EPA/600/R-98/041
                                                 Apnl 1998
  Dehydrogenation Catalyst for
         Production of MtBE
                      by
          Josephine M. Hill, Randy D. Cortright,
           Dale F. Rudd, and James A. Dumesic
            University of Wisconsin, Madison
                 Madison, WI 53706
                   CR-822939
                  Project Officer

                   Paul Randall
             Sustainable  Technology Division
       National Risk Management Research Laboratory
                Cincinnati, OH 45268
NATIONAL RISK MANAGEMENT RESEARCH LABORATORY
      OFFICE OF RESEARCH AND DEVELOPMENT
     U.S. ENVIRONMENTAL PROTECTION AGENCY
              CINCINNATI, OH 45268

-------
                                        Notice

The U.S. Environmental Protection Agency through its Office of Research and Development
partially funded the research described herein under Cooperative Agreement CR-822939 to the
University of Wisconsin - Madison.  It has been subjected to the Agency's peer and administrative
review, and has been approved for publication as an EPA document. Mention of trade names or
commercial products does not constitute endorsement or recommendation for use.

-------
                                        Foreword
The U.S. Environmental Protection Agency is charged by Congress with protecting the Nation's
land, air, and water resources.  Under a mandate of national environmental laws, the Agency
strives to formulate and implement actions leading to a compatible balance between human
activities and the ability of natural systems to support and nurture life. To meet this mandate,
EPA's research program is providing data and technical support for solving environmental
problems today and building a science knowledge base necessary to manage our ecological
resources wisely, understand how pollutants affect our health, and prevent or reduce
environmental risks in the future.

The National Risk Management Research Laboratory is the Agency's center for investigation of
technological and management approaches for reducing risk from threats to human health and the
environment.  The focus of the Laboratory's research program is on methods for prevention and
control of pollution to air, land, water and subsurface resources;  protection of water quality in
public water systems; remediation of contaminated sites and ground water; and prevention and
control of indoor air pollution. The goal of this research effort is to catalyze development and
implementation  of  innovative, cost-effective environmental  technologies;  develop scientific  and
engineering information needed by EPA to support regulatory and policy decisions; and provide
technical support and information transfer to ensure effective  implementation of environmental
regulations and strategies.

This publication has  been produced as part of the Laboratory's strategic long-term research plan.
It is published and made available by EPA's Office of Research and Development to assist the
user community and to link researchers with their clients.
                                          E. Timothy Oppelt, Director
                                          National Risk Management Research Laboratory

-------
                                         Abstract

The objectives of this project were to better understand the effect of different catalyst preparation
parameters, the effect of different catalyst treatment parameters, and the mechanism of deactivation.
Accordingly, catalysts were made using various preparation methods and with a range of platinum
(Pt) and tin (Sn) contents. Catalysts were also tested under a variety of reaction conditions.
Catalysts were characterized using kinetic studies, chemisorption, transmission electron microscopy
and temperature programmed oxidation.  The results indicate that Pt/L-zeolite  catalysts are not
selective  without  Sn.  Sn reduces  the size of the Pt ensembles thus  inhibiting the  hydrogenolysis,
isomerization and coking reactions. At high Sn loadings, or after long periods  at high temperature
(-873 K) the  surface Pt ensembles become enriched with Sn and metal sintering occurs; thus, the
number of exposed Pt atoms and subsequently the activity, are decreased.  It is difficult to reverse
the results of these processes. Reacting isobutane at a lower temperature (-798 K) in the absence
of feed hydrogen reduces the Sn enrichment and the metal sintering.  At these conditions the main
mechanism of deactivation is coking, a process which is relatively easy to reverse with flowing
oxygen at a moderate temperature.

This report was submitted in fulfillment of grant CR-822939 by J.A. Dumesic and co-workers
under the sponsorship of the United States Environmental Protection Agency.  This report covers
a period from October 1994 to October 1997.
                                             IV

-------
                                   Table of Contents

Foreword	iii

Abstract	   iv

List of Figures	vi

List of Tables	viii

Acronyms and Abbreviations	,	ix

Acknowledgments  	•
   Chapter  1  Introduction	1

   Chapter  2  Conclusions  	•>
   Chapter  3  Recommendations  	   • ^

   Chapter  4  Experimental Procedures  	  *
     Catalyst Preparation  	 5
     Kinetics/TPO apparatus	   5
     Chemisorption apparatus	   7
     Transmission Electron Microscopy  	   ^
     Description of Measurement Parameters  	   7
     Quality Assurance 	 8

   Chapter 5 Results and Discussion  	  	  11
     Catalyst Preparation	  	   11
     Catalyst Treatment	  	   17
     Catalyst Deactivation	  	  22
     Summary of Preferred Conditions  ...  	  29
     Discussion	  	  29

References	34
                                            V

-------
                                     List  of Figures

1  - 1.   The production of MtBE from methanol and isobutane.. [[[  .2

4-1.   Schematic of kinetics  apparatus [[[   6

4-2.   Schematic of reactors [[[    6

4-3.   GC calibration curves  for methane,  ethane, propane  and n-butane                        9

4-4.   GC calibration curves  for isobutane  and isobutene [[[ 9

4-5.   Duplicate kinetic  runs [[[  10

5-1.   Effects of preparation methods of 1:2 Pt/Sn/L zeolite on isobutane conversion at 873 K, 1
         atm, and a H2/i-C4H10 ratio of 2 [[[   12

5-2.   Effects of preparation methods of 1 :2 Pt/Sn/L- zeolite on isobutane dehydrogenation
         selectivity at 873 K,  1 atm, andaH2/i-C4H10ratio of 2 [[[ 12

5-3.   Effects of platinum content on isobutane dehydrogenation reaction rates at 873 K, 1 atm,
         and a H2/i-C4H10 ratio of 2 over Pt/Sn/L- zeolite  (0.83 wt% Sn) ....................................... 13

5-4.   Effects of platinum content on isobutane dehydrogenation reaction rates at 873 K, 1 atm,
         and a H2/i-C4H10 ratio of 2 over Pt/Sn/L-zeolite (0.42 wt% Sn) ......................................  13

5-5.    Effects of tin content  on isobutane dehydrogenation  reaction rates at  873  K,  1 atm, and a
         H2/i-C4H10 ratio of 2 over Pt/Sn/L-zeolite (0.1 wt% Pt) [[[ 14

5-6.   Effects of tin content on isobutane dehydrogenation reaction rates at 873 K,  1 atm, and a
         H^i-C4H10 ratio of 2 over Pt/Sn/L-zeolite (0.5 wt% Pt) [[[  14

5-7.   Comparison of the isobutane dehydrogenation rates at 873 K, 1 atm, and a H2/i-C4H10 ratio
         of 2 [[[  15

5-8.   Comparison of the isobutene selectivity at 873 K, 1 atm, and a H2/i-C4H10 ratio of 2 ......... 16

5-9.   Comparison of the isobutane conversion at 873 K, 1 atm, and a H2/i-C4H10 ratio of 2.. ..... 16
5-10. Comparison of the isobutane dehydrogenation rates at 873 K, 1 atm, and a fL/i-C^H^ ratio

-------
5-14. Comparison of the isobutane dehydrogenation rates at 798 K over 1:2.5 Pt/Sn/L-zeolite
        with different feed streams	  21

5-15. Comparison of the isobutane conversion at 798 K and in i-C4H10 only over 1:2.5 Pt/Sn/L
        andPt/L.	  	  21

5-16. Comparison of the isobutane conversion at 798 K and in i-C4H10 only over 1: 1.4 Pt/Sn/L-
        zeolite andl:2.5 Pt/Sn/L-zeolite	  22

5-17. Comparison of the TPO spectra after 4 h reaction at 873 K ..1	  24

5-18. Isobutane conversions and dehydrogenation selectivities at 798 K in isobutane	26

5-19. Comparison of the TPO spectra after 4 h reaction at 798 K,  1 atm, and in isobutane only for
        L-zeolite-  and silica-supported catalysts	  26

5 - 20. Isobutene turnover frequencies at 798 K in isobutane	28

5-21. The activity of 1:2.5 Pt/Sn/L-zeolite at 873 K, 65 1  Torr He and 109 Torr iC4H10 before and
        after regeneration in 25% 0, in He at 300ฐC for 120 minutes	 28

5 - 22. Effects of regeneration on isobutane dehydrogenation reaction	 29
                                             vn

-------
                                      List  of Tables




5-1. CO uptake on fresh catalysts and after reaction at 873 K in 2: 1 H2:iC4Hio	 23



5-2. CO uptake versus reduction temperature for Pt/L	  23




5-3. CO uptakes and extents of carbon deposition on catalysts	  24




5-4. Isobutane uptake on catalysts	   25




5-5.  Preferred preparation and treatment  conditions..	  29
                                            vin

-------
            Acronyms  and  Abbreviations
LaB6            Lanthanum Boride
MtBE           Methyl  tert-butyl  ether
QA              Quality assurance
SCCM          Standard cubic  centimeters per minute
TEM            Transmission electron microscopy
TOF            Turnover  frequency
TPO            Temperature  programmed oxidation
WHSV          Weight  hourly  space velocity
                            IX

-------
                              Acknowledgments

We would like to acknowledge and thank Jane Bare and Paul Randall, EPA Project
Officers, for their help in this project.

Several undergraduates in Chemical Engineering at the University of Wisconsin have
helped with data collection. We would like to acknowledge the help of Janice
Schneider, Kyle Lewis, Lisa Gracyalny,  Mike Recupero and Willie Wiyatno.
                                        X

-------
                                          Chapter 1
                                        Introduction

Industrial catalytic reaction pathways for the production of methyl text-butyl ether (MtBE), an
important motor fuel oxygenate, typically involve 1) the production of synthesis gas from
hydrocarbon feedstocks, 2) the conversion of synthesis gas into methanol, 3) the production of
isobutene  by the  cracking, alkylation, isomerization and  dehydrogenation  of light  hydrocarbons, and
4) the conversion of methanol and isobutene into MtBE (see Figure 1-1). The weak link in this
sequence of steps  is  the  production of the isobutene, a presently inefficient reaction that requires a
chromium-based catalyst that must be regenerated frequently (approximately every half hour).
Chromium is a heavy metal that is known to be toxic.  Substantial advances in pollution prevention
could be achieved by the discovery of a clean and efficient catalytic technology for the production
of isobutene. The annual productions of MtBE and isobutene were 17.62 and 3.23 billion Ibs  in
1995[1].

Our research group has recently invented a new class of dehydrogenation catalysts that is capable
of achieving these pollution prevention goals in our laboratory.  This new catalyst, the UW
Dehydrogenation Catalyst, is  highly selective and  extremely active for the dehydrogenation  of
isobutane  to isobutene, with the formation of virtually no pollutant by-products, such  as methane,
ethane, propane, n-butane and carbon  on the  catalyst.  Specifically, the catalyst achieved a greater
than 99% selectivity to isobutene at an average conversion of 55% over a 10 day period. This is
better than that achieved by the commercial catalyst used in UOP's Oleflex process - 92%
selectivity and 50% conversion[2]. Furthermore, this non-chromium catalyst requires only
infrequent  regeneration.  A patent  will be issued for this new catalytic technology  through the
Wisconsin Alumni Research Foundation in the spring of 1998.

The favorable catalytic properties of the UW Dehydrogenation Catalyst are a result of a unique
combination of metals, modifier,  and support. Each of the four components of the Pt/Sn/K-L-
zeolite dehydrogenation catalyst is  important in  allowing  this  system  to be highly selective and
stable for isobutane  dehydrogenation. Platinum (Pt) catalyzes the removal of hydrogen from
isobutane to produce isobutene. The addition of tin (Sn) reduces the size of the surface platinum
ensembles and suppresses isomerization and hydrogenolysis reactions. The presence of potassium
(K) neutralizes the support, which  inhibits the  acid-catalyzed isomerization  and coking reactions;
moreover, potassium suppresses the competing isomerization and hydrogenolysis reactions, and
enhances the dehydrogenation rate. The L-zeolite micropore structure supports small Pt/Sn
particles and perhaps also stabilizes adsorbed isobutyl species.

While it is known that each of the four  components of the UW Dehydrogenation Catalyst are
necessary, it is not known how the various catalyst preparation  and  treatment  parameters influence
the reactivity of the  Pl/Sn/K ensembles  in the L-zeolite.  For example, it is known  that the addition
of both tin  and potassium are necessary to produce a selective  and  active dehydrogenation catalyst,
but the  optimum  amounts of these  components  have not yet  been established.  In addition, different
process conditions may  require different catalyst formulations. While the UW Dehydrogenation
Catalyst exhibits good  stability under reaction conditions, it still deactivates. It is necessary,
therefore, to  also  investigate the mechanism for  deactivation.

The objectives of this project were  to better understand  the effect of different  catalyst  preparation
parameters, the effect of different catalyst treatment parameters,  and the mechanism of deactivation.
Accordingly, catalysts were made using various preparation methods and with a range of Pt and Sn
contents. Catalysts were also tested with a variety of reaction conditions. The parameters by which
the performance of the catalyst was measured were isobutane activity, dehydrogenation selectivity
and stability. Chemisorption measurements and transmission electron microscopy (TEM) analysis,

-------
as well as kinetic studies, were used to characterize the catalysts. Temperature programmed
oxidation (TPO) was used to determine the extent of carbon deposition on the catalysts after
reaction.

                               n-Butane
                                   I
                             Isomerization
                                            3
                                     Isobutane
                          Dehydrogenation
                           iC4H10->IC4H8 + H2
                                    I
                        Methanol     Isobutene
                                    1
                            MTBE Process
Hydrogen
Figure 1-1. The production of MtBE from methanol and isobutene.

-------
                                          Chapter 2
                                         Conclusions

The metal particles of the Pt/Sn/K-L UW Dehydrogenation Catalyst are contained within the pores
of the L-zeolite.  This  catalyst is very selective for  isobutane dehydrogenation because the Sn
reduces the size of the Pt ensembles, thus hindering hydrogenolysis, isomerization and coking
reactions. Potassium also reduces the size of the Pt ensembles and neutralizes the support,
inhibiting the  acid catalyzed isomerization and  coking reactions.

In preparation of the catalyst, Sn must be introduced prior to Pt and no acidity should be introduced
by the metal precursors.  A Pt/Sn ratio of 1:2.5 and a Pt content of 0.5 wt% were found to be
optimal in terms of activity, selectivity and stability. Pt/Sn ratios lower than 1:2.5 (i.e. the addition
of more  Sn) had better stability  but did not significantly improve the selectivity and did decrease the
activity. With  less Sn the catalyst is not very selective.

In the pretreatment of the catalyst, a slow heating rate of less than 2 K/min should be used to obtain
higher activities. Calcination of the catalyst during pretreatment results in significant particle
sintering.

A feed of isobutane only (i.e. in the absence of hydrogen) can be used for the reaction at 798 K but
the stability of the catalyst is related to its selectivity.  For  example, the unmodified Pt/L  catalyst is
not very selective and  its activity  decreases to  essentially zero within minutes on  stream  in isobutane
only.  Conversely, the highly selective 1:2.5 Pt/Sn/L catalyst maintains its activity in isobutane for at
least 20 h on  stream.

At 873 K and in the presence of hydrogen, the Pt/Sn/K-L catalysts deactivate by coking, enrichment
of the surface with tin, and/or sintering of the metal particles. At 798 K and in the absence of
hydrogen, the major mode of deactivation for the Pt/Sn/K-L catalyst is coking.  The addition  of Sn
and K to supported Pt catalysts improves the resistance of these catalysts to coking. Both additives
decrease the size of the Pt surface ensembles, thereby suppressing the hydrogenolysis,
isomerization and coking reactions. The presence of Sn may also help to transport coke
precursors from  the  active sites  to  the support.

-------
                                         Chapter 3
                                    Recommendations

Based on the deactivation studies, the UW catalyst is most suited for use at temperatures of 798 K
and below. Low temperature applications for this catalyst may include the dehydrogenation of
higher paraffins. This application would be limited by the size and shape of paraffins that can fit
into the pores of the Pt/Sn/K-L-zeolite catalyst. Another application, in which Sn would play an
active role in the  chemistry  as  opposed to having only a  geometric effect  (i.e. reducing the  size of
the Pt ensembles), is the hydrogenation of saturated aldehydes to saturated alcohols.

The role of potassium should be further investigated. It may be possible  to take advantage of
potassium to reduce the carbon deposition on a working catalyst. With less carbon deposition the
catalyst  will have to be regenerated less  frequently.

-------
                                         Chapter 4
                                Experimental  Procedures
Catalyst  Preparation
The standard procedure for preparing a Pt/Sn/K-L-zeolite catalyst is as follows. The K-L-zeolite
(Tosoh) was calcined in dry air at 873 K for 20 h, dried in flowing He at 673 K for 2 h, and
transferred to a dry  nitrogen atmosphere. Tin  was then added to the L-zeolite using incipient wetting
impregnation with a solution of tributyltin acetate (Aldrich) in methanol. The catalyst was then
dried in air at 390 K for 2 h and treated at 573 K in a flowing mixture of 25 mol-% oxygen in
helium for 1 h. The Sn/L-zeolite was then transferred to a dry nitrogen atmosphere and the
platinum added using an aqueous solution of tetraammineplatinum nitrate (Aldrich). After
treatment at 533 K in a flowing mixture of 25 mol-% oxygen in helium for 1 h, the Pt/Sn/L catalyst
was reduced in hydrogen at 873 K for 6 h (temperature ramp of 2 K/min), transferred to a glass vial
in a dry nitrogen atmosphere and stored for further use. The concentrations of the Sn and Pt
solutions were adjusted to obtain the desired weight  percents of  each metal. The  absolute amount
of Pt was varied between  0.1 and 0.5 wt%, while the  Pt/Sn ratio was varied between  1:0 and 1:14.

Variations in this procedure have  been used to determine the effect of certain preparation variables.
Each of the following variations was done independently:  1. the Sn precursor was changed from
tributyltin acetate in methanol to  an aqueous  solution of stannic chloride, and 2. the order  of
impregnation of the metals was reversed (i.e. Pt  before Sn).

To determine the effect of the  support on the deactivation of the  catalyst,  silica-supported  platinum
catalysts were prepared  by ion-exchange, using the  method of Benesi et al. [3].  After ion-exchange,
the Pt/silica catalysts were filtered, washed with  deionized water and dried  overnight in air  at 390 K.
A portion of the catalyst was treated in flowing oxygen at 573 K for 2 h, followed by reduction in
flowing hydrogen at 673 K for 2  h.  Tin was  added to the remaining uncalcined Pt/silica catalyst by
evaporative impregnation of a tributyltinacetate/pentane solution,  after which the catalyst was dried
overnight in air at 390 K.  Potassium was added to some of the Pt/Sn/silica catalyst by incipient
wetness impregnation with an aqueous  solution  of KOH, and the catalyst was then dried in air at
390 K. The Pt/Sn/silica and Pt/Sn/K/silica catalysts were calcined with flowing oxygen at 673 K
for 2 h and  subsequently reduced in flowing hydrogen at 773 K  for 7 h.


Kinetics/TPO apparatus

The kinetics and TPO apparatus is constructed of stainless-steel, with an attached Gas
Chromatograph (GC) and Residual  Gas Analyzer (RGA). Figure 4-1 is a  schematic of the kinetic
apparatus.  The gas  out can be flowed to  the  RGA or directly vented if the mass spectrometer is not
used. The inlet gases are flowed through mass flow controllers  (Hastings). Isobutane (Liquid
Carbonic, 99.5%) is passed over  a bed of reduced nickel on alumina at 333 K to remove sulfur
impurities. Hydrogen (Liquid Carbonic) is flowed through a Deoxo unit (Engelhard) and a bed of
molecular sieves (13X) at 77 K. Helium (Liquid Carbonic) is flowed through Cu turnings at 423 K
and a bed of molecular sieves (13X) at 77 K. The reactor inlet and outlet gases are analyzed by an
HP-5890 GC with a TCD detector and  a  30 ft, 23% AT Chromosorb PAW column at 333 K. The
RGA is a Quad 25 OB Residual Gas Analyzer with an electron multiplier detector. The RGA is
usually operated at pressures of 10~3 Pa with an electron energy of 70 volts, and an emission current
of 150 [lamps.  Calibration of the Mass 44 signal is done using a gas mixture  of 5% CO, in He
(AGA).

-------
Figure 4-2 is a schematic of the reactors used. Both reactors are made from quartz. The U-tube
has a lower surface area and thus, minimizes gas phase reactions. The fritted reactor has a larger
catalyst capacity.
                                                                          Gas Out
 He
     DeOx
 1—'     S    k
        (liq.N2)
                                           Reaction
                                            Gas Out
                                              Furnace
iC4
Ni/AI2O3
OxyTrap
Figure 4-1. Schematic of kinetics apparatus.
                                a.
Figure 4-2. Schematic of reactors: a. U-tube, b. fritted reactor.

-------
Chemisorption apparatus

The vacuum apparatus used for chemisorption measurements is constructed from glass, with a
stainless steel dosing section and a Baratron capacitance manometer. Carbon monoxide (Liquid
Carbonic) is purified by passage through a heated quartz wool trap at 573 K to remove metal
carbonyl compounds, and then dried by passage through molecular sieves at 77 K.  Isobutane
(Liquid  Carbonic) is purified  with  three freeze/pump/thaw  cycles immediately  before starting
adsorption measurements. The gas uptake amounts ((imol/g) are calculated by extrapolating the
plateau of the isotherm to zero  pressure.


Transmission  Electron Microscopy

Transmission  electron  microscopic  studies  were  conducted  using a high resolution  microscope
(Phillips CM 200 equipped with a  LaB6 filament) at a point resolution of 0.2 nm and a
magnification range of 25x to l,600,000x. A spot size of 120 nm was used to minimize damage to
the specimens.  Samples were prepared by grinding  the catalyst in a mortar and pestle with filtered
water (18 Mohm deionised water). A drop of the solution was then placed on specially prepared
copper grids  (1000  mesh, Ted Pella Inc.).  Before adding the drop of solution, the grids were
dipped in an  adhesive solution made by  dissolving the adhesive from double sided tape in naphtha.
The sticky grid  with the drop  of catalyst  solution was then placed over a hot plate to evaporate the
water. The resulting grid contains many particles which overhang the edge of the grid bars and are
transparent to the electron beam.

Description of Measurement  Parameters

The following list is a definition of the terms used  to describe the experimental conditions and
results.

Pt-Sn ratio-   Wp'   * 118-69
Pt^n ratio.   ^^
Weight hourly space velocity (WHSV):
                                       fT' * v
                                       r   xiC.Ht
                      I   F" * r     \
                      I    r   -*ic H
Isobutane conversion:   1-—7-^—ijs- *100%


                                         K + kGCiCH
-------
where:
       F     inlet flow rate (seem)
       Fฐ     outlet flow rate (seem)
       GC^mo GC area of isobutane peak
       k,K   conversion factor for the GC
       M     mass of catalyst (g)
       msites   moles of sites per gram of catalyst (mol/g) (as defined by CO adsorption)
       nt     number of carbon atoms in the product compound
       wn    mass of Pt per gram of catalyst (g/g)
       wSn   mass of Sn per gram of catalyst (g/g)
       Xj     mole fraction  of products (methane,  ethane, propane,  propene, isobutene, butenes)
       xiC4Hซ  mฐle fraction  of  isobutene
       x;c4Hio mole fraction of isobutane


Quality  Assurance

Quality assurance (QA) objectives have been outlined for this project  in order to  ensure that
comparisons can be made between kinetic runs. These quality assurance objectives have been met
and thus, there exists a high degree of confidence in the conclusions drawn from this study.

Calibrations of the equipment (e.g.  mass flow  controllers, GC and Mettler balance) have been done
on a regular basis. Figures 4-3 and 4-4  show representative GC calibrations for the C1-C4
hydrocarbons.

As an additional verification of the  performance of the equipment, carbon balances are calculated
and checked for each kinetic run. In all cases,  the carbon balance was 100+-2.5%. The carbon
balance is given  by

                      _,  ,   ,  ,        Moles of Carbon to Reactor  ^,__„,
                      Carbon balance = —	——	———— * 100%
                                    Moles of Carbon out of Reactor

Duplicate kinetic runs have been done. Figure 4-5 is representative of the results from duplicate
kinetic runs.

This has been an exploratory project designed to investigate the effects of a number of process and
catalyst preparation variables on the performance of the UW Dehydrogenation Catalyst system. As
such,  there should be no  legal or regulatory ramifications if the QA objectives were not  met.

-------

.016
0014
f~
g 0.012
& 0.01
^0.00*
3 0.006
0.004
0.002
ฐ I


y
	 -y


/
\//'-
iiii
) 5

n nnndists j_ 1 QnROo riPv R
— v.vvv'tlooo + I .yUD• n .0
^ . -ฐ .---•'''
• _• *
.-*'
• ,ป
i i , , , i i , , , i .
105 1106 1.5 106
GC Area
Ooonfi
0.99998:
OQQQdl-
. yyyij .
1
i
-j
-
—
2 1








O6

Figure 4-3. GC calibration curves for •, methane, Q, ethane, 0, propane and •, n-butane.
        0.4

       0.35


        0.3
     c
    •S 0.25

    | 0.2


    I 0.15

        0.1

       0.05

          0
y = -0.0117 + 6.279e-09x R= 0.9966
   = -0.002368 + 6.603e-09x R= 1
            0      1  10'  2 10'  3  10'  4  10'  5 10' 6  10

                                  GC Area
Figure 4-4. GC calibration curves for •, isobutane and G, isobutene.

-------
         20
    c
    o
    
-------
                                          Chapter 5
                                  Results  and  Discussion
Catalyst  Preparation
It has been suggested that the favorable properties of Pt/Sn/L catalysts stem partially from the
interaction between the metals and the pores of the zeolite[4, 5]. Thus, one of the first tasks on this
project was to determine if indeed the metals  were within the pores of the zeolite. The TEM was
used to examine a 1:2.5 Pt/Sn/L catalyst. Very few particles were visible on the external surface of
the catalyst. It is then assumed that because the catalyst is active but particles are not visible, that the
metals must be within the pores. Subsequently, each catalyst was examined in the TEM following
preparation as a quality control  check.  If significant numbers of particles were visible then the
catalyst would be rejected and a fresh sample prepared.

The activity profiles (turnover frequencies (TOP)  versus  time plots) are  obtained using
approximately 5 mg of catalyst in a U-tube reactor with an isobutane conversion of less than 15%
of the equilibrium  value  for  the dehydrogenation reaction.  The TOF is on a total Pt basis unless
otherwise stated. Conversion and selectivity profiles are obtained using between 0.2 and 1 g of
catalyst in a U-tube or  fritted reactor with an  initial isobutane conversion close to or at the
equilibrium value for the dehydrogenation  reaction.

The effect of tin precursor and impregnation sequence is shown in Figures 5- 1 and  5-2. The
WHSV was 13.2 h'1 for these experiments and the catalysts  contained 0.5 wt% Pt.  When the
catalyst was prepared using a stannic chloride precursor  rather than tributyltinacetate, both the
conversion and selectivity decrease (see Figures  5-1 and 5-2).  The chlorine from the SnCl4
solution introduces  acidity to the  catalyst, which  promotes the isomerization and hydrogenolysis
reactions, thus lowering the selectivity of the catalyst. One of the features of K-L-zeolite is that it is
a neutral support.

The effect of impregnation sequence is also  shown in Figures 5-1 and  5-2. When  Sn is
impregnated first, the activity of the  catalyst is  significantly higher than  when  Pt  is impregnated
first. The latter catalyst has better selectivity for the first 5 hours on stream, after which time the
selectivities are similar. It is likely that when Sn is impregnated second, it blocks a large fraction of
the Pt sites and  the Pt that is  exposed is in  very small ensembles.   Small Pt ensembles are desirable
because the dehydrogenation reaction  is  structure  insensitive  while the  undesired side reactions,
isomerization and hydrogenolysis, are structure sensitive. Hydrogenolysis, isomerization and
coking reactions require several adjacent Pt surface sites[6, 7]. Thus, the dehydrogenation reaction
will be favored  on  small  Pt ensembles.

The effects of Pt and Sn content on activity were  studied by varying both the absolute amount of Pt
and the PttSn ratio. The results are shown in Figures 5-3 to 5-6. At low PtSn ratios the activity
decreases significantly.  The  catalyst activity is stable,  although low, with a sufficient amount  of Sn.
A Pt content of 0.5 wt% appears to be optimal for the catalyst in terms of activity.
                                               11

-------
         0s-
         o
         c
              100
80
60
                    3D OO   O
          8
                 o
                                                         n  an
0
                                         10         15
                                            Time (h)
                                                 20
25
Figure 5-1. Effects of preparation methods of 1:2 Pt/Sn/L zeolite on isobutane conversion at 873
    K, 1 atm, and a H2/i-C4H10 ratio of 2; sequential impregnation of: 0, Sn then Pt; •, SnQ4
    then Pt; and Q, Pt then Sn.
                                   Q   Q
                                          nnn
                                          10         15
                                            Time (h)
                                                 20
25
Figure 5-2. Effects of preparation methods of 1:2 Pt/Sn/L-zeolite on isobutane dehydrogenation
       selectivity at 873 K, 1 atm, and a H2/i-C4H10 ratio of 2; sequential impregnation of: 0, Sn
       then Pt; •, SnCl4 then Pt; and Q, Pt then Sn.
                                           12

-------
               20
          0>
          O
          CD
         Q
                 0
                              n   n
                              ,  I  .  .
          0
10          15
  Time (h)
                                                                 20
25
Figure 5-3. Effects of platinum content on isobutane dehydrogenation reaction rates at 873 K, 1

       atm, and a H2/i-C4H10 ratio of 2 over Pt/Sn/L-zeolite (0.83 wt% Sn); •   ,0.5 wt% Pt (1:2.7

       Pt/Sn); and Cl, 0.1 wt% Pt (1:14 Pt/Sn).
LL

O


 C
 O
          ง>:
          CD
          Q
               20
                15
                 0
                                             Time (h)
                                                                 20
                                                                   25
Figure 5-4. Effects of platinum content on isobutane dehydrogenation reaction rates at 873 K, 1

atm, and a H/i-C^ ratio of 2 over Pt/Sn/L-zeolite (0.42 wt% Sn);  •   ,0.5 wt% Pt (1: 1.4 Pt/Sn);

and Cl, 0.1 wt% Pt (1:7 Pt/Sn).
                                            13

-------
          03
          O)
          o
          0)
         Q
               20
               15
               10
                 0
                   0
10         15
  Time  (h)
Figure 5-5. Effects of tin content on isobutane dehydrogenation reaction rates at 873 K, 1 atm,
and a H/i-C4H10 ratio of 2 over Pt/Sn/L-zeolite (0.1 wt% Pt); O, 0.15 wt% Sn (1:2.5 Pt/Sn); •,
0.42 wt% Sn (1:7 Pt/Sn); and Q, 0.83 wt% Sn (1:14 Pt/Sn).
               20
         O
          O)
          
-------
Based on the results of the previous studies with various Pt and Sn contents, three  catalysts were
prepared and studied in more detail (see also the  results  in the Catalyst Deactivation  section). The
catalysts all contain 0.5 wt% Pt and Pt/Sn ratios of 1:0 (Pt/L), 1:1.4 (1:1.4 Pt/Sn/L) and 1:2.4
(1:2.4 Pt/Sn/L). The activities of these catalysts are shown in Figure 5-7. Note that the TOF is on
a per site basis, where the number of sites was determined by CO chemisorption. Without Sn the
Pt/L catalyst is  much less active for the production of isobutene.

The  selectivities of these  catalysts are  compared in Figure 5-8  and the  corresponding conversions  in
Figure  5-9.  Although the  isobutane conversions are similar over all three catalysts, the selectivities
are different. Note that the reactor configuration used for these experiments was designed for the
TPO experiments and significant gas phase reactions may have occurred. Therefore, the absolute
selectivities  are  lower than if the  experiments had been designed to minimize gas phase reactions (as
in Figure 5-2).   It  is evident that a certain amount of Sn is required to achieve a reasonable
selectivity as the unpromoted Pt catalyst is  very unselective  to  isobutene formation  and instead,
methane, ethane, propane and coke are formed.
            100
    O
             60
             40
          a.
    "O ,_ -
    t  ^
    Q
0      0.5      1       1.5     2      2.5
                          Time (hr)
                                                                       3.5      4
Figure 5-7. Comparison of the isobutane dehydrogenation rates at 873 K, 1 atm, and a
ratio of 2 over A, Pt/L-zeolite, •,1:1.4 Pt/Sn/L-zeolite and •   ,1:2.4 Pt/Sn/L-zeolite.
                                                15

-------
        100
    c
    05


    JD
    O
   E    20  I
    0
   —        c

             0     0.5    1
                                              .4....V..A ..
1.5     2     2.5     3
    Time  (h)
                                                            3.5    4
Figure 5-8. Comparison of the isobutene selectivity at 873 K, 1 atm, and a H2/i-C4H10 ratio of 2
over A, Pt/L-zeolite, •, 1:1.4Pt/Sn/L-zeoliteand•   ,1:2.4Pt/Sn/L-zeolite.
    c
    O

    'w

    c'
    CO
         100



          80



          60



          40



          20
                               (}
            0



              0    0.5    1     1.5    2     2.5    3    3.5     4
                                     Time (h)
Figure 5-9. Comparison of the isobutane conversion at 873 K, 1 atm, and a
over A, Pt/L-zeolite, •, 1 : 1 .4 Pt/Sn/L-zeolite and •, 1:2.4 Pt/Sn/L-zeolite.
                                                                            ratio of 2
                                          16

-------
Catalyst Treatment

The properties of the catalyst are affected by both the pretreatment (i.e.  the reactivation during
which the catalyst is taken from being exposed to air at room temperature to  being fully reduced at
the temperature  of reaction) and  the  reaction  conditions.  It has been  established that supported
metal catalysts must  be heated slowly to avoid sintering the metals and to avoid damage to the
support by the evaporating absorbed water (from exposure to the air) [8-10]. A study was thus
undertaken to determine how best to pretreat or  reactivate the Pt/Sn/L catalysts.

The following pretreatments were used with a 1:2.5 Pt/Sn/L catalyst (0.5 wt% Pt):

1. The catalyst was heated in a 10% H2 in He flow from room temperature (RT) to 400 K over 30
min, held for 1 h at 400 K, heated to 873 K over 2 h and then held at  873 K for 1 h in H2 (4 h 30
min  pretreatment).

2. The catalyst was heated in a 10% Hj in He flow from room temperature (RT) to 873 K at 2
K/min and then held at 873 K for Ih in Hj (6 h  pretreatment).

3. The catalyst was heated in He to 400 K over  1 h, held at 400 K for 2 h, heated to 533 K over 1 h
10 min, held at  533  K for 1 h, cooled to RT, and then reduced by heating from RT to 873  K over 16
h in a 10% H2 in He flow, held at 873 K for 1 h in H2 (22 h pretreatment).

4. The catalyst was heated in 25% 0, in He to 533 K over 2 h, held for  1 h at 533 K, cooled in He
to RT, and then reduced by heating from RT to 873 K over 16 h in a  10% H2 in He flow and held at
873 K for 1 h in H2 (20 h pretreatment).

The kinetic  measurements were performed with  approximately 5 mg of catalyst in a quartz U-tube
reactor.  After each pretreatment, the  samples were exposed to a 2:  1 ratio of F^ to isobutane  at 873
K and the activity measured. For TEM analysis, approximately 0.04 g of catalyst was placed in a
fritted reactor.  Instead of being exposed  to reaction conditions  after the  pretreatment, the  samples
for the TEM were cooled to RT in He and passivated with a flowing  stream of 0.2% 0, in He for
30 min.

The 1:2.5 Pt/Sn/L catalyst is essentially free of particles after preparation. After the pretreatments,
however,  particles are visible on the catalyst.  The number and size of particles  depends on the
pretreatment. The  pretreatment involving calcination before  reduction,  treatment #4, resulted in the
greatest amount  of particle migration and agglomeration with  80% of the catalyst having evidence of
particles, while the drying at 533 K followed by a  16 h reduction  (treatment #3) resulted in the least.
In the latter case, metal particles were only visible on approximately 10-20% of the catalyst. A few
more particles were visible on the 6 h pretreated (treatment #2) sample  and  considerably more
particles  were visible  on the 4 h 30 min pretreated (treatment #1) sample, with 50% of the catalyst
having evidence of particles.

The activities of the  catalyst after the pretreatments are compared in Figure 5- 10. Included in this
comparison,  is the activity of the  catalyst immediately after preparation (i.e.  after the 6 h reduction  at
873 K, the reaction was  started without cooling  to RT and  without exposing the  catalyst to air).  The
catalysts  that have been passivated and reactivated are  significantly less active than the catalyst that
had not been exposed to air. It has also been observed that if the Pt/Sn  catalysts are exposed to air
for extended periods  after the Sn or Pt impregnation  during preparation, significant  particle
agglomeration occurs. This agglomeration does not happen on a Pt catalyst without Sn, which
suggests the oxidized Sn species may be the cause for the metal agglomeration.
                                               17

-------
The activity of the catalyst pretreated for 4 h 30 min (treatment #1) is lower than that of the other
pretreatments. In this pretreatment, the catalyst is heated at 4K/min from 400 K to 873 K. A
decrease to 2K/min in the  6 h pretreatment appears to improve the activity  of the catalyst.
Decreasing the heating rate to 0.6K/min and preceding the reduction by drying at 533 K (22 h
pretreatment) does not seem to significantly improve the  activity, even though fewer particles  were
visible on the catalyst. Also, the activity of the calcined sample (20 h pretreatment) is  higher than
that of the catalyst pretreated for 4 h 30 min, even though significantly more agglomeration
occurred on the former catalyst. The TEM analysis may be somewhat misleading because the
particles that were seen were  probably  only a small fraction of the total  number  of particles in the
catalysts.

The other pretreatment variable that has been studied is the time for which the  catalyst is reduced
immediately before reaction. A 1:2.7 Pt/Sn/L and a 1:1.4 Pt/Sn/L catalyst were reduced for 1 or 16
h and their dehydrogenation activities compared (see Figures 5-1 1 and 5-12). A significant loss in
activity occurs when either catalyst is  reduced for 16 h compared with  1 h. This loss of activity
could be due to sintering of the metal particles or enrichment of the Pt particles with Sn. A 1:2.5
Pt/Sn/L catalyst was held at 873 K for 24 h in He and then reacted with 651  Torr He and 109 Torr
isobutane (see Figure 5-13).  In He, the reduction of the Sn should be significantly reduced. Again,
a decrease in activity occurs after the high temperature soak. This suggests that the temperature is
causing some sintering of the metals. TEM examination of a Pt/Sn/L catalyst after a 16 h soak at
873 K in H,, confirms that particle agglomeration does occur.

Finally, the reaction  conditions have been varied.  Figure 5- 14 is a comparison of the activity of a
1:2.5 Pt/Sn/L catalyst at 798 K in a pure isobutane feed, a 2: 1 hydrogen/isobutane feed and a 2: 1
helium/ isobutane feed.  The  activity is lower in the presence  of hydrogen  because the reaction is
negative order in hydrogen (i.e. the hydrogen  inhibits the production of isobutene). The catalyst,
however,  is more stable in the presence of hydrogen. In fact, a Pt catalyst without Sn has essentially
no activity when used in the absence of hydrogen (see Figure 5-15). A 1:1.4 Pt/Sn/L  catalyst is
active for isobutane conversion for a few hours on stream only, compared to a 1:2.5 Pt/Sn/L catalyst
which maintains reasonable conversion for 20 h on stream (see Fig 5-16). Thus, the catalyst must
be highly  selective to be used with an isobutane feed stream that does not contain hydrogen.
                                               18

-------

U_
P
1 —
C
O
C
0
O)
o
C
-o
JZ
o>
Q



ou
25

20
\n 15


10
1 W

5

n
\j

ti i i i i i i

A^
AAA A
* * A -
•
'HH^M^^^^^A
^^^^^^^^^^^V^r ^t ^& ^B ^m ^1
0AA__ ~ • V ^^
-I-J- *^^ ^ • • A A
++-H-++ + + I T '
T T -f +
.

1 1 1 1 1 1 1













0 0.5 1 1.5 2 2.5 3 3.5 4
Time (h)
Figure 5-10. Comparison of the isobutane dehydrogenation rates at 873 K, 1 atm, and a H,/i-
C4H10 ratio of 2 over 1 :2.5 Pt/Sn/L-zeolite. Pretreatment: A, none, • , 22 h, • , 6 h, •, 20 h, and +,
4 h30

U_
O
H
c
0
"•4— •
a
o
•o
>

-------
o
 c
 o
    O)
    0)
    Q
          20
          15
            0
          0
                                                    o0
                                      Time (h)
                                                           20
                        25
Figure 5-12. Effects of reduction time on isobutane dehydrogenation reaction rates at 873 K,
atm, and a H/i-C4H10 ratio of 2 over 1: 1.4 Pt/Sn/L-zeolite (0.5 wt% Pt, 0.42 wt% Sn); 0, 1 h
reduction and •   ,16 h reduction.
    LL
    O
    C
    o
    0)'
    o
    (D
    Q
        15
          0
              1111
                        1 1 1, 1,  , .  . ,,  1 1 1, 1, .  , ,
                 0.5
                             1.5
2.5
                                    Time  (h)
3.5
Figure 5-13. Effect of He soak on isobutane dehydrogenation reaction rates at 873 K, 1 atm, and
a 651 Torr He and 109 Torr i-C4H10 over 1:2.5 Pt/Sn/L-zeolite (0.5 wt% Pt, 0.83 wt% Sn); •, 1 h
reduction, and O, 24 h He soak.
                                           20

-------
II
o
c
g
ca
C
CD
0)
o
k_
TJ
<—

Q


J
'c/T
"co
rt
"S
"co

IL
^_ -
'co




roo i
80

60


40


20


0 t
_ • • • • i • • • • i • • • • i • • ' ' i ' ' ' ' i ' ' ' ' i ' ' ' ' i ' ' ' ' .
L ~
?AAAA :
P AAAAA A A A A A :
-•flJMH ••^B • • _ • T
••ง5 • • • • i
"**••• ~
	
•
— ""
.
', ... 1 .... 1 .... 1 .... 1 , ... 1 .... 1 .... 1 ... I
i n n i i <; 9 71 3 ? 1 i
                                       Time  (h)
Figure 5-14. Comparison of the isobutane dehydrogenation rates at 798 K over 1:2.5 Pt/Sn/L-

zeolite with different feed streams. A i-C4H10 only, •  ,2:1 H2/i-C4H10, and •   ,2: 1 He /i-C4H10.
        100
     ง  80  -

     'w
     CD
     >  60

     o
     O
     CD  40
         20  -
           0
J-
L
n
i i 1 i . i i ! . . i i 1 i .. • 1 i i i i 1 • • i • 1 i • i • ^
D n n n n n c
! ,1 ! ! . 1 k. 1 ! ill 1 1 .^ 1 1 . 1^1 1 . 1 lj 1 1 1 -A
             0     0.5     1     1.5     2      2.5     3     3.5     4
                                     Time  (h)
 Figure 5-15. Comparison of the isobutane conversion at 798  K and in i-C4H10 only over A, Pt/L-
 zeolite, and Q, 1:2.5 Pt/Sn/L-zeolite.
                                            21

-------

CcT
tf^
c
o
'eo
CD
c
o
0
CD
^™ซ
OS
"ZJ
.Q
O
W



1 UVJ


80


60


40
1

20

<
o
\j
C

1 1 1 1 1 1 1 1 1 1 • • 1 • 1 •

1
_ -
.
.
_ —
•
.
— ~

^
- "" "•

•
^_ :
•M . . *f i . . . . i . , , . I , . . i
) 5 10 15 2
Time (h)
Figure 5-16. Comparison of the isobutane conversion at 798 K and in i-C4H10 only over '
Pt/Sn/L-zeoliteandl  ,l:2.5Pt/Sn/L-zeolite.
1,1:1.4
Catalyst Deactivation

It is evident that deactivation is occurring in the activity profiles of the Pt/Sn/L-zeolite catalysts (for
example, see Figures 5-10 and 5-  11). As previously mentioned, coking, metal sintering and Sn
enrichment of the Pt particles can all  contribute to the deactivation of the catalyst.  Accordingly, the
catalysts were analyzed with TPO and TEM after 4 h of reaction at two conditions - 873 K in a 2:  1
hydrogen to isobutane feed stream and 798 K in a pure isobutane feed stream. The former
condition is similar to  commercial  conditions,  and the latter condition maximizes  coking while
minimizing  sintering.

For the  experiments at 873  K, approximately 0.75 g of catalyst was loaded into a quartz sample cell
which could be transferred between the kinetics/TPO apparatus  and  the chemisorption apparatus
without  exposing the catalyst to air. The catalyst was first reduced in Hj by heating overnight in a
 10% H2 in He flow to 873 K, held at 873  K for 1 h in H,, and evacuated for 30 min at 673 K. The
CO uptake was then measured at 298 K, after which the catalyst was heated to 873 K over 2 h in Hj
and treated for 4 h under reaction conditions (Hj to iC4H10 ratio of 2 at 873 K).   After purging the
catalyst with He at 673 K for 1 h to remove adsorbed Hj, the catalyst was evacuated at 298 K  for 2
h and the CO uptake was measured. Finally, the catalyst was heated to 873 K at  10 K/min in a 2%
O2 in He flow to obtain the TPO spectrum. Any carbon that has accumulated on the catalyst during
reaction will be removed as  CO, when the catalyst is heated in the presence of oxygen. During the
collection of the TPO spectrum, the exit gas stream from the catalyst was analyzed using a mass
spectrometer which monitored the mass 44 signal (mass 44 corresponds to the signal for CO2).
Calibration of the  mass 44 signal was achieved using a gas mixture of 5% CC>2 in He (AGA).
                                             22

-------
At the low space velocity of these experiments, the isobutane conversions  were initially  equal to
75%, 65% and 55% for the 1:1.4 Pt/Sn/K-L, Pt/K-L, and 1:2.4 Pt/Sn/K-L, respectively, and these
values decreased to -40% after 4 h on stream. The CO uptakes before and after reaction are given
in Table 5-1. An increase in the amount of Sn in the catalyst causes a decrease in the number of
sites on the fresh catalyst, as seen in other studies[ 1 1-13]. The CO uptake is low on the Pt/L
catalyst, in comparison with previous studies  of this catalyst system[ 14]. Therefore, CO uptakes
were measured on a Pt/L catalyst that was reduced at 673,798 and 873 K for 1 h (temperature ramp
of 2K/min). The corresponding CO uptakes are listed in Table 5-2. Higher temperatures of
reduction lead to  a decrease in the  CO uptake, which accounts for the low uptake on the fresh  Pt/L
catalyst. Similar results have been observed by Vaarkamp and co-workers[ 10]. After reaction, the
CO uptake decreases on  all catalysts;  however, the relative decrease in the number of sites is smaller
as the percentage of Sn in  the catalyst increases.

Table 5-1: CO uptake on fresh catalysts and after reaction at 873 K in 2:  1 H^iC^H,,).

Catalyst                                   CO Uptake (|imol/g)*
                                     Fresh               After Reaction
"1:2.4 Pt/Sn/L
1:1.4 Pt/Sn/L
Pt/L (II)
5.0
7.5
7.8
3.4
4.1
1.2
*error is +0.5 (imol/g
Table 5-2: CO uptake versus reduction temperature for Pt/L.

Reduction Temperature	CO Uptake (|imol/g)
673 K                                            21
798 K                                             15
873 K                                            10
The TPO spectra of the Pt/L, 1:1.4 Pt/Sn/L and 1:2.5 Pt/Sn/L catalysts after 4 h of reaction in
isobutane and hydrogen at 873 K are shown in Figure 5- 17. Peaks occur at approximately 680 and
780 K for the Pt/L and 1: 1.4 Pt/Sn/L catalysts. The TPO spectrum of the 1:2.5 Pt/Sn/L catalyst has
a broad peak between 600 and 760 K.  The amount of carbon on each catalyst is given in Table 5-3.
Increasing the amount of Sn in the catalyst decreases the amount of carbon deposited under these
reaction  conditions.

Analyses of the  catalysts by TEM indicate that sintering of the metal  particles occurs after reaction
at 873 K. Before the  reaction, the L-zeolite-supported samples have few particles  visible on  the
external surface of the zeolite, and these visible particles are 1-2 nm in size. The 1:1.4 Pt/Sn/L
catalyst looks similar to the 1:2.5 Pt/Sn/L catalyst before and after reaction. After reaction some
particles, up to 5 nm in size, are visible. However, even though more particles are visible after
reaction, the majority  of the zeolite specimen (-70%) does not show visible  particles. The Pt/L
catalyst is similar to the Pt/Sn/L catalysts  before reaction (i.e. few particles are visible), but after
reaction particles are visible on a significant fraction of the zeolite (-50%).  The average size of the
particles appears to be smaller (e.g., 1 nm) on the Pt/L catalyst  after reaction compared  to the
Pt/Sn/L sample.
                                              23

-------
Table 5-3: CO uptakes and extents of carbon deposition on catalysts.
Catalyst


Bsnf:n'W-O't-873 . K in iC^il(JH2
1:1.4 Pt/Sn/L
Pt/L(H)
Reaction at 798 K in iCJilo
1:2.5 Pt/Sn/L
Pt/UD
l:lPt/Sn/silica
l:l:3Pt/Sn/K/silica
Pt/silica
CO Uptake*
(jimol/g)
5.0
7.5
7.8
6.1
15
16
10
18
Carbon
(p.mol C/g)
90
170
500
260
560
770
210
2400
Deposition+
(|j.mol C/(imol site)
18
22
64
43

45
21
132
* error is +_0.5 ^imol/g
'error is +20%
     ,0)

      E
      o
      c
      0
      o
     w
       CM
     o
     o
                 Pt/L

     1:1.4Pt/Sn/

1:2.5 Pt/Sn/
               300      400      500      600       700
                                     Temperature (K)
                                       800
900
Figure 5-17. Comparison of the TPO spectra after 4 h reaction at 873 K, 1  atm, and
ratio of 2 for Pt/L-zeolite, 1:1.4 Pt/Sn/L-zeolite and 1:2.5 Pt/Sn/L-zeolite.
The uptake of isobutane was measured on the 1:2.5 Pt/Sn/L catalyst before and after reaction. The
adsorption of isobutane occurs primarily on the L-zeolite  (the  Pt/Sn/L sample contains  a relatively
small number of surface Pt sites), and this uptake is thus a measure of the accessible interior
surface area of the catalyst. The change in uptake is negligible within the error of the experiment
(see Table 5-4).  This  results is consistent  with the observation that only a small  amount of carbon
is deposited on the catalyst during 4 hours of reaction.
                                             24

-------
Table 5-4. Isobutane uptake on catalysts.

Catalyst                                       Isobutane Uptake ((imol/g)*
_ Fresh _ After Reaction
L-zeolite                                       713
Reaction at 798 K in iCJiw
Pt/L(I)                                        603                    528
Reaction at 873 K in iCfl^ H2
1:2.4 Pt/Sn/L _ 568 _ 545
* error is +15
Measurable  catalytic activity for isobutane conversion in the absence of Hj is maintained by all
catalysts for reaction times of 4 h, except for the Pt/L- zeolite catalyst (see Figure 5- 18). The
isobutane conversion of the 1:2.5 Pt/Sn/L catalyst is lower than that of the Pt/Sn/silica and
Pt/Sn/K/silica catalysts because the reactor contained only 3 [imol of sites compared to 8 and 5
(imols of sites,  respectively, for the silica-supported catalysts. The activity  of the Pt/L catalyst
decreases to essentially zero within 30 minutes on stream. For all catalysts except Pt/Sn/K/silica,
the selectivity  increases significantly between  1 and 30 minutes on stream,  as shown in Figure 5-18.
This increase in selectivity may be  caused by the formation  of coke which decreases the size  of the
Pt ensembles, thus suppressing the isomerization and hydrogenolysis  reactions compared to the
dehydrogenation reaction.

Figure 5-19 shows TPO spectra for catalysts after 4 h reaction at 798 K in isobutane. The TPO
spectrum of the Pt/silica catalyst has peaks at 7 10 and 790 K, while the 1: 1 Pt/Sn/silica catalyst
displays broad peaks at 690,760 and 820 K. The spectra of Pt/L and  1:2.5 Pt/Sn/L have peaks
near 680 and 780 K, while the 1: 1:3 Pt/Sn/K/silica has a peak around 680 K. The peak at 360 K in
the spectrum of the 1:2.5 Pt/Sn/L catalyst is from weakly bound carbon on the  surface, that can be
removed by purging the catalyst with He at 673 K for 1  h. Comparing Pt/silica with Pt/Sn/silica,
and comparing Pt/L with Pl/Sn/L, it appears that Sn suppresses the extent of coke formation and it
appears to alter  the relative areas of the high  and low temperature peaks,  (i.e., decrease the ratio of
the 680 peak to the 780 K peak). The addition of K to the Pt/Sn/silica catalyst significantly
decreases the amount of carbon deposited on  the catalyst.
                                              25

-------
   c
   a
   >
   c
   o
   O
   CD

   I
      100
       80
       60
40
20
   -    0
        V
        o
        a
              0.5
v
a
                ^
                 a
                   u
           1.5    2    2.5    3
          Time on Stream (h)
                                         3.5
 A    A    :     x  11    %   i    j
'    •     VI..M..;..!
j.... .^,,, i j-.... +... .^ •.. .^ •. • ปj • • •  ~
100

90  ฎ
   •<
80  I
   (Q
    CD
70  |
    6'
60  =
                                                    CD
                                   50

                                   40

                                   30
Figure 5-18. Isobutane conversions (filled symbols) and dehydrogenation selectivities (open
symbols) over 1:2.5 Pt/Sn/K-L (circles), Pt/K-L (squares), 1:1 Pt/Sn/silica (diamonds), 1:1:3
Pt/Sn/K/silica (triangles), and Pt/silica (inverted triangles) at 798 K in isobutane.
     CD
     .4—ป
     'SP
    m
     C
     0
     LU
         0.2
        0.15
         0.1
        0.05
      CM
     O
     o
                       Pt/L
             1:2.5 Pt/Sn/L
         1:1 PVSn/Silica
              . .  . .1;1,:3 Pt/Sn/K/SiNca,
            300    400     500     600    700     800     900
                            Temperature (K)
Figure 5-19. Comparison of the TPO spectra after 4 h reaction at 798 K, 1 atm, and in isobutane
only for L-zeolite- and silica-supported catalysts.
                                  26

-------
The total carbon accumulated on the catalysts  was determined from the area under the TPO spectra,
where the baseline for integration was taken to be the CO, evolution at 300 K. Table 5-3 lists the
amounts of carbon deposited on the basis of catalyst mass and also based on the number of sites
determined by CO chemisorption on the fresh catalysts. The order of decreasing amount of carbon
per site is Pt/silica ป1:1 Pt/Sn/silica > 1:2.5 Pt/Sn/L > Pt/L >1:1:3 Pt/Sn/K/silica.

To  obtain  deactivation profiles  corresponding  to  the above  reaction conditions,  approximately  0.005
g of catalyst was loaded into a quartz U-tube reactor with quartz wool.  The deactivation  profiles for
the 1:2.5 Pt/Sn/L and Pt/L catalysts are compared in Figure 5-20, at conditions for which the
isobutane conversion was  less than 10% of the equilibrium  value. The Pt/L catalyst deactivates
rapidly,  whereas  the Pt/Sn/L catalyst has a high initial  activity, deactivates over the first 30 minutes
on stream,  and then continues to deactivate at a slower rate.  The  deactivation behavior of the
Pt/Sn/L catalyst is typical for many catalysts[ 15-17].

Analyses of the  catalysts by TEM before and after reaction in isobutane at 798 K indicate that little,
if any, particle agglomeration occurs during the  reaction. Essentially no particles  are  visible on the
L-zeolite-supported catalysts after reaction at 798 K. The silica-supported catalysts have particles
visible with diameters from 2-5 nm on  the  support before and after reaction.  Isobutane uptake was
measured on the Pt/L catalyst before and after reaction (see Table 5-4). After reaction, the  uptake
decreases by 75 |imol/g. This result indicates that carbon deposition had occurred within the pores
of the zeolite, and this pore blocking may account for the rapid decline with time in catalyst activity.

Regeneration of the Pt/Sn/L catalyst has been attempted. Figure 5-21 shows  the activity of a 1:2.5
Pt/Sn/L catalyst at 873 K in 65 1 Torr He and 109 Torr isobutane. After 23 hours on stream, the
catalyst was cooled to 573 K and held at this temperature for 2 hours in a flowing stream of 25%
O2 in He.  The catalyst was then purged with He for 30 minutes, heated to 873 K at 8 K/min in 5%
H2 in He, held for 2h in H2 at 873 K and exposed to reaction conditions.  Only a fraction of the
initial activity  was recovered.

A different regeneration scheme was successful in reestablishing the performance of a 1:2.7
Pt/Sn/L catalyst, provided that the catalyst was reduced for 16 h prior to the initial reaction and after
each regeneration. Figure 5-22 shows isobutane dehydrogenation reaction rates over fresh 1:2.7
Pt/Sn/K-L-zeolite (0.5 wt% Pt) at 873 K, 1 atmosphere, and a Hj iC4H]0 molecular ratio of 2.0.
This catalyst was pretreated with H2 at 873  K for 16 h before reaction and was allowed to deactivate
over 80 h  at reaction conditions.   The catalyst was then regenerated with a 10% 0, in He mixture
for 2 h at 533 K, cooled to room temperature, reduced with a 10% H2 in He mixture  by increasing
the temperature from room temperature to 873 K over 8 h, and then further reduced in pure H^ at
873 K. After this regeneration, Figure 5-22 shows that the catalyst exhibited  essentially the same
dehydrogenation  activity and  stability.  After running for 50 h, the catalyst was regenerated using
the same procedure for a second time. After the second regeneration, the catalyst exhibited slightly
lower dehydrogenation reaction rates compared to the fresh catalyst. The catalyst was allowed to
deactivate for 24 h after this second regeneration and then was regenerated a  third time, The third
regeneration followed  a procedure  similar to the first two regenerations, but the  catalyst was
regenerated with 10% 0,  in He mixture for 2 h at 573 K instead of 533 K. The catalyst  was then
exposed to isobutane at 873 K in the absence of hydrogen. The catalyst was regenerated a fourths
time using the same procedure  as  the third regeneration. After this fourth regeneration, the catalyst
exhibited slightly lower dehydrogenation activity  compared  to the fresh catalyst. However,  there
was little change in the performance of the catalyst between the  second  and fourth regenerations.
                                               27

-------


it-
CD
CD
CD
"tj
.a
o
CO


1 UU
80 '
60
tn.
40



20
n i
." ' ' ' ' ' ' ' ' " ' ' ' 	 	 ....,...._
L
:ฐฐ0ฐooOoo 0 n :
0 ฐ 0 0 o (
.
-
I
.
_
— •"
tmmm ฑmim*t i i i ซl i i i id i i i i W ' i ' 4' ' ' ' 4' ' ' ' 1
                    05     1
                                      Time (h)
     35
Figure 5 - 20. Isobutene turnover frequencies at 798 K in isobutane over O, 1:2.5 Pt/Sn/K-L, and
•, Pt/K-L.
                                   10         15
                                      Time (h)
                                                              • o
20
25
 Figure 5-21. The activity of 1:2.5 Pt/Sn/L-zeolite at 873 K, 651 Torr He and  109 Torr iC4H10
 before and after regeneration in 25% 0, in He at 573 K for 120 minutes; •, fresh catalyst and 0,
 regenerated  catalyst.
                                            28

-------

1 1
LL.
O
c
o
1-5
0 C-
o
•o
.c
CD
Q
8
7
6
5
4
3
2
1
            0
                                       -I—1—I—I—I—I—I—I—[-
               0
20
40          60
  Time  (h)
80
100
Figure 5 - 22. Effects of regeneration on isobutane dehydrogenation reaction rates at 873 K, 1 atm,
and a H^/i-C^ ratio of 2 over 1:2.7 Pt/Sn/L-zeolite (0.5 wt% Pt, 0.83 wt% Sn); •, fresh catalyst;
0, regenerated catalyst (1st Regeneration); A, regenerated catalyst (2nd Regeneration); and 0,
regenerated catalyst  (4th Regeneration).
Summary  of Preferred  Conditions

The  following table lists  the preferred preparation and pretreatment conditions  for the
dehydrogenation of isobutane using the UW Dehydrogenation Catalyst.

Table 5-5: Preferred preparation and treatment conditions for the catalyst.

Variable	Preferred Value	
PtSn Ratio
Pt Content
Sn Precursor
Impregnation Order
Pretreatment
                      1:2.5
                    0.5  wt%
               Tributyl tin acetate
                Sn first, Pt second
              Heating rate < 2K/min
                 No calcination
          1 h reduction in EL, at rxn temp
Discussion

The goal of this project was to learn more about why the UW Dehydrogenation catalyst is such a
good  catalyst for isobutane  dehydrogenation and how the  preparation  and pretreatment parameters
could be chosen to achieve  maximum performance from the catalyst. It is known that small Pt
                                              29

-------
ensembles are essential to achieve high activity, selectivity and stability[ 11]. Our results indicate
some of the factors which affect whether the catalyst will have and/or maintain  small Pt ensembles.

The  first task in this project was to determine the location of the metal particles  on the zeolite
support.  The TEM analysis indicated that the particles  were within the pores of the zeolite after
preparation. The TEM results support other work which showed that Pt/Sn/L catalysts  were 50
times more active than Pt/Sn/silica catalystsf 18]. One  of the differences between  silica and L-
zeolite is their pore structure. The silica used in the study was non-porous. L-zeolite, however, has
a unidimensional pore structure. It has been suggested that when the reaction occurs within the
pores of the zeolite, favorable  interactions between reaction intermediates and the pore walls can
occur, thus promoting  the rate of reaction.

As well  as the location of the  metals, the interaction between the metals is also  important.  Tin is
necessary to make the catalyst  selective (see Figure  5-8) but the  exact role  of tin is not  fully
understood[19]. It is likely that the interaction between the Pt and Sn changes during preparation,
during pretreatment and over the course of  the reaction as the environment to which the  catalyst  is
exposed changes. Previous work on  these catalysts has shown that platinum and tin form an alloy
after high temperature reduction[5, 6]. In the preparation of the catalyst, the final step is a 6 hour
reduction at 873 K, however, the catalyst is  then exposed to air when it is  transferred from the
storage vial to the reactor. Work on  model Pt/Sn systems on silica films has shown that oxidation
will  destroy a Pt/Sn alloy and cause  Pt and Sn to separate[20]. The pretreatment is designed to
reduce the metals and reform the alloy. It is shown from the pretreatment  studies herein that the
pretreatment method does affect  the  performance of the  catalyst.  In particular,  the heating rate and
atmosphere are important (see Figure 5-10). Calcination at 533 K before  the reduction results in
significant  sintering  of the  catalyst particles. This result is in contrast to the preferred treatment for
Pt/L  catalysts  in  which a calcination  before  reduction is required to achieve high dispersions[21-
23].  The higher activity  observed with a  slower  rate of heating is  consistent with the results of other
studies[8,9].

In a related investigation[ 11], it was shown that increasing the amount of tin relative to platinum on
K-L-zeolite above a 1:2.5  Pt/Sn ratio, marginally increases the dehydrogenation selectivity, but  at
the expense of decreased catalyst activity and stability. The  results of this  study confirm those
results (see Figures 5-3 to 5-6). If the Sn content is too high, then the activity decreases
significantly. In terms  of selectivity,  however, the Sn content  cannot be too low or the catalyst is not
highly selective  (see Figure 5-8). The selectivity of the catalyst  is also reduced  if acidity is
introduced (see Figure 5-2). If the catalyst is less selective, it will also be less stable. The TPO
results show that the addition of Sn improves the resistance of the K-L-zeolite-supported Pt
catalysts to the formation of coke during reaction at 873 K (Figure 5-17 and Table 5-3).  In
addition, Table 5-1 shows that the higher coke content blocks a higher percentage of sites for CO
adsorption on the Pt/K-L catalyst than on the Pt/Sn/K-L catalysts. Tin improves the resistance of
the catalysts to coking by  reducing the Pt  ensemble size, thus preventing  the undesired  reactions
which lead to coke formation.

Although the amount of carbon deposited on the catalyst has an important effect on catalytic
activity,  it is also  important to  address the location of the carbon (e.g., whether carbon is on the
active sites or on the support) [24]. It has been shown that the lower temperature peaks in TPO
spectra  generally  correspond to carbon on  Pt, while the higher  temperature peaks  correspond to
carbon on the supportf 15, 17, 24-26]. Liwu et al. [ 17] studied carbon deposition on Pt/A^Qj and
Pt/Sn/A^Os after the dehydrogenation of n-butane. They observed that Sn decreased the ratio of
the carbon deposited on the metal to the  total carbon deposited.  In  agreement with this  behavior,  it
has been observed in the  studies herein, that the addition of Sn decreases the amount of carbon
associated with Pt on both the silica- and L-zeolite-supported Pt catalysts. In addition, the CO
                                               30

-------
uptakes before and after reaction (see Table 5-1) indicate that fewer sites are lost on the catalyst as
the Sn content is increased.

It has been suggested that Sn transports coke from the active sites to the support on Pt/Sn/alumina
catalystsf 17, 27, 28]. Lieske et al. [27] proposed a drain-off mechanism to explain the stabilizing
effect of Sn on Pt/alumina catalysts. The presence of Sn decreases the size of the surface Pt
ensembles so that the hydrocarbon cannot readily form multiple carbon-metal bonds. Because the
adsorbed species is  attached less strongly to the metal surface, it is more mobile and can more
easily migrate to the support. This effect is evident by the change in the ratio of the peak heights in
the TPO spectra with the addition of Sn to the catalysts.

The ability of Sn to improve the resistance of the  Pt catalyst to carbon  accumulation becomes more
evident when the reaction is run with an isobutane feed in the absence of feed hydrogen. In the
absence of hydrogen, carbonaceous species can accumulate on the surface of the catalyst and block
the active sites. In agreement with this behavior, the TPO results shown in Figure 5-19 and Table
5-3 indicate  higher levels of coke after reaction in  the absence of hydrogen at 798 K compared to
reaction in the presence of  hydrogen at 873 K.  A  Pt/L catalyst deactivates  to near zero conversion
in minutes (see Figure 5-20), a 1: 1.4 Pt/Sn/L catalyst deactivates to essentially zero conversion in 4-
5 hours on stream, while the highly selective 1:2.5 Pt/Sn/L catalyst maintains over 20% conversion
(equilibrium conversion for these conditions is 42%) for 20 hours on stream in isobutane at 798 K
in the  absence of feed hydrogen (see  Figure 5-16). The TEM results showed that little, if any,
particle agglomeration occurs during the reaction  at 798 K in the absence of hydrogen and so
coking is the  main mechanism of deactivation.  In the presence of hydrogen the stability of the
catalyst is better but the activity  will be lower because the reaction is negative order in hydrogen (see
Figure 5-14). Hydrogen is produced by the dehydrogenation reaction and the reactor configuration
can affect the product stream (see Figure 5- 15).

The deactivation studies on the silica catalysts indicate that K, as  well as Sn, improve the resistance
of the catalyst to carbon deposition (see Figure 5-19 and Table 5-3). As well as decreasing the size
of the surface Pt ensembles, K may also hinder the dehydrogenation of adsorbed species. It has
been shown that the addition of K to Pt/Sn/silica decreases  the CO uptake but increases the H2
uptake[29].  Barbier [30] has shown that the addition of K to alumina catalysts decreases the peak
temperature  in the TPO spectrum by  100 K, indicating that  the carbon formed on the K-modified
surface is more weakly bound. It is possible that the ion-exchanged potassium present within the
L-zeolite may also help to  suppress the formation  of highly dehydrogenated carbonaceous species.
This behavior may at least  partially  explain why the peaks in the  TPO spectrum are at lower
temperatures for the Pt/K-L sample compared to the Pt/silica catalyst.

It has been shown in previous studies that Pt/L catalysts are resistant to coking during reactions
with hydrocarbons  such as  n-hexane[ 14,3  1,321.  In our experiments, the Pt/K-L catalyst
accumulates  much less carbon than  the Pt/silica catalyst,  but the  isobutane  conversion over Pt/K-L
decreases to essentially zero within several minutes on stream. The L-zeolite support has
unidimensional  pores with diameters of 0.7  1 nm,  and our isobutane adsorption studies  (see Table
5-4) indicate that some of the  pores  become blocked during isobutane dehydrogenation over Pt/K-L
in the absence of hydrogen. Thus, the coke that forms on the Pt/K-L catalyst may be present near
the pore mouths, and thereby block Pt sites deeper within  the pores of the  L-zeolite from reaction
with isobutane. In the aforementioned  studies with n-hexane, the reaction was conducted in  the
presence of hydrogen, with hydrogen to hydrocarbon ratios between 5.5 and 8[ 14, 31, 32]. In
agreement with these previous studies,  the Pt/K-L  catalyst maintains  its activity for isobutane
conversion in  the presence  of hydrogen,  although this catalyst is  not selective for isobutene
production (see Figures 5-8 and 5-9). Hydrogen probably removes the highly dehydrogenated
carbonaceous species from the surface as by-products (e.g., hydrogenolysis products), such that the
                                              31

-------
selectivity of Pt/K-L to isobutene is low.  On the Pt/Sn/K-L  catalysts,  the deactivation profiles  were
similar with and without hydrogen. This behavior indicates that the presence of Sn is sufficient to
prevent the formation of highly dehydrogenated surface species that lead to the formation of coke
and other unwanted by-products.

The results of the  16 h reduction  and the  deactivation studies indicate that the extent  of Pt/Sn
interaction increases during the course of the reaction. After a long reduction in H2, the activity of
the catalyst has decreased (see Figures 5-11 and 5-12). In fact, the catalysts exhibited a similar rate
of isobutene production after treatment in hydrogen for 16 h as catalysts that were treated for 1 h in
hydrogen and were subsequently exposed to hydrogen and isobutane for 16 h. Mossbauer
spectroscopy has shown that Pt and Sn form alloy clusters in L-zeolite, but the majority of the Sn is
present as Sn2+[5], Studies on Pt/Sn/alumina [ 19,331 and Pt/Sn/NaY-zeolite[ 13] have shown that
the proportion of tin in metallic particles increased with metal particle  size. Furthermore,  Merlen  et
al. [19] have suggested that metallic tin cannot be stabilized in small particles.  Thus, the Sn2+ in the
Pt/Sn/K-L-zeolite catalyst may slowly become reduced in the presence of hydrogen during
isobutane dehydrogenation at 873 K and migrate to the Pt/Sn alloy particles, resulting in surface
enrichment with metallic Sn and the loss  of accessible Pt sites. Thus, although the presence of Sn
improves the resistance  of the  catalyst to coke formation, it may adversely affect the sintering of the
catalyst, since larger particles were visible  on the  Pt/Sn/K-L catalysts compared to the  Pt/K-L
catalyst after reaction at 873 K. The 24 h He treatment at 873  K(see Figure  5-13) indicates that
some activity is lost in a non-reducing atmosphere which suggests that particle sintering occurs at
high temperatures. TEM analysis of catalysts held at 873 K for extended periods (over 16 h)  in He
or H2 (no iC4H10) indicates that sintering of the metal particles does occur.

The regeneration scheme that was successful for the Pt/Sn/L catalyst required that the catalyst be
reduced for  16 h prior to the initial exposure to isobutane and after each regeneration (see Figures
5-21 and  5-22). As discussed above, a longer reduction leads to increased Pt and Sn interaction
but also some metal sintering.   The  regeneration in oxygen removes the carbon and also oxidizes the
Sn, destroying the Pt/Sn alloy, which may have become highly surface enriched in Sn  over the
course of the reaction. A long reduction,  therefore is needed to reform the Pt/Sn  alloy. It is
unknown  whether  this  regeneration  method redispersed the metals that had agglomerated  on the
external surface of the catalyst.

In summary, the Pt/Sn/K-L-zeolite UW Dehydrogenation Catalyst is a highly selective catalyst for
isobutane dehydrogenation to isobutene.  Through this project our research team  has been able to
understand the interactions between  the components of the catalyst - namely,  Pt, Sn,  K and L-
zeolite. Sn is necessary to make  the catalyst selective by reducing the size of the Pt surface
ensembles, thus  hindering the  competing  reactions,  such  as  isomerization, hydrogenolysis and
coking,  which require large Pt ensembles; K  neutralizes the support; and L-zeolite provides a
support in which  the Pt can be highly dispersed  by  depositing the metals within the pores.  Typical
industrial conditions employ hydrogen in the feed stream to minimize the amount of coking. In the
presence of hydrogen,  however, the equilibrium  dehydrogenation conversion  is reduced and thus
high temperatures are employed  to maintain the conversion. At 873 K in a 2:1 hydrogen to
isobutane  feed stream,  the  catalyst deactivates  by coking,  metal agglomeration  and surface
enrichment of the Pt particles  with Sn. It has been found, however, that Pt/Sn/L catalysts can be
operated for isobutane dehydrogenation at 798 K in the absence of feed hydrogen and maintain a
reasonable conversion,  provided that the catalyst  is highly selective. At these conditions, the
temperature  is low enough that metal  agglomeration and  Sn enrichment are  significantly
suppressed. Without hydrogen, the catalyst is more susceptible to coking but if the catalyst is
highly  selective, excessive  coking does not occur. The coking that does occur  is much easier to
reverse than the metal sintering which occurs at higher temperatures. Thus, running the reaction in
isobutane only at 798  K is very feasible for Pt/Sn/L-zeolite catalysts.  Based on the results of this
                                               32

-------
study, this catalyst may also  be applicable for the  selective  dehydrogenation of higher paraffins to
olefins. Further study of this application is being undertaken in our laboratory.
                                                33

-------
                                        References

 1.     Kirschner, E.M., Growth of Top 50 Chemicals Slowed in 1995 From Very High 1994 Rate.
 C&EN,  1996. April 8: p.  16-20.

 2.     Gregor, J.H., J.E. Zimmermann, and P.R. Cottrell, The Oleflex Process: Evolution,
 Commercialization and Optimization of a Dehydrogenation Technology. DGMK Conference -
 Selective Hydrogenations and Dehydrogenations, 1993. : p. 189-200.

 3.     Benesi.H A R.M.  Curtis and H.P. Studer, Preparation of Highly Dispersed Catalytic
 Metals. Journal of Catalysis, 1968.  10: p. 328-335.
 4.      Derouane, E.G.  and D.J. Vanderveleen, Structural Recognition and Preorganisation in
 Zeolite Cat
 45: p. L15-
Zeolite Catalysis: Direct Aromatisation ofn-hexane on zeolite L-based Catalyst. Appl. Catal., 1988.
      "  " 5-L22.
 5.      Cortright, R.D. and J.A. Dumesic, L-zeolite-supported platinum andplatinumltin catalysts
for isobutane  dehydrogenation. Applied Catalysis A: General, 1995.129:  p. 101-115.

 6.      Cortright, R.D. and J.A. Dumesic, Microcalorimetric, Spectroscopic,  and Kinetic Studies of
 Silica-Supported Pt and PtISn Catalysts for Isobutane Dehydrogenation. Journal of Catalysis,
 1994. 148: p. 771-778.

 7.      Dautzenberg, P.M., et al.,  Conversion of n-Hexane over Monofunctional Supported and
 Unsupported PtSn Catalysts. J.Catal., 1980. 63: p.  119-128.

 8.      Gallezot, P., et al, Location and Structure of Coke Deposits on Alumina-Supported
 Platinum Catalysts by EELS Associated with Electron Microscopy. J.Catal., 1989. 116: p. 164-
 170.

 9.      Park, S.H., M.S. Tzou, and W.M.H. Sachtler, Temperature Programmed Reduction and Re-
 oxidation of Platinum in Y-zeolites. Applied Catalysis, 1986.24: p. 85-98.

 10.     Vaarkamp, M., et al., Hydrogen as a Modifier of the Structure and Electronic Properties of
 Platinum in Acidic Zeolite LTL: A Combined Infrared and X-ray Absorption Spectroscopy Study.
 J.Phys.Chem.,  1995.99: p.  16067-16075.

 11.     Cortright, R.D., thesis entitled Kinetic, Microcalorimetric, and Spectroscopic Studies  of
 Supported Pt and Pt/Sn Catalysts for Reactions of Isobutane, Department  of Chemical
 Engineeringl994, University of Wisconsin - Madison: Madison, p. 230.

 12.     Garetto, T.F., A. Borgna, and C.R. Apestegufa, Effect of Alloying on  the Sulfur Resistance
 of Bimetallic Pt-based Catalysts, in Catalyst Deactivation, B. Delmon  and G.F. Froment, Editors.
 1994, Elsevier  Science B.V.: New York. p. 369-376.

 13.     M&iaudeau, P., et al., Studies on PtxSny Bimetallics in NaY. J.Catal., 1995. 154: p. 345-
 354.

 14.     Sharma, S.B., et al, Microcalorimetric 13C-NMR Spectroscopic  and Reaction Kinetic
 Studies of Silica- and L-zeolite-Supported Platinum Catalysts for n-Hexane Conversion. J.Catal.,
 1994. 150: p. 234-42.
                                             34

-------
 15.    Barbier, J., Coking of Reforming Catalysts, in Catalyst Deactivation 1987, B. Delmon and
G.F. Froment, Editors. 1987, Elsevier Science Publishers B.V.: Amsterdam, p. 1-19.

 16.    Fouche, V., P. Magnoux, and M. Guisnet, Coke Formation and Deactivation ofPt-Ultra-
stable Zeolite HY and PtH-Mordenite Catalysts During Hydrogenation of Benzene. Applied
Catalysis, 1990.58: p.  189-198.

17.    LAWJ., L., et at., Dynamic Process of Carbon Deposition on Pt and Pt-Sn Catalysts for
Alkane Dehydrogenation. Applied Catalysis, 1990.67: p. 1 1-23.

 18.    Cortright, R.D., et al. Reactions oflsobutane and Isobutylene over Silica- and L-zeolite-
Supported Pt/Sn and Pt/Sn/K Catalysts, in 11 th International Conference of Catalysis - 40th
Anniversary. 1996. Elsevier Science B.V.

19.    Merlen,ii efjjW^aracterisation of Bimetallic Pt-SnlAl2O3 Catalysts: Relationship
between Particle Size and Structure. Journal of Catalysis, 1996.159: p. 178-188.

20.    Chojnacki, T.P. and L.D. Schmidt, Microstructures of Pt-Sn and Rh-Sn Particles on SiO2.
Journal of Catalysis, 1991. 129: p. 473-485.

21.    Foger, K. and H. Jaeger, Redispersion  of Pt-zeolite Catalysts with Chlorine. Applied
Catalysis, 1989.56: p.  137-147.

22.    Karnpert, F.W.H., et al, Influence of Preparation Method on Metal Cluster Size of
Pt/ZSM-5 Catalysts as Studied with EXAFS Spectroscopy. J.Phys.Chem, 1990. 94: p. 8574-
8578.

23.    Vaarkamp, M., et al., Hydrogen as a Modifier of the Structure and Electronic Properties  of
Platinum in Acidic Zeolite LTL: A Combined Infrared and X-ray Absorption Spectroscopy Study.
J.Phys.Chem., 1995.99:  p.  16067-16075.

24.    Mieville,  R.L., Coking Characteristics of Reforming Catalysts. J.Catal, 1986.100:  p. 482-
488.

25.    Parera, J.M., N.S. Figoli, and E.M. Traffano, Catalytic Action of Platinum on Coke
Burning.  J.Catal., 1983.79:  p. 481-484.

26.    Barbier, J., G. Corro, and Y. Zhang, Coke Formation on Platinum-Alumina Catalyst of
Wide Varying Dispersion. Applied Catalysis, 1985.  13: p. 245-255.

27.    Lieske.H., A-.Sฃrkany, and J. Volter, Hydrocarbon adsorption and coke formation on
Pt/Al2O3 and Pt-Sn/A12O3 Catalysts. Applied Catalysis,  1987.30:  p. 69-80.

28.    Bat-ills, O.A., A.  Holmen, and EA.  Blekkan, Propane Dehydrogenation over Supported
Platinum Catalysts: the influence of  Tin on the Coking Properties, in Catalyst Deactivation,  B.
Delmon and G.F. Froment, Editors. 1994, Elsevier Science B.V.: New York.

29.    Cortright R.D. and J. A. Dumesic, Effects of Potassium on Silica-Supported Pt and Pt/Sn
Catalysts for Isobutane Dehydroganation. J.Catal.,  1995. 157: p. 576-583.

30.    Barbier, J., Deactivation of Reforming Catalysts by Coking- a Review.  Applied Catalysis,
 1986. 23: p. 225-243.
                                             35

-------
31.    Iglesia, E. and J.E.  Baumgartner. A Mechanistic Proposal for Alkane Dehydrocyclisation
Rates on PtIL-zeolite. Inhibited Deactivation ofPt sites within Zeolite Channels, in 10th
International Congress on Catalysis. 1992. Budapest, Hungary: Elsevier Science Publishers B.V.

32.    Lane, G.S., F.S. Modica, and J.T. Miller, Platinum/Zeolite Catalyst/or Reforming n-
Hexane:  Kinetic and Mechanistic Considerations. Journal of Catalysis,  1991.129: p. 145-158.

33.    Huang, Z., etui., Transmission Electron Microscopy and Energy Dispersive X-ray
Spectroscopy Studies of Pt-Sn/y-Al2O3 Catalysts. Journal of Catalysis, 1996. 159: p. 340-352.
                                             36

-------