. Foram obtidos melhores resultados quanto à acessibilidade e redutibilidade do cobalto
no caso da sílica para um pH mais ácido, apresentando, ainda assim, a formação de silicatos de cobalto.
Deverá ser realizado um estudo mais detalhado quer à sílica quer à alumina relativamente aos
parâmetros de síntese. Deverá ser igualmente testada a incorporação de um metal promotor de modo
Abstract ................................................................................................................................................... vii
Acknowledgements ................................................................................................................................. ix
Table of Contents .................................................................................................................................... xi
Table List ............................................................................................................................................... xiii
Figure List ............................................................................................................................................... xv
Glossary ................................................................................................................................................ xvii
Figure 7- Main synthesis pathways to mesostructured materials. [21] ................................................. 15
Figure 8 - Spray - Dryer (Büchi B 290) working principle and material's structuration mechanism.
Adapted from [20] .................................................................................................................................. 18
Figure 9 – Calcination profile for the synthesized supports. ................................................................. 23
Figure 10 - Calcination profile for JFR014 and JFR016. ....................................................................... 25
Figure 11 - Calcination profile for JFR019, JFR021, JFR022, JFR023, JFR024 and JFR037. ............ 25
Low angles XRD Organization of the porosity of the support (mesostructuration)
XPS Degree of oxidation of various elements
TEM
Information of the support: organization of the porosity of the support
(mesostructuration).
Information on the metallic phase: form in which the cobalt is, size of the
cobalt particles
SEM Information of the support: average size and morphology of the aerosol
elementary particles
30
3.3.1. N2 adsorption-desorption
This method is widely used for studying textural properties of catalysts. Adsorption equilibrium is
represented by isothermal plots, which describe the adsorbed quantity as function of the equilibrium
pressure P of the gas in contact with the solid. Actually, relative pressure P/P0 is used instead of P,
where P0 is the saturated vapor pressure of the adsorbate at the measurement temperature (≈ 77 K in
case of nitrogen). Therefore, an adsorption – desorption isotherm consists in measuring the quantity of
gas that is adsorbed on (or desorbed from) the surface of the solid, at a given temperature. [29]
Nitrogen is the most common molecule used in this technique. However, other molecules can be
used for specific purposes: for instance, argon, a very small monoatomic gas, is more appropriated for
the study of microporous samples, or krypton, where the low saturated vapor pressure (≈ 2 torr) can be
used to measure small adsorbed quantities precisely (for specific surface areas below 1 m2/g). [29]
Before measuring the adsorbed quantity, a degassing (or pre-treatment) stage is carried out in order
to eliminate the compounds adsorbed on the surface of the sample (H2O, CO2, etc.). The isotherms are
obtained by gradually increasing the pressure, where the small pores are filled first. The gas condenses
in successively larger pores until a saturated vapor pressure level is reached at which the entire porous
volume is saturated with liquid. Furthermore, the adsorption – desorption phenomena is highly suitable
for the study of samples where the pore size is in the mesoporous domain. [29]
The desorption isotherm is not often superimposed over the adsorption isotherm, which shows up
as a hysteresis phenomenon. This fact happens due to capillary condensation, which corresponds to a
phase transition effect caused by the interactions with the surface of the solid where the gas phase
abruptly condenses in the pore, accompanied by the formation of a meniscus at the liquid – gas
interface.
In Figure 14 it is described a typical isotherm of a mesostructured solid with a very good level of
organization. Moreover, the pore size distribution of a mesostructured solid is very narrow, as the pore
diameter is very uniform.
Figure 14 – Typical isotherm for a mesostructured solid. [29]
The main parameters obtained through this technique are the specific surface area, pore size
distribution, specific porous volume and information on the structure (pore shape, interconnection, etc.).
These information can be given by several models, where the BET model is used to determine the
specific surface area, the BJH method is used in the pore size distribution, while the t-plot method
consists in comparing the adsorption isotherm of a given solid in terms of the adsorbed thickness, what
makes possible to characterize the micro- and mesoporosity.
31
In this study, these analyses were performed in a Micromeretics ASAP 2420 equipment.
3.3.2. Temperature Programmed Reduction (TPR)
This method is used to evaluate the catalyst reducibility and the oxidation degree of the active
phase. The hydrogen consumption is followed as a function of temperature. TPR is extremely attractive
due to its high sensitivity to chemical changes induced by a catalyst promoter or by the support. [29]
This technique consists in measuring the consumption of hydrogen while heating a catalyst with a
linear temperature rate under continuous gas flow. TPR profiles do not provide direct information about
the modification of the catalyst structure, because hydrogen consumption could be attributed to different
reduction processes. In this study, the TPR analysis were performed in a Mircrometrics Auto Chem II
2920. The catalyst, during this analysis, was under a mixture of 5% H2 in air, at 58 cm3/min, and the
temperature was raised up to 1000°C, at a rate of 5°C/min. The hydrogen consumption is followed
through a thermal conductivity detector (TCD), and, it can be seen that, during the reduction process,
several products such as H2O, CO2 or CO are formed.
3.3.3. X-Ray Diffraction (XRD)
XRD is often used for identification of cobalt crystalline phases and evaluation of the crystal size
using the Debye - Scherrer equation (Eq. 15), where, β is the angular breadth of a diffraction line, C is
a constant, λ is the X-ray wavelength, L is volume – averaged size of crystallites, and finally, θ is the
Bragg angle.
The peak breadth, β, can be either calculated through the full-width half-maximum (FWHM) or by
the “integral width”, which corresponds to the area under diffraction peak divided by peak maximum.
Therefore, is often some uncertainty in measuring the size of cobalt crystallites with the Debye - Scherrer
equation, because the β definition will determine the value of the Bragg constant (C).
For very small and very large crystallites, it has been noticed a low accuracy in measuring the
crystallite sizes. The XRD technique is not very sensitive to the presence of small crystallites of cobalt
oxides (< 2 – 3 nm), since the peaks get too broad to be identified and measured. Broadening of the
XRD lines is caused by structural imperfections of the sample. [5]
The values measured during an analysis are the relative intensity levels of the diffraction lines,
corresponding to beam intensities. Moreover, the measurement of the scattering angle is done in lattice
planes which are identified using indices in three dimensions: h, k and l. These last three are related to
the directions of axes defining the crystalline system. [29]
Therefore, in order to follow the Bragg condition (Eq. 16), there is no possibility of scattering planes
with a spacing less than half the wavelength. Where n is an integer, λ is the wavelength, d is the plane
spacing, and finally, θ is the Bragg angle.
𝛽 =𝐶𝜆
𝐿 cos 𝜃
(Eq. 15)
𝑛𝜆 = 2𝑑 sin 𝜃 (Eq. 16)
32
These analyses, ranging from 5 to 72° (2 𝜃), were performed in a PANalytical X’Pert Pro equipment
with the reflection configuration.
3.3.4. Low angles X-Ray Diffraction
Low angles X-Ray diffraction analysis is often used to obtain parameters such as size, morphology
and distribution of particles. Normally, X-ray diffraction (XRD) covers angles in the range of 10° to 100°,
corresponding to a typical wavelength of 0.1 nm, to interatomic distances. Low angles XRD concerns
angles under 5°, thus, distances are greater than a nanometer. [29]
Therefore, low angles XRD processes at a higher resolution and scattering strength while
producing the scattering peaks, making it a suitable method for characterize mesostructured materials.
These analysis were performed in a PANalytical X’Pert Pro equipment in the transmission configuration,
and the analysis range was from 0.2 to 10° (2 𝜃). Moreover, the analysis were done without beam-stop,
in order to prevent the detector saturation. Also, it was used a Cu0.2 attenuator.
3.3.5. X-ray Photoelectron Spectroscopy (XPS)
XPS allows the characterization of the external surface layers of the catalyst (5 to 10 nm) [35] and
the degree of oxidation or electronic state of various elements (chemical neighborhood). The sample
that will be characterized is bombarded by an X-ray photon beam, thus, electrons of different elements
are emitted in terms of number and energy by an appropriate detector. The measured kinetic energy is
directly connected to the electron biding energy by the photoelectric effect. [29]
The characterization in this study was carried out in an ESCA KRATOS Axis Ultra spectrometer
with a monochromatic source of Al. The obtained spectrum are compared with references in order to
identify the signals.
3.3.6. Transmission Electronic Microscopy (TEM)
TEM provides detailed information concerning the composition and structure of heterogeneous
catalysts with real-space resolution down to the atomic level.
Once a sample is crossed by an electron beam, this may be partially adsorbed or deflected. A
certain fraction of these electrons, and those that have not been deflected, are combined to form an
image. Therefore, the use of transmission electron microscopy is based in controlling the electrons
involved in image formation. Thus, the image of the sample depends on the electron – matter
interactions. One can note that atoms with high atomic number scatter more electrons and at larger
angles than those with lower atomic number. Therefore, a particle of heavy metal on a light oxide
support, for instance silica or alumina, will appear dark. [29]
This characterization was carried out in a JEM 2100F equipment. To improve the contrast between
the support and the metallic phase, it is easiest to analyze it in a dark – field image, by selecting the
plan (220) for metallic cubic cobalt and the plan (440) for cobalt oxide (Co3O4).
3.3.7. Scanning Electronic Microscopy (SEM)
This characterization technique allows, along with TEM, to a local chemical and textural
characterization of catalysts. The principle of operation is very close to the one used in TEM. Therefore,
the image is obtained through the interaction of material’s and the electron beam, also known as electron
33
probe, with energies between 0.5 and 35 kV. However, when using SEM microscopes, the image
resolution is limited by the size of the electron probes. [29]
SEM provides information about the size, shape and 3D arrangement of the catalyst particles.
Nevertheless, SEM is not able to fully identify the intrinsic structure of the catalyst, thus, TEM should be
performed to fully characterize the pore structure.
In the present study, this type of analyses was performed with a Supra 40 equipment, with no pre-
treatment of the sample.
This page was intentionally left blank
35
4. Results and Discussion
Generic remarks
As the final application of these catalysts is the FT synthesis, it is more suitable the use of an
ultrasonic nozzle than a two fluid nozzle (normally used at the laboratory scale). Indeed, it produces
larger elementary particles. Therefore, the produced elementary particles would be closer to the ones
of the industrial catalysts that have elementary particles ranging from 80 to 100 µm.
In the following paragraphs, first are described the silica results and after alumina results. One
should take into account, that the main priority of this study is not the synthesis of a mesostructured
solid, but the synthesis of a mesoporous solid, since the FT process does not require mesostructured
catalysts.
On one hand, alumina matrix is the most used support in the FT synthesis. On the other hand,
there is more information in the literature concerning silica sol – gel chemistry and also IFPEN has a
vast experience in synthetizing silica based materials through the aerosol process. Therefore, a more
detailed study was performed on silica matrixes.
Supports were firstly tried to produce instead of trying to produce directly a catalyst, in order to
know if it was possible to synthesize it and also to characterize the elementary particles obtained through
the ultrasonic nozzle, which had never been used at IFPEN.
All the synthesized supports and catalysts were characterized through N2 adsorption – desorption
isotherm and the latter was also characterize through TPR. Besides these analysis, other
characterization techniques were applied when needed.
36
4.1. Silica
The following paragraphs tackle the synthesis of supports and the introduction of cobalt onto a
silica matrix, using as inorganic precursor TEOS.
4.1.1. Supports
Comparison with the two-fluid nozzle, generic characteristics and trials reproducibility
First of all, a trial was done with the two-fluid nozzle (JFR003) as it was the usual nozzle used in
the laboratory at IFPEN, as previously mentioned. This trial allowed a comparison between the two –
fluid and the ultrasonic nozzle, making it possible to evaluate the major differences between these two
equipments.
Concerning trials with the ultrasonic nozzle, two trials (JFR005 and JFR006) with the same spray
– drier parameters were performed in order to evaluate the trials reproducibility. Moreover, two samples
(JFR006 and JFR007) were collected along one trial, in order to evaluate if there were any differences
in the samples collected in the begging and the end of one trial.
Textural properties
Concerning the supports textural properties (BET surface and pore size distribution) nitrogen
adsorption – desorption isotherms were performed, which are described in Figure 15.
Figure 15 – Nitrogen adsorption – desorption as a function of relative pressure and pore size distribution determined by BJH adsorption for JFR003, JFR005, JFR006 and JFR007.
In Table 8 it is described the surface area, the porous volume and the pore diameter for all the
previous samples.
0
50
100
150
200
250
0 0,2 0,4 0,6 0,8 1
Ad
so
rbe
d
Vo
lum
e (
mL
/g)
Relative Pressure
JFR003
JFR005
JFR006
JFR007
0
0,02
0,04
0,06
0,08
0,1
0,12
0 10 20
dV
/dD
vo
lum
e (
mL
/g)
Average pore diameter (nm)
37
Table 8 - BET surface, porous volume and pore diameter for JFR003, JFR005, JFR006 and JFR007.
Comparing the isotherms of JFR003 and JFR005, one can see that the solid obtained with the two
– fluid nozzle presents a hysteresis loop with more vertical lines than the solids obtained with the
ultrasonic nozzle. Thus, JFR003 seems more organized than JFR005. However, low angles XRD and
TEM analysis should be performed to fully verify this statement.
Furthermore, the isotherms shape of JFR005 and JFR006 are very similar, hence, it is possible to
conclude that the trials are reproducible. Likely, the isotherms shape of JFR006 and JFR007, are very
similar too, so it is possible to conclude that there is no change along time in solids belonging to the
same trial.
Higher surface areas are achieved with the two – fluid nozzle which is due to an increase in the
mesoporous volume.
Silica particles size distribution
In order to characterize the elementary particles morphology produced through the ultrasonic
nozzle, SEM analysis were performed in JFR005, JFR006 and JFR007.
As is possible to conclude through Figure 16 the synthesized silica particles are spherical, and as
one can see, the particles present the same shape throughout the trials, showing once more the trials
reproducibility (Figure 16 (A) and (B)). Furthermore, looking at the SEM micrographs (B) and (C) one
can conclude once more that there are no significant changes along the same trial.
Reference Porous type BET surface (m2) Porous volume (mL/g) Pore diameter (nm)
JFR003
Mesoporous
183 0.37 7.8
JFR005 114 0.20 7.8
JFR006 101 0.18 7.9
JFR007 107 0.18 7.9
(A) (B)
(C)
38
Figure 16 – SEM micrographs for (A) JFR005, (B) JFR006 and (C) JFR007.
Moreover, it was not possible to precise if the elementary particles are hollow as there are no
damage elementary particles.
For each sample it was done a particle’s size distribution, which are shown in Figure 17.
Figure 17 – Particles size distribution for JFR005, JFR006 and JFR007.
0
0,05
0,1
0,15
0,2
0 10 20 30 40 50
Fre
qu
en
cy (
vo
lum
e)
Elementary particles diameter (µm)
JFR005
0
0,05
0,1
0,15
0,2
0,25
0 10 20 30 40 50 60
Fre
qu
en
cy (
vo
lum
e)
Elementary particles diameter (µm)
JFR006
0
0,05
0,1
0,15
0,2
0,25
0 10 20 30 40 50 60
Fre
qu
en
cy (
vo
lum
e)
Elementary particles diameter (µm)
JFR007
(A) (B)
(C)
39
Regarding these histograms, it is possible to find the same results as before concerning the trials
reproducibility.
Table 9 - Results of the elementary particle size determinate by SEM.
The particle size distribution presents for the three supports a wide range of particles size, as it can
be seen in the histograms, since the particles size range from 2 to 60 µm. However, the maximum of
population is found for 30 and 35 µm.
As previously mentioned, the ultrasonic nozzle was used instead of the two – fluid nozzle due to
the final application of these catalysts in the FT process. It was expected to have the majority of
elementary particles ranging from 80 – 100 µm, and as one can see the medium elementary particle
diameter is around 30 µm, which value is acceptable for testing.
Organization of the porosity of the support
In order to confirm the existence of a mesostructured solid, low angles XRD was performed which
is shown in Figure 18. The presence of a peak at 0.75° (2θ) for the three samples confirms the existence
of an organized structure.
One can see that there are not three peaks as it is characteristic in mesostructured solids with a
very good level of organization. There are two hypotheses that can justify this absence: it can be
characteristic of samples where parts of them are very well organized and other parts do not present
organization at all. It can also be characteristic of samples where the entire sample is organized but not
with a good level of organization. Moreover, one can suppose that the three samples have a worm-like
structure (second option) and it will be confirmed thanks to a TEM analysis.
Moreover, the above assumption seems to be true, as the TEM analysis of a catalyst (later
described) reveals an organization throughout the sample, however with a low level of organization,
which is characteristic of worm – like structures.
Reference
Elementary particles medium
diameter (µm)
Elementary particles minimum
diameter (µm)
Elementary particles maximum
diameter (µm)
Standard deviation (µm)
JFR005 34.18 2.15 52.68 14.28
JFR006 34.51 2.34 59.38 17.40
JFR007 29.99 2.83 56.55 16.51
40
Figure 18 – Low angles XRD patterns for JFR005, JFR006 and JFR007.
4.1.2. Cobalt catalysts
It was synthetized a silica matrix containing 15 wt.% of cobalt concerning the total mass of solid
(JFR016). This solid was obtained by atomizing a solution which contained HCl and a pH, approximately,
of 2. This synthesis consisted mainly in introducing the cobalt precursor solution into the solution that
was used to produce the support. This solid was tested in a FT catalytic unit, as it will be later described.
The following paragraphs describe a preliminary study on the calcination temperature in order to
determine the best conditions to favor the formation of active catalytic species.
One should note, that at this point it is not possible to know if the solids with cobalt loaded are
catalysts or not, as they were not catalytic tested. However, as a simplification they will be referred as
“catalysts”.
Calcination temperature influence
JFR014 and JFR016 were calcinated at different temperatures, 550°C and 350°C, respectively.
The first one experimented the same calcination temperature as the support, and JFR016 was
calcinated with a lower temperature in order to prevent the formation of cobalt silicates, which are hardly
reducible species. [5]
Textural properties
Hereby are presented several analysis results to evaluate the influence of the calcination
temperature on the textural properties of the support.
JFR005 (silica matrix) is only represented in Figure 19 so it can be seen the differences in
Figure 19 - Nitrogen adsorption – desorption as a function of relative pressure and pore size distribution determined by BJH adsorption for JFR005 (support reference), JFR014 and JFR016.
In Table 10 is summarized the surface area (BET surface), the porous volume and the pore
diameter.
Table 10 - BET surface, porous volume and pore diameter for JFR005, JFR014 and JFR016.
The support isotherm shape is very close to the one of the catalysts. Comparing the t-plot of the
catalyst (JFR014) and the support (JFR005), which have the same calcination temperature, a higher
microporous volume is achieved for the catalyst (0.054 cm3/g), while the support (JFR005) presents a
microporous volume of 0.015 cm3/g. In terms of microporous surface area, JFR014 has 131 m2/g and
JFR005 has 42 m2/g. Therefore, the introduction of cobalt nanoparticles induces a higher microporosity,
and also a higher mesoporosity, as there was an increase in the porous volume along with a decrease
in the pore diameter. (Find the t-plot in appendix A)
Moreover, the isotherms shape of both catalysts is very similar. It was confirmed by the t-plot that
a higher microporosity is found in JFR014, which was calcinated at 550 °C. This is expected because
0
20
40
60
80
100
120
140
160
180
0 0,2 0,4 0,6 0,8 1
Ad
so
rbe
d
Vo
lum
e (
mL
/g)
Relative Pressure
JFR005 (Support)
JFR014 (Catalyst, calcination at 550°C)
JFR016 (Catalyst, calcination at 350°C)
Reference Porous type BET surface (m2) Porous volume (mL/g) Pore diameter (nm)
JFR005
Mesoporous
114 0.20 7.8
JFR014 226 0.26 6.6
JFR016 164 0.26 6.5
0
0,01
0,02
0,03
0,04
0 5 10 15
dV
/dD
vo
lum
e (
mL
/g)
Average pore diameter (nm)
42
with a higher calcination temperature small voids in the microporosity domain are created in the
surfactant hydrophilic part. Figure 20 illustrates this phenomenon.
Figure 20 - Scheme representing the solid surfaces calcinated at 350°C and 550°C. Adapted from [20]
Both catalysts present a very similar porous size distribution. For JFR014 the maximum population
is found at 6.6 nm while for JFR016 is found at 6.5 nm, therefore both solid have mesopores.
The support (JFR005) presents higher pore diameters than the catalysts (JFR014 and JFR016).
This smaller pore diameter could be due to a different micelle – inorganic precursor interaction.
Organization of the porosity of the silica matrix
To attest the existence of a mesostructured solid, low angles XRD was performed which is shown
in Figure 21. For JFR014 and JFR016 it was not found a long distance organization because no peak
at small angles was observed.
It is also possible to conclude that the introduction of cobalt onto the support has made solids with
no long distance organization, due to the impact of cobalt species on the chemical reactions inducing
the mesostructuration process.
Besides that, the calcination temperature does not have influence on the mesostructuration.
Vµ pore: 0.054 cm3/g
Mesoporous volume: 0.11 cm3/g
Vµ pore: 0.015 cm3/g
Mesoporous volume: 0.12 cm3/g
43
Figure 21 – Low angles XRD patterns for JFR014 and JFR016.
Finally, the calcination temperature was set up as 400°C, an intermediate value that accomplishes
a higher surfactant removal and trying to avoid the formation of cobalt silicates. Also, there is no interest
in creating micropores for the final application of these catalysts, because the long produced
hydrocarbons chains would have difficult leaving the catalyst.
Catalytic test
Moreover, before having all the analysis results, a part of JFR016 (before being dried or calcinated)
was reduced under hydrogen flow, and tested in a FT catalytic test unit (slurry reactor). After 16 hours,
there was no activity, in another words, there were no products formation. As there were no other
catalytic tests in this work, no setup or reaction condition are presented here.
Therefore, two main hypotheses were proposed for the catalyst lack of activity: first, the presence
of chlorine could affect the catalyst activity, and second, it could exist a strong interaction between the
silica precursor and the cobalt precursor, consequently affecting the cobalt dispersion. In the fowling
paragraphs is described in detail each hypothesis.
1) Presence of chlorine in the catalyst
The presence of chlorine in FT catalysts has showed a significant decrease in activity. This lack of
activity could be due to the poising of several surface sites by Cl atoms. [30] This could be related with
the chlorine strong electronegativity, which prevents the CO dissociation on the catalyst. Moreover, it
was found that the effect of chlorine atoms on the catalyst was slowly reversible.
1 4 2 59 4 i*0.7 - F ile : C 1 5 K 2 58 3 .ra w - T yp e : P S D fa s t-s can - S ta rt: 0 .2 1 2 ° - E n d : 4.99 2 ° - S te p : 0 .01 7 ° - S te p tim e : 20 0 . s - T em p. : 2 5 °C (R o om ) - T im e S ta rte d : 0 s - 2 -Th e ta : 0.21 2 ° - Th e ta : 0.10 5 ° - A u x1: 0 .0 - A ux2 : 0.0 - A u x3 :
1 4 2 38 0 - F i le : C 15 K 2 5 8 0. raw - Typ e: P S D fa s t-sca n - S ta rt: 0. 21 2 ° - E n d: 4 .9 9 2 ° - S te p: 0 .0 1 7 ° - S tep t im e: 2 0 0. s - Te m p .: 2 5 °C (Ro o m ) - T im e S ta rte d : 0 s - 2-T he ta: 0 .2 1 2 ° - T he ta: 0 .1 0 5 ° - A u x1 : 0 .0 - A u x2: 0 .0 - A u x3 : 0 .0 -
1 4 0 97 1 - F i le : C 15 K 2 5 8 1. raw - Typ e: P S D fa s t-sca n - S ta rt: 0. 21 2 ° - E n d: 4 .9 9 2 ° - S te p: 0 .0 1 7 ° - S tep t im e: 2 0 0. s - Te m p .: 2 5 °C (Ro o m ) - T im e S ta rte d : 0 s - 2-T he ta: 0 .2 1 2 ° - T he ta: 0 .1 0 5 ° - A u x1 : 0 .0 - A u x2: 0 .0 - A u x3 : 0 .0 -
1)
1 4 2 13 9 - F i le : C 15 K 2 5 8 5. raw - Typ e: P S D fa s t-sca n - S ta rt: 0. 21 2 ° - E n d: 4 .9 9 2 ° - S te p: 0 .0 1 7 ° - S tep t im e: 2 0 0. s - Te m p .: 2 5 °C (Ro o m ) - T im e S ta rte d : 0 s - 2-T he ta: 0 .2 1 2 ° - T he ta: 0 .1 0 5 ° - A u x1 : 0 .0 - A u x2: 0 .0 - A u x3 : 0 .0 -
1 4 1 41 6 - F i le : C 15 K 2 5 8 2. raw - Typ e: P S D fa s t-sca n - S ta rt: 0. 21 2 ° - E n d: 4 .9 9 2 ° - S te p: 0 .0 1 7 ° - S tep t im e: 2 0 0. s - Te m p .: 2 5 °C (Ro o m ) - T im e S ta rte d : 0 s - 2-T he ta: 0 .2 1 2 ° - T he ta: 0 .1 0 5 ° - A u x1 : 0 .0 - A u x2: 0 .0 - A u x3 : 0 .0 -
B la n c - F i le : C1 5 N 25 8 4 .ra w - T yp e : P S D fa s t-s can - S ta rt: 0 .2 1 2 ° - E n d : 4.99 2 ° - S te p : 0 .01 7 ° - S te p tim e : 2 0 0 . s - T em p . : 2 5 °C (R o om ) - T im e S ta rte d: 0 s - 2 -Th e ta : 0 .21 2 ° - Th e ta : 0 .10 5 ° - A u x1 : 0 .0 - A ux2 : 0.0 - A u x3 : 0 .0 - D i
1 4 2 1 39 - L ef t A n g le: 0 .5 8 0 - R ig h t A n gle : 1 .1 9 8 - L e ft In t. : 5 5 9 8.1 00 C o u nts - Rig h t I nt .: 12 1 1 .4 2 9 Co u n ts - P e aks : 0 - P a ram s : 0 - W e ig ht : -1 .0 0 0 - kA 2 Ra tio : 0.5 - R e l. : 1 .8 0 8 % - T h .R .: 1 .5 46 % - R .In t. : 0 . 00 0 %
In order to prevent this effect HCl was replaced for HNO3 in the initial solution. The following
paragraphs describe the effect of HNO3, either on the supports and catalysts.
Textural properties
In Figure 22, one can see the nitrogen adsorption – desorption isotherms for the supports (JFR005
(HCl) and JFR028 (HNO3)) and for the catalysts (JFR016 (HCl) and JFR021 (HNO3)). One can note that
JFR005 and JFR016 are just represented as references, since they were synthetized with HCl.
Moreover, JFR016 was calcinated at 350°C and JFR021 at 400°C. A different calcination
temperature induces changes in the textural properties, but it was made the choice to compare them.
Figure 22 - Nitrogen adsorption – desorption as a function of relative pressure and pore size distribution determined by BJH adsorption for JFR005 (reference), JFR016 (reference), JFR028 and JFR021.
Concerning the isotherms shape (supports and catalysts isotherms), one can conclude that, there
are no significant changes, except that JFR021 presents a hysteresis loop with more straight lines,
leading to a more organized structure, which is further confirmed.
In Table 11 it is summarized the surface area (BET surface), the porous volume and the pore
diameter for all the previous samples.
0
50
100
150
200
250
300
350
0 0,2 0,4 0,6 0,8 1
Ad
so
rbe
d
Vo
lum
e (
mL
/g)
Relative Pressure
JFR005 (Support, HCl)
JFR016 (Catalyst, HCl)
JFR028 (Support, HNO3)
JFR021 (Catalyst, HNO3)
0
0,02
0,04
0,06
0,08
0,1
0 5 10
dV
/dD
vo
lum
e (
mL
/g)
Average pore diameter (nm)
45
Table 11 - BET surface, porous volume and pore diameter for JFR005, JFR028, JFR016 and JFR021.
Both supports either with HCl (JFR005) or HNO3 (JFR028) present a different porous size
distribution.
Concerning the supports, one can conclude that the introduction of HNO3, increases (variation of
13%) the surface area along with an increase in the porous volume and in the pore diameter, possibly
due to a different micelle – inorganic precursor interaction.
Regarding the catalysts, using HNO3 instead of HCl increases significantly (variation of 103%) the
surface area as well as the porous volume, and reduces the pore diameter. This change in the pore
diameter could be due to the presence of cobalt species on the chemical reactions inducing the
mesostructuration process.
Organization of the porosity of the silica matrix
Concerning the support (JFR005) and catalyst (JFR016) synthetized with HCl, the organization of
the porosity of the support has changed with the introduction of cobalt, since JFR005 was an organized
structure and JFR016 did not present any organization.
Regarding the samples that were synthetized with HNO3, the organization of the porosity of JFR028
(support) should be confirmed thanks to low angles XRD analysis, and the organization of JFR021
(catalyst) was confirmed thanks to the same analysis (Page 49, Figure 25).
Therefore, one can conclude that the presence of HCl or HNO3 has influence in the
mesostructuration of the catalysts, probably due to a different interaction of the cobalt species on the
chemical reactions inducing the mesostructuration process.
Reference
Porous type
BET surface (m2)
Porous volume (mL/g)
Pore diameter
(nm)
Supports JFR005 (HCl)
Mesoporous
114 0.20 7.8
JFR028 (HNO3) 129 0.24 8.4
Catalysts JFR016 (HCl) 164 0.26 6.5
JFR021 (HNO3) 345 0.46 7.1
46
Cobalt reducibility
The cobalt reducibility as well as accessibility was studied by TPR analysis. In Figure 23, it is the
TPR profile for both samples (JFR016 and JFR021) after calcination. After calcination both samples
were violet (which normally corresponds to the Co nitrate precursor coloration).
As previously mentioned, in chapter 2, the cobalt oxide reduction to metallic cobalt happens in two
reactions. The first one (Eq.10) corresponds to the reduction of Co3O4→CoO, which normally happens
between 100 - 350°C and the second reaction (Eq.11) corresponds to the reduction of CoO→Co°, which
normally happens between 400 - 600°C.
As one can see in Figure 23, for JFR016 there is a major peak at 796°C and for JFR021 there are
two main peaks, the first at 842°C and the second at 899°C.
Both samples present peaks at high temperatures which does not correspond to the reduction of
CoO→Co°, since there is no peak corresponding to the reduction of Co3O4→CoO (between 100-350°C)
and as it is at a much higher temperature than usual (400-600°C). Actually, these peaks may suggest
the existence of cobalt silicates which are due to a strong interaction between the support and cobalt
oxide. These species are hardly reducible, thus, it can explain the existence of peaks at elevated
temperatures.
At this point, we can conclude that the hypothesis aforementioned of strong Co-O-Si interactions
may be responsible for the lack of catalytic activity due to difficulties in the Co reduction.
Furthermore, using HCl (JFR016) or HNO3 (JFR021) does not have influence in the cobalt
reducibility. However, as previously stated, the presence of Cl atoms can affect negatively the activity
of the catalysts in the FT synthesis. Thus, from now on, all the catalysts were synthesized with HNO3 in
the initial solution.
Figure 23 – TPR profile for JFR016 and JFR021.
0
0,2
0,4
0,6
0,8
1
1,2
0 200 400 600 800 1000
Hyd
rog
en
co
ns
um
pti
on
(mL
/g c
ata
lys
t)
Temperature (°C)
JFR016 (HCl)
JFR021 (HNO3)
47
2) Strong interaction between the silica precursor and the cobalt precursor
As it can be seen in Figure 23, hardly reducible cobalt species were being formed, which supports
the hypothesis of a strong interaction between silica precursor and the cobalt precursor. These oxides
(cobalt silicates) are often amorphous, which makes it harder to characterize those using conventional
techniques, such as XRD. A low cobalt content and a high surface area favor the formation of hardly
reducible oxides. Mixed cobalt – silicium or aluminum oxides can be formed during the catalysts
preparation, oxidative and reductive pretreatments, and in the course of FT reaction. [5]
Therefore, in this case if there is a strong interaction between cobalt and TEOS in the solution,
cobalt in the final powder could be ultra – dispersed which leads to cobalt silicate and/or very small
cobalt particles that are very hard to reduce, as seen in chapter 2.
In order to change the interaction between silica precursor (TEOS) and the cobalt precursor (cobalt
nitrate), were made three solutions at different pH. The formed oligomers are mainly linear and easily
condensable in an acid media (0<pH<2) and ramified in a less acid media (pH>2). The silica PZC is
approximately at pH=2, hence a solution was made approximately at a pH=PZC (as in the usual
procedure), and the other two at a pH above and below the PZC.
Consequently, changing the pH of the solution would affect the size and the dispersion of the cobalt
particles. Atomizing a solution with a pH below the PZC provides a silica surface positively charged,
hence repulsion will occur between Co2+ ions and the surface leading to a “non – homogenous
distribution” (creating perhaps cobalt domains and hence larger particles of cobalt species). On the other
hand, atomizing a solution with a pH above the PZC provides a silica surface negatively charged, and
the Co2+ ions will be “homogenous distributed”.
Consequently, the following samples were prepared: pH=1 (JFR019), pH=2 (JFR021) and pH=3
(JFR022).
Textural properties
Hereby is described the pH influence on the textural properties. In Figure 24 it is described the
nitrogen adsorption – desorption isotherms for the three samples (JFR019, JFR021 and JFR022), and
JFR016 is only represented as a reference (this solid was synthesized through an atomization solution
which contained HCl and present a pH approximately of 2).
48
Figure 24 - Nitrogen adsorption – desorption as a function of relative pressure and pore size distribution determined by BJH adsorption for JFR016 (reference), JFR019, JFR021 and JFR022.
In Table 12 it is summarized the surface area (BET surface), porous volume and pore diameter for
the previous samples.
Table 12 - BET surface, porous volume and pore diameter for JFR016, JFR019, JFR021 and JFR022.
Concerning the isotherms shape, one can see that JFR019 (pH = 1) is slightly different from the
others. JFR019 presents a hysteresis loop with more vertical lines, indicating the existence of a more
organized solid, which is confirmed by TEM analysis, further presented. The shape of JFR021 (pH = 2)
presents a hysteresis loop with more straight lines than JFR022 (pH = 3), but less pronounced than
JFR019.
All the three samples (JFR019, JFR021 and JFR022) present different porous size distributions.
This is due to different interactions between the surfactant and the inorganic species induced by the
different pH.
0
50
100
150
200
250
300
350
0 0,2 0,4 0,6 0,8 1
Ad
so
rbe
d V
olu
me
(m
L/g
)
Relative Pressure
JFR016
JFR019 (pH = 1)
JFR021 (pH = 2)
JFR022 (pH = 3)
Reference Porous type BET surface
(m2) Porous volume
(mL/g) Pore diameter
(nm)
JFR016
Mesoporous
164 0.26 6.5
JFR019 (pH = 1) 338 0.38 5.2
JFR021 (pH = 2) 345 0.46 7.1
JFR022 (pH = 3) 258 0.35 6.3
0
0,03
0,06
0,09
0,12
0 5 10 15
dV
/dD
vo
lum
e (
mL
/g)
Average pore diameter (nm)
49
JFR019 and JFR021 present higher surfaces areas than JFR022 due to a higher mesoporosity, as
JFR021 presents a higher porous volume, and, JFR022 presents a slightly higher porous volume, but a
much smaller pore diameter.
Therefore, is possible to conclude that regarding the textural properties, atomizing solutions with
different pH leads to solids with different properties.
Organization of the porosity of the silica matrix
In order to study the influence of the pH on the mesoporosity, low angles XRD analysis was
performed.
As one can see in Figure 25, for the sample JFR022 (pH = 3) there is no long distance organization.
While for JFR019 (pH = 1) and for JFR021 (pH = 2), it was confirmed the existence of a mesostructured
organization.
This can be explained, since a higher condensation rate is expected at pH = 1 and pH = 3. However,
JFR019 (pH = 1) is more organized than JFR022 (pH = 3) due to a better interaction between the
surfactant and the inorganic molecular precursor, as for each pH, the silica surface is differently charged.
The mesoporosity was also confirmed by TEM analyses, which are described later on.
Furthermore, one can conclude that the pH of the atomized solution has influenced in the
mesostructuration process, as it was expected.
Figure 25 - Low angles XRD patterns for JFR019, JFR021 and JFR022.
1 4 2 59 4 i*0.7 - F ile : C 1 5 K 2 58 3 .ra w - T yp e : P S D fa s t-s can - S ta rt: 0 .2 1 2 ° - E n d : 4.99 2 ° - S te p : 0 .01 7 ° - S te p tim e : 20 0 . s - T em p. : 2 5 °C (R o om ) - T im e S ta rte d : 0 s - 2 -Th e ta : 0.21 2 ° - Th e ta : 0.10 5 ° - A u x1: 0 .0 - A ux2 : 0.0 - A u x3 :
1 4 2 38 0 - F i le : C 15 K 2 5 8 0. raw - Typ e: P S D fa s t-sca n - S ta rt: 0. 21 2 ° - E n d: 4 .9 9 2 ° - S te p: 0 .0 1 7 ° - S tep t im e: 2 0 0. s - Te m p .: 2 5 °C (Ro o m ) - T im e S ta rte d : 0 s - 2-T he ta: 0 .2 1 2 ° - T he ta: 0 .1 0 5 ° - A u x1 : 0 .0 - A u x2: 0 .0 - A u x3 : 0 .0 -
1 4 0 97 1 - F i le : C 15 K 2 5 8 1. raw - Typ e: P S D fa s t-sca n - S ta rt: 0. 21 2 ° - E n d: 4 .9 9 2 ° - S te p: 0 .0 1 7 ° - S tep t im e: 2 0 0. s - Te m p .: 2 5 °C (Ro o m ) - T im e S ta rte d : 0 s - 2-T he ta: 0 .2 1 2 ° - T he ta: 0 .1 0 5 ° - A u x1 : 0 .0 - A u x2: 0 .0 - A u x3 : 0 .0 -
1)
1 4 2 13 9 - F i le : C 15 K 2 5 8 5. raw - Typ e: P S D fa s t-sca n - S ta rt: 0. 21 2 ° - E n d: 4 .9 9 2 ° - S te p: 0 .0 1 7 ° - S tep t im e: 2 0 0. s - Te m p .: 2 5 °C (Ro o m ) - T im e S ta rte d : 0 s - 2-T he ta: 0 .2 1 2 ° - T he ta: 0 .1 0 5 ° - A u x1 : 0 .0 - A u x2: 0 .0 - A u x3 : 0 .0 -
1 4 1 41 6 - F i le : C 15 K 2 5 8 2. raw - Typ e: P S D fa s t-sca n - S ta rt: 0. 21 2 ° - E n d: 4 .9 9 2 ° - S te p: 0 .0 1 7 ° - S tep t im e: 2 0 0. s - Te m p .: 2 5 °C (Ro o m ) - T im e S ta rte d : 0 s - 2-T he ta: 0 .2 1 2 ° - T he ta: 0 .1 0 5 ° - A u x1 : 0 .0 - A u x2: 0 .0 - A u x3 : 0 .0 -
B la n c - F i le : C1 5 N 25 8 4 .ra w - T yp e : P S D fa s t-s can - S ta rt: 0 .2 1 2 ° - E n d : 4.99 2 ° - S te p : 0 .01 7 ° - S te p tim e : 2 0 0 . s - T em p . : 2 5 °C (R o om ) - T im e S ta rte d: 0 s - 2 -Th e ta : 0 .21 2 ° - Th e ta : 0 .10 5 ° - A u x1 : 0 .0 - A ux2 : 0.0 - A u x3 : 0 .0 - D i
1 4 2 1 39 - L ef t A n g le: 0 .5 8 0 - R ig h t A n gle : 1 .1 9 8 - L e ft In t. : 5 5 9 8.1 00 C o u nts - Rig h t I nt .: 12 1 1 .4 2 9 Co u n ts - P e aks : 0 - P a ram s : 0 - W e ig ht : -1 .0 0 0 - kA 2 Ra tio : 0.5 - R e l. : 1 .8 0 8 % - T h .R .: 1 .5 46 % - R .In t. : 0 . 00 0 %
The cobalt reducibility as well as accessibility was studied by TPR analysis. In Figure 26, it is the
TPR profile for JFR019, JFR021 and JFR022 after calcination.
It is important to refer that JFR019 after calcination was black, which is the characteristic color of
cobalt oxide (Co3O4), whereas JFR021 and JFR022 were violet after the same post – treatment.
Figure 26 - TPR profile for JFR019, JFR021 and JFR022.
As one can see in Figure 26, for JFR019 there is a first peak at 375°C and a second peak at 816°C.
As the first peak is at a higher temperature than usual, it could correspond to the reduction of
Co3O4→CoO and/or to the decomposition of residual NOx groups (exothermic reaction). These NOX
groups are due to the cobalt precursor solution (cobalt nitrate) and the HNO3 used to acidify the solution,
because JFR019 is the more acidic solution so the more NO3- concentrated. [5] If one decomposes the
second peak, a part of it might correspond to the CoO→Co° reduction. And the majority of this peak, as
it happens at an elevated temperature may suggest, once more, the existence of cobalt silicates which
are due to a strong interaction between the support and cobalt oxide.
Regarding JFR021 it shows two peaks at 842°C and 899°C, and, JFR022 presents one peak at
825°C. These peaks may only suggest again the existence of cobalt silicates due to a strong interaction
between silica and cobalt oxide.
A XRD analysis was performed in order to characterize the crystalline cobalt oxide phases present
in the three catalysts.
Oxide phases
In order to know in which phase is the cobalt XRD analysis were performed. The results are
described in the following XRD diagrams. In Figure 27 A) the green line is refered to another sample
not mentioned in this part.
0
0,2
0,4
0,6
0,8
1
1,2
1,4
1,6
0 200 400 600 800 1000
Hyd
rog
en
co
ns
um
pti
on
(mL
/g c
ata
lys
t)
Temperature (°C)
JFR019 (pH = 1)
JFR021 (pH = 2)
JFR022 (pH = 3)
51
Figure 27 - XRD diagrams: A) JFR019 and B) JFR021 and JFR022.
As one can see in Figure 27, for JFR019 it was found a small peak of cobalt oxide (Co3O4) which
is in agreement with the TPR analysis. These particles were measured at 36.8° (2θ) and they have a
size of 22 nm.
Moreover, for all the three samples it was found a peak of metallic cobalt and this was not expected
since the samples were only calcinated, and not reduced.
It was expected after the conclusions taken by the TPR, which revealed the possible existence of
cobalt silicates, to found a peak of cobalt silicates in the XRD diagrams. The absence of this peak may
be justified by the cobalt silicates being amorphous, thus does not appear on the XRD diagrams.
Therefore, a XPS analysis was performed in JFR019 (which presented a first peak in TPR and a peak
of Co3O4 in XRD), and in JFR021. This analysis was not done in JFR022 since it present very similar
results to the ones of JFR021. The results from XPS are summarized in Table 13, which describe in
which form is the cobalt.
140971 - 142139
0 4-00 5 -96 5 6 (A) - C ob alt - C o - Y : 8 .4 7 % - d x b y: 1 . - W L: 1 .5 4 06 - C u bic - a 3.54 4 30 - b 3 .5 44 3 0 - c 3.54 4 30 - a lp ha 9 0 .0 00 - b eta 90 .0 0 0 - g a m m a 9 0.00 0 - Fa ce -ce ntered - Fm -3m (2 25 ) - 4 - 44 .5 2 37 - I/ Ic PD F 7 .3 - F7 =10 0 0(0 .0
0 0-00 9 -04 1 8 (A) - C ob alt O x id e - C o 3 O 4 - Y : 9.05 % - d x b y : 1. - W L : 1 .54 06 - C u b ic - a 8 .0 84 0 0 - b 8 .0 8 40 0 - c 8 .0 84 0 0 - a lp h a 90 .0 0 0 - b e ta 9 0.00 0 - ga m m a 9 0.00 0 - Fa ce-ce n te re d - F d-3 m (22 7 ) - 8 - 5 2 8.29 8 - F1 9= 3 4(0 .0 2 30 ,
O p e rat io n s: S li ts F ixed | Im po rt
1 42 1 39 - Fi le : F1 5K 15 6 85 .raw - Typ e: PS D fas t -scan - S tar t: 5 .0 17 ° - End : 7 1.99 5 ° - S te p: 0 .0 3 3 ° - S tep t im e : 49 9 .7 s - Te m p .: 2 5 °C (R oo m ) - T im e S ta rte d : 0 s - 2-T he ta : 5 .0 17 ° - Th eta: 2 .5 08 ° - Au x1 : 0.0 - Au x2 : 0.0 - Aux3 : 0.
O p e rat io n s: S li ts F ixed | Im po rt
1 40 9 71 - Fi le : F1 5K 15 6 82 .raw - Typ e: PS D fas t -scan - S tar t: 5 .0 17 ° - End : 7 1.99 5 ° - S te p: 0 .0 3 3 ° - S tep t im e : 49 9 .7 s - Te m p .: 2 5 °C (R oo m ) - T im e S ta rte d : 0 s - 2-T he ta : 5 .0 17 ° - Th eta: 2 .5 08 ° - Au x1 : 0.0 - Au x2 : 0.0 - Aux3 : 0.
Lin
(C
ps)
0
1
2
3
4
5
6
7
8
2-Theta - Scale
6 10 20 30 40 50 60 70
142380 -142594
0 4-00 5 -96 5 6 (A) - C ob alt - C o - Y : 8 .4 7 % - d x b y: 1 . - W L: 1 .5 4 06 - C u bic - a 3.54 4 30 - b 3 .5 44 3 0 - c 3.54 4 30 - a lp ha 9 0 .0 00 - b eta 90 .0 0 0 - g a m m a 9 0.00 0 - Fa ce -ce ntered - Fm -3m (2 25 ) - 4 - 44 .5 2 37 - I/ Ic PD F 7 .3 - F7 =10 0 0(0 .0
0 0-00 9 -04 1 8 (A) - C ob alt O x id e - C o 3 O 4 - Y : 9.05 % - d x b y : 1. - W L : 1 .54 06 - C u b ic - a 8 .0 84 0 0 - b 8 .0 8 40 0 - c 8 .0 84 0 0 - a lp h a 90 .0 0 0 - b e ta 9 0.00 0 - ga m m a 9 0.00 0 - Fa ce-ce n te re d - F d-3 m (22 7 ) - 8 - 5 2 8.29 8 - F1 9= 3 4(0 .0 2 30 ,
O p e rat io n s: S li ts F ixed | Im po rt
1 42 5 94 - Fi le : F1 5K 15 7 39 .raw - Typ e: PS D fas t -scan - S tar t: 5 .0 17 ° - End : 7 1.99 5 ° - S te p: 0 .0 3 3 ° - S tep t im e : 49 9 .7 s - Te m p .: 2 5 °C (R oo m ) - T im e S ta rte d : 0 s - 2-T he ta : 5 .0 17 ° - Th eta: 2 .5 08 ° - Au x1 : 0.0 - Au x2 : 0.0 - Aux3 : 0.
O p e rat io n s: S li ts F ixed | Im po rt
1 42 3 80 - Fi le : F1 5K 15 6 86 .raw - Typ e: PS D fas t -scan - S tar t: 5 .0 17 ° - End : 7 1.99 5 ° - S te p: 0 .0 3 3 ° - S tep t im e : 49 9 .7 s - Te m p .: 2 5 °C (R oo m ) - T im e S ta rte d : 0 s - 2-T he ta : 5 .0 17 ° - Th eta: 2 .5 08 ° - Au x1 : 0.0 - Au x2 : 0.0 - Aux3 : 0.
Lin
(C
ps)
0
1
2
3
4
5
6
7
8
2-Theta - Scale
6 10 20 30 40 50 60 70
JFR019
(A)
(B)
Co°
Co3O4
JFR021
JFR022
Co°
52
Table 13 - Results from XPS analysis for JFR019 and JFR021.
The results from XPS analysis, confirmed the existence of cobalt oxide in JFR019, and it reveals
the existence of Co2+ in both samples. XPS analysis is only able to specify that there is a specie in the
catalyst that presents an oxidation state of Co2+, but it could most likely be cobalt silicates.
Furthermore, it is interesting to note that no Co° was observed by XPS.
Organization of the porosity of the silica matrix and cobalt dispersion
In order to study with more detail the silica mesoporosity and the cobalt phase, TEM analysis were
performed. The following pictures show TEM micrographs for the previous three samples.
Figure 28 - TEM micrographs for the support of JFR019.
Regarding the silica matrix, as one can see in Figure 28, the porous have a worm-like shape. Also,
TEM showed that the pore size was approximately 5 nm, what is in agreement with the value taken from
N2 adsorption-desorption isotherm.
Figure 29 - TEM micrographs for metallic phase of JFR019.
Reference Co2+ Co3O4
JFR019 55,3% 44,7%
JFR021 100% -
Co3O4
Co°
53
The metallic phase (Figure 29) is present in the form of particles which are heterogeneously
distributed and ultra – dispersed onto the matrix.
Moreover, there are zones where cobalt particles were not visible. However EDS analysis showed
that Co is also present. Images in a dark – field mode did not show any diffraction, what makes possible
to conclude that cobalt, in these zones, is in an amorphous form. Besides that, cobalt was found in two
crystalline forms: cobalt oxide (Co3O4) and metallic cobalt (Co°).
For JFR019 was done a particle size distribution. The cobalt oxide particle size distribution in
volume is represented in Figure 30, and the size range is very broad going from about 10 to 50 nm.
Figure 30 – Cobalt oxide particles size distribution in volume for JFR019.
Concerning the samples JFR021 and JFR022 (Figure 31), globally they present the same
morphology. The oxide matrix presents a mesoporous domain as expected, and the structuration was
still happening, leading to a worm – like structure. The metallic phase was present in an ultra-dispersed
form, and, as seen in JFR019, it was found an amorphous phase rich in cobalt.
Figure 31 - TEM micrographs for the supports of: (A) JFR021 and (B) JFR022.
Conclusions
To conclude, it was found that JFR019 presented the most organized structure since it was
synthesized under the most acidic conditions, as previously stated.
Furthermore, these results are in agreement with the initial hypothesis: there is a peak in TPR at
elevated temperatures which is due to existence of cobalt silicates, however this species are not
detectable on XRD due to their non – crystallinity, and finally TEM analysis show that the cobalt in its
majority is in an amorphous form.
To sum up, there is a strong interaction between silica and cobalt, and changing the pH was not
enough to change the Si-O-Co interaction. However, with the solution at pH = 1 (JFR019), better results
were achieved, therefore more acidic conditions can be tested, to improve the solids obtained with TEOS
as inorganic precursor.
Silica conclusions
It was possible to load cobalt on a silica oxide matrix by spray – drying. Therefore, a solid with
cobalt already loaded (15 wt.% concerning the final catalyst mass) was synthetized by atomizing a
solution that had TEOS as inorganic precursor and acidified water with HCl at pH = 2. The resulting solid
was tested under a catalytic unit and there was no product formation. In order to modify the lack of
catalytic activity, the pH of the initial solution was modified.
First of all, the water was acidified with HNO3 instead of HCl to avoid chlorine atoms that can poison
active sites of the catalyst. It was supposed that the lack of activity was not mainly due to the presence
of Cl but to cobalt silicate. Also, the presence of HCl or HNO3 influences the mesostructuration process
of the solids. However, as chlorine atoms can affect the activity of the FT catalysts, all the solids started
to be synthetized with acidified water with HNO3.
Moreover, it was notice by TPR analysis the existence of peaks at high temperatures. This could
indicate the presence of cobalt silicates, which are hardly reducible species. [5]
In order to change the Si-O-Co interaction solutions with different pH were atomized. Three
solutions were made at pH = 1, pH = 2 and pH = 3. For all the samples different textural properties were
achieved (surface area, porous volume and pore diameter). Also, it was concluded that the pH has
influence on the mesostructuration process, as only the samples at pH = 1 and pH = 2 were
mesostructured. Moreover, it was concluded that the Si-O-Co interaction was only modified at pH =1,
however not enough to prevent the formation of cobalt silicates.
55
4.2. Alumina
Study of Aluminum chloride as inorganic precursor
The following paragraphs tackle the synthesis of alumina and cobalt alumina solids, using as
inorganic precursor AlCl3·6H2O. As previously mentioned, this study is not so detailed as the one of
silica, due to a larger knowledge and experience at IFPEN in synthetizing silica based materials by spray
- drying.
4.2.1. Supports
Generic characteristics and spray – drier parameters influence
JFR009 was synthesized in order to characterize the alumina supports produced by the ultrasonic
nozzle. Moreover, JFR010 was synthesized with a faster feed pump rate (2.1 mL/min) in order to know
the influence of this parameter in the support textural properties and to know if further atomizations could
proceed with this feed pump rate. Finally, the feed pump rate was set up at 2.1 mL/min.
Textural properties
Concerning the supports textural properties nitrogen adsorption – desorption isotherms were
performed, which are described Figure 32.
Figure 32 - Nitrogen adsorption – desorption as a function of relative pressure and pore size distribution determined by BJH adsorption for JFR009 and JFR010.
In Table 14 it is summarized the surface area (BET surface), the porous volume and the pore
diameter for all the previous samples.
0
50
100
150
200
250
300
350
0 0,2 0,4 0,6 0,8 1
Ad
so
rbe
d V
olu
me
(m
L/g
)
Relative Pressure
JFR009
JFR010
0
0,02
0,04
0,06
0,08
0 5 10 15 20dV
/dD
vo
lum
e (
mL
/g)
Average pore diameter (nm)
56
Table 14 - BET surface, porous volume and pore diameter for JFR009 and JFR010.
In Figure 32, one can see that the isotherm shape of JFR009 and JFR010 is far different from the
typical isotherm for a mesostructured solid, and it is possible to state that this solid is probably not
mesostructured. Furthermore, this statement is confirmed by low angles XRD analysis, which is
described in the following paragraphs.
Concerning the difference in the isotherm shape of JFR009 and JFR010 and the differences in the
pores diameter, a more detailed study on alumina oxide matrix should be performed in order to justify
these differences. But it seems that the organic molecule (surfactant) does not play the role of templating
agent as with silica matrix, because of the much lower pore size.
Only JFR009 was characterized with more detail, because it was synthesized in the same
conditions as silica supports. Thus, it is possible to make comparisons between them.
Morphology of the aerosol elementary particles
In order to characterize the aluminum elementary particles morphology produced through the
ultrasonic nozzle, SEM analysis was performed in JFR009.
Figure 33 - SEM micrographs for JFR009.
Comparing Figure 33 with Figure 16 (page 37), one can see that alumina elementary particles have
a more fragile aspect than silica elementary particles. This can be due to different hydrolysis,
condensation and micellization reactions, and also, because alumina is not mesostructured (described
in Figure 34). Also, it is possible to see that alumina elementary particles are less spherical than the
ones of silica, and, that alumina elementary particles are hollow inside, as they are crushed, which did
not happen with silica elementary particles.
Reference Porous type BET surface (m2) Porous volume (mL/g) Pore diameter (nm)
JFR009 Mesoporous
248 0.41 1.8
JFR010 258 0.47 3.2
57
Organization of the porosity of the support
In order to confirm the existence of a mesostructured solid, low angles XRD was performed which
is shown in Figure 34. It is possible to conclude that alumina is not mesostructured as there is no peak
in the low angles XRD analysis.
Figure 34 - Low angles XRD patterns for JFR009.
As previously mentioned, this lack of organization is not restricting in the final application of these
catalysts. The priority is to have a porous solid, which is the case but with a lower pore diameter.
4.2.2. Cobalt catalysts
Cobalt catalysts on an alumina matrix were synthesized containing 15 wt.% of cobalt, concerning
the final mass of catalyst. Following the same strategy as in silica, the presence of chloride in the catalyst
should be avoided. However, to understand the influence on having HNO3 instead of HCl, a trial was
done where the atomization solution contained with HCl (JFR023) and another where it contained HNO3
(JFR024). Also, some trials were made with aluminum nitrate (Al(NO3)3) as inorganic precursor, to avoid
the presence of chlorine atoms in the final solid.
One should note, that at this point it is not possible to know if the solids with cobalt loaded are
catalysts or not, as they were not catalytic tested. However, as a simplification they will be referred as
In Figure 35, one can see the nitrogen adsorption – desorption isotherms for the supports (JFR010
and JFR025) and for the catalysts (JFR023 and JFR024). One can note that JFR010 is just represented
as an oxide matrix reference, since it only differs in the presence of HCl or HNO3.
Figure 35 - Nitrogen adsorption – desorption as a function of relative pressure and pore size distribution determined by BJH adsorption for JFR010 (reference), JFR023, JFR025 and JFR024.
In Table 15 it is summarized the surface area (BET surface), the porous volume and the pore
diameter for all the previous samples.
Table 15 - BET surface, porous volume and pore diameter for JFR010, JFR025, JFR023 and JFR024.
As one can see in Figure 35, the isotherms shape change significantly with the presence of HCl or
HNO3. As previously mentioned in the supports sub – chapter, the supports (JFR010) synthetized
through an atomization solution which contained HCl are not mesostructured, and likelihood, the
catalysts which contained HCl (JFR023) are not also. Moreover, the isotherm shape of the support which
contained HNO3 is very different from the typical isotherm for a mesostructured solid, thus is reasonable
0
50
100
150
200
250
300
350
0 0,2 0,4 0,6 0,8 1
Ad
so
rbe
d V
olu
me
(m
L/g
)
Relative Pressure
JFR010 (Support, HCl)
JFR023 (Catalyst, HCl)
JFR025 (Support, HNO3)
JFR024 (Catalyst, HNO3)
Reference
Porous type
BET surface (m2)
Porous volume (mL/g)
Pore diameter (nm)
Supports JFR010 (HCl)
Mesoporous
258 0.41 3.2
JFR025 (HNO3) 260 0.45 3.1
Catalysts JFR023 (HCl) 65 0.15 2.9
JFR024 (HNO3) 121 0.19 1.8
0
0,02
0,04
0,06
0,08
0,1
0 5 10
dV
/dD
vo
lum
e (
mL
/g)
Average pore diameter (nm)
59
to affirm that both supports and catalysts containing HNO3 are not mesostructured. One should take into
account that to verify this statement a low angles XRD and TEM analysis should be performed.
Concerning the supports (JFR010 and JFR025), one can conclude that the introduction of HNO3,
almost does not modify the surface area, porous volume or pore diameter. Regarding the catalysts
(JFR023 and JFR024), using HNO3 instead of HCl increases significantly (variation of 86%) the surface
area. However there is a reduction in the pore diameter along with a reduction in the porous volume
with the introduction of HNO3 in the catalyst, instead of HCl.
The incorporation of cobalt on the alumina matrix changes significantly the textural properties,
reducing considerably the surface area, the porous volume and the pore diameter, whereas it is the
contrary with silica matrix. Therefore, the textural properties of the obtained solids are not satisfactory
for the application of these catalysts in the FT process. Hence, a further study should be performed in
order to improve these textural properties.
Cobalt reducibility
The cobalt reducibility as well as accessibility was studied by TPR analysis. In Figure 36 it is the
TPR profile for JFR023 and JFR024.
Figure 36 - TPR profile for JFR023 and JFR024.
After spray drying both solutions, the collected powder was light blue, which is the typical color of
cobalt aluminates. In order to confirm this supposition, a TPR analysis was performed.
For both samples there is only one peak at high temperatures which does not correspond to the
reduction of CoO→Co°, since there is no peak corresponding to the reduction of Co3O4→CoO (between
100-350°C) and as it is at higher temperature than usual (400-600°C).
Therefore, these peaks may suggest the existence of cobalt aluminates, which is in agreement with
the powder color, due to a strong interaction between the oxide matrix and cobalt oxide.
To completely avoid the presence of chlorine atoms in the final catalyst aluminum nitrate (Al(NO3)3)
was tested as inorganic precursor.
0
0,2
0,4
0,6
0,8
1
1,2
1,4
1,6
1,8
0 200 400 600 800 1000
Hyd
rog
en
co
ns
um
pti
on
(mL
/g c
ata
lys
t)
Temperature (°C)
JFR023 (HCl)
JFR024 (HNO3)
60
Study of Aluminum nitrate as inorganic precursor
In the following paragraphs it is described the synthesis of cobalt catalysts on an alumina matrix,
with 15 wt%., concerning the final catalyst mass.
Textural properties
In Figure 37 is described the nitrogen adsorption – desorption isotherms for JFR037 and JFR024.
One should note that the latter is only represented as a reference, as the inorganic precursor is the only
difference between the both samples.
Figure 37 - Nitrogen adsorption – desorption as a function of relative pressure and pore size distribution determined by
BJH adsorption: JFR024 and JFR037.
In Table 16 it is summarized the surface area (BET surface), the porous volume and the pore
diameter for all the previous samples.
Table 16 - BET surface, porous volume and pore diameter for JFR024 (reference) and JFR037.
Comparing the isotherms shape it is possible to affirm that JFR037 is not mesostructured, however
a low angles XRD and TEM analysis should be performed to confirm this.
0
20
40
60
80
100
120
140
0 0,2 0,4 0,6 0,8 1
Ad
so
rbe
d V
olu
me
(m
L/g
)
Relative Pressure
JFR024 (AlCl3)
JFR037 (Al(NO3)3)
Reference Porous type BET surface (m2) Porous volume (mL/g) Pore diameter (nm)
JFR024 Mesoporous
121 0.19 1.8
JFR037 32 0.11 12
0
0,01
0,02
0,03
0 10 20 30
dV
/dD
vo
lum
e (
mL
/g)
Average pore diameter (nm)
61
Concerning the differences in the pore diameter and in the surface area, a more detailed study on
should be performed in order to justify these differences, but it seems that with Al(NO3)3 the organic
molecule (surfactant) plays a role in building the porosity as it is wider than with AlCl3 precursor.
The increase in the pore diameter is positive, as the produced hydrocarbons chains would be
desorbed from the catalyst more easily. However, the porous volume it is not satisfactory for the FT
process, hence a more detailed study should be done to improve the textural properties.
Cobalt reducibility
The cobalt reducibility as well as accessibility was studied by TPR analysis. In Figure 38 it is the
TPR profile for JFR037 and JFR024. It is important to refer that after calcination, both samples were
black, which is the characteristic color of cobalt oxide (Co3O4), instead of violet for JFR024.
Figure 38 - TPR profile for JFR024 and JFR037.
As one can see, in Figure 38, the TPR profile for the samples obtained with Al(NO3)3 as inorganic
precursor have changed significantly.
If one decomposes the peaks of JFR037 it is found a first peak around 415°C, a second peak at
615°C and finally a third peak at 895°C.
For JFR037 sample the first peak likelihood corresponds to Co3O4→CoO reduction, whereas the
second peak may correspond to CoO→Co° reduction. Concerning the first peak, one can note that it is
at a more elevated temperature than usual, which might be explained through the decomposition of
residual NOX groups (exothermic reaction). This group exists due to the presence of NO3 in the inorganic
precursor, cobalt precursor and acidified water.
Finally, JFR037 presents a third peak that probably correspond to the existence of cobalt
aluminates. To confirm this, a more detailed characterization should be performed.
Alumina conclusions
The loading of cobalt onto an alumina matrix was possible. Moreover, it is believed that none of the
synthetized alumina based solids are mesostructured (due to the low angles XRD for the support and
to the isotherms shape).
0
0,2
0,4
0,6
0,8
1
1,2
1,4
1,6
0 200 400 600 800 1000
Hyd
rog
en
co
ns
um
pti
on
(mL
/g c
ata
lys
t)
Temperature (°C)
JFR024 (AlCl3)
JFR037 (Al(NO3)3)
62
To follow the same strategy as in silica, two solids with cobalt already loaded (15 wt.%, concerning
the final catalyst mass) were synthetized through two solutions containing HCl and HNO3 and aluminum
chloride as inorganic precursor. The loading of cobalt in the alumina matrix decreases the surface area,
porous volume and the pore diameter. One should note, that the achieved textural properties are not
satisfactory, therefore a further study should be performed to improve these. Moreover, for both cases
it was noticed a strong interaction between cobalt and alumina precursor, as in TPR analysis, there were
peaks at elevated temperatures.
In order, to avoid the presence of chlorine atoms in the final solid, aluminum nitrate was tested as
inorganic precursor.
One should note that the change in the aluminum precursor allows a modification in the interaction
between alumina and cobalt, hence, the reduction of Co3O4 → Co° was achieved.
63
5. Conclusions and future perspectives
The aim of this work is the synthesis of active FT catalysts by spray – drying. At this point, is not
possible to conclude if the synthetized solids are catalysts, as they were not catalytic tested. It is only
possible to affirm that it was possible to synthetize, by spray – drying, solids with a cobalt loading of 15
wt.% onto a silica and alumina matrix.
Concerning the solids synthetized onto a silica matrix, it was noticed that all of them present very
good textural properties for final application in the FT process. Nevertheless, the existence of cobalt
silicates was observed, due to high cobalt dispersion that lead to a strong interaction between the
inorganic precursor (TEOS) and cobalt. Therefore in order to reduce this interaction the pH of the
atomization solution was changed.
Only at pH = 1 better results were achieved, as there was a reduction of Co3O4 →Co°. However, it
was noticed the formation of cobalt silicates. Nevertheless, a more detailed study should be done, to
really evaluate which parameter could change the interaction between silica and cobalt. Also, it should
be understood the presence of metallic cobalt in samples that were not reduce. Depending on the
catalytic tests results, it should be understood where the cobalt nanoparticles are located on the support
and it could be tried the loading of a promoter metal in order to see if the cobalt reducibility was improved.
Moreover, concerning the solids synthetized onto an alumina matrix, it was noticed that the textural
properties are not satisfactory for the application in FT process. In solids produced with aluminum
chloride, as inorganic precursor, it was notice the existence of cobalt aluminates. Another aluminum
precursor, aluminum nitrate, was tested: the cobalt reducibility was improved since there was a reduction
of Co3O4 →Co° but, there were still cobalt aluminates. Though, concerning alumina matrixes a more
detailed study should be performed in order to improve the textural properties and also to completely
understand several differences that were reported in the results chapter.
This page was intentionally left blank
65
6. References
[1] BP Energy Outlook 2035, January 2014;
[2] B. Anantharaman, D. Chatterjee, S. Ariyapadi, R. Gualy, Hydrocarbon Processing (2012),
pp.47-53;
[3] C. Boissière, D. Grosso, A. Chaumonnot, L. Nicole, C. Sanchez, Advanced Materials, Vol.23