Luận văn thạc sĩ synthesis of supported bismuth molybdate catalyst and the application in selective oxidation of propylene to acrolein

‘The application of supported bismuth molybdate catalysts in selective propylene oxidation results is provided a better understanding about the role of lattice oxygen and surface area of

MINISTRY OF EDUCATION AND TRALNING

HANOI UNIVERSITY OF TECHNOLOGY

Trương Duc Duc

SYNTHESIS OF SUPPORTED BISMUTH MOLYBDATE CATALYST AND THE APPLICATION LN SELECTIVE OXIDATION OF PROPYLENE TO

ACROLEIN

Tổng hợp vật ligu Bismuth Malybdate trén chất mang và ứng dụng làm xúc

tác cho phần ứng oxi hoa chon loc propylene thanh acrolein

Major: Chemical engineering

MSC, TILESIS

Supervisor:

Dr Le Minh ‘Thang Invited supervisor : Assoc Prof Anders Riisager

Hanei, 10/2009

based catalysts ‘Iwo synthesis directions this leads to the synthesis have been carried out, One is based on the idea of reducing the particle size of catalyst,

nanometer sized y-Bi,MoO, crystals with dimension of 50nm were synthesized by

hydrothemal treatment Another is based on the idea of impregnating bismuth

xnolybdale on high surface area supports, silica supported B-BizMo;Oy and zirconia supported B-JizMo)Op catalysts with various percentage of B-3iMo)Os loading

contents were therefore synthesized Both works aimed to investigate the influence

of surface area of catalyst on their catalytic activities in the selective oxidation of propylene to acrolein

In this thesis, bismuth molybdate catalysts and their applications in the selective oxidation of propylene to acrolein were presented very detail ‘The Mars and van Krevelen mechanism as well as the important role of lattice oxygen were also discussed in detail

Synthesis of mesoporous ZO; support by hydrothermal method was also a studying task since the synthesis of mesoporous 410) ordinary work and requires a special

attention in order to obtain a high surface area and a stable pore structure

‘The application of supported bismuth molybdate catalysts in selective propylene oxidation results is provided a better understanding about the role of lattice oxygen and surface area of catalyst Both surface area and lattice oxygen of catalyst play as key roles in the reaction Supported bismuth molybdate samples possess both high surface area and high oxygen mobility resulted in an extraordinary increase of

catalytic activities Among supports, ZrO, is mare efficient due to the presence of

synergy effect

Thục T.D - 2009

“Tôm tắt

Nội dung của luận văn nảy là tổng hợp vật liệu xúc tác trên cơ sở bismuth xmmolybdate có diện tích bề mặt riêng cao Hai hướng tổng hợp vật liệu dã dược thực hiện Hướng thứ nhất dựa trên ý tưởng làm giâm kích thước hạt xúc tác, tỉnh thế y-

Bị jMoO; có kích thước hạt khoảng 5Onm đã được tổng hợp thành công bằng phương pháp kết tỉnh thủy nhiệt Hướng thủ hai dụa trên ý trẻởng mang bismuth mmolybdate lên các chất mang có bể mặt riêng cao, các xúc tac bismuth molybdate ngAm tâm trên siliea và zirconia đã được điều chế với các hàm rong chất mang, khác nhau Cä hai hướng đều nhằm đế khảo sát ảnh hưởng của diện tích bê mặt riéng của xúc táo đến hoạt tính xúc tác của nó trong phản ứng oxi héa chọn loc

Kết quả nghiên cửu ứng đụng vật liệu xúe tác trên cơ sở bismulh rnolybdate có điện tích bê mặt riêng cao trong phân ứng oxi hớa chọn lọc propylene đã cung cấp thêm những hiểu biết rõ hơn về vai trỏ của oxy mạng lưới và bỀ mặt riêng của xúc tác Cả

ác đều đồng vai trô quyết định

điệu tích bể mặt riêng và oxy mượng lưới của xúc

trong phân ứng Các mẫu xúc tác ngâm lâm bismuIh miolybdate có cã diện tích bề

„mặt riêng lớn và oxy mạng lưới linh động dẫn dễn hoạt tính xúc tác tăng đáng kể Trong hai loai chat mang nay Thủ ZrÔ; có hiệu quả hơn bởi vì sự xuất hiện của hiệu tửng hiệp trợ xúc tác giữa chất mang với pha phan ung

hydrocarbons” cooperated between Hanoi University of Technology (HUT) and Technical University of Denmark (DIU) ‘fhe work presented herem has been carried out at Lab of Petrochemicals and Catalysis Materials, Adsorption, Faculty of Chemical Kngineering, Hanoi University of ‘Technology under the supervision of

Dr Le Minh ‘Thang ‘the work is also performed in the cooperation with Center for Sustainable and Catalysis Chemistry, ‘fechnical University of Denmark (DTU), Denmark,

First and foremost, 1 would like to thank my supervisor Dr Le Minh ‘Thang, for her exceptional patience with me At the beginning, I was “clumsy” and made a lot of my own mistakes, but in spite of this, she always gives me a patience and provides valuable guidance and support But the most valuable thing I have learnt from her is “how to do science” [ also would like to thank Dr Nguyen Hong Lien, director of Lab of petrochemicals and catalysis, adsorption materials for her support and patience as well as the great opportunity which she offered me to work in her lab during the last two years

During my master study, I also have a chance to be trained for six months at Center for Sustainable and Catalysis Chemistry, Denmark On this period, I leamed many things about Supported Ionic Liquid Phase Catalysts (SILP) for the hydroformylation of ethylene, a new research area, which provided very useful complimented knowledge for my research direction

Iam deeply grateful to Prof Rasmus Fehrmann in the DTU Kemi Centre for Catalysis and Sustainable Chemistry for giving me the opportunity to work at his lab and his kindly guidance, support and palience lo me during my stay in Denmark Tam also thamklul io Associate Prof Anders Riisager who gave me invahuble

thoughtful insights, advice, support, discussions and encouragements from the first

days I came to DTU

Tam very gratefully for the kindness and help from all my colleagues and partuers I must thank Mse Nguyen Ha Hanh lor her friendly guidance and support

in the time T joined the project T woutd especially thank Post Doe Olivier Nguyen

Thục T.D - 2009

Van Buu, Post Doc Eduardo Garcia Suarez, Post Doc Jianmin Xiong from CSC lab for their friendship and generous help during the tne | stayed in Denmark,

Spccial gratitude must be given to all the graduate students with whom I have collaborated in this project at Hanoi University of ‘Technology for thew great assistance and cooperation Vo Hoang ‘Tung, Ho Si Dang, Le ‘the Duy, Nguyen Loang Lai | enjoyed working together with all of you!

1 am also grateful to the Danida Fellowship Cente (DSC) for financial support, 1 especially want to express my sincere gratitude to Prof, Vu Dao ‘Thang, Prof Le Van Hieu for their support, encouragement and useful discussions

‘Yo my mother, father, mother-in-law, father-in-law, brothers and my best friends my warmest thanks for their love, encowagement and support during, all the years of my education

1 sincerely thank my best friends Christian Juel Adamsen, his wife Marie Lahn gigaard and their little angel daughter Alberte for their true love, warm friendship and encouragements

‘The most important of all, 1 would like to thank ray wife, Thanh Thuy, with all my love Her endless love, support, understanding and patience to me were immeasurable More than anything else, her love has carried me through the many

challenges | faced during my graduate years

‘Truong Due Due Hanoi, 10.2009

‘Table 1.2 Some examples of nmulti-component BiMo based catalysts

Table 1.3 Apparent Activation Euergics of Partial Oxidation of Propylene to Acrolein over Bismuth Molybdate Catalysts [98]

Table 2.1 Raw chemicals for synthesis of unsupported and supported bismuth molybdates

Table 2.2 Equivalent amounts of chemicals corresponding to different samples

‘Table 2.3 Strong lines corresponding to different phase of bismuth molybdates and cubie phase zirconia

‘Table 2.4; 12 bands (cm) of bismuth molybdates

Table 2.5: Raman frequencres in cm” of malybdate species

Table 2.6: Raman frequencies of bismuth molybdates

Table 2.7 Retention time of sume products and relaled compounds

Table 3.1 Summary of synthesized zirconia samples

Table 3.2 Summary of synthesized bismuth molybdate samples

Bi siles (Adapted from van den Elen and Ricok |1] J)

Thục T.D - 2009

Figure 1,2 The unit cel! structure of a-BizMo;O1 (left figure) and atom map of the anit cell (right figure) (Adapted from van den Elzen and Rieck)

Figure 1.3 A unit cell of P-BisMo, (right figure) and atom map of the unit cell (left figure), (Adapted from H.-Y Chen and A.W Sleight [16))

Figure 1.4 Representation of the [-Bi:Mo0» structure projected along (010) [17]

‘The figure shows clusters of MoO), tetrahedral to form Mo,Qyg ‘The small circle represents a bismuth atom

Figure 1.5 From left to right, a unit cell of y-Bi,MoU, and its atom map ‘The figure

is adapted from Teller ct al |21)

Figure 1.6 Phase equilibrium diagram of the bismuth molybdate system

Figure 1.7 Molooule structure of acrolein

Figure 1.8 ‘the reaction paths of the partial oxidation of deuterium-labelled propylene, Z-KD/KH, while 1 and 3 are the numbers of carbon atoms where hydrogen is abstracted [46]

Figure 1.9 The reaction paths of the formation of side produets [46]

Figure 1.10 A schematic of the Mars and van Krevelen mechanism on bismuth aolybdate catalysts [88]

igure 1.11 A schematic of bridges and doubly bonded oxygen ions on bismuth aolybdate catalysts 172]

igure 1,12 A schematic of steps in propylene oxidation into acrolein over bismuth anolybdate catalyst [82]

Figure 1.13 A reaction mechanism of propylene oxidation into acrolein, showing

acid-base and redox steps [72]

Figure 1.14 Schematic PT phase diagram of ZrO, [105]

Figure 1.15 Three phases of 710,

Figure 1.16 Schematic diagram of the mechanism of preparation of mesoporous

‘ZO, (suggested by Guorong Duan et al [108D)

Figure 2.2 Diagram of soxhlet extraction

Figure 2.3 A X-ray generated tube (a) theory diagram (b) a real tube

Figure 24: illustrates how diffraction of X-rays by crystal planes allows one to derive lattice by using Bragg relation (a) and real XRD parton (b)

Figure 2.5: Principle of infrared absorption

Figure 2.6: Principle of Raman seaticring

Figure 2.7 Some typical types of isotherm

Figure 2.8 The BRT model of mullilayers adsorplion

Figure 2.9: The BET plot

Figure 2.10 The interaction between the primary electron beam and the sample in an

cleciron microscope Ivads lo a number of detectable signals

Figure 2.11 Schematic diagram of a rarmission electron mieroscope

Figure 2.12 Diagram of reactor setup

Figure 3.1 Influence of H,0/7r in gel on surface area of products

Figure 3.2 XRD parllen of zirconia prepared by hydrothermal (7.6)

Figure 3.3 Absorption — desorption curve and pore distribution curve of

synthesized ZrO, sample (7.6)

Figure 3.4 XRD pattern of ZrO, sample at different calcination temperatures

Figure 3.5 Influcnee of calcinalions temporature on surface arca of samples

Figure 3.6 SEM image of zirconia calcinated at $80°C (Z7)

Figure 37 XRD pallens of samples (M2 — M6) al various crystallizing

temperatures

Figure 3.8 SEM images of samples at various crystallizing lemperaLures

Thục T.D - 2009

Figure 3.9 XRD pattems of samples (M4, M7, M8, M9) at various calcination

temperatures

Figure 3.10 SEM images of samples at various calcination temperatures

Figure 3.11, FY-Raman of samples (M4, M7, M8, M9) at various calemation

temperatures

Figure 3.12 TG/DSC diagram of y-Bi.MoO, sample (M4)

Figure 3.13 XRD parttens of samples at various pEI value in gel

Higure 3.14, X-ray diffraction pattems of samples synthesized by sol-gel and hydrothermal methods

Hig 3.15, 1"l-Raman spectra of samples synthesized by sol-gel and hydrothermal methods

Higure 3.16, I-SLUM images of samples synthesized by sol-gel and hydrothermal methods

Higure 3.17 Specific surface area (Spr1) of samples prepared by solo-gel and hydrothermal methods

Hig 3.18 Conversion of propylene (a) and rate of acrolein formation (b) over samples synthesized by sol-gel and hydrothermal methods

Figure 3.19 XRD palterns of zirconia supported B-BizMo, samples (a) and silica supported [J-BiaMo;Os samples (b)

Figure 3.20 TT-Raman điagrarns of ziroonia sưpporied B-DizMo;Q; samples (a) and silica supported B-BizMo;Ok samplss (b)

Figure 3.21 FT-IR spectra of silica supported 0-Bi,Mo,O» samples

Figure 3.22 SEM images of $iQ, support (a) 10%beta/SiO, sample (b) and 40%bela/SiO, sample (c)

Figure 3.23 SEM image of 40%beta/SiO, sample

Figure 3.24 TEM image of 10%beta/SiO,

Jigure 3.25 Mechanism of B-BipMo2U5 formation on the surface of $10, support

Figure 3.26 SEM images of 10%beta/Z1O, (a) and 40%bete/Zr0, (b)

Figure 3.27 A B-BiyMo,y formation on ZrO, support mechanism

Figure 3.28 The absorption — desorption curve and pore distribution curve of 109beta/SIO;

Figure 3.29 The absorption — desorption curve and pore distribution curve of 40%beta/SIO;

Figure 3.30 conversion of propylene in selective oxidation af propylen over silica supported beta bismuth molybdates and pure beta bismuth molybdate

Figure 3.31 rate of acrolein formation in selective oxidation of propylen over silica supported beta bismuth molybdates and pure beta bismuth molybdate

Figure 3.32 TPRO diagrams of pure SiO, 10%hola/SiO,, 40%bota/SiO, and pure B= Bison,

Figure 3.33 Conversion of propylene in selective oxidation of propylen over

vivconia, supported beta bistnuth molybdales samples and pure bela bismuth

40% belw/770, samples al different emperalure

Figure 3.36 Reaction rate for propylene consumption at 500°C of samples with difleronl percentage of bela mixing wilh 770, and $i, supporls

Figure 3.37, Comparison of activation energy of pure [-BizMo,0», P-BiyMo,0» impregnated on SiO, and ZrO, samples

Thục T.D - 2009

L1.1 Crystal Structure of Bismuth Molybdates

L13 Relationship between structure of Bismuth Molybdate and catalytic

1.1.4 The catalytic performance of bismnth molybdate

1.2 Selective oxidation of propylene †o acrolein

1.2.1 Acrolein and its production

12.2 Thermodynamic of selective oxidation of propylene to acrolein

12.3 Catalysts for sclective oxidation of propylenc

12.4 The Reaction Mechanisms of Selective Oxidation of Propylene to Acrolein

on Bismuth Molybdate Catalysts scenes

1.2.5 Kinetics of Propylene Oxidation to Acrolein

1.3 Several catalytic supports

3.1 Mesoporous zirconium dioxide

£3.1.2 Synthesis and application of zirconia in catalysis field

L3.2 Amorphous silica

1T.1.1 Synthesis of Bismuth Molybdate with nano-size crystal

1I.4.2 Application in this fhesis sen hhheheeeeereeeeeselÔ

1.5 Physical absorption for the determination of surface area and pore distribution

13

1.8 Thermal analysis

TI.8.1 Principle

IL 9 temperature-programmed reoxidation {TPROI 5

IL9.1 Principle

11.9.2 Application in this thesis

IL 10 Evaluation of catalytic activity

TIT.1.E Tnfiuence ðŸ HyQ/Z⁄r volưme ra[i0, << che

UL.1.2 Influence of crystallized temperature

H13 Influence of calcination temperature on surface area and phase

composition of samples 0 :c0ccccecececesesssseeeeesenensceeeensreneesenear

TH.2 Synthesis and characterization of nano gama bismuth molybdate 87 UL2.4 The influence of crystallizing temperature

TIL.2.2 The influence of calcination temperature

THT.2.3 The infinence of pH on pháse campasilion of samples 94 UL2.4 Comparison of bismuth molybdate synthesized by sol-gel and

1H.3.Synthesis, characterization and cafalytic activities of supported

Bismuth molybdate system in selective oxidation of propylene 100

3.1 Compositional characterization of bismuth molybdate on supports 100 TI1.3.2 Distribution of hismuth molyhdate an supports 108

TI¥.3.3 Catalytic activities of beta bismuth molybdate impregnated on SiO,

INTRODUCTION

More than one third of worldwide chemical products, are produced by catalysed reactions with oxides types of catalysts [1] Among the catalysts, selective oxidation can be considered the most typical example of metal oxide-type materials

as heterogeneous catalysts, which produce around one quarter of total organic chemicals worldwide Acrolein has broad industrial and agricultural applications and had been produced by silica supported sodium catalysing vapaur-phase condensation of acetaldehyde and formaldehyde at temperatures between 300°C and 320°C [2] Large-scale production of acrolein process was later commercialized by SOMO (Standard Oil of OIG) in 1950s using the Ileame and Adams’ catalytic process

The use of tnetal oxide catalysts began wilh the realisation of the need to

cmveri low molecular-weight hydrocarbons, such as nalurat gas and refinery off- gas, to largor molecules of higher valuc Especially, the conversion of light hydrocarbons inlo products contaming oxygen or ofher hetero aloms produces important intermediates for the petrochemical industry A number of such intermediates are derived (rom propylene, ¢.g., acrylonitrile, acrolein, acrylic avid, and propylene oxide It was Heame and Adams [3] in 1948 who first reported that

cuprous oxide could selectively oxidize propylene into acrolein with a yield of

about 50% at propylene/ oxygen ratios of about one However, the yield of acrolein

on cuprous oxide was low at that time ‘therefore, it was necessary to find the other catalytic systems with higher effective catalytic performance In 1959, Idol, and in

1962, Callahan et al found that bismuth/molybdenum catalyst has higher catalytic performance in selective oxidation of propylene to acrolein than that obtained in the cuprous oxide system Later, Veatch and co-workers [4] discovered that bismuth molybdate-based catalysts (bismuth phosphomolybdates) are more superior to ouprous oxide The bismuth molybdate-based catalysts have since become the amajor catalysts in commercial processes worldwide to produce acrolein [5,6,7,8]

Since the discovery of the process, a huge amount of reports of studies on the

process have bean published, mainly focusing on reaction mechanisms and calalyst

characterivation lo reveal the myslery belund the selective oxidation The most

important findings are the [act that the acrolein was formed via the formation of

allyl intermediate on the catalyst surface and the reaction uses lattice oxygen In tenns of bismuth molybdate, it was also found that only three phases of pure bismuth molybdate, namely œ, B, and y phases (œ-BiyMosOks, B-BizMo;Os, and y- Ti;MoOs, respectively) are active and selective for the reaction

Main disadvantage of bismuth molybdate is its extremely low specific

surface arca (typically ~0.1 to 1.0 m?.g?) duc to its bulk structure Tt could be the

reason for low conversion of light olefin aver bismauth molybdates However, the seleclivily of đesired produets is quite high resulted in bismulh inolybdales are Uill the moat effeotive catalysis for sclective oxidation of olefin Recently, some aulhors reported that they suecced in synthesis of bismuth molybdate with surface arca of 2

— 4m’.g [9], but it was still low Few aticmpls have been done to increase the surface area of bismuth molybdate by impregnation of bismuth molybdate (from 0.5%wt to 5%wt) on silica gel [10] but results showed that catalytic performance decreased significantly due to low selectivity of desired products Supported bismuth molybdates were belicved that its catalytic activitics were not cffective because it was very difficult to obtain desired phases of bismuth molybdates on the surface of supports by impregnation and because of low ion conductivity of most supports ‘Therefore, the understanding of influence of specific surface area on the catalytic activities as well as reaction mechanism has been still not clear up to now

The research in this thesis was designed to enlarge the surface area of bismuth molybdates which achieved an improved understanding in the effect of specific surface area on catalylic perfonnance of bismuth molybdates and kinetics Two directions were offered One is to reduce bismuth molybdate crystal particle to nanometer sized one resulting in increasing the surface arca On the other band, the research also aimed Lo impregnate bismuth molybdate on some kind of high surface

area supports to distribute better catalytic sites [or ihe selective oxidation of

gmopylene to acrolem High conductive support was also chosen 19 stucly sincc it

‘was belived that lattice oxygen plays an important role in the catalytic reaction

Thục T.D - 2009

CHAPTER I: LITERATURE REVIEWS

In this chapter, we consider in detail the stwucture of bismuth molybdates and the theories behind catalytic selective oxidation of propylene to acrolein over bismuth molybdate catalysts ‘This will contains historical background of the reaction discussion in depth of the bismuth molybdates, how they play as catalytic role in selective oxidation of propylene to acrolein and the background of the kinetic as well as the mechanisms of the reaction

1.1 Bismuth Molybdates

L141 Crystal Structure uf Bismuth Molybdates

The fact that only certain phases of known bismuth molybdate are active as

selective oxidation catalysts means that the bulk structure determines the activity of

bismuth molybdale Before discussing lhe relation between structure and

activity, the crystal structures of the three phases of active bismuth molybdate,

Scheelite crystalline structure is buill from stacking up of Ca?! arn WO, ions

Tn the case of a-BiMo3Ojg, Ca?! is replaced by Bi” ion while WO," is replaced by MoO,” ion Beth WO," and MoO,” have the same structure To maintain the charge balanee within the crystal dus to the replacernent of a? ion (Ca) with a’

ion (Bi"), the structure contains an ordered arrangement of vacancies in the Bi

positions corresponding Lo the formation of BiyMo;Q)2 Figure 1.1 shows the

projection of the crystal along b axis and the bismuth vacation sites on the structure

Figure 1.1 Bismuth vacant sites in ø-Bi¿Mo;Ox; along b axis projection ‘the solid lines show the a-Bi.Mo,Oy) unit cell while the dashed line is a unil cell of scheelite Large solid circles represent occupied Bi sites while the small circles are the empty

Bi siles (Adapted from vant den Flven and Rieck [11])

All MoO, (ctrahedron in a-BryMo3Q12 are in Mo2Qg fonn [11] There are two

forms of Mc,0s, one of which possesses a central symmetry while the other does not, In addition, all Bi have 8 oxygen neighbours However, the Bi-O distances are

not equal and range from 2.12 to 2.93 A ‘The unit cell structure of ø-Bi;MoyOis

according to van den Elzen and Rieck is shown in Figure 1.2

ii) 8-BuMo;O,

Growing a single crystnl o[ B-BuMo;Os to the sive suitable for single

crystalline X-ray diffraction was not an casy task This is lhe reason why the erystal

structure of B-BiMo,Og was not well underslood unlil 1975 In 1974, Chen [13]

grew a single crystalline B-RiyMo,Os, bul the crystal was twinned and not suitable

for accurate X-ray structure analysis Duc to the difficulties, van den Elzen and

Rieck did a comprohensive powder X-ray diffraction analysis on a high purity B-

BigMo,0y and obtained its stracture

Thục T.D - 2009

Figure 1.2 The unit cell structure of o-BizMo;0)) (left figure) and atom map of the

umt cell (right figure) (Adapted from van den Elzen and Rieck)

The crystal was prepared according to the method developed by Batist [14]

Van den Elzen and Rieck noted that Bi** had octahedral coordination to oxygen

anions at various distances as in the alpha phase, while Mo® had tetrahedral coordination with oxygen van den Elzen and Rieck further proposed that Mo tetrahedral might be strongly distorted and linked together in pairs as in the alpha phase

The structure of -BiyMo;Os 1s built with square clusters of four MoO,

tetrahedrons to become Mo404¢ They bound together by Bi ions located halfway on the axis passing through the centre of the squares [15] Some Bi cations are surrounded by eight oxygen ions from the MoO, tetrahedrons There are also some oxygen ions associated only with Bi cation in the coordination sphere of others Thus, the structure may be visualised as composed of rows of oxygen ions,

connected only to Bi cations, running parallel to the (Mo,0y6)-Bi-(Mo4Oj¢) units

These resemble the ribbons of Bi,O from the Bi,O layers in koechlinite However,

not all Bi sites are filled One in every four Bi sites is empty The structure may be thus represented by Bi(BiyO;)(Mo¿Ons), where the first Bi is associated only with

the (Mo,©,¢) units and the Bi cations in parentheses are bonded to the oxygen

atoms associated only with Bi as well as to those shared with the (Mo,4Oj.) units The structure of a unit cell of B-Bi,Mo,O, and its crystal representation along the

(010) face are shown in Figure 1.3 and 1.4, respectively

Figure 1.4 Representation of the B-BizMo,Oo structure projected along (010) [17]

The figure shows clusters of MoO, tetrahedral to form Mo,Ojs The small circle

represents a bismuth atom

iii, y-Bi,Mo0, The gamma phase of bismuth molybdate is found naturally as_ mineral koechlinite The structure was firstly determined by Zemann in 1956 [18] In 1973, van den Elzen and GD Rieck [19] re-determined the structure using single crystal X-ray diffraction because they found that the crystal model from Zemann was inaccurate due to a large linear absorption coefficient as a result of irregular and

Duc T.D - 2009

21

poorly defined specimen Using the single crystal X-ray diffraction technique, they

found that the actual symmetry was Pca21 rather than Cmea as suggested by

Zemann However, the model developed by van den Elzen and Rieck has two

oxygen atoms that are separated by only 2.2 A in O-O distance The distance is a

physically unrealistic feature, because the closest distance allowable for two-

unbounded oxygen is 2.4A, based on the effective ionic radii of oxygen [20] The

structure of y-BiyMoO, in Zeeman and van den Elzen and Rieck’s work were determined from crystals of natural koechlinite mineral

The crystalline structure model developed by Teller et al [21] and Theobald et

al [22] was based on Zeeman and van den Elzen and Rieck’s work It contains

series of layers made up of (Bi,02)," and (MoO,),2” octahedron units connected by O® ions in the arrangement of (Bi,O,)*", 0”, (MoO,)"", O*, An overview of the

structure is given in Figure 1.5

Figure 1.5 From left to right, a unit cell of y-Bi,MoOg and its atom map The figure

is adapted from Teller et al [21]

Unlike the alpha and beta phases, the gamma phase of bismuth molybdate has three polymorphs, namely, y, y’, andy” The structure discussed above is the lowest

temperature polymorph At 570°C, the y polymorph is reversibly turned into the `

polymorph and further to the irreversible y” polymorph if it is heated to 670°C [23,

Many methods have been applied 1o prepare bismuth molybdates like solid

stalc, co-precipiation, sol-gel, hydrothermal, spray drymg In (his parl, ihoy are

discussed briefly some advantages and disadvantages of these methods

i, Solid state technique:

It is the oldest method for preparation of bismuth molybdate The procedure is

quite simple, in which Bi:O; and MoO; are mixed in stoichiometric proportion

After that, mixture is calcinated at the tempratire range of 5G0-650°C to gain

desired phase bismuth molybdate

A.Wantanabe et al [26] prepared Polycrystalline y(H)-Bi,MoO, from starting

materials BiO, and MoOs, both 99.9% pure, which were thoroughly mixed in stoichiometric proportion under ethanol After drying, the mixture was heated in a

covered platinum crucible at 655°C for 40 br Using HRTEM, they confirmed that

the high firing temperature was need to make order structure of bismuth molybdate

RN Vannier et al [27] prepared sample by solid state reaction from

stoichiometric amounts of Bi,0; and MoO; oxides in a covered gold crucible,

calcination procedure was carried out after that They found new structural Bi-Mo

Mixed Oxides with a Stricture Based on [By2O.4)., Columns but il is very complicated and nol to be easy to contrel

In general, solid state technique has some disadvantages like high calcination

temperature, difficult to make homogeneous composition which is very important to pain desired phases, product possesses low surface area (less than 2 m°g”) in common

it, Co-precipitation technique

Co-precipitation method is a popular method to prepare bismuth molybdate

‘because its procedure is quite simple and it allaws to easily control composition of

products The starling reactants are typically a solution of Bi(NO3)3 5 A,O in waler

and a suitable quantity of HNO; and a solution of (NH1)sMo;Os, 4H2O in water The solution of bismuth nitrate is added to ammonium molybdate or vice versa pH

Thục T.D - 2009

23

is adjusted in the range of about 1-9 by adding NH; or HNOs, depending on each author ‘The precipitate is filtered and dried or the liquid is directly evaporated,

Trecipitates are calcined at some piven temperature, normally from 450-580°C

Ji Chul Jung et al [9] successful im preparation of bismulh molybdate by coprecipilalion followed the above procedure: A known amount of bismuth nitrate

(BiCNO3);) was dissolved in distilled water that had been acidified with

concentraled nilrie acid The solution was then added dropwise inlo an aqueous

solution containing a known amount of ammonium molybdate ((NH,)qMo;02.)

under vigorous stirring, During the co-precipitation step, the pH of the mixed solution was precisely controlled using known amounts of amunonia solution ‘The pil values were maintained at 1.5 and 3.0 in the preparation of o-DisMoxOy, and y- BizMoOg, respectively, After the resulting solution was vigorously stitred at room temperature for | h, the precipitate was filtered to obtain a solid product The solid

product was dried overnight at 110 °C, and it was then calcined at 175 »C for S hin

a stream of air to yield the final form The products both showed very high

crystallinity, They also successed in prepared bismuth molybdate with surface area

of 1.9 and 3.5 m?.g" which has quite high surface area compare to another publics before

Zhou Bing ef al [28] also prepared bismuth molybdate by precipitation but at Jugh pH value (pH — 7) and various calcination temperature They found that the

aiore calcination temperature increased, the lower surface area of products gained

‘The surface area of product decreased from 1.2 m’.g” to 0.8 nt”.27 when increasing

temperature from 420°C to 570°C

Nianxe Song et al [25] tried to prepared three phases of bismuth molybdate

by coprecipitation method but they just achieved small surface area products

(abont 2 m? ø”) and high impurities in products

One of disadvantages for (his method is the [iltration of the preeipilant from

the solution, this stop is very casy te make loss of Bi or/and Me ion which leads to

gain desired phase bistauh molybdate difficully Almost results showed thal very

low surface area products were achieved by co-previpitation,

iii, Sol-gel method

Sol-gel is still new method for the preparation of bismuth molybdates It has attracted a lot of attention of scientist recently, because sol-gel method has a lot of advantages like: easily controlling the composition of gel to gain the desired phase

of bismuth molybdates; synthesis conditions are simple, easily to put other components into gel to make multiphase oxides so on In common, gel was prepared by following procedure: The precursor solutions were prepared from aqueous solutions of (NII4)eMo;Oz4 (solution A) and Bi(NO,),TINO; (solution B)

A volume of solution A was slowly added into an equivalent amount of solution B cenvesponding to the desired Ri/Mo molar atomic rabos, and concentrated HNO, was continucusly added in order to preserve the high acidity of the mediunr and to prevent the precipitation of bismuth molybdates Citric acid was added into solution

in order to resull in complexes The ablained solutions were gellyfied at 60-80°C

then dried

al 110°C (or 2h, and the spongy solid precursors obtained were crushed, Powders

unlil the gels were completely formed The transparent yellow gels wer

obtained after the gelation were dircetly calcined in an air flow at 580°C for 2h with the beating rate of 10°CAnim |30, 31, 33, 33]

MT Le et al [32,33] succeed in preparation of «,, y pure bismuth molybdates

by sol-gel method using citric acid as a complexes reagent The surface area of the products were depended on the amount of added citric acid The highest surface

area reached to 12 m*.g' However, the disadvantage of this method is the burning

of citne acid which caused a local overheat 1 results in sintering and aggregation

E Godard et al, H-G Lintz ct al |34, 35] also tried to prepare bismuth smolybdates by sol-gel followed the above procedure Instead of gelation by heat in air, the water was evaporated under vacuum in a revolving flask at 30°C until a viscous solution was obtained After that, this solution was dried at 80°C under vacuum for 16 h in order to obtain a spongy precursor which was crushed, decomposed at 300°C for 16 h and calcined at 170°C for 18 h The products had high crystallinity [lowever, they found that it was very difficult to achieve high

surface area products The surface area of all of products were lower 2 m” g!

Tn summary, sol-gel is a good method to prepare desired bismuth molybdate phases easily However, it seemed to be very difficult to achieve high surface area

Thục T.D - 2009

bismuth molybdates because desired bismuth molybdate phases were only formed

during calcimations, but calcination could lead to aggregation

iiii, Hydrothermal method

Hydrothermal method is very famous in synthesis of zeolite and porous

amalerials but, recently, the hydrothernal method has been also applied to prepara

tismuth molybdates An outstanding advantage point of hydrothermal was desired

bismuth molybdate phase formed dircotly in crystallization stage Therefore, aggregation could be prevented during calvinations

IlonglIua Li et al [36] suggested preparation of bismuth molybdates following

the procedure: Hi(NO¿);.5IIyO was dissolved into concentrated LINO solution,

while stoichiometric amounts of (NIL)Mo0.4.411,0 (Bi/Mo = 2/3, 2/2 and 2/1 for

œBbMG¿Ob, REBiMo;Os mủ y-Biy MoO, respectively) was dissolved in deionived

waler The lwo solutions were mixed under vigorous slirring The pH of the mixture

was adjusted to a specific value with ammonium hydroxide (1:1 volume fraction)

Afler stirring for 0.5 h, this mother iquor was poured into a Teflon-hned slainloss steel autoclave until 80% of the volume of the autoclave was occupied After that,

the autoclave was scaled into a slainiess steel tank and kept al 180 °C for 24h Ther

the reactor was cuoled Lo room Lemperaiure naturally Obtained samples were

collected and washed with deionized water and dried at &0 SỞ tn air In the ol

the P-phase, it was necessary to calcined at 560 °C since the pure material was formed only after further heat treatment of the hydrothermally treated sample o-

BiyMo;Qr, was preparod sux

formed at high value of pH (7-9) B-PizMoxOs just only formed at pH of about 3 with high purities Obtained products possessed quite small particle size, their

ssful al pH lower than 3 while y-BisMoO, easily

surface areas were 10.9 m?.g? and 16.4 mg? afler calcination

Andrew M Beale ct al [37] prepared three phases of bimuth molybdates by hydrothermal method Using An KDXRD and combined XRD/XAS study they confimed that the a and y bismuth molybdate phases form directly after hydrothermal treatment whereas the B phase need to be applied to an additional heating to finally form the monophasic material This method also yielded materials with a higher surface area (about 10 m’p')

Therefore, hydrothermal method could be one of prospective methads to

improve specific surface area of bismuth molybdates

Spray drying methnd:

in air at different temperatures from 200 to 650°C: for 5h & phase were prepared successfully even at temperatures as low as 200°C while pure y phase was only oblained when calcined al 550°C and pure f phase at, 600°C Spray drying scomed

to be very convenient method, which confinns itself as amore rapid way {a prepare

bismuth molybdates, and results in higher purity of final products because the use of

a homogeneous precursor solution, where BP” and Mo" ions interacted with cach other according to an accuracy sct by the stoichiometric ratio of the solution which

is carried over inlo the fine spray dried powder In addition, the reproducibility of spray drying is very high and it is not effected by external factors (the

coprecipitalion of unwanted species, the lost of ions in ihe Mtraic 30 on )

However, final products still have low surface area ( Sper of œ, B, y bismuth molybdates were about l5 m°g”) because of aggregation occurred during calcinations

11.3 Relationship between structure of Bismuth Molybdate and catalytic activity

According to the literature on phase equilibrium studies [39, 40, 41, 42] as shown in lig 1.6, there are eight phases known on binary bismuth molybdenum oxides, prepared by high temperature synthesis Among the eight phases, only three

of them are active and selective toward portial oxidation of olefin Linljac et al

Thục T.D - 2009

Figure 1.6 Phase equilibrium diagram of the bismuth molybdate system

Among the family of active bismuth molybdenum catalysts, several arguments have been raised in deciding which one has the best activity and selectivity At least there are three opinions about the order of activity of the ø, B, and y phases The

first comes from Beres et al [43], Millet et al [44] and also Monnier and Keulks

[45] who found that the S-phase was the most active and selective, followed by a

and y (B>a>y) The second comes from Batist et al [14] who observed that y phase was equally good as the beta phase (B= y >a) Yet another opinion from more recent

studies by Cullis et al [7] who observed that the order of propene activity decreased

in the order of a>B>y whereas the order of selectivity was B>o>y

In all active bismuth molybdate catalysts, lattice oxygen plays the key role in

the catalytic activity for partial oxidation of propylene The role of lattice oxygen

has been proven by tracer studies using '*O [46, 47] as well as photoelectron

characterization of the catalyst surface [48, 49] The importance of lattice oxygen in

the catalytic selective oxidation led to the conclusion that an active catalyst has to

be able to provide lattice oxygen for oxidation

Batist et al [14] attempted to correlate the Bismuth-Molybdate activity and

selectivity with the bulk crystal structure In their opinion, the catalytic activity for

selective oxidation of olefin was coumected with the comer-sharing Mo-O

octahedron, while the catalytic activity was largely absent in compounds containing edge-shared octahedron ‘hey believed that in the 1:1 bismuth to molybdenum ratio compound, Mo ion has an octahedral coordination, which allows a high catalytic activity In addition, they suggested that the interlinking of Mo- octahedron by edge sharing is responsible for the low activity of the 2:3 molybdate Llowever,

detailed studies [11, 12] have shown that in the 2:3 compound, the molybdenum ion

is surrounded by a distorted tetrahedron Infrared and ultraviolet investigations by

kumar and Ruckenstein [50] revealed the presence of both tetrahedral and octahedral oxomolybdemum species on the surface of the 2:1 and 1:1 active

Bismuth- Molybdates

The importance of cdge sharing Mo-O octahedron was also reported in

Haber’s review [51] When the reaction ocourred on the bismuth molybdate surface,

the uptake of oxygen from the surface caused oxygen vacancies The oxygen

vacaticies gencrated clectroplihe oxygen, which they led to iolal oxidation

products The oxygen vacancies, suprisingly, were not found on the reacting surface The absence of clecrophilic surface oxygen is now believed to be causad by the so-called “shear plane”, where the corner-sharing Mo-O tumed into the edge- sharing Mo-O The ability of bismuth molybdate to sorve the change of the sharing

is believed to be the source of their selectivity for partial oxidation reaction

Buttrey et al [52] mentioned that the event that is taking place on the surface could lead to the structural reorganization ‘they investigated the structure of Big03.nMoO), orystal, covering n=3 (a phase), n=2 (B phase} and n=1 (y phase) to

find the relation between the structure and catalytic activity and concluded that all

active phases were actually derived from the fluorite structure They further conchided that fluorite structure was possibly responsible for the accommodation of

either cation or anion vacancies in the crystal framework for the n = 2 and 3 and

determined the catalyst activily toward the catalylic selective oxidation of olefin For the phase with n = 1, fluorite-related structure did not exist and its activity was

duc lo the facile palh of lattice oxygen (O* ) diffusion through MoO, and BisO)

layers [53] Bultrey et al [52], however, confessed thal the structure identified in

ther study might be different from the real structure under actual reaction

Thục T.D - 2009

2

conditions This raises a question about the role of the structure of the catalyst under

the reaction condition ‘his will be further investigated in more detail in this thesis

L1.4 Catalytic performance of bismuth molybdate

Although bismuth molybdate have been the best calalyst for selective

oxidation of ole(in so far, but the catalytic aetzvitics bas beer still very low Fansuri,

LL etal studied the selective oxidation of propylene to acrolein on B- BiyMoy(y and they found the conversion of propylene just reached about 4% at 450°C with the best sample M.'.Le et al [31 jstudied the oxidation of propylene to acrolein for very long time, they also showed that conversion of propylene was only 3.8% at 425°C over 0.1 gram of pure phase bismuth molybdates Hence, improving the catalytic activities of bismuth molybdate containing materials has attracted much attention from scientists, recently This section briefly summarized researches which aimed

to improve the catalytic activities of bismuth molybdates recently

M.T.Le et al; Ji Chul Jung et al [30, 54] found the synergy effect between o- BigMo;012 and y-BiyMoOg for the selective oxidation of propylene to acrolein The synergistic effect of these two catalysts was explained by a remote control iechanism, in which oxygen species formed on the j-BisMoQg migrals onto Ihe surface of the œ-ii,IMo;O; to create aotive sites Therefore, it is believed that the synergistio effeot of the œ-I;Mo;Oi; and -BiyMoOs catalysts in the selective

oxidalion of propylene was dus to a combmahom of the facile oxygen mobility of y-

ByMoQ, catalyst and the abundant adsorption sites of o- BiM]oj0); catalyst for propylene This synergistic cffcct could help increase the selectivity significantly in comparison with that of pure œ and y phases at the same reaction conditions

On the other hand, many attempts have been done to find out the multiphase

oxides systems (Bi, Mo, Co, Ti, Sn, Sh, Pb, Se, Si, .O) which was believed be

able to increase the activities of bismuth molybdate containing catalysts for

selective oxidalion of propylene [55] Tnsertion of various clements in bismuth

molybdate system which could make new materials and perhaps, good catalysis are made Because modifiers affect the formalion of new phases, increase the

proportion of interfacial regions with a certain structure, and promote the growth of

molybdate microcrystals in such a way that faces active in sclective oxidation are in

large execss, Furthermore, modifiers can change the acid-base properties of the

catalyst and the rate of electron and oxygen transfer O V Udalova et al [55]

studied Cog gMoj Fe, sBips o.25Sbo1Kp1O, catalysts for selective oxidation of

propylene to acrolein Ihe results obtained provide insights into the role of the components of the catalyst CoMo(, forms the structural framework of the catalyst fron molybdate can be stabilized on CoMoO, as P-phase As its content is increased, the catalyst gains activity but its selectivity declines 13ismuth molybdate

is responsible for the selectivity of the process When present in small amounts, MoO, raises the selectivity, binds free oxides, and converts reduced molybdates

into their oxidized forms lixcess molybdenum trioxide causes a dramatic fall in the catalytic activity Potassium and antimony decrease the catalytic activity, but even small amounts of these elements raise the selectivity of the catalyst Chromium can

substitute for iron atoms in the mukicomponents catalyst Ni, Mn, and Mg substitute for Fe in iron molybdate to decrease the catalytic activity M.T.Le et al

[31] suggested adding SnO, and ZrO, on beta phase bismuth molybdate to increase

conductivity of sample With 10%SnO, adding, the conductivity inoreased

significantly which led to very strong synergy effects resulted in increasing catalytic

performance

Recently, some scientists have focused on increasing surface area of bismuth snolybdate catalysts by many ways [10, 56, 57] In order to overcome the surface area limitation of bismuth molybdates, two main directions, supported bismuth smolybdates on supports and nanometer sized bismuth molybdates, have beon

investigated Yo-Han Han et al [10] studied lattice oxide ion-transfer effect in

silica-supported y phase bismuth molybdate catalysts for selective oxidation of propylene to acrolein The highly dispersed bismuth molybdate catalysts on silica

were found to be intrinsically active but poorly selective to acrolein When they

increased the loading amount the oxidation activity drastically increased The poor

acrolein selectivity of this catalyst was improved by continuous use in the catalytic

oxidation for making the particle size of the dispersed bismuth molybdate larger

The catalylic activity and selectivily were litle influenced by the loading amount The results demonstrate that, for the activity and selectivity, bismuth molybdate

catalysis need to be of a cerlain particle sive which ean provide sulTicient lattice

oxide ions during the catalytic redox cycle Ashish Bhakoo el al [56] studied supported binary oxide monolayers 10%BiyMo;0,2/TiO, The selectivity al 500°C

Thục T.D - 2009

31

is as good as that of unsupported ơ-BiyMoyOx;, implying that a large reservoir of lattice oxide ions is not a prerequisite for a selective oxidation catalyst, although the yield is only moderate Chao Xu ct al [$7] succeed in preparation of 7-BiyMoO,

nanoplates by hydrothermal with support of PVP (poly vinyl pyrrolidone) as

surfactant The resulls indicated that as-prepared BizMoOQ, product bad a typical plate-like structure with a thickness range of 100-150 nm This results promised to

open the ability Lo inorcase surface arca of bismuth molybdates

In summary, it could be said that researches about improving catalytic activities of bismuth molybdates have been achieved some of great successes However, most of them just have been focused on changing composition of bismuth molybdate contained catalyst systems which aimed to create synergy effects as well

as to change the rate of electron and oxygen transfer so on A new idea is to increase the specific surface area of pure bismuth molybdates or supported bismuth molybdate catalysts has still attracted less attentions ew authors have reported some initial results about improving surface area of bismuth molybdates [10, 56, 57], Although increasing the surface area of bismuth molybdates has seemed not bring the excellent results, the initial data gave us a chance to carry on studying

1.2 Selective oxidation of propylene to acrolein

L2.1 Acrolein and its production

Tn order to understand the importance of acrolein products, il is necessary lo

take a look about it Acrolein was first prepared in 1843 by Redtenbacher by the dry distillation of fal [58] Commercial production of acrolein began in Gormany in

1942, by a process based on the vapour-phase condensation of acetaldehyde and

formaldeliyde This method was used until 1959 when a process was introduced for

producing acrolein by vapourphase oxidation of propylene [59] Several catalysis have been used in the vapour-phase oxidation of propylene, including cuprous

oxide, bismulht molybdate and antimony oxide [60] All commercial production of acrolein is curently based on propylene oxidation [59] In 1975, global production

of acrolein was about 59 000 tonnes [60] The worldwide capacity for production of xefined acrolein is about 1 13 000 tonnes per year |61[ Acrolein is produced by three companies each in Japan and the United States of America and by one

company each in France and Germany [62]

glutaraldehyde, pyridines, tetrahydrobenzaldehyde, allyl alcohol and glycerol, 1,4 butanedial and 1,4-butenediol, 1,3-propanediol, DL-glyceraldehyde, flavours and fragrances, polyurethane and polyester resins [59, 63] According to the report from Occupational Safety and Health Administration of US Department of Labour, 50%

of acrolein production is for glycerine, 25% for methionine, and the rest for other applications ‘the most important direct use of acrolein is as a biocide: It is used as a herbicide and to control algae, aquatic weeds and molluses in reciroulating process water systems It is further used to control the growth of microorganisms in liquid fuel, the growth of algae in oil fields and the formation of slime in paper

manufacture Acrolein has been used in leather tanning and as a tissue fixative in histological work [59, 60, 61] Acrolein is widely known as a biocide Because of

its biocidal activity, acrolein is commonly applied as an aquatic herbicide and

slimicide The molecular structure of acrolein is shown in Figure 1.7 Acrofein has

twa conjugative unsaturated carbon bonding, one from the vinyl group and the other

from aldehyde group Due to the existence of these groups, acrolein possesses

reactions characteristic of both an unsaturated and an aldehyde compounds Some

examples of reaction characleristics of acrolein are Diels-Alder reaction between

two acrolein molecules, addition to carbon-carbon double bond, polymerisation, and

yeaciion with amine compounds [2]

Basic chemical and physical properties (@) Description: Colonrless to yellowish liquid with extremely acrid, irritating odour (b) Boilng-point: $2.5-53.5 °C

(¢) Melting-point: -86.9 °C Figure 1.7 Molecule struoture of acrolein (gy pyansity: 0.8410 at 20 °C/4 °C

Ht

c=c-c<

Thục T.D - 2009

33

The activity of the aldchyde group to attack an amine-containing molecule makes it reactive to a protein molecule ‘ihe reaction underlies the activity of

acroleii as an anti microbial and biocide, where the reaction of acrolem with protein

on a cell wall can cause damage to the cell and kill it Acrolein is very toxic and

flammable It also undergoes polymerization easily and exothennally ‘the polymerization can be initiated by light, heat, air or peroxides It is also polymerized in the existence of alkaline solution such as amines, ammonia, and caustic soda or by mineral acids such as sulphuric acid Acrolein polymerization, initiated by alkaline or acids, is very exothermic and no inhibitors can stop the polymerization once it is initiated Acrolein reacts with water and therefore addition

of water to stored acrolein must be avoided, as the acrolein-containing water layer is

particularly prone to polymerization Acrolein vapour polymerises upon condensation [2]

1.2.2 Thermodynamic of selective oxidation of propylene to acralein

For industrial applications, aorolein is commercially produced by heterogeneously catalysed gas-phase oxidation of propylene [2] ‘The catalytic partial oxidation of propylene to form acrolein follows the reaction equation below

‘The reaction is exothermic and produces 340.8 kJ of heat per mol of reacted propylene The Gibbs free energy (AG) shows that the reaction will spontaneously

occur, once the reaction is initiated

CH,CH CH,+0, »CH, CHCHO+H,0

AH? = 340.8 kJ mol-1

AG? = -180.19 kF mol-1 Although the reaction is energetically spontaneous, acrolein is not the anly

produclk when propylene is reacted with oxygen Several other products such as

CO,, CO, acetaldehyde, formaldehyde carboxylic acids ete can also form Some

possible products and their thermodynamic paramoters are listed in Table 1.1

‘the thermodynamic parameters shown in ‘fable 1.1 reveal that the formation af side products (such as CO, and CO) is more thermodynamically favourable than the

formation of acrolein

‘Table 1.1 ‘Thermodynamic parameters of the formation of other propylene oxidation

Cylis(e) + O2(g) > Cal L,O20) + HạO(Ð -550.2293 | -404.21

CyHe(g) | O2(g) > CsHyO) | HOC) -338.7959 | -426.24

(* data taken from Grasselli |64])

12.3 Catalysts for sclective oxidation of propylene

Commercial production of acrolein by heterogeneous calalytic selective

oxidalion of propylene started im 1948, when cuprous oxide calalysis were chosen

for Shell Chemical Company [65] The acrokein yield from this process, however, was less then 50% from the total input of propylene, which [ef enormous opportunily (o develop better catalysts than cuprous oxide In the 1950s, SOHTO

initiated an exlensive research info calalylic vapour phase oxidalion of unsaluraled

hydrocarbons to produce useful products such as acrolem, acrylonitrile,

molhaerylonilrile, and acrylic avid [6] The research was developed based on the

Lewis concept | 66|, in that lattice oxygen of reducible metal oxide would serve as a betler oxidant for hydrocarbon selective oxidation dum the gas phase oxygen Several catalysts, either single metal oxide or a mixture of two or more metal oxides were tested for their activity towards acrolem formation from propylene In the early stages of the development, SOHIO found that some mixtures of metal oxides gave good selectivity as well as yield of acrolein, However, the catalysts did not work properly in the existent of molecular oxygen in the feed stream The extensive

research led the SOHIO team to discover bismuth-molybdate, which had better

selectivity for acrolein than cuprous oxide [2], even when the feed stream contained smotecular oxygen Since then, the bismutl-molybdate system has been significantly improved by adding some metal oxides to form multicomponent catalysts These catalyst systems have even better activity and selectivity for the oxidation of

Thục T.D - 2009

propylene to acrolein, Some cxamples of bismuth-molybdate based multi- component catalysts are presented in Table 1.2

Table 1.2 Some examples of multi-component BiMo based catalysts

yoquirern I, the catalysts must have the abilily to perform internal oxidation -

reduction processes Copper oxide is one of many examples of metal oxides that have such ability It is also the only single component metal oxide that has good activity and selectivity for paitial oxidation of olefin In fact in cuprous oxide, the

copper exists in threc-oxidation statc, namely Cu motal, Cu” and Cu”, allowing

internal reduction-oxidation ‘These ious constitute a muxkwe of Cu-Cu,0-CuO

system Ybat means the selective oxidation of propylene occurs on two kinds of active sites (CuO and CuO sites) An electron diffraction study on a thin layer of copper film showed that the cuprous oxide is responsible for the activity of the catalyst [46] Bettahar [72] reported in a review that a stoichiometric or copper rich CuO is the most active catalyst for the selective oxidation of propylene to acrolein while oxygen rich Cu,0 favours CO, and 11:0, which is in agreement with the report by Krenzke [46], Lixtensive studies on copper oxide catalysts led to the conclusion that metal oxides catalysts for selective oxidation reactions should

contain two metal ion species in the form of M'M"O, The M! ion is normally a

heavy metal and M™ from transition metal groups, which have variable oxidation

numbers, From this requirement, tin antimonite, wanum antimonite and bisnvuth ainolybdate are among the most active and extensively studied binary metal oxide catalysts Individual metal oxides do not have sufficient activity to convert propylene to acrolein, but mixtures can have very good activity and selectivity

Lranium antimonite, as reported by Krenzke [46], has excellent selectivity

towards acrolein over a wide range of reaction temperatures There are two phases

found in this catalyst system, namely USb,Oy» and USbO, [73] where only SbV and

‘UV species are contained in the catalyst The first phase is selective and active for acrolein [ortnalion and hence, ils consentralion determines Ihe overall performance

of calalysl, while the second phase has lower seleclivily towards acrolein Matsuura

[74] reported thal the catalyst has special oxygen mobility, which favours surface

reaction of propylene partial oxidation 10 acrolein However, due to ils bighly

radioactive nature, Lis calalyst has litle scope for commercial application

‘Tin antimonite is a mixture of SnO, and a solid solution of ø-Sb;O and Sb,0, where none of them, as individual, is active as a catalyst for propylene oxidation to acrolein [75] It has been reported that the addition of only 6 8% Sb to tin(LV) oxide increased the conductivity by three orders of magnitude [46] ‘his is

attributed to the substitution of Sb** for Sn‘* ions in the tin(IV) oxide lattice It is

believed that the catalytic activity of these mixed oxides is related to the presence of the solid solution and the Sb;0, species lying on the surface of Sn,

Rismuth molybdates are the second most studied catalyst after copper oxide

and serve as (he main ingredient in almost all commercial calalysis for propylene

oxidalion to acrolein There are several known phases of binary bismuth molybdates

bul onily three phases are active and selective for the partial oxidation of propyl

Œ

to acroleim |Š, 76] The ralios of Bì to Mo in these catalysis are within the range of

2:3 and 2:1, namely o-BisMo;012, f-BrMo20», and y-Bi2MoO5

1.2.4 The Reaction Mechanisms of Selective Oxidation of Propylene to Acrotein

on Bismuth Molybdate Catalysts

The mechanism of propylene oxidation has beew investigaled through three important components related to the reaction: propylene, oxygen and the catalyst

system Art oxidation reaclion may be considered as lwo groups of reactions:

electrophilic oxidation, which proceeds through the activation of oxygen and

Thục T.D - 2009

nucleophilic oxidation, which involves hydrocarbon activation in the first step and aucleophilic oxygen insertion in one of the further steps ‘The details were given below

i, Propylene Activation

It is probably Adams and Jennings [3] who first started the studies of the reaction mechanisms of propylene partial oxidation to acrolein over bismuth moybdate catalysts In they work, Adam and Jemning used propylene labelled with deuterium and used a kinetic isotope effect analysis to find out the probability of deuterium atoms being abstracted relative to hydrogen atoms The conclusions of their work were:

1 The first slep in the oxidation of propylene is the abstraction of a hydrogen alom from the methyl group and the process is (he rate-deiermining step

2 The hydrocarbon intermediate formed aiter the abstraction is a symmetrical structure, probably similar to the a-allylic species ‘This species is not cyclic because no deuterium was found in the middle of carbon atoms of the resulting acrolein

3 Propylene underwent two successive hydrogen abstractions before the addition of oxygen

The existence of the symmetrical intermediate was also found by Voge ct al

177] by using propylene labelled with “C on a cuprous oxide catalyst at 300°C ‘The

same conelusion was also made by Krenzke [46] Report on studies where allyl radicals were generated in-situ, either by allyl iodide [78] or from gas phase radicals

1791, confirmed the existence of the symmetrical allyl intermediate Figure 1.8 gives the reaction path of the propylene activation via the formation of allylic intermediate

of the products [16]

Transient response and spectroscopy methods have also been used to prove the existence of the allylic radical Kobayashi [80] used a transient response method ona bismuth molybdate with Bi‘Mo ratio of about 1.0 and comprised of alpha, beta and gamma phases, and revealed the existence of the allylic intermediate The allylic intermediate was also reported by Martir and Lundsford [79] using an EER spectroscopy method, and by Schultz and Beauchamp [81] using a photoelectron

spectroscopy The study by Carvazan et al [82] using an FT-IR spectroscopy

technique on a multiphase Bi, Mo, and Co catalyst and proposed a mechanism which is in agreement, with those of Cullis and Huckwall [7] and Rettahar et.al [72]

A more recent study by Ono, Ogata, and Kuczkowski [83] using labelled oxygen and microwave spcetroscopy alxo supporled the existence of allylic inlermnediale

Several studies [84, 85] ou the formation of side products also confirmed the allylic intermediate mechanism The side products normally accompany the selective oxidation of propylene to acrolein Keulks and Daniel [86] investigated the

oxidation of MC labelled and unlabelled acrolein They found that the carbon

dioxide is formed almost exclusively from the further oxidation of acrolein The

Thục T.D - 2009

39

same result was also found carlicr by Russian roscarchors |84, 85] Figure 1.9 shows

the reaction path of the formation of some side produets

Figure 1.9 The reaction paths of the formation of side products [46]

All of the above studies lead to the same conclusion that the selective

oxidation of propylene occurs through the formation of the allylic radical The formation of allylic intermediate in the reaction of partial oxidation of propylene over bismuth molybdate catalysts has been widely accepted although the isolation

of the allylic and allyl-peroxiradicals was only proven in 198] by Martir and Lundsford [79] in their matrix isolationEPR studies on o-Bi,Mo,0,, and y-

BiMoO;

ii, Oxygen Insertion

The step after propylene activalion is the abstraction of second hydrogen and

the oxygen insertion into allyl group The abstraction of the second hydrogen is unlikely to oceur af the same line as the allylic [ormation is taking place because the reaction is cnergetically rate-determiming stop Studics using sclf-gencrating allylic species [87, 78] have showed ihal the secrmd hydrogen was abstracted from

the allylic intermediate

catalysts, Burrington et al [80] used allyl alcohol-1,l-d2 and -3,3-d2 in their investigation ‘hey conclded that the insertion of oxygen occurs before the abstraction of the second hydrogen and facilitated by the presence of a C—O bond

Sa far, no mention has been made of the source of oxygen for oxygen

insertion into the allylic intermediate For the formation of acrolein, there are two

possible sources of oxygen One involves the use of lattice oxygen and the other an

adsorbed form of molecular or gas phase oxygen The first mechanism is referred to

as the redox mecharism where the calalyst ilself acs as the oxidising agent Tn Uns

mechanism, molecular oxygen serves only to re-oxidise the reduced catalyst The

second type of oxygen reacts with the allylic intermediate to form hydroperoxide

The hydroperoxide then decomposes lo acrolein and water

‘The redox mechanism was first proposed by Mars and van Krevelen [88] and

dbas since been known as the Mars and van Krevelen mechanism In early 1954, Mars and van Krevelen [89] concluded that the catalytic oxidation of hydrocarbon

took place in two steps: (a) the reaction between hydrocarbon and the oxide in

which the latter is reduced and the former is oxidised, followed by (b) reoxidation

of the reduced catalyst by gaseous oxygen to the original state of activity and selectivity The Mars and van Krevelen mechanism is schematically given in Figure

1.10, The concept where the lattice oxygen of a reducible metal oxide could serve

as a useful oxidising agent for hydrocarbons was actually the basis of the early work

at SOIT[O which led to the development of bismuth molybdate catalyst [6]

Tho usefulness of lattice oxygen as Ihe source of oxygen for oxidation reactions is dus Lo Ihe fact that the oxygen is relalively casy to be removed from Ihe lattice Burlamacchi ct al 190] gave ovidence that the lattice oxygen ions on y- BigMoO, surface could be easily removed by applying high vacuum al lomperalures

above 300°C Fattore et al |91| obseived that acrolein could be formed without any gasvous oxygen over bismuth molybdate and several olher metal oxides catalysts

Thục T.D - 2009