Embed Size (px)
Citation preview
SURFACE EFFECTS ON HOMOGENEOUS ORGANIC REACTIONS IN MICROREACTORS
By
Abhinav Jain (B.Tech (Chemical Engg.), National Institute of Technology
Karnataka, India)
Submitted to the Department of Chemical & Biomolecular Engineering in partial fulfillment of the requirements for the
Master of Engineering in Chemical Engineering
at the
NATIONAL UNIVERSITY OF SINGAPORE
August 2010
© National University of Singapore 2010. All rights reserved.
Author…………………………………………………………………….................. Abhinav Jain
Department of Chemical & Biomolecular Engineering August 19th, 2010
Certified by………………………………………………………………………………..........
Saif A. Khan Assistant Professor of Chemical & Biomolecular Engineering,
Thesis Supervisor (National University of Singapore). Certified by………………………………………………………………………......................
Dr Levent Yobas Assistant Professor at Dept of Electronic & Computer Engineering,
Thesis Supervisor (Hong Kong University of Science and Technology).
ii
To my grandfather Chain Sukh Das
and my family for passing on
immense knowledge and courage
iii
Acknowledgement
The concept of one man army, one person solely moving a hill to bring a change or answer an unanswered question is long gone. As like most of us, I needed a team to compensate for my weaknesses and guide me in my research endeavor. This was my team-
Supervisors: Dr Saif A. KHAN at the Department of Chemical and Biomolecular Engineeirng, National University of Singapore and Dr Levent YOBAS, formerly with A*STAR Institute of Microelectronics, Singapore. Thank you both for being amazing guides. Dr Khan; you have been a continuous source of inspiration and motivation for me. You ignored my blunders and look through to my intentions. Encouragement given by you to think more creatively and learn from mistakes cannot be substituted in my life. Dr Yobas; your motivation to explore fascinating world of microfabrication brought my thoughts to the real world.
Mentors: Dr Md. Taifur Rahman, Singapore MIT Alliance and Dr Kang Tae Goo, A*STAR Institute of Microelectronics. Both Dr Rahman and Dr Kang played a very crucial role in shaping this thesis work. Dr Rahman; you always welcomed my queries and advised me on small yet significant hurdles I faced during the experimentation. You were always there to talk to not only as a mentor but also as a good friend. Dr Kang; you shared your experience in microfabrication and facilitated my work at the Institute. I cannot imagine fabricating microreactors without your support.
Co-workers: Pravien, Suhanya, Zahra, Sophia, Annalicia, Anna, Kasun G., Vaibhav and Daniel Sutter. You guys made this happen. Daniel working with you was fun and exciting. Your observations and reasoning made while working together later helped me in cracking the bubble-problem in the UV spectrometer. Kasun your assistance in performing experiments is highly appreciated. Time spend with you in lab is a wonderful memory. Pravien and Suhanya, thanks for all your moral support and assistance in performing experiments and proof-reading the thesis. It was great arguing with you. Zahra, Sophia, Annalicia, Anna, Vaibhav, Pravien and Suhanya; you guys made my stay in laboratory fascinating. The Karaoke songs we sang together, group lunch we went out every Friday and movies we watched together were moments to savor.
Facilitators: Ms Sylvia Wan, Jamie Seo, Ms Novel at National University of Singapore and Ms Trang, Ms Sarah, Dr Teo, Mr Lawrence at A*STAR Institute of Microelectronics. Thank you all for your kind support in procuring consumables and assisting in microreactor characterization and fabrication.
Friends: Suresh, Naresh, Arun, Vinayak, Michael, Suvankar, Max, Anoop, Joon, Evan, Miti and Thaneer. Thank you all for your continuous support in making my stay in Singapore an amazing chapter of my life. Miti, thanks for helping me out on various fronts zillion of times. I’m lucky to have you, Piyush and Veer in Singapore.
Family: My parents, uncles and aunts, brothers, cousins and in-laws, nephews and nieces, and grandparents. I cannot imagine coming so far in life without your encouragement and support. You are the light of my life.
iv
I also gracefully acknowledge the Department of Chemical and Biomoleular Engineering, National University of Singapore, for providing an opportunity and financial assistance to pursue my Master degree. I thank A*STAR Institute of Microelectronics for providing their facilities for my research work.
Abhinav Jain August 19th, 2010, Singapore.
v
Table of Contents
Acknowledgements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .iii
Table of Contents.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii
List of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . x
List of Figures. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .xi
List of Schemes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii
1 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Microreactors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1.1 General background on Microreactors. . . . . . . . . . . . . . . . . . . . 1
1.1.2 Types of Microreactors. . . . . . . . . . . . . . . . . . . . . . . . . . . . .2
1.1.3 Transport properties in Microreactor . . . . . . . . . . . . . . . . . . . . . 7
1.1.4 Microreactors in Action . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.2 Organic Synthesis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.2.1 Heterogeneous reactions . . . . . . . . . . . . . . . . . . . . . . . . . . .10
1.2.2 Homogeneous reactions. . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.3 Microreactors for Organic Synthesis . . . . . . . . . . . . . . . . . . . . . . . 11
1.3.1 Heterogeneous reactions in microreactors . . . . . . . . . . . . . . . . . 12
1.3.2 Homogeneous reactions in microreactors . . . . . . . . . . . . . . . . . 14
1.4 Enhancement of reaction rates: the missing link & Motivation . . . . . . . . . 15
1.5 Structure of Thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
1.6 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2 Methodology. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.1 Inside a Microreactor. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.1.1 Effect of Temperature. . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.1.2 Effect of Pressure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.1.3 Effect of Surfaces. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.1.4 Conclusion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.2 Designing the Experiment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .26
2.2.1 Selection of Chemistry . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.2.2 Selection of Microreactor . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .31
2.4 References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
vi
3 Silicon Microreactors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .34
3.1 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.1.1 Silicon Microreactors and Chemical Engineering. . . . . . . . . . . . . .34
3.2 Microfabrication of Silicon Microreactor. . . . . . . . . . . . . . . . . . . . . 36
3.2.1 Development of Protocol. . . . . . . . . . . . . . . . . . . . . . . . . . .39
3.2.2 Development of Photolithography mask. . . . . . . . . . . . . . . . . . . 39
3.2.3 Fabrication Steps. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
3.3 Interconnecting Microreactor. . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
3.3.1 Solder-based interconnects. . . . . . . . . . . . . . . . . . . . . . . . . . .45
3.3.2 O ring-based interconnects . . . . . . . . . . . . . . . . . . . . . . . . . .47
3.3.3 Sealant-based interconnects. . . . . . . . . . . . . . . . . . . . . . . . . 48
3.4 Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .49
3.5 References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
4 Experimentation and Observations. . . . . . . . . . . . . . . . . . . . . . . . . . . 52
4.1 Experimental setup. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
4.2 Experimental Protocol. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
4.3 Sampling and detection. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
4.3.1 UV-Vis Spectrometry. . . . . . . . . . . . . . . . . . . . . . . . . . . . .55
4.3.2 GC-FID Analysis. . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . 59
4.4 Experimentation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
4.4.1 Silicon based microreactors. . . . . . . . . . . . . . . . . . . . . . . . .61
4.4.2 Polymer based microcapillaries. . . . . . . . . . . . . . . . . . . . . . .62
4.5 Results and Discussions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
4.5.1 Same surface-to-volume ratio. . . . . . . . . . . . . . . . . . . . . . . .64
4.5.2 Same material but different surface-to-volume ratio. . . . . . . . . . . . 66
4.6 Error Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .67
4.7 Heterogeneity and Organic reactions. . . . . . . . . . . . . . . . . . . . . . . . 69
4.7.1 On Water reactions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
4.7.2 Surfaces and Organic reactions. . . . . . . . . . . . . . . . . . . . . . . .71
4.8 Microreactors and Organic Reactions revisited. . . . . . . . . . . . . . . . . . . 72
4.9 Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
4.10 References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
5 Summary and Outlook. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
5.1 Principal Thesis Contributions. . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
vii
Appendix A. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . .81
Appendix B. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . .88
viii
Summary
This thesis focuses on microreactors used for single-phase organic reactions and their effect
on the chemical transformation. Microreactors are defined as micro-structured flow vessels
in which at least one of the geometric dimensions is in micrometer size range. In recent
years the area has seen extensive development, especially for studying and performing
organic syntheses by both academia and industry.
Microreactor technology promises superior control, safety, selectivity and yields in
chemical transformations. High surface-to-volume ratio achieved in microreactors enables
excellent heat and mass transfer rates by facilitating better transport of reacting species and
properties. Although they are relatively expensive to fabricate and have limited capabilities
to handle solid reactants, higher yield obtained and minimal waste generation makes the
overall chemical synthesis economically viable. One of the striking features of using such
reactors for both homogeneous and heterogeneous organic synthesis is dramatic
improvement in reaction rates and yields compared to conventional macro sized reaction
vessels such as bench-top flask. It is argued that this increase is a direct outcome of
enhanced transport properties (heat and mass) realized in microreactors. This enhancement
accelerates reaction rates, yield and selectivity by shifting diffusion controlled reaction
system to kinetically-controlled reaction regime. The argument is valid for heterogeneous
chemical reactions where overall reaction rate is limited by transfer of chemical species
across phases, or where the reaction rate is a strong function of temperature. However in
principle, factors such as inter-phase heat and mass transfer should not affect course of
well-mixed quasi-isothermal homogeneous reactions. Thus, the observed increase in
reaction rate for homogeneous chemical reaction in microreactors has sparked a debate
regarding their reaction mechanism in the research community.
In this work we attempt to analyze this deviation in theoretical and observed experimental
reaction parameters by hypothesizing the increase in reaction rate as a direct consequence of
appreciable participation of reactor walls (surfaces) in a microreactor. In other words, we
hypothesize that homogeneous reactant experiences significant participation of reactor walls
due to high surface-to-volume ratios. This leads to higher chemical transformation; in effect
‘heterogenizing’ a homogeneous reaction. The hypothesis is investigated by performing
single-phase organic reaction experiments in micro-capillary reactors of different materials
and internal cross-sectional areas. We compared the conversion of reactants in
microreactors of different materials with same surface-to-volume ratio and vice-versa.
ix
The outcome of our study indicates higher conversions in the microreactors as compared to
an equivalent synthesis in a macro-scale system with noticeable difference with different
material of construction. However a firm conclusion could not be derive due to errors
associated with the measurements. Furthermore, we attribute the observed increase in yield
is due to participation of reactor surfaces, as in light of similar phenomena observed in ‘on-
water’ and ‘on-surface’ reaction studies.
x
List of Tables
Table 2.1 – Microcapillaries and their surface-to-volume ratios.
Table 3.1 – Etching of Silicon wafers.
Table 3.2 – Surface-to-volume ratios for the designed microreactors.
Table 4.1 – Flow rates for both Silicon microreactors and Polymer Microcapillary.
Table 4.2 – Developed method used for GC-FID analysis.
Table 4.3 – Retention time of reactants and products.
Table 4.4 – Chemical structure and repeated units in the polymeric material.
xi
List of Figures
Figure 1.1 – Microchannels generated by wet etching of a stainless steel foil.
Figure 1.2 – Typical Selective Layer Melting fabrication process layout.
Figure 1.3 – FlowStart, a commercial microreactor platform for chemists.
Figure 3.1 – Different types of silanol groups with hydrogen bonding.
Figure 3.2 – Isotropic and Anisotropic etching of a masked surface.
Figure 3.3 – Reactive Ion etching process.
Figure 3.4 – Design microreactor with extended surface.
Figure 3.5 – Extended surface and rectangular channel (all units in mm).
Figure 3.6 – Microfabrication steps and microreactor cross sections.
Figure 3.7 – Microfabrication steps and microreactor cross sections.
Figure 3.8 – Soldering metal ferrules with a silicon microreactor on a hot plate.
Figure 3.9 – Delamination of deposited metal layer on microreactor along the dicing lines.
Figure 3.10 – O-ring based microreactor packaging.
Figure 3.11 – Microreactor packed in a epoxy based sealant.
Figure 4.1 – Block diagram of the experimental setup.
Figure 4.2 – a) Silicon microreactor with optical fiber based online UV-vis analysis. b) Inset showing the optical fibers running inside the microreactor.
Figure 4.3 – Cross-sectional view of microreactor depicting misalignment problem.
Figure 4.4 – Fabricated microcross UV flow cell.
Figure 4.5 – Signal intensity affected by bubbles in the flow system recorded over time at wavelengths of 450 nm(black), 400 nm(magenta) and 240 nm(blue); flow rate = 20 ml/min.
Figure 4.6 – A gas chromatogram of a chemical mixture obtained using a Flame ionization detector; intensity is plotted against time.
Figure 4.7 – Delamination of epoxy from a silicon microreactor.
Figure 4.8 – Images of the patterned surface of a silicon wafer during microfabrication.
Figure 4.9 – GC-FID chromatogram for a sample.
xii
Figure 4.10 – Plot of conversion in microreactors with surface-to-volume ratio of 7874 m2/m3.
Figure 4.11 – Plot of conversion in microreactors with surface-to-volume ratio of 15748 m2/m3.
Figure 4.12 – Plot of conversion in microreactors with surface-to-volume ratio of 22857 m2/m3.
Figure 4.13 – Plot of conversion in Radel R microreactors and batch system.
Figure 4.14 – Plot of conversion in PEEK microreactors and batch system.
Figure 4.15 – Plot of conversion in FEP microreactors and batch system.
Figure 4.16 – On water reactions in comparison to the neat and aqueous homogeneous reactions.
Figure 4.17 – Mechanism for reaction of carboxylic amino acid on SiO2 surface.
Figure 4.18 – Mechanism for reaction of benzoquinone with methyl indole on a surface bearing hydrogen bonds.
xiii
List of Schemes
Scheme 1.1 – Reaction between ethane and chlorine.
Scheme 1.2 – Reaction between a vitamin intermediate in hexane with conc. sulfuric acid.
Scheme 1.3 – Suzuki reaction between phenylboronic acid and 4-bromobenzonitrile in oxolane-water mixture.
Scheme 1.4 – Suzuki reaction between 3-bromobenzaldehyde and 4-fluoro-phenyl boronic acid.
Scheme 2.1 – Coupling reaction between 1,4 benzoquinone and 2-methyl indole.
Scheme 2.2 – Coupling reaction between 1,4-benzoquinone and a propanethiol.
Scheme 2.3 – Coupling reaction between 2,5-Dichloro-1,4-benzoquinone and 2-methyl indole.
Scheme 3.1 – Condensation reaction between a silane and an isolated silanol group.
Scheme 4.1 – Tautomerism in benzoquinone.
1
1. Introduction
This introductory thesis chapter outlines microreactors and organic reactions in general. The
discussion starts with types, fabrication approach and properties of microreactors, followed
by introduction to organic reactions and application of microreactors in organic synthesis.
The discussion provides a firm foundation to the investigation carried out in this work, and
sets stage for the hypothesis outlined in the later sections of this chapter.
1.1 Microreactors
Microreactors are miniature reaction vessels for carrying out chemical reactions in which at
least one of the lateral dimensions is less than a millimeter and are also known as
microstructured reactors or microchannel reactors. In the simplest form, it is a
microchanneled flow confinement designed to carry out chemical transformations.1 A
microreactor in practice may comprise of a single or multiple chemical unit-operations to
carry out execute a desired reaction-engineering task. Depending on the application, a
microreactor can also be integrated with microsensors, microactuators and microflow-
switches to generate a “micro total analysis system”.2
1.1.1 General background of Microreactors
Microreactor technology has tremendously grown in past decade affecting nearly all
domains of science and technology. It is a relatively young technology with interesting
developments happening each day. Such developments have resulted in commercial market
ready products for diagnostics and syntheses purposes. Their unique ability to provide
enhanced heat and mass transfer rates further make them a suitable candidate to carry out
chemical and biological reactions with high yields and selectivity.
The development of microreactor technology dates back to the 1980s when a unique patent
on building a microstructured system for chemical processes was published in East
2
Germany.3 In the year 1989, the Forschungszentrum Karlsruhe, Germany presented the first
micro-heat exchanger and identified its potential for chemical systems.4 Similar works were
carried out in early 90s at Pacific Northwest National Laboratory, USA to harness potential
applications for energy sector. By the late 90s, researchers around the globe started
recognizing potential of the technology and the area showed an exponential growth since
then.5
1.1.2 Types of Microreactors
Microreactors are generally classified on the basis of material of construction. The type of
material used for construction influences physical properties of microreactors such as
hydrophilicity, zeta-potential, solvent compatibility, operating temperature and pressure
range, durability and fabrication cost.6,7,8 Based on the material of construction,
microreactors can be further classified as-
1.1.2.1 Metal based Microreactors
These reactors are chosen for applications involving high temperature and pressure. The
choice of metal for construction range from noble metals such as silver, platinum, rhodium
to their alloys with copper, titanium, stainless steel, nickel, etc.9,10 The microfabrication
methodology to process and manufacture microreactors in metals has been widely adopted
from semi-conductor device processing technology. One of the following techniques or
their combinations is employed to carry-out complete microreactor fabrication.
Etching – Etching is a process by which a material is weathered away and patterned by
selectively exposing it to an etching agent. Photolithography is the most common technique
used for patterning the surface of the material. In general the removal is a chemical process
in which the etching agent removes the exposed metal. There are two types of etching
techniques, dry etching and wet etching. Dry etching uses reactive gases or plasma to ‘eat-
away’ exposed surfaces. Wet etching uses corrosive chemical solution in place of gases or
3
plasma and is relatively cheaper than dry etching techniques. Figure 1.1 shows
microchannels generated by wet etching in a stainless steel foil.9
Figure 1.1 – Microchannels generated by wet etching of a stainless steel foil
Micromachining – Noble metals chemically resistant and are difficult to pattern using
etching agents. Precision micromachining is the most preferred choice to pattern such
metals. Micromachining can be performed by spark erosion, laser machining or mechanical
precision machining using diamond-tip tools. However there is a limitation to the
dimensions which can be processed using micromachining and depends upon the material,
technique and machine. Also, the surface smoothness of the processed patterned depends on
the type of technique employed.9
Selective Laser Melting (SLM) – Although this is one of the most expensive
microfabrication techniques, the process allows generation of full three dimensional
microstructures. In this technique, a thin layer of metal powder is distributed on the base
structure. Using a high power focused laser beam, the surface is patterned according to a 3D
CAD model. The high temperature generated by the focused beam melts and patterns the
metal on the layer. The process is repeated to give a full 3D structure. Figure 1.2 outlines a
selective laser melting process.12,13
4
Figure 1.2 – Typical Selective Layer Melting fabrication process layout13
Bonding methods- Fabricated micropatterns are assembled and bonded together to generate
a microreactor. The surface may be electropolished before assembling to have nanometer
scale surface smoothness. High precision is required in aligning as misalignment may lead
to poor or unusable microreactors. For metal based microreactor, diffusive bonding at high
temperature is the most preferred choice for bonding. This process involves bonding the
patterned laminas together under high vacuum, temperature and mechanical pressure.
1.1.2.2 Glass and Silicon based Microreactors
Glass and silicon based microreactors are extensively used in engineering systems. Ease of
fabrication, solvent compatibility, fabrication process flexibility and capability to operate at
higher temperature and pressure makes them suitable candidate for research and
development. Furthermore, extensive knowledge and expertise from semiconductor
fabrication industry is available for microfabrication of glass and silicon. Microreactors in
these materials are manufactured in following ways:
Etching- Etching is widely used microfabrication technique for glass and silicon. Dry
etching technique uses reactive gases or their plasma to preferentially ‘eat-away’ glass or
silicon. The pattern to be etched is masked using a photoresist or by generating a chemically
inert layer so that only the patterned surface is exposed to the reactive environment. Based
5
on type of etchant used, two types of etching can be achieved, i.e., isotropic etching or
anisotropic etching. In isotropic etching, the etching direction is not influenced by the
crystal lattice plane of the material and the rate of etching is uniform in all directions. In
anisotropic etching, the rate of etching is non-uniform and varies with the crystal lattice
plan of the material. Depending upon the type and process parameters for dry etching (using
reactive ions or plasma), isotropic or anisotropic etching can be achieved for both glass and
silicon. Details of plasma assisted dry etching (a.k.a. Deep reactive Ion Etching) is
discussed in chapter three.
Glass and silicon can be isotropically or anisotropically wet-etched. Glass can be
isotropically wet etched using aqueous hydrogen fluoride (generally 10%). Silicon has an
interesting etching characteristic. It gives an isotropic etch when etchant used is aqueous
solution of Hydrogen Fluoride, Nitric Acid and Acetic Acid. However, the etching of
Silicon is anisotropic when potassium hydroxide is used as etchant. KOH preferentially
attacks <100> plane of silicon crystal, giving rise to a V-groove when <110> plane of
silicon is exposed to the etchant.9 Anisotropic etching is useful for generation of special
structures such as filters in the microchannel.
Micropowder blasting – It is a micro-abrasion process in which an abrasive is impinged
using compressed air. It is analogous to sandblasting which is used for polishing and
cutting. In this technique, a masked surface is exposed to a stream of abrasive material
striking the patterned surface with a very high momentum. The high energy microabrasive
powder bombards and removes the exposed surface, leaving behind a patterned surface.
Bonding methods- Special bonding methods are used for bonding silicon and glass
micropatterns. Anodic bonding is one of the most popular techniques and is typically used
for bonding glass and silicon surfaces together. In this process, both the surfaces are kept in
close contact at a temperature of about 400~500°C and direct current between 700-1000V is
applied. The high temperature makes the glass conduct sodium ions and the applied voltage
6
drifts these ions across the contact into the silicon surface. The silicon atoms thus form a
strong chemical SiO bridge between the glass and silicon surfaces.
Fusion bonding is another bonding technique used for bonding two silicon or glass surfaces
together. In this process, surfaces are made hydrophilic by chemical treatment with aqueous
solution of ammonium hydroxide and hydrogen peroxide. The surfaces (laminas) adhere to
each other due to van der Waals interaction. For silicon-silicon bonding, the combined
microreactor is heat treated in an oxidizing kiln at around 1050°C for an hour. In case of
glass-glass bonding, the combined laminas are kept between 400~500°C for several hours.11
1.1.2.1 Polymer based Microreactors
Polymers are extensively used for manufacturing microreactors these days. The most
important advantages of polymeric microreactors compared to all other types are – ease of
fabrication, handling and patterning, lower overall manufacturing, ease of fluidic
interconnections. Microreactors in polymers are fabricated using one of the following
procedures-
Hot embossing – In this technique, a micropattern is embossed on surface of a polymeric
material such as PMMA (poly methyl meta acrylate), polycarbonate and polystyrene using
hot-press die. The technique enables high throughput and is relatively inexpensive
compared to other techniques. However, it may suffer from irregular and defective
patterning.
Extrusion – This technique generates thin microcapillary tubings like microreactors. In this
technique, long microcapillaries are extruded from a plastic-melt through a micro-nozzle.
The generated capillaries are widely used in commercial and industrial applications
including High Performance Liquid Chromatography (HPLC).
Soft lithography and patterning- Soft lithography and patterning is one of the most popular
microfabrication techniques among researchers. In this technique, a micropattern is
lithographed in photo-curing epoxy. SU-8® is one of the most widely used negative photo
7
curing epoxy. The generated microstructure acts as a negative mold and is used for rapid
generation of microreactors in elastomers such as PDMS (poly dimethyl meta siloxane) and
Poly-Urethane.14
1.1.3 Transport properties in Microreactors
In comparison to conventional reactors, the dimensions of microreactors provide very high
surface-to-volume ratios. In other words, same amount of chemical flowing through a
microreactor will see more ‘wall’ of the microreactor than when flowing through a
conventional reactor. Mathematically,
(1.2) 3(1.1) 2
LV
LA
∝
∝
(1.4) 1
1
(1.3) 1
>>
<<
∝∴
VALwhen
LVA
where, A is internal surface area and V is volume of a microreactor. Thus for a microreactor,
surface-to-volume ratio (or specific surface area) is between 10,000 m2/m3 to 50,000 m2/m3
whereas it is 100 m2/m3 for a conventional macroscopic systems.15 The enhanced specific
surface area also results in high heat-transfer coefficient of up to order of 10 kWm-2K-1,
resulting in very rapid heating and cooling rates.16 It also enables us to physically carry out
a chemical reaction in a microreactor at quasi-isothermal conditions with a well-defined
residence time. Furthermore, rapid heat-transfer rate eliminates generation of hot-spots in a
microreactors which reducing by-products formation, enhances yield of a reaction, and
enables execution of highly temperature-sensitive and exothermic reactions.
Mass transfer in microreactors is another important transport property which makes them an
attractive choice over conventional systems. In comparison to conventional systems, mixing
time in a microreactor (micromixer) is typically of the order of milliseconds. Smaller axial
dimensions and enhanced contact area results in a very small diffusion time. Reactions can
8
also be quenched in milliseconds, giving ability to isolate intermediate products and
precisely control yield in a multi-step reaction system. Thus, microreactors have shown turn
out as the preferred choice when it comes to fast reactions.
Interestingly, microreactors have proven to be useful for multi-phase flows. Conventional
systems provide very limited contact area, making interphase transfer slower. In a
microreactor the specific interface area can reach up to 50000 m2/m3 for liquid-liquid
systems and up to 20000 m2/m3 for gas-liquid systems.
Single-phase fluid flow in a microreactor is characterized by a low Reynolds number. The
flow is laminar with Reynolds number of less than 1000 and most of the mixing occurs by
diffusion and secondary flows and transport of materials is essentially through diffusion. If
spatial features or active mixers are not used in microreactors, there will be negligible
turbulence-based mixing. According to Fick’s law of diffusion the diffusive flux J is,
(1.5) cDJ ∇−=
where, c is the concentration of a diffusing entity, D is the coefficient of diffusion and V is
the gradient operator. Time taken for a molecule to diffusion through a distance x will be,
(1.6) 2
Dxt =
Now for diffusion controlled reactions, decreasing the diffusion distance for a molecule will
decrease the time factor by power of 2. Therefore, a reaction in 10-2 cm diameter
microreactor will happen 10000 times faster than in a 1 cm diameter vial. This dramatic
reduction in reaction time has been one of the most important features of research in
microreactor technology. The mixing in a microreactor can be enhanced by incorporating a
micro-mixer or by incorporating segmented slugs of inert gases or liquids.17,18
9
1.1.4 Microreactors in Action
In recent years, microreactors have become a subject of interest for chemical process
companies such as BASF, Lonza, Novartis, BP chemicals and Degussa. These companies
have extensively developed chemical processes involving several aspects of the technology.
It has been estimated that about 50% of reactions in fine chemicals and pharmaceuticals
industry can benefit from continuous processes based on microreactor technology.19
Recently, a team at Lonza received the prestigious Sandmeyer Prize-2010 for their key
achievements in design and manufacturing of microstructured devices, including laboratory
studies describing pharmaceutical reactions in microreactors and the successful transfer of
processes to commercial production.20 This prize is generally given to chemists for their
contribution in advancement of chemistry, and awarding such prize to a process team
clearly indicates significant potential of the technology for advancement of chemistry.
Furthermore, substantial impact has been made by microreactor technology in synthesizing
and screening of potential drug candidates which otherwise is a capital and labor-intensive
task.21
1.2 Organic Reactions
Organic reactions are chemical reactions involving (or producing) organic compounds.
Reactions such as addition reactions, elimination reactions, substitution reactions, pericyclic
reactions, rearrangement reactions and redox reactions comprises of such organic
reactions.22 For example, following reaction between ethane and chlorine shown in scheme
1.1 is an example of an addition reaction.
CH2 CH2 + Cl ClCl Cl
H H
Scheme 1.1
10
These reactions are responsible for production of man-made chemicals such as drugs,
plastics, food additives and fabrics. In fact organic molecule and dyes are now been used for
development of dye-sensitized solar cells, which may in future replace silicon-based solar
cells. Based on type of phases involved in an organic reaction, the reactions can be
classified as homogeneous or heterogeneous organic reaction.
1.2.1 Heterogeneous Organic Reactions
Heterogeneous organic reactions comprise a class of organic reactions in which reactants
are present in two or more physical phase–solid and gas, solid and liquid, or two immiscible
liquids. In these types of reactions one or more reactant may undergo chemical change at an
interface.23 A reaction involving solid catalyst and gaseous reactants is an example of
heterogeneous organic reaction. These reactions can either be a diffusion controlled reaction
or a kinetically controlled reaction. In diffusion controlled reactions, the overall rate of
reaction are limited by diffusion of reacting species between phases.24 Thus, rates of
reaction can be increased by enhancing diffusion (or availability) of reacting species.
However, in kinetically controlled reactions the rates of reaction are not affected by mass
transfer of the species and can only be altered by changing reaction parameters.25 These two
factors determine whether a reaction rate will be accelerated by enhancing transport of
chemical species (i.e. by mixing etc.) or by changing the reaction parameters of a reaction
(i.e. by changing temperature, activation energy etc.). This information is useful for analysis
and usability of microreactors for chemical reactions.
1.2.2 Homogeneous Organic Reactions
‘Homogeneous’ organic reactions are organic reactions in which all reactants exist in same
phase (for example, reaction between two chemical species in a miscible liquid). Similar to
heterogeneous reaction systems, homogeneous reactions are also either a diffusion
controlled reaction or a kinetics controlled reaction. However, for diffusion controlled
reactions the intra-phase diffusion governs the overall rate of reaction. In kinetics controlled
11
homogeneous reactions, rates of reaction can only be altered by changing reaction
parameters.
1.3 Microreactors for Organic Synthesis
As discussed briefly in earlier sections, microreactors have promising applications in
organic syntheses. Some of the key features which make this technology a hot technology
for organic syntheses are –
• Significantly low reagent handling. Compared to conventional diagnostics and
synthesis systems, geometric dimensions of microreactors enable lesser reagent
handling and waste generation, which in turn lowers the operation costs. This
unique feature of microreactors is very beneficial for expensive and labor-intensive
drug discovery processes.
• Faster analysis, response time, and safer operation. Smaller diffusion distances and
higher surface-to-volume ratio enables rapid cooling or heating of reacting species.
This enables superior detection and process control, making notoriously unsafe (and
runaway) reactions to be carried out even in a laboratory.
• Compactness. Large scale integration allows accommodation of several processes
in a small footprint.
• High-throughput and scale-out capability. High-throughput for analysis and
syntheses can be easily achieved by massive parallelization of microreactors. Thus
eliminating engineering difficulties encountered with scaling up of a conventional
process.
• Lower fabrication costs. Microreactor based systems are generally cheaper when
compared to conventional systems.
12
• Safer to operate. Compared to conventional reactor system, compact design and
high heat and mass transfer rate of microreactors make them safer to operate.
Microreactors have promising benefits however their applications are limited by some of
the following key factor–
• High research and process development cost.
• Surface interactions and flows. Physical and chemical effects such as capillary
forces, surface roughness, and chemical interactions with material of construction
are dominant at microscale. Thus, these effects make operation of such reactors
difficult.
• Low signal-to-noise ratio. Due to geometric limitations of integrating a sensor in an
integrated-microreactor will generally have lower signal-to-noise ratio.
Several named organic reactions and processes have been realized in microreactors so
far.26,27,28,29,30,31,32,33,34,35 Furthermore, the technology has found its application in industrial
and laboratory systems for applications such as drug-screening and organic
syntheses.36,37,38,39 Some key developments in the area of microreactors for organic synthesis
are briefly discussed in following sections.
1.3.1 Heterogeneous reactions in microreactors
Heterogeneous reactions are an integral part of an organic synthesis process. For example,
several organic reactions require a solid catalyst phase on which reacting species diffuse in,
react, and diffuse back in bulk medium. Diffusion of reacting molecules in an immiscible
liquid system across phase boundaries in presence of phase-transfer catalyst is another such
example. These heterogeneous reactions are mainly diffusion-controlled reactions.
Increasing surface-to-volume ratio for such reactions increases overall contact area for the
phases to interact.16 Thus, reaction rates for heterogeneous reactions are generally higher in
microreactors than conventional macro-scale system.
13
In order to carry out heterogeneous reactions in microreactors, factors such as flow-
behavior and clogging-issues are taken into account, which eventually calls for specially
engineered systems. Following are some of the engineered system for heterogeneous
reactions in microreactors.
On-Bead and Monolith Systems – ‘On-bead’ synthesis became popular by Merrifield’s work
on polystyrene matrix for peptide synthesis which eventually led to solid phase organic
synthesis and polymer-assisted solution synthesis.40 In this process, beads are functionalized
by a suitable reagent or catalyst which promotes the reaction among the reactants. However,
earlier polymer support suffered from problems such as partial solubility, mechanical
weakness, and broad range of particle sizes. Most of these problems were solved using an
inorganic matrix to support the organic resin.41 The remaining shortcomings of ‘on-bead’
systems (such as packing problem) were eliminated in monoliths. Monoliths are continuous
phase of porous material that can be used without generating high backpressure observed
with fine particles.42
Non-catalytic reactions – Several non-catalytic heterogeneous organic reactions have also
been successfully optimized using microreactors. For example, a rapid liquid-liquid
biphasic exothermic reaction to form a vitamin intermediate was benefited by using
microreactors.5 The reactant phase (hexane) was immiscible with concentrated sulfuric acid
phase in which the intermediate product will eventually shift. The formed product is
temperature sensitive and would quickly generate by-products, giving a lower yield of only
70% in a semi-batch industrial process. The same reaction when carried out in microreactor
system with a micro-mixer and a heat exchanger gave 80~85% yield. The reaction scheme
is outlined in scheme 1.2.
14
Reactant in hexane
conc. H2SO4Product in H2SO4
by_product 1 by_product 2 by_product 3
Intermediate in H2SO4
Scheme 1.2
Catalytic reactions – Several heterogeneous catalytic reactions have been investigated in
microreactors. Greenway and co-workers have reported Suzuki reaction between
phenylboronic acid and 4-bromobenzonitrile in oxolane-water mixture with 1.8%
palladium/silicon dioxide as the catalyst, which was immobilized on the microreactor
surface. A 10% higher yield was obtained in this microreactor system than compared to
conventional batch reactor. The reaction scheme has been outlined in following scheme
1.3.43
CNBOH
OHCNBr+
1.8% Pd/SiO2
THF/H2O
25 min
Scheme 1.3
1.3.5 Homogeneous reaction in microreactors
Homogeneous reactions are the organic reactions in which all the reacting species are
present in a single fluidic phase. As the intra-phase diffusive timescale for reacting species
is very small (order of few seconds), a reaction occurs with fast mixing and high
concentration homogeneity. This influence both catalytic and non-catalytic reactions in
microreactors and has been discussed below.
Catalytic reactions – Catalytic reactions in homogeneous microreactor system have been
shown to drastically enhance yield of a desired product. For example, Suzuki reaction
between 3-bromobenzaldehyde and 4-fluoro-phenyl boronic acid in presence of dissolved
15
Pd catalyst has given 90% yield in a microreactor than just 50% yield in conventional
system.44 The schematics has been outlined in scheme 1.4.
BOH
OHF Br
CHO
+ [Pd(PPh3)4]
NaOH, DMFF
CHO
Scheme 1.4
Non-catalytic reactions – Non-catalytic homogeneous reactions have also been shown to
accelerate in microreactors. For example, Ahmed and coworkers have shown to enhance
hydrolysis of p-nitrophenyl acetate in a microreactor.45
In the above discussion we saw how organic synthesis has benefitted from utilization of
microreactors. However, in many cases (especially homogeneous reactions) it is difficult to
explain the improvement of yield by using microreactors. The following section analyzes
these observations in details and sets up the stage for the thesis.
1.4 Enhancement of reaction rates: the missing link &
motivation
As discussed in previous sections, microreactors can influence reaction rates and yield of an
organic reaction. It is arguably valid to say that one of the key reasons for increase in yield
for heterogeneous reactions is the improvement of heat and mass transfer rates. High
surface-to-volume ratio and smaller diffusive time scale ensure that both heat and mass
transfer occur rapidly with little side-products. Difficulty arises when we try to examine
yields for homogeneous chemical reactions and compare it with an equivalent well-mixed
conventional reactor, and leaves us with questions –“Why yield of a well-mixed
homogeneous chemical reaction much higher in a microreactor than in a conventional
reaction, even though the diffusive time scales can be comparable in both cases? Do the
surfaces of a microreactor have something to do with this increase?”.
16
These observations have baffled several researchers and have sparked a debate in the
scientific community on the possible cause of such spectacular increase.46 Furthermore, the
concept of obtaining higher reaction rate and yield in a microreactor has brought many
commercial platforms in the market to obtain higher yield for synthetic chemists in recent
years. One such commercial platform is shown in Figure 1.3.47
Figure 1.3 FlowStart, a commercial microreactor platform for chemists
These platforms are gaining popularity among the chemists as now they can obtain higher
yield and selectivity for a tedious and time-consuming organic syntheses reaction. However,
full potential of the technology cannot be harnessed without a proper and deep
understanding of factors influencing an organic reaction in microreactor.
Systematic studies conducted by Ueno et al. and Ahmed et al. provide a firsthand insight on
such enhancements.34, 45 The investigations were primarily limited to analyze enhancement
of mass-transfer rates as the major cause for reaction rate enhancement. However,
enhancement of yield and reaction rates for homogeneous reactions cannot be explained by
improved mass-transfer rates. Studies indicate that even for well-mixed conventional and
microreactor system, the yield (and also reaction rate) is high in microreactor system. 45
This motivated us to consider surfaces as the potential contributor to the observed
enhancement. Our belief was partially based on the fact that surfaces (especially silicon
17
dioxide) have shown to increase reaction rates and yield for some organic reactions, and
partially on the fact that in previous studies most of the physical and chemical factors
remained same for both conventional and microreactor system other than surface-to-volume
ratio.49 Thus, we focused our investigation on surfaces (or walls) of microreactors. This was
done by choosing two classes of microreactors, i.e. both silicon and polymeric
microreactors with varying surface-to-volume ratio. Silicon microreactors have a native
silicon dioxide layer on their surface. These reactors were designed such that they have
variable surface-to-volume ratio for same volume, and were fabricated at A*STAR’s
Institute of Microelectronics, Singapore. Polymeric micro-capillaries were obtained from
commercial sources and were considered for the study owing to their availability and
flexibility in terms of material of construction, and presence of chemical groups on their
surfaces.50
1.5 Structure of Thesis
The thesis consists of five chapters. The first chapter gives a brief overview about the
microreactors and microreactor technology, organic reactions and their importance to
industries and society, benefits of carrying out organic synthesis reaction in a microreactor
and motivation of the current study.
Chapter 2 outlines the methodology developed in this thesis to analyze effects of surfaces
on yield of a chemical reaction for microreactors. Selection of a model chemistry and
design of experiments are covered in this chapter.
Design and fabrication of silicon microreactors is covered in chapter 3. This chapter
describes the fabrication steps and protocols followed in development of silicon
microreactors.
Chapter 4 presents the experiments carried out to back the hypothesis. The necessary
experimental procedures are described in details and the section closes by discussing
18
interim conclusions derived from the experiments. Later sections of the chapter talks about
surface effects observed on organic reactions in other research findings. And finally
experimental results of the study are analyzed in light of enhancement effects observed in
the outlined research findings.
Finally, chapter 5 summarizes the thesis work and the observations made. This chapter also
outlines the contributions and suggestions which could further lead to deeper understanding
of the enhancement effects in a microreactor.
19
1.6 References
1. Watts, P. & Wiles, C. Recent advances in synthetic micro reaction technology. Chemical communications 2007, 443-467(2007).
2. Reyes, D.R. et al. Micro Total Analysis Systems. 1. Introduction, Theory, and Technology. Analytical Chemistry 74, 2623-2636(2002).
3. Löhder, W. & Bergann, L. Akademie der Wissenschaften der DDR. (1986).
4. Schubert, K. et al. Herstellung und Test von kompakten Mikrowärmeobertregern. Chemie Ingenieur Technik 61, 172-173(1989).
5. Jähnisch, K. et al. Chemistry in microstructured reactors. Angewandte Chemie (International ed. in English) 43, 406-446(2004).
6. Arulanandam, S. & Li, D. Determining [zeta] Potential and Surface Conductance by Monitoring the Current in Electro-osmotic Flow. Journal of Colloid and Interface Science 225, 421-428(2000).
7. Lee, J.N., Park, C. & Whitesides, G.M. Solvent compatibility of poly(dimethylsiloxane)-based microfluidic devices. Analytical chemistry 75, 6544-6554(2003).
8. Murphy, E. et al. Solder-based chip-to-tube and chip-to-chip packaging for microfluidic devices. Lab on a Chip 7, 1309–1314(2007).
9. Madou, M. Fundamentals of microfabrication. (CRC Press LLC. London, UK:1997).
10. Brandner, J. et al. Microreactors: Epoch-Making Technology for Synthesis. MCPT, 2001 75-87(2003).
11. Wirth, T. Microreactors in Organic Synthesis and Catalysis. (John Wiley & Sons, Inc., New York: 2008).
12. Mills, P.L., Quiram, D.J. & Ryley, J.F. Microreactor technology and process miniaturization for catalytic reactions--A perspective on recent developments and emerging technologies. Chemical Engineering Science 62, 6992-7010(2007).
13. http://commons.wikimedia.org/wiki/File:Selective_laser_melting_system_schematic.jpg
14. Fujii, T. PDMS-based microfluidic devices for biomedical applications. Microelectronic Engineering 61-62, 907-914(2002).
15. Lerou, J.J. et al. Microsystem technology for chemical and biological microreactors. Dechema monographs 132, 51(1996).
16. K. Schubert, W. Bier, J. Brandner, M. Fichtner, C. Franz, G.L. Process Miniaturization - IMRET 2: 2nd International Conference on Microreaction Technology (New Orleans, USA, 1998) 88(1998).
20
17. Günther, A. et al. Transport and reaction in microscale segmented gas--liquid flow. Lab on a Chip 4, 278-286(2004).
18. Hessel, V., Löwe, H. & Schönfeld, F. Micromixers--a review on passive and active mixing principles. Chemical Engineering Science 60, 2479-2501(2005).
19. Roberge, D.M. et al. Microreactor technology: A revolution for the fine chemical and pharmaceutical industries? Chemical Engineering & Technology 28, 318-323(2005).
20. http://www.lonza.com/group/en/company/news/newsreleases/lonza_team_receives.html
21. Dittrich, P.S. & Manz, A. Lab-on-a-chip: microfluidics in drug discovery. Nat Rev Drug Discov, 5, 210-218(2006).
22. Kurti, L. & Czako, B. Strategic applications of named reactions in organic synthesis. (Elsevier/Academic Press: Boston, 2005).
23. http://www.britannica.com/EBchecked/topic/264278/heterogeneous-reaction%20
24. Atkins, P. Physical Chemistry. 825-828(WH Freeman & Co: 1998).
25. Fox, M.A. & Whitesell, J.K. Organic chemistry. (Jones and Bartlett Publishers: Sudbury, US: 2004).
26. Watts, P. & Haswell, S.J. The Application of Microreactors for Small Scale Organic Synthesis. Chemical Engineering & Technology 28, 290-301(2005).
27. Greenway, G.M. et al. The use of a novel microreactor for high throughput continuous flow organic synthesis. Sensors and Actuators B-Chemical 63, 153-158 (2000).
28. Schwalbe, T., Kadzimirsz, D. & Jas, G. Synthesis of a Library of Ciprofloxacin Analogues By Means of Sequential Organic Synthesis in Microreactors. QSAR & Combinatorial Science 24, 758-768(2005).
29. Organic, M.C. Catalysis in Capillaries by Pd Thin Films Using Microwave-Assisted Continuous-Flow Organic Synthesis (MACOS)**. Reactions 2761 -2766(2006)
30. Hessel, V. et al. Aqueous Kolbe-Schmitt synthesis using resorcinol in a microreactor laboratory rig under high-p-T conditions. Organic Process Research & Development 9, 479-489 (2005).
31. Shi, G.Y. et al. Capillary-based, serial-loading, parallel microreactor for catalyst screening. Analytical Chemistry 78, 1972-1979(2006).
32. Srinivas, S. et al. A scalable silicon microreactor for preferential CO oxidation: performance comparison with a tubular packed-bed microreactor. Applied Catalysis A: General 274, 285-293(2004).
21
33. Sugimoto, A. et al. The Barton reaction using a microreactor and black light. Continuous-flow synthesis of a key steroid intermediate for an endothelin receptor antagonist. Tetrahedron Letters 47, 6197-6200(2006).
34. Ueno, M. et al. Phase-transfer alkylation reactions using microreactors. Chemical communications 936-937(2003).
35. Murphy, E.R. et al. Accelerating reactions with microreactors at elevated temperatures and pressures: profiling aminocarbonylation reactions. Angewandte Chemie (International ed. in English) 46, 1734-1737(2007).
36. Acke, D.R. & Stevens, C.V. A HCN-based reaction under microreactor conditions: industrially feasible and continuous synthesis of 3,4-diamino-1H-isochromen-1-ones. Green Chemistry 9, 386(2007).
37. Watts, P. & Wiles, C. Synthesis of Analytically Pure Compounds in Flow Reactors. Chemical Engineering & Technology 30, 329-333(2007).
38. Chambers, R.D. et al. Elemental fluorine. Part 16. Versatile thin-film gas-liquid multi-channel microreactors for effective scale-out. Lab on a chip 5, 191-198(2005).
39. Becht, S. et al. Micro Process Technology as a Means of Process Intensification. Chemical Engineering & Technology 30, 295-299(2007).
40. Merrifield, R.B. Solid Phase Peptide Synthesis. I. The Synthesis of a Tetrapeptide. Journal of the American Chemical Society 85, 2149-2154(1963).
41. Atherton, E. et al. A physically supported gel polymer for low pressure, continuous flow solid phase reactions. Application to solid phase peptide synthesis. Journal of the Chemical Society, Chemical Communications 1981, 1151-1152(1981).
42. Baumann, M. et al. Azide monoliths as convenient flow reactors for efficient Curtius rearrangement reactions. Organic & biomolecular chemistry 6, 1587-1593(2008).
43. Greenway, G.M. et al. The use of a novel microreactor for high throughput continuous flow organic synthesis. Sensors and Actuators B: Chemical 63, 153-158(2000).
44. Hessel, V., Hardt, S. & Löwe, H. Chemical micro process engineering: fundamentals, modelling and reactions. (John Wiley & Sons, UK: 2004).
45. Ahmed, B., Barrow, D. & Wirth, T. Enhancement of Reaction Rates by Segmented Fluid Flow in Capillary Scale Reactors. Advanced Synthesis & Catalysis 348, 1043-1048(2006).
46. Valera, F.E. et al. The Flow's the Thing..Or Is It? Assessing the Merits of Homogeneous Reactions in Flask and Flow. Angewandte Chemie (International ed. in English) 14, 2478–2485(2010).
47. Future Chemistry: http://www.futurechemistry.com/ (Last accessed on 1st August, 2010).
22
48. Ahmed, B., Barrow, D. & Wirth, T. Enhancement of reaction rates by segmented fluid flow in capillary scale reactors. Advanced Synthesis & Catalysis 348, 1043–1048(2006).
49. Basiuk, V.A. Organic reactions on the surface of silicon dioxide: synthetic applications. Russian Chemical Reviews 64, 1003-1019(1995).
50. Hornung, C.H. et al. A microcapillary flow disc reactor for organic synthesis. Org. Process Res. Dev 11, 399-405(2007).
{Bibliography}
23
2. Methodology
This chapter describes the strategy developed for analyzing enhancement of reaction rates
and yield for homogeneous reactions in microreactors. We approached the problem by first
listing out possible physical and chemical factors which may affect the reaction rates and
yield. We then analyze these factors and highlight the key factors which are most likely to
be causing the observed enhancement. Based on these selected set of factors, we develop an
experimental strategy to analyze the observed effect.
2.1 Inside a Microreactor
The key characteristics of microreactors which make them unique from conventional
reaction systems are rapid mixing times, large surface-to-volume ratio, and enhanced heat
and mass transfer rates. Thus, some of these factors may be effecting an organic reaction
such that the yield and selectivity are predominantly influenced in a microreactor.
Now if we look at the rate of chemical reaction for a closed constant-volume system, it is
proportional to change in concentration of chemical species participating in the reaction per
unit time. For a reaction with reactants A and B producing products C and D with
stoichiometric coefficients a, b, c and d respectively,
(2.1)
the rate of chemical reaction will be mathematically defined as:
(2.2)
where [X] denotes the concentration of the chemical species X.1 Alternatively, the rate of
reaction can also be given by law of mass action and is defined as,
(2.3)
dDcCbBaA +→+
dtDd
ddtCd
cdtBd
bdtAd
arate ][1][1][1][1
==−=−=
βα ][][ BAkrate =
24
where k is the rate constant and α, β are the order of the reaction with respect to the reacting
species A and B. The order of a reaction is an experimentally determined quantity and in
some cases could be equal to the stoichiometric ratio of the respective chemical species.
Furthermore, rate constant is defined by Arrhenius equation as:
(2.4)
where A is frequency factor, T is temperature, Ea is activation energy, and R is universal gas
constant . Activation energy here plays a very decisive role in influencing the kinetics of a
reaction as it is an exponential facto and can be easily influenced by presence of catalyst or
physical state of system.
Looking at the system from a macroscopic level, these parameters can be analyzed from
kinetic and thermodynamic perspective by studying the effect of temperature, pressure, and
physical state (or surfaces) to provide a simpler and clearer picture. The following sub-
sections analyses these factors to evaluate their effect in enhancing the reaction rates and
yield.
2.1.1 Effect of Temperature
Temperature can significantly influence rate and yield of a chemical reaction as it appears
as an exponential term in the Arrhenius equation (2.4). Due to high heat-transfer rate in a
microreactor (~10 kWm-2K-1), reactions can be potentially be carried out at quasi-isothermal
conditions in a microreactor than compared to a conventional flask system.2 However most
of the reactions showing enhancement effects have moderate heat of reactions which
conclude that even the conventional flask systems were operating near the desired
temperature conditions. Therefore, any effect of heat transfer rates which can cause such
noticeable change in yield and reaction rates can be ruled out.
−
=RT
E
Aek
a
25
2.1.2 Effect of Pressure
Pressure is another factor which may affect reaction rates and yield of a chemical reaction
by increasing activity of a reaction system. Pressures were just above the atmospheric
pressure (~1 atm) in most of the microreactor systems in which enhancement effects were
observed. These slight pressure-differences are inadequate to give any notable change in
reaction rates and yield of such reactions. Furthermore, pressure will have very little effect
on rate constant for condensed-phase reactions (i.e., solid or liquid).3, 4 Thus, it is safe to
rule out the effect of pressure for enhancement of organic reaction rates in microreactors.
2.1.3 Effect of Surfaces
Surfaces on the other hand can potentially affect course of a chemical reaction. Surface-to-
volume ratio in a microreactor is typically about 10,000 m2/m3 than compared to 100 m2/m3
achievable in a conventional reactor of.5, 6 In other words, chemical species see more of
microreactor surface than surface of a conventional reactor such as flask for a given volume
of reactants. There are several ways in which surfaces in microreactor can influence
reaction rates, such as (i) surfaces can act as a catalyst or a co-catalyst and help to
promoting reaction rates, (ii) surface energy of the surfaces can influence the enthalpy and
entropy of reaction system, thereby influencing the rate constant of a reaction.7, 8
Furthermore, surfaces can enhance reaction rates by creating ‘heterogeneity’ in
homogeneous reaction system (‘on-water’ reactions).9, 10, 11 Thus, surfaces seems to be the
most promising candidate responsible for enhancing reaction rates and yield.
2.1.4 Conclusion
Our preliminary investigation reveals that that surface effects should be the predominant
factor influencing and enhancing a reaction in a microreactor. Therefore, in this thesis we
restrict ourselves to the study of surfaces on homogeneous organic reaction for rate
enhancement, and accordingly design experiments to verify this hypothesis.
26
2.2 Designing the experiment
In previous section we found surfaces to be the most able factor which can influence
reaction rates and yield. The effect was examined by carrying out a systematic experimental
study using a model homogeneous reaction and microreactors with different materials and
surface-to-volume ratio respectively.
2.2.1 Selection of the Chemistry
The model reaction system for this study was chosen and optimized according to the
following guidelines:
a) The model chemical reaction should have moderate rate of reaction in batch system
(t1/2~30 min). Fast reactions may have inadequate interaction with microreactors’
surface and slow reaction will require longer sampling and analysis time.
b) Solvent for the model reaction should be relatively mild and non-corrosive in
nature. This is to ensure that surface properties do not alter by corrosive nature of
the solvent.
c) Reaction rate and yield can be easily quantified using an analysis technique which
require little amount of sample (in ml).
For the model chemical system, a coupling reaction system was envisioned as a promising
system. A coupling reaction is a reaction in which two organic molecules join together,
forming a new carbon-carbon bond.12 To start with the study, coupling reaction between
1,4 benzoquinone and 2-methyl indole was considered (scheme 2.1).13
O
O
NH
CH3
+NH
CH3 O
O
THFHCl , r.t.
2.1 a 2.1 b 2.1 c
Scheme 2.1
27
1 g each of 1,4 benzoquinone ( 98%, Sigma Aldrich, USA) and 2-methylindole (98%;
Aldrich, USA) were taken in a test-tube, and were dissolved in 5ml of tetrahydrofuran
(99.9%; Sigma Aldrich, USA). The content was transferred to a 25ml round-bottom flask
with an air-condenser. To the reaction system, 1 ml of conc. hydrochloric acid (31% Merck,
Germany) was added and the setup was maintained at 60°C for 2 hours in a silicon-oil bath.
The product of the reaction (2.1 c) was an intense purple colored compound and could be
easily quantified in a UV-Vis spectrometer. The reaction was also carried out at room
temperature (about 25°C), however the reaction was very slow and color change indicating
formation of product was observed only after leaving the reaction system overnight. This
limited the scope of the reaction as a desirable model chemical system.
Other reaction candidates were analyzed which can react even at the room temperature.
However due to its ability to form colored product, the benzoquinone ring system was
preferred as one of the reactant. An online sub-structure search for benzoquinone analogs
were carried out on ACS’s Scifinder®.14 Reaction between 1,4-benzoquinone and a thiol
was identified as another prospective model reaction (scheme 2.2).15
O
O
+ r.t.
2.2 b
CH3
SH
OH
OH
SCH3
H2O
2.2 a 2.2 c
Scheme 2.2
2 g of 1,4-benzoquinone (98%, Sigma-Aldrich, USA) was taken in a test-tube and was
dissolved in 10 ml of Acetone (99.5% Sigma-Aldrich, USA). The content was transferred
to a 20 ml sampling vial and to it 500 ml of 1-propanethiol (99% Sigma-Aldrich, USA) was
added. 1 ml of water was then added to the sampling vial and it was left for 15 mins to
react. The solution started turning red from faint yellow, and after half hour dark red
28
solution was obtained. However, 1-propanethiol (2.2 a) has strong obnoxious odor which
makes it very difficult to handle. Thus, this reaction scheme was discarded.
Again, the reaction in scheme 2.1 was reanalyzed. This time however different analogs of
1,4-benzoquinone were considered for accelerating the reaction rates. In principle an
electron withdrawing groups in 1,4-benzoquinone can accelerate the nucleophilic attack by
making it more electron deficient. 2,5-Dichloro-1,4-benzoquinone is one such analog with
two strong electron-withdrawing groups (chloride ion). Hence it was used to carry out the
reaction (scheme 2.3).16
NH
CH3 +
O
O
Cl
Cl
N CH3
O
O
Cl
Cl
4a 4b 4c.1
Scheme 2.3
1.5g of 2,5-Dichloro-1,4-benzoquinone (98%; Aldrich, USA) and 1g of 2-methylindole
(98%; Aldrich, USA) were taken in a test-tube, and were dissolved in 5ml of
tetrahydrofuran ( 99.9%; Sigma Aldrich, USA). The content was transferred to a 20ml
sampling vial and was left at room temperature for reaction to happen. The solution turned
purplish-red within half hour from faint red color, indicating formation of the product.
The final product obtained was analyzed using Aluminum backed Thin-layer
chromatography plate (‘Al Sil G/UV’, Whatman, UK). The sample resolved well with ethyl
acetate-methanol mixture (Both HPLC grade, VWR LLC, USA). A batch of product
mixture was synthesized and the chemical entities were separated using a column
chromatography. A mixture of both single-addition (2.3 c), and double-addition products
(2.3 d and 2.3 e) were obtained which were later confirmed using NMR analysis.
2.3 a 2.3 b 2.3 c
29
N CH3
OH
OH
Cl
Cl
NCH3
N CH3
O
O
Cl
Cl
NCH3
2.3 d 2.3 e
Therefore, direct coupling reaction between 2,5-Dichloro-1,4-benzoquinone and 2-
methylindole was chosen as the model chemical reaction system.
2.2.1.1 Solvent Optimization
The reported reaction between 2,5-Dichloro-1,4-benzoquinone and 2-methylindole was
carried out in tetrahydrofuran (THF). THF is a good aprotic solvent and is widely used in
laboratories and industries for synthesis. However it is corrosive toward plastics and has
detrimental effects on their polymer matrix.17 Thus, a substitute for THF was investigated to
carryout reaction outlined in scheme 2.3.
The reactions between 2,5-Dichloro-1,4-benzoquinone (4b) and 2-methyl indole (4b) were
carried out simultaneously in following solvents (all HPLC grade): tetrahydrofuran,
acetonitrile, 2-propanol, dimethyl sulphoxide(DMSO), dimethyl formamide (DMF),
acetone, DI water and ethyl acetate. Both the reactants (40 mg of 4a and 36 mg of 4b) were
dissolved in 3ml of each solvent. 1ml of each solution was added in respective sampling
vials. To the samples, 1 ml of hydrochloric acid (31% Merck, Germany) was added and
the samples were left at room temperature for 1 hr. The samples were then analyzed
using thin-layer chromatography (TLC). Each sample was spotted on a pre-cut TLC plate
(Whatman, UK; ‘Al Sil G/UV’) and the chromatogram was developed using HPLC grade
methanol and ethyl acetate (10:1) as mobile phase for one hour. The TLC plate was dried in
air for 15 mins and was analyzed under Florescent light source (ENF-240C/FBE,
30
Spectronics, USA). The comparison was done by analyzing the area of the spot on the TLC
corresponding to the chemical species (retention time). The study revealed that reaction in
acetonitrile is comparable to tetrahydrofuran. Acetonitrile is non-corrosive to most
polymers and do not give any unwanted side products which were observed in case of
dimethyl sulphoxide. Thus, acetonitrile was chosen as the solvent for carrying out the model
chemical reaction.
2.2.1 Selection of Microreactors
Two factors were considered while selecting microreactors for the study--material of the
microreactors and internal surface-to-volume ratio of the microreactor. For the same, both
custom-made silicon microreactors and polymeric microcapillaries were considered. Silicon
microreactors were designed and fabricated at A*STAR’s Institute of Microrelectronics,
Singapore. The detail of fabrication procedure are discussed in the next chapter.
Microcapillaries are a popular choice among researchers and engineers. They are relatively
inexpensive, readily available and cheaper to replace than a chip-based microreactors.18
Microcapillaries made of Radel R (Polysulfone), polyether ether ketone (PEEK) and
fluorinated ethylene propylene (FEP) with different inner diameters and 1/16” outer
diameter were purchased from Upchurch® Scientific, USA. Inner diameters were chosen
such that for the same material, we have different surface-to-volume ratios. The surface-to-
volume ratio can be easily calculated using following equation:
(2.2)
where d is the internal diameter, r is the radius and l is a arbitrary length of a
microcapillary. The chemical structure of the materials used and their surface-to-volume
ratios of microcapillaries used are tabulated in the following table-
dlr
rlVolumeSurface /4
22
==π
π
31
Table 2.1 Microcapillaries and their surface-to-volume ratios
Material Chemical Structure Internal
Diameter
Effective
surface-to-
volume ratio
(approx.)
Radel R SO
OO
O
n
508 mm 7874 m2/m3
254 mm 15748 m2/m3
PEEK OO
O
n
508 mm 7874 m2/m3
254 mm 15748 m2/m3
175 mm 22857 m2/m3
FEP
F
F
F
F
F
F
F
CF3n
508 mm 7874 m2/m3
229 mm 17467 m2/m3
175 mm 22857 m2/m3
2.3 Summary
In this chapter we design a methodology to experimentally analyze the effect of surfaces in
a microreactor. Model chemistry was chosen and optimized according to the experimental
requirements and convenience. We described the selection process for selecting right set of
microreactors. In the next chapter, we also discussed the design and fabrication of silicon
microreactors used in this study.
32
2.3 References
1. McNaught, A.D. & Wilkinson, A. Compendium of Chemical Terminology. (Blackwell Scientific Publications, UK: 1997).
2. K. Schubert, W. Bier, J. Brandner, M. Fichtner, C. Franz, G.L. Process Miniaturization - IMRET 2: 2nd International Conference on Microreaction Technology (New Orleans, USA, 1998).
3. Isaacs, N.S. Physical organic chemistry. (Prentice Hall, USA: 1995).
4. Ahmed, B., Barrow, D. & Wirth, T. Enhancement of reaction rates by segmented fluid flow in capillary scale reactors. Advanced Synthesis & Catalysis 348, 1043–1048(2006).
5. Lerou, J.J. et al. Microsystem technology for chemical and biological microreactors. Dechema monographs 132, 51(1996).
6. Jahnisch, K. et al. Chemistry in microstructured reactors. Angewandte Chemie International Edition 43, 406–446(2004).
7. Basiuk, V.a. Organic reactions on the surface of silicon dioxide: synthetic applications. Russian Chemical Reviews 64, 1003-1019(1995).
8. Piloyan, G. & Bortnikov, N. Influence of the surface energy on the kinetics of chemical reactions of mineral nanoparticles. Doklady Earth Sciences 432, 690-692(2010).
9. Jung, Y. & Marcus, R.A. On the Theory of Organic Catalysis "on Water." J. Am. Chem. Soc. 129, 5492-5502(2007).
10. Narayan, S. et al. "On water": unique reactivity of organic compounds in aqueous suspension. Angewandte Chemie (International ed. in English) 44, 3275-3279(2005).
11. Blackmond, D.G. et al. Water in organocatalytic processes: debunking the myths. Angewandte Chemie (International ed. in English) 46, 3798-3800(2007).
12. March, J. Advanced organic chemistry: reactions, mechanisms, and structure. (Wiley: New York, 1985).
13. Niu, F. et al. Promotion of organic reactions by interfacial hydrogen bonds on hydroxyl group rich nano-solids. Chemical Communications 2008, 2803–2805(2008).
14. http://www.cas.org/products/sfacad/index.html (Last accessed on 1st August 2010).
15. Yadav, J. et al. Organic synthesis in water: Green protocol for the conjugate addition of thiols to p-quinones. Journal of Molecular Catalysis A: Chemical 274, 116-119(2007).
33
16. Zhang, H. et al. “On Water”-Promoted Direct Coupling of Indoles with 1,4-Benzoquinones without Catalyst. European Journal of Organic Chemistry 2006, 869-873(2006).
17. Müller, H. Tetrahydrofuran. Ullmann's Encyclopedia of Industrial Chemistry (Wiley-VCH Verlag GmbH & Co., Germany: 2002)
18. Hornung, C.H. et al. A microcapillary flow disc reactor for organic synthesis. Org. Process Res. Dev 11, 399-405(2007).
{Bibliography}
34
3. Silicon Microreactors
3.1 Introduction
Silicon has remarkable physical and chemical properties which make it a very useful
material for micro-machining and fabrication. It is the principal constituent of
semiconductor devices, and is used primarily in mono-crystalline form. In fact, several
methodologies used for fabrication of microreactors and micro-total analysis systems
(mTAS) are based on physical and chemical processes designed and developed for
fabrication of electronic chips.1 Thus, micro-machining of silicon utilizes immense
technical know-how acquired by semiconductor industry.
3.1.1 Silicon Microreactors and Chemical Engineering
Silicon has a high melting point of 1414°C and excellent thermal conductivity of
149 Wm−1K−1 at 300K. This conductivity is threefolds higher compared with stainless steel
(45 Wm−1K−1).2 In a microreactor, higher thermal conductivity is advantageous in rapid
cooling and heating, thermal quenching of chemical reactions and elimination of hot spots.
It has a low coefficient of thermal expansion (2.6 µm·m−1·K−1 at 25 °C) which exerts
relatively less internal thermal stresses at elevated temperature than other popular materials
of construction such as polymers and metals.3 Young’s modulus for silicon is 185 GPa
which makes it as stiff as wrought iron (190 GPa) and suitable for construction of high
pressure reaction vessels.4 Furthermore, silicon is chemically stable towards most chemicals
due to formation of a thin and rigid layer of silicon dioxide. This layer protects the reactive
silicon underneath and comprises of silanol groups which makes the surface hydrophilic.
The type of silanol group residing on a silicon surface depends on temperature and its
history of chemical treatment.5 Figure 3.1 outlines three types of such groups.
35
Si
O O O
OH
Si
O O
OH
SiO
O O
OH
SiO O
OH
OH
Isolated Silanol groups Vicinal Silanol groups Germinal Silanol groups
a b c
Figure 3.1 Different types of silanol groups with hydrogen bonding
These silanol bonds further increase the scope of silicon microreactors. They can be
chemically modified to impart hydrophobicity to the silicon-surfaces, or to immobilize
application-specific chemicals such as enzymes or catalysts. Generally, silanes are used for
such surface modifications.6 A condensation reaction between a silane and isolated silanol
group is given in Scheme 3.1.
Scheme 3.1
Silicon’s unique surface and mechanical ability makes it an excellent material of
construction for microreactors designed to carry out chemical engineering processes. These
microreactors have been used for applications such as catalyst screening, fuel cell analysis
and construction, controlled growth of nanomaterials (quantum dots), PCR amplification,
and safer on-demand analysis and synthesis of explosive precursors.7,8,9,10
Si
O O O
OH
3.1 a
+ SiH3 RSiH2
R
Si
O O O
+ OH2
3.1 b 3.1 c
36
3.2 Microfabrication of Silicon Microreactor
Silicon microreactors are fabricated by performing several micro-machining steps on a
monocrystalline silicon wafer. Micro-machining is carried out by etching processes as these
processes offer very high geometrical resolution. The surface is masked using masking
layers of suitable materials which do not react with the etchant. Two types of etching
technology are available for micro-machining: anisotropic etching and isotropic etching.
Anisotropic etching is micro-machining of a surface by physical and/or chemical
weathering such that the etching rate is not uniform in all directions (or planes). On the
other hand in an isotropic etching the etching rate is independent of the direction (or plane)
of the surface and is uniform in all directions.1 Figure 3.2 outlines the difference both of
these etching techniques
Figure 3.2 Isotropic and Anisotropic etching of a masked surface.
The etching technology features and masking layers used for micro-machining silicon are
outlined in Table 3.1.
Table 3.1 Etching of Silicon wafers
Etching Technology Etching Process Feature Masking layer
Anisotropic
etching
Reactive Ion
Etching (eg. DRIE)
Combination of physical and
chemical dry etching
technique
SiO2, Si3N4,
Photoresist
Wet chemical KOH preferentially etches SiO2, Si3N4
37
etching <100> plane of mono-
crystalline silicon
Isotropic etching
Wet chemical
etching
Achieved by HF : HNO3 :
CH3COOH ; High etching
rates
SiO2, Si3N4,
Plasma Etching
Combination of physical and
chemical dry etching
technique
SiO2, Si3N4,
Photoresist
Reactive ion etching is a dry etching process in which plasma of reactive ions is used for
etching a patterned surface. In this etching process, the etching gas is filled at low pressure
(~100 mTorr) in a cylindrical etching chamber. The substrate (surface) is electrically
isolated from the rest of the chamber and is connected to a radio-frequency (RF) power
source. Low pressure and RF power source result in formation of very dense plasma in the
chamber which accelerates in the electric field toward the substrate. The high energy
reactive ions bombard the surface and anisotropically dislodge the atoms from the
substrate.1 A typical reaction ion etching setup is shown in Figure 3.3.
For this study, microreactors were fabricated using Deep Reaction Ion Etching (DRIE)
process.11 This process is highly anisotropic and results in steep trenches with aspect ratios
of 20:1 or more. The patterned wafer is etched by bombarding plasma of reactive gases
which weathers the exposed silicon surface. DRIE can be achieved by employing cryogenic
process or Bosch process.12
38
Figure 3.3 Reactive Ion etching process
In cryogenic process, the patterned wafer is maintained at -110°C to slow down the
isotropic etching rate caused by the reactive ions in the plasma. As the ions strike
perpendicular to the wafer surface, only the upward facing surface is etched.
In Bosch process (also known as time-multiplexed etching), vertical structures are etched by
alternatively repeating between two modes:
1. First mode is standard isotropic plasma etch using sulfur hexafluoride (SF6) where
ions impinge the surface vertically.
2. In the second mode, a chemically inert passive layer is deposited (generally using
C4F8).
The passive layer protects the entire surface from chemical attack. However, the vertical
bombarding of ions on the surface etches the passive layer on the horizontal surface more
than that on the sides of a feature. These modes are repeated several times with each mode
lasting for several seconds. The overall effect of the etching results in high aspect ratio
patterned surface.
39
3.2.1 Development of Protocol
In order to analyze the effect of surfaces in enhancing organic reactions in microreactors,
silicon microreactors were designed such that for the same internal volume they had
different surface-to-volume ratios. This was achieved by incorporating extended surfaces
inside the microchannel and is discussed along with the development of photolithography
mask. Microfabrication protocol was developed at A*STAR Institute of Microelectronics
(IME), Singapore for 8” double-side polished silicon wafers. Solder-based chip-to-world
approach as described by Murphy et al. was adopted for packaging (fluidic interconnects).13
The detailed micro-machining protocol used at IME with wafer state is given in appendix
A.
3.2.2 Development of Photolithography mask
Both the front and back surfaces of the wafers were patterned. The front pattern gave us the
required fluidic channels, and the back pattern was meant for generating etch-through holes
for fluidic interconnects. These patterns were designed using Autodesk’s AUTOCAD®
designing software.14 The patterns were sent to Infinite Graphics Pte Ltd, Singapore for
printing emulsion based transparencies. Third dimension (depth) is process dependent and
is decided by etching rate.
The front surface was etched to give 170 mm deep trenches, and the pattern generally
consisted of a long rectangular microchannel of 200 mm width. Depending on the desired
surface-to-volume ratios, extended surface elements were replaced with the equivalent
length rectangular microchannel element. A pattern with rectangular channel and extended
surface is shown in Figure 3.4.
40
Figure 3.4 Design microreactor with extended surface
The extended surfaces were designed such that each curved element had almost 1.04 times
the volume of an equivalent length rectangular microchannel. However, the internal surface
area of the whole element was 1.81 times that of an equivalent rectangular microchannel.
Figure 3.5 outlines the features of the curved element and the rectangular element. The
excess volume was corrected by reducing the equivalent length of the rectangular
microchannel.
Figure 3.5 Extended surface and rectangular channel (all units in mm)
Five microreactors were designed with increasing number of extended surface elements
such that their surface-to-volume ratios increase while their volumes remain the same
(Table 3.2). These microreactors’ pattern was set to fit on an eight inch silicon wafer and
41
sent for emulsion transparency printing. The image of the actual pattern is shown in
appendix A.
Table 3.2 Surface-to-volume ratios for the designed microreactors
Reactor No. Surface-to-volume ratio m2/m3 (approx.)
1 12720 2 12275 3 11829 4 11384 5 9604
3.2.3 Fabrication Steps
For fabrication Double-side-polished silicon wafers (<100> with 725 +/-20 mm thickness)
were obtained from SVM Microelectronics Inc., USA. Facilities at A*STAR Institute of
Microelectronics, Singapore were used for micro-fabrication. The steps followed to carry
out fabrication of a single wafer are as follows:
1. A silicon dioxide layer of 1.5 mm was deposited on the silicon wafer using
Novellus® CVD equipment.
2. The front surface of the wafer was spin-coated with a positive photoresist of
thickness 10 mm and was baked for 10 min.
3. The wafer from Step 2 was photo-lithographed on EVG® Aligner using the
designed emulsion transparency mask.
4. The photo-lithographed wafer was rinsed with lithography developer.
Subsequently, the wafer was washed in iso-propyl alcohol and DI water, and spun-
dried.
5. The wafer was descummed (plasma cleaned) using Plasmatherm® plasma cleaner
and the exposed silicon dioxide layer was etched with a target depth of 1.5 mm on
Lametcher® etcher.
42
6. The wafer was inspected under scanning electron microscope and the first silicon
etch was performed with a target depth of 170 um on STS® MEMS etcher.
7. The wafer was checked for the target depth, and then was taken for photoresist
stripping.
8. After stripping-off the photoresist, the wafer was cleaned in hot Piranah (3:1
mixture of conc. sulfuric acid and hydrogen peroxide) at 95°C for 20 mins to
remove any trace amount of photoresist.
9. The backside of the wafer was then spin coated using positive photoresist.
10. The wafer was aligned with the front side using the ‘hair-lines’ pattern and photo-
lithographed on EVG® Aligner using the designed backside emulsion transparency
mask.
11. The photo-lithographed wafer was developed, rinsed, washed and dried as done in
Step 4.
12. The front side of the wafer was attached to a dummy silicon wafer using pressure
sensitive Nitto-Denko® double-sided heat tape.
13. Vacuum and mechanical pressure was applied to this compound-wafer to
completely bond the wafer on EVG® Anodic Bonder.
14. The compound-wafer was descummed in Plasmatherm® cleaner, and etch-through
was performed with a target depth of about 650 mm on STS MEMS® etcher.
15. Once etching was complete, the dummy wafer was detached from the main wafer
by heating the compound-wafer on a hot plate maintained at 180°C.
16. A batch of five wafers was processed in the first round.
17. The photoresist was chemically stripped-off using acetone at the stripping bay and
the wafer was subsequently washed and dried.
18. Silicon dioxide layer on the wafer was removed by using buffered oxide etch (BOE:
NH4F + HF). The process was monitored visually by looking at the color of the
43
layer and once complete, the wafer was plunged in DI water. This reaction is fast
and typically take 15~20 sec.
19. The wafer was further cleaned in hot Piranah for 20 mins, rinsed in DI and spin-
dried.
20. Pyrex-glass wafer was cleaned in Piranah for 20 mins and were bonded to front side
of the wafers via anodic bonding on EVG® anodic bonder.
21. Back of the wafer (silicon side) was sputter-coated with titanium (500Å), copper (2
mm) and gold (1000Å) respectively.
22. The wafer was than diced along the dicing lines to release individual microreactors.
The whole microfabrication process and the cross-sectional views of an etched
wafer are shown in Figure 3.6.
44
Figure 3.6 Microfabrication steps and microreactor cross sections
45
3.3 Interconnecting Microreactor
The fluidic interconnects enable leak-proof entry and exit of reactants and products
respectively. These interconnects are very critical in the operation of a microreactor as, they
provide material exchange with the macroscopic world, and along with the microreactor
governs the maximum operating pressure. Fluidic contacts may also facilitate effective heat
transfer from microreactor as well.15 Figure 3.7 shows a metal packaging incorporated with
a heat exchanger for a microreactor.
Figure 3.7 Microfabrication steps and microreactor cross sections
Interconnects used for a microreactor depends on chemical nature of reactants and products,
the physical nature of the microreactor surface, and operating pressure and temperature the
syste We used and tried following interconnection techniques for the For the fabricated
silicon microreactors, following interconnection techniques were used in this research work.
3.3.1 Solder-based interconnect
Similar to joining electronics component, solder-based interconnect is suitable when the
operating pressure is quite high as it has been reported to withstand pressures up to 200
bar.13 However, its operation limit is governed by melting point of solder-metal used.
46
Another requirement of this interconnect is metallization of microreactor surface around
fluidic opening so that solder can form an alloy and hold both the fluidic tubing and
microreactor together. In this work, we use this technique to interconnect metalized surface
of the silicon microreactors with steel microtubing.
Microreactor was placed on a hot plate maintained at 180°C such that metalized back faces
away from the hotplate. A thin layer of lead-free solder (Multicore, Malaysia; EN 29453)
was soldered using a generic soldering iron. Thin layer of solder was also bonded around
brass metal ferrules (1/16” OD, Swazlok®) using phosphoric acid as flux. The ferrules were
quickly placed up on metalized and pre-soldered fluidic opening on the hot-plate using
tweezers. The temperature of the hotplate was reduced to 150°C and the microreactor was
removed from the hot plate. Pre-cut stainless steel microtubes were inserted in the ferrules
and soldered to the metal ferrules using phosphoric acid as a flux. Figure 3.8 shows a
microreactor with metal ferrules soldered using a hot plate.
Figure 3.8 Soldering metal ferrules with a silicon microreactor on a hot plate.
One of the major problems with such interconnect is collapse of other soldered joints while
soldering itself. Silicon is an excellent conductor of heat. While soldering locally at one
point, rapid heat transfer rises temperature of the whole chip. This weakens a soldered joint
between other ferrules and microreactor by melting solder-metal between them. The
47
detached metal ferrule cannot be re-soldered as already metals from the metalized-layer
were dissolved in the previously soldered solder-metal.
In order to overcome the problem with the metal layer, a thicker metal layer was deposited
during the sputtering process itself (fabrication step 21). However, proper adhesion of the
metal layer with silicon surface could not be achieved. The metal layer delaminated around
dicing lines and could be easily peeled off indication very poor adhesion (Figure 3.9).
Figure 3.9 Delamination of deposited metal layer on microreactor along the dicing lines
As the solder based interconnection methodology failed for the designed silicon
microreactor, other interconnection techniques were explored to establish fluidic contact
and reuse the microreactors. The metalized layer was dissolved in aqua regia (1:3 mixture
of nitric acid and hydrochloric acid) to expose a clean back surface.
3.3.2 O-ring casing based interconnect
In this type of interconnect, a microreactor is sandwiched between a metal or polymer
housing (chuck) with o-rings sealing the fluidic openings of the microreactor and the
housing. The o-rings hermetically seal the microreactor with fluidic interconnect on the
housing. This type of interconnect have following pros and cons-
Pros:
It is more convenient to replace a clogged and damaged microreactors mounted
with this type of interconnect than other kind of interconnects.
It can operate a microreactor at higher temperature.
3 mm
48
Cons:
The chuck is expensive and tedious to machine.
This type of interconnect impart additional mechanical stress on microreactor.
Casing for our microreactors was designed using Cocreate® CAD software, and was
fabricated using computer-aided machining (CAM) at the Department of Chemical and
Biomolecular Engineering workshop, National University of Singapore. PEEK was used as
the material of construction and Teflon was used for making o-rings. A packaged
microreactor with fluidic contacts is shown in Figure 3.10.
Figure 3.10 O-ring based microreactor packaging
3.3.3 Sealant-based interconnect
Sealant-based interconnect is the most common type of fluidic interconnect methodology
used for prototyping. In this methodology, microtubing is sealed to fluidic openings of a
microreactor with an appropriate sealant such as epoxy. It is a simple and cost effective
technique, and very useful for rapid prototyping. However, this type of interconnect is only
useful for handling non corrosive chemicals at low pressures.
For the interconnection between the fluidic opening of the designed microreactor and
microtubing, the back surface of the microreactor was descummed in Plasma cleaner (PDC-
32G; Harrick Plasma, USA) for 10 mins at high settings. This was performed to completely
remove any dust and organic impurities which will then ensure better adhesion.
Microtubing with ferrule was manually aligned with fluidic opening of the microreactor and
49
the sealent (5 mins Epoxy, ITW Devcon, USA) was applied. This was performed for all
fluidic opening and the microreactor was left overnight for complete curing. Figure 3.11
shows a silicon microreactor glued with microtubing.
Figure 3.11 Microreactor packed in a epoxy based sealant
3.4 Summary
In this chapter we discussed the design and fabrication of silicon microreactors. The
fundamentals of microfabrication for silicon were discussed with detail pertaining to
fabrication of the microreactor, and fluidic-interconnect methodologies used in this work
was also discussed. In the next chapter, the experimental part of the thesis will be discussed.
50
3.5 References
1. Madou, M. Fundamentals of microfabrication. (CRC Press LLC. London, UK:1997).
2. Shanks, H.R. et al. Thermal conductivity of silicon from 300 to 1400 K. Physical Review 130, 1743-1748(1963).
3. Okada, Y. & Tokumaru, Y. Precise determination of lattice parameter and thermal expansion coefficient of silicon between 300 and 1500 K. Journal of Applied Physics 56, 314(1984).
4. Wortman, J.J. & Evans, R.A. Young's modulus, shear modulus, and Poisson's ratio in silicon and germanium. Journal of Applied Physics 36, 153(1965).
5. Nawrocki, J. The silanol group and its role in liquid chromatography. Journal of Chromatography A 779, 29-71(1997).
6. Jal, P.K., Patel, S. & Mishra, B.K. Chemical modification of silica surface by immobilization of functional groups for extractive concentration of metal ions. Talanta 62, 1005-1028(2004).
7. Ajmera, S.K. et al. Microfabricated cross-flow chemical reactor for catalyst testing. Sensors and Actuators B: Chemical 82, 297-306(2002).
8. Yen, B.K. et al. A Microfabricated Gas–Liquid Segmented Flow Reactor for High-Temperature Synthesis: The Case of CdSe Quantum Dots. Angewandte Chemie (International ed. in English) 34, 5447 -5451(2005).
9. Trau, D. et al. Genotyping on a complementary metal oxide semiconductor silicon polymerase chain reaction chip with integrated DNA microarray. Analytical Chemistry 74, 3168-3173(2002).
10. Srinivas, S. et al. A scalable silicon microreactor for preferential CO oxidation: performance comparison with a tubular packed-bed microreactor. Applied Catalysis A: General 274, 285-293(2004).
11. Klaassen, E.H. et al. Silicon fusion bonding and deep reactive ion etching: a new technology for microstructures. Sensors and Actuators A: Physical 52, 132-139(1996).
12. Jansen, H.V. et al. Black silicon method X: a review on high speed and selective plasma etching of silicon with profile control: an in-depth comparison between Bosch and cryostat DRIE processes as a roadmap to next generation equipment. Journal of Micromechanics and Microengineering 19(3), (2009).
13. Murphy, E.R. et al. Solder-based chip-to-tube and chip-to-chip packaging for microfluidic devices. Lab on a Chip 7, 1309-1314 (2007).
14. http://usa.autodesk.com/adsk/servlet/pc/index?siteID=123112&id=13779270 (Last accessed on 1st August, 2010).
51
15. Jähnisch, K. et al. Chemistry in microstructured reactors. Angewandte Chemie (International ed. in English) 43, 406-446(2004).
{Bibliography}
52
4. Experimentation and Observations
In this chapter, we discuss the experiments performed to analyze effect of surfaces on with
our chosen chemistry in a microreactor. The results of the experiments and inference
derived are discussed in the later part of the chapter.
4.1 Experimental Setup
There were three modules in our experimental setup--pumping module, microreactor
module and detector module. Each of these modules were connected using PTFE
microtubing (OD 1/16” and ID 254 μm; Vici Instruments, USA) and microfittings
(Upchurch Scientific, USA). A block diagram of the experimental setup is shown in Figure
4.1.
Figure 4.1 Block diagram of the experimental setup
Pumping module comprised of a syringe pump (PHD-2000; Harvard Apparatus, USA). The
reactants for the model chemistry were filled in 3 ml Luer-lock syringes (HS7687863;
Terumo Medical Corp., USA) and were loaded on the syringe pump. The syringes were
connected to the PTFE microtubing via female Leur adaptors (P-658; Upchurch Scientific,
USA) together with flangeless nut (F-336N; Upchurch Scientific, USA). The other ends of
tubing connected the pumping module to the microreactor module.
Microreactor module consisted of selected microreactors with fluidic packaging
(interconnect) through which reactants and product could be transported thru microreactors.
For the silicon microreactors, O-ring based interconnect packaging with was used for
Microreactor Module
Si mreactor / Polymer
microcapillary
Pumping Module
Syringe pump
Detector Module
UV-Vis / GC-FID
mtubing
53
connecting the PTFE tubing. In case of sealant (epoxy) based interconnects, microtubing
was glued with PEEK ferrule (F-162; Upchurch Scientific, USA) to establish proper fluid
flow. For carrying out experiments using polymer microcapillaries, a micro-mixing tee with
silica frit (CM1XPK; Vici Instruments, USA) was used. Silica frit in the tee imparts
vigorous mixing of reactants before they enter microcapillary.
Detector module is the name given to part of the setup which registers conversion achieved
in performing the chemical reaction in microreactors. Two approaches were adopted for
detecting the chemical conversion—continuous-online monitoring and thermal quenching
followed by offline characterization—which are further discussed in section 4.3. UV-Vis
spectroscopic characterization and quantification technique was used for continuous-online
monitoring. Offline characterization was performed using a gas chromatograph with flame
ionization detector (GC-2010; Shimadzu Corp., Japan).
4.2 Experimental Protocol
All experiments were performed at room temperature of about 23°C. Dilute reactant
solutions with low molarity were prepared and used. This was to ensure that neither the
surface of a microreactor is oversaturated with reactant molecules, nor the intensively
colored product brings anomalies in UV-Vis spectroscopic analysis. 2,5-dichloro-1,4-
benzoquinone was used as the limiting reagent.
Two 10 ml standard flasks were cleaned with dish washer and deionized water (18 MΩ),
followed by rinsing twice with acetone to remove any trace amount of organic entities.
They were blow-dried by using compressed air gun. 0.040 g of 2,5-dichloro-1,4-
benzoquinone (98%; Aldrich, USA) was weighed on a weighing paper and was
subsequently transferred to one of the 10 ml standard flask. 0.020g of Mesitylene (805890;
Merck, Germany) was added to this standard flask as an internal standard. To this
acetonitrile (HPLC grade dried over molecular sieve; VWR, USA) was added. The salts
were dissolved and the solvent was filled up to the mark. 0.0354 g of 2-methyl indole (98%;
54
Aldrich, USA) was weighed on a weighing paper and was transferred to another 10 ml
standard flask. Again, the salts were dissolved in acetonitrile and the solution was filled up
to the mark.
The two reactant solutions were filled in disposable syringes (3ml, HS7687863; Terumo
Medical Corp., USA), and the syringes were mounted on a syringe pump (PHD-2000;
Harvard Apparatus, USA). Using fluidic adapters and PTFE tubing the syringes were
connected to the microreactor module.
Calculation of flow rates
Flow rates for each experimental runs were calculated such that residence time of the
reactant mixture within the microreactors is 15 min. In case of polymeric microcapillaries,
lengths were fixed at 18 cm. Thus, the required flow rates for the reactants were calculated
using the following mathematical expression:
(4.1)
For a circular capillary equation (4.1) will be,
(4.2)
where f is required flow rate from the syringe pump, d is internal diameter and l is length of
a microcapillary, n is number of reactor inlets and t is the residence time required (here 15
min). The calculated flow rates are given below in Table 4.1.
Table 4.1 Flow rates for both Silicon microreactors and Polymer Microcapillary
Microreactor type Feature Flow rate (ml/min)
Silicon Microreactor Same internal volume 0.6384
Polymer
Microcapillary
ID 508 mm 1.218
ID 254 mm 0.3046
ID 229 mm 0.2467
residence
*nalcrossectiotime
lengthAreaf =
ntldf
4
2π=
55
ID 175 mm 0.1445
Starting protocol
For the experiments, following starting protocol was used for all of the studies-
1. The reactants were pumped at 10 times the required flow rate for first 10 min to
ensure that both the reacting solutions reach the microreactor.
2. The flow rates were adjusted back to the required value and the system was left
undisturbed to stabilize for 30 min.
4.3 Sampling and detection
The conversion in a microreactor was calculated with respect to 2,5-dichloro-1,4-
benzoquinone by comparing its concentration in reactant and product samples respectively.
As mentioned earlier, both UV-Vis Spectroscopic analysis and GC-FID chromatographic
analysis were used. However due to system limitations and problems with executing
experiments, only gas chromatography was finally used for analysis. Pure product samples
were isolated using column-chromatography (as mentioned in section 2.3) and were used
for standard reference. There structures were confirmed by comparing NMR data with the
available literature.1
4.3.1 UV-Vis Spectroscopy
Continuous analyses of the products were carried out initially using an optical guide pre-
etched in silicon microreactors. The assembly is based on work by Jackman et al.2 A
polyimide coated optical fiber (100 mm, FIBER-100-UV; Ocean Optics Inc., USA) was
taken and was slipped inside PTFE tubing (254 mm, 1/16” OD; Upchurch Scientific, USA)
to provide protection from mechanical wear. One end was left as it is with a bout 2 cm bare
optical fiber protruding out from the PTFE tubing. To the other end a SMA optical
56
connector was glued using bare fiber adapter kit (BFA-KIT; Ocean Optics Inc., USA). In
total two such optical fiber modules were fabricated. The protruding optical fiber was
slipped in the optical guide of a microreactor. High pressure grease was applied to ensure
that the chemicals do not leak from the surrounding orifice. The optical fiber connection
with a microreactor is shown in Figure 4.2.
Figure 4.2 a) Silicon microreactor with optical fiber based online UV-vis analysis. b) Inset showing
the optical fibers and the microreactor.
The SMA connector end of one fiber was connected to a UV light source (DH-2000; Ocean
Optics, USA) and another SMA connector end was connected to an optical fiber-based UV-
vis spectrometer (USB 2000; Ocean Optics, USA).
While performing experiments, no UV signals were detected by the spectrometer. However
tapping of the microreactor was momentarily giving some low signals which indicated
shortcoming in the present engineered design. A thorough analysis of the system revealed
two possible factors behind such erratic operation, (i) the dimensions of guideways, and (ii)
the ends of the inserted bare optical fiber. The dimensions of the guide-way for optical fiber
were 200 mm by 170 mm. However, diameter of the bare optical fiber with polyimide
coating was only 130 mm. This leaves a lot of room to transmitter and collector optical
fibers for misalignment. Optical signals will only be collected by the collector optical fiber
57
when both transmitter and collector optical fibers are aligned exactly face-to-face. An
illustration of the optical fiber in the guide-way is shown in figure 4.3.
Figure 4.3 Illustration of the optical fiber in the guide-way of a microreactor.
The second possible reason for low signal strength is the surface roughness of the optical
fiber end thru which light travels. Rough surfaces can cause substantial electromagnetic
scattering due to which the signal strength reaching the collector fiber will be very less.
After repeated fabrication, the problem could not be solved. Thus, a custom made online
UV flow-cell was machined in Teflon to overcome the misalignment problem and increase
signal strength. Accessories for the flow cell such as quartz windows, teflon o-rings and
microfitting were used from flow injection analysis kit (FIA-1000-Z ; Ocean Optics Inc.,
USA). The fabricated flow-cell had an optical path length of 900 mm, and arrangements to
fit in 400 mm standard optical fiber cable with SMA connectors. Figure 4.4 shows an
image of the fabricated UV flow cell with the accessories.
Figure 4.4 Fabricated microcross UV flow cell
58
This flow-cell was attached just at the exit of a microreactor. For an experiment, acetonitrile
(HPLC grade, VWR Inc., USA) was used as a blank reference. The detection system works
fine except for the fact that the operation is severely affected by the presence of bubbles in
the flow. Figure 4.5 shows a UV signal affected by the presence of bubbles in the flow
system at a high flow rate of 20 ml/min.
Figure 4.5 Signal intensity affected by bubbles in the flow system recorded over time at wavelengths
of 450 nm(black), 400 nm(magenta) and 240 nm(blue); flow rate = 20 ml/min
Bubbles are generated by dissolution of dissolved gases in the reactant solution. These
bubbles are squeezed and acquire ellipsoidal shape while flowing in a micro-capillary.
However, when these bubbles reach the optical opening of the flow-cell, they expand and
become spherical. The optical opening acts as a gutter and the liquid flow through it without
disturbing the bubble. Gas bubbles have lower refractive index than the liquid itself and
thus acts as a diverging concave lens. This diverge the optical signals reaching the collector
fiber and the intensity of the optical signal falls down. The variation destabilizes base line
and spectrum of the UV spectrometer and makes detection impossible.
In light of problems associated with the UV-Vis spectroscopic detection, measurement and
quantification was carried out by gas chromatography.
59
4.3.2 GC-FID Analysis
Gas chromatography is an analytic technique for separating and analyzing compounds that
can be vaporized without decomposition. In this technique, the moving phase (or "mobile
phase") is an inert carrier gas such as helium and the stationary phase is a microscopic layer
of liquid or polymer on an inert solid support.3 The separated entities are analyzed using
detectors such as mass spectrometers (MS) or flame-ionization detectors (FID).
Flame Ionization detector works by analyzing ions generated by combustion of a sample.
The positive electrode is connected to the nozzle head where the flame is produced. The
negative electrode is in the form of a tubular electrode and is positioned above the flame. In
FID, a small hydrogen-air flame burns at temperatures high enough to pyrolyze organic
compounds in a sample, producing positively charged ions and electrons. The ions
generated by pyrolysis of the sample at the positive electrode are attracted to the negative
tubular electrode, inducing an electrical current. The electrical current is measured with a
high-impedance ammeter and is plotted against time to give a chromatogram. The measured
current is proportional to the reduced carbon atoms in the flame. Passage of inert carrier gas
through the detector does not produce any signal as the gas cannot be ionized. The first peak
in the chromatogram generally corresponds to the organic solvent in the sample and the
remaining two peaks are the electric current signals obtained for the other two components
present in the sample.4
The detector is sensitive to the mass of the organic sample rather than the concentration.
However it should be noted that the relative area under the curve for the two different
organic entities cannot be directly correlated to their concentration.
4.3.2.1 Method development
Method development is essential for error-free identification and quantification of chemical
species present in an organic sample. In method development operation parameters are
optimized to give stable and reproducible results.
60
For the experiments the method was developed for analyzing 2-methyl indole, 2,5-dichloro-
1,4-benzoquinone and Mesitylene. Mesitylene was used as an internal standard. The
reaction generates one major product and two by-products, both of which show very poor
sensitivity in flame ionization detectors. Thus, conversion of 2,5-dichloro-1,4-benzoquinone
was considered for analyzing effects of surfaces. Samples of the reactants and products
were made in acetonitrile (HPLC grade; VWR Inc., USA) and were injected individually to
obtain retention time reading using the GC-FID with fused silica capillary column (Rxi 5sil
MS; Restek Inc., USA). Split ratio of 20:1 was maintained at the injector port and the
carrier gas flow rate was set at 40 ml/min. Helium was used as the carrier gas and Nitrogen
was used as the make-up gas.
The initial column temperature was set at 80°C with a ramp of 20°C per min to reach
250°C, which was further maintained for 5 mins. By trial-and-error an optimum method
was developed to obtain maximum resolution of peaks with minimum operation time. The
final method developed and used is outlined in table 4.2.
Table 4.2 Final method used for GC-FID analysis
S.No Rate (°C/min) Temperature (°C) Hold Time (min)
1 - 120.0 0.30
2 10.00 205.0 5.00
4.3.2.2 Sampling and Analysis
The samples were obtained by quenching the products at 0°C obtained from experiments
described in section 4.4, and were immediately analyzed by manually injecting them in the
gas chromatograph. 0.8ml of each sample was injected into the injector port of the GC
system using a gas-tight syringe from about 2ml of the sample collected in GC sampling
vials maintained at 0°C.
61
4.4 Experimentation
4.4.1 Silicon based microreactors
Experiments were carried out for both custom-made silicon microreactors and polymer
microcapillaries. However after repeated unsuccessful attempts with silicon based
microreactors, the experiments involving silicon microreactors were dropped.
These silicon microreactors suffered inadequate fluid-flow. Already failure of solder-based
interconnect has been previously discussed in chapter 3. With the O-ring based fluidic
interconnect packaging, the fluids started leaking within 15 to 20 min of operation. Initial
speculation for failure was attributed to possible engineering errors incurred while micro-
machining of the packaging. The experiments were continued by using sealant-based fluidic
interconnects packaging. However the interconnect started leaking within minutes of
operation and within half hour cured epoxy layer delaminated completely (figure 4.7).
Figure 4.7 Delamination of epoxy from a silicon microreactor
Finally, a generic high-strength cement based epoxy was used for interconnect and water
was pumped through the microreactor. It was observed that even after 30 min of pumping,
no water came out from the microreactor and after about an hour of pumping, the syringe
pump stalled. Stalling of the pump happens in situations when flow resistance is very high.
Such a high resistance is only possible when microchannels are blocked. After analyzing
previously taken microscopic images at A*STAR Institute of Microrelectronics, it was
62
discovered that most of the micropatterns have sufficient irregularities to cause blockage
and render microreactors unusable. Figure 4.8 shows images of sections of the silicon
wafers after DRIE with broken microchannels.
Figure 4.8 Images of the patterned surface of a silicon wafer during microfabrication. The
microchannels are broken or irregular in all 3 images (a,b and c).
Such irregularities are known to arise at two places during microfabrication-- DRIE step
and photolithography step. In DRIE step, the problem arises when debris formed during the
DRIE settles in the channel and eventually leads to improper etching. In photolithography
step, irregularities arise because of emulsion transparency mask. The occurrence and
possibility of irregularities due to this was confirmed by analyzing other wafers
photolithographed using the same mask and procedure.
4.4.2 Polymer based microcapillaries
Experimental studies with polymer microcapillaries were performed using the developed
experimental protocol. Gas chromatogram for the samples was used to calculate conversion
of 2,5-dichloro-1,4-benzoquinone in a microreactor. A typical chromatogram obtained for a
sample is shown in Figure 4.9.
63
Figure 4.9 GC-FID chromatogram for a sample
An important fact to note at here is that contrary to mass spectrometer detector (MS), the
area under the curve for compounds analyzed by FID may not represent relative
concentration in a sample. This is because the response of flame ionization detector for a
chemical entity depends on the presence of unoxidized carbon atoms and mass of the entity.
The retention time for the compounds is given in Table 4.3.
Table 4.3 Retention time of reactants and products.
Chemical Species Retention time (approx.)
Acetonitrile 2.7 min
Mesitylene 3.6 min
2,5-dichloro-1,4-benzoquinone 5.6 min
Product peak (major product) 6.6 min
2 methyl indole 7.1 min
Internal Standard
Mesitylene
Reactant– Benzoquinone
Solvent Peak
Chromatogram
Product Peak
Reactant– 2 methyl indole
64
Conversion for 2,5-dichloro-1,4-benzoquinone was calculated using following
mathematical expression-
(4.3)
where AreaBenzoquinone is the area under the curve for benzoquinone and Areamesitylene is the
area under the curve for mesitylene in a chromatogram. The ratio of areas for sample and
reactant gives the total consumption of 2,5-dichloro-1,4-benzoquinone in a reaction as
mesitylene does not participate in a reaction and is always conserved. The experimental data
is tabulated in appendix B.
4.5 Results and Discussions
The conversion achieved in the microcapillaries were calculated using the above formula
and was compared with conversion obtained in a batch system following the same
experimental protocol.
4.5.1 Same surface-to-volume ratio
Following graphs were obtained for microreactors with different material of construction
and having similar surface to volume ratio.
Figure 4.10 is a plot of conversion achieved in microreactors with surface-to-volume ratio
of 7874 m2/m3 (internal diameter of 508 mm). The conversion is higher for Radel R in
comparison to the conversion achieved in batch system. However, the error bars are over
lapping for the PEEK and FEP.
100
reactantmesitylene
neBenzoquino
samplemesitylene
neBenzoquino
1 ×
−
Area
Area
Area
Area
65
Conversion
0
5
10
15
20
25
30
Radel R 508 PEEK 508 FEP 508 BATCH
Microcapillaries with ID (in um)
Con
vers
ion
(%)
Figure 4.10 Plot of conversion in microreactors with surface-to-volume ratio of 7874 m2/m3
Although, the trend is similar for microreactors with surface-to-volume ratio of 15748
m2/m3 (internal diameter about 250 mm), the standard deviation for the conversion in way
too high in comparison to batch system that it is very difficult to derive any clear
conclusion. The plot is shown in Figure 4.11.
Conversion
02468
10121416
Radel R 254 PEEK 254 FEP 229 BATCH
Microcapillaries with ID (in um)
Con
vers
ion
(%)
Figure 4.11 Plot of conversion in microreactors with surface-to-volume ratio of 15748 m2/m3
In case of surface-to-volume ratio of 22857 m2/m3 (internal diameter about 175 mm), the
overall conversion is higher for FEP and PEEK compared to batch with FEP having non-
overlapping error bars (Figure 4.12). Microreactor for Radel R was not available for the
surface-to-volume ratio required in this set of experiment.
66
Conversion
0
5
10
15
20
25
30
PEEK 175 FEP 175 BATCH
Microcapillaries with ID (in um)
Con
vers
ion
(%)
Figure 4.12 Plot of conversion in microreactors with surface-to-volume ratio of 22857 m2/m3
4.5.2 Same material but different surface-to-volume ratio
It was interesting to see that the conversion of 2,5-dichloro-1,4-benzoquinone reduces with
increasing surface to volume ratio for Radel R (figure 4.13).
Conversion
0
5
10
15
20
25
30
Radel R 508 Radel R 254 BATCH
Microcapillaries with ID (in um)
Con
vers
ion
(%)
Figure 4.13 Plot of conversion in Radel R microreactors and batch system
In case of PEEK, the conversion declines initially with minimum conversion achieved in
microreactors with surface-to-volume ratio of 15748 m2/m3 (internal diameter of 250 mm).
Again, the data obtained data for PEEK is not conclusive as the error bars are overlapping
(Figure 4.14).
67
Conversion
0
5
10
15
20
25
PEEK 508 PEEK 254 PEEK 175 BATCH
Microcapillaries with ID (in um)
Con
vers
ion
(%)
Figure 4.14 Plot of conversion in PEEK microreactors and batch system
Similarly, in case of FEP microreactors the conversion is highest for surface-to-volume
ratio of 17467 m2/m3 (internal diameter of 250 mm). Here the error bar does not overlap
with the batch system and is shown in Figure 4.15.
Conversion
0
5
10
15
20
25
30
FEP 508 FEP 229 FEP 175 BATCH
Microcapillaries with ID (in um)
Con
vers
ion
(%)
Figure 4.15 Plot of conversion in FEP microreactors and batch system
4.6 Error Analysis
We believe that the principal reason for such high deviation observed in the experimental
data is because of error associated with the analytical instrument. Standard deviation
measured for set of eight repeated injections in GC-FID was 6.2% and was calculated by
68
manually injecting same solution of 2,5-dichloro-1,4-benzoquinone within a span of 3
hrs. Values for error calculations were obtained using equation (4.4).
(4.4)
Now, if we look at the final conversion obtained in a microreactor using equation
(4.3), it involves four measured parameters. Thus, considering that each parameter
can deviate by about 6.2%, the deviation in the value of calculated conversion can
be significant.
Possibilities for error caused by loss of some chemical species into the environment
was also analyzed by performing blank experiments without 2-methyl indole. In
place of 2-methyl indole, pure acetonitrile was used and chemical entities (i.e. 2,5-
dichloro-1,4-benzoquinone and mesitylene) dissolved in acetonitrile were made to flow
through the microcapillaries. The experiment indicated no loss of chemicals in the
microcapillaries; however again, significant standard deviation was observed indicating that
the error associated with the analytical instrument is the prime cause of the observed
deviation.
Another interesting observation derived from our experimental data was that the standard
deviation was minimum for a batch system than a continuous microreactor flow
experiments. We believe that there are two more factors contributing to the error. Firstly,
the product is collected in a cold sampling vial manually. Although same starting and
sampling protocols were followed, the process is prone to human errors. Secondly, several
plasticizers are used in extrusion manufacturing of polymer based microcapillaries. These
plasticizers can leach-out and influence a reaction, causing deviation in experimental
results.
The errors associated with the study are significantly high to give a clear insight on surface
effects for homogeneous reactions in microreactors. In the next section we look at plausible
samplemesitylene
neBenzoquino
Area
Area
69
causes working behind the observed effect in light of established theories and experimental
observations derived from scientific literature. The theories are an attempt to explain the
trends, and they provide us an insight on the observed enhancement effects.
4.7 Heterogeneity and Organic reactions
Some experimental observations indicate that chemical nature of surfaces as well as
surface-to-volume ratio in a microreactor does affect the kinetics of a chemical reaction.
However the exact science working on the reaction kinetics is unclear. In this section, we
take a closer look at chemical kinetics and discuss enhancement effects observed in other
experimental and theoretical studies. We analyze the experimental results in light of the
known enhancement effects and present a tentative hypothesis for the observed change in
kinetics for our model chemistry in microreactors.
We currently believe that a surface influences an otherwise homogeneous reaction by
participating in it. A homogeneous chemical reaction system ‘sees’ significant participation
of material in reactor walls due to high surface-to-volume ratios, which in turn generate
significant ‘heterogeneity’. Furthermore, the surface can participate in a reaction by either
stabilizing transition states as observed in ‘on-water reactions’, or second by influencing
enthalpy and entropy of a reaction.
4.7.1 On Water reactions
‘On water’ reaction is a group of organic reactions in which reaction in carried out in an
emulsion system with water as an immiscible phase. This reaction group exhibits an
unusually high reaction rate compared to the same reaction in an organic solvent alone or in
a dry media reaction.5,6,7,8,9 Although, water generally acts as a dormant phase in the
reaction system with no active mass transfer occurring between the organic and aqueous
phases, the very presence of it influences a reaction rate. Several research groups have
70
analyzed these ‘on water’ reaction systems closely theoretically or experimentally, and have
put forward several theories to explain it.
Marcus and co-workers have shown that in some ‘on water’ systems, enhancement of
reaction rate (upto 5 folds) can be achieved because of hydrogen bonds protruding in
the organic phase.10 They argue that heterogeneity is crucial for large rate
enhancements since the acceleration was found to be less dramatic if conducted in
homogeneous aqueous solution. In their ‘on water’ system model, one in every 4
hydrogen bonds in the water phase protrude in the organic phase and forms stronger
hydrogen bonds with the transition state than with the reactants. However, in a
homogeneous reaction system within water, the effect is not prominent as the
transition state ‘see’ parallel hydrogen bond which are not as effective in stabilizing
the transition state as perpendicular ones protruding in oil phase. Figure 4.16
compares the highlighting difference in both cases.
Figure 4.16 On-water reactions in comparison to the neat and aqueous homogeneous reactions.10
Thus, heterogeneity generated by oil-water interface can dramatically influence the kinetics
of a reaction.
Some researchers have proposed that the observed ‘on water’ effect of water is primarily
caused by preventing deactivation of a transition state than promotion of activity.11,12 Their
experimental observations showed that water suppresses the formation of proline
71
oxazolidinones in proline-mediated aldol reactions. In either case, it is safe to say that the
heterogeneity generated in a chemical reaction system is capable of influencing the
kinetics of the reaction.
4.7.2 Surfaces and Organic reactions
Presence of seemingly inert surfaces could also influence kinetics of a chemical reaction.13
For example, oxidation of alcohols to aldehydes and ketones using potassium permanganate
gives higher yield when carried out in presence of silicon dioxide.14 Although silicon
dioxide here acts as an inert support, its mere presence affects the yield of reaction.
Bromination of biphenyl, 4-bromobiphenyl, and 4-nitrobiphenyl showed that reaction does
not occur in carbon tetrachloride as solvent if silica gel is absent. Another interesting
example is cyclodehydration of carboxylic g-,d-, and e-aminoacids to the corresponding
lactams on refluxing with toluene in presence of silica gel or aluminium oxide.15 In this case
the surface of the oxide acts as a catalyst. The mechanism of the reaction is outlined in
Figure 4.17.
Figure 4.17 Mechanism for reaction of carboxylic amino acid on SiO2 surface15
72
Thus, there are cases in which surfaces have influenced chemical kinetics of organic
reactions. Similarly, high surface-to-volume ratio in a microreactor may be a prime factor
working towards enhancing organic reactions.
4.8 Microreactors and Organic Reactions revisited
In previous section it was shown how heterogeneity generated at an inter-phase (or surface)
can affect kinetics of an organic reaction. Research findings in which reaction kinetics were
influenced by presence of a surface were also discussed. In this section an attempt has been
made to explain trends and results of the experimental studies conducted, and provide a
tentative hypothesis for influence of surfaces on homogeneous organic reactions in
microreactor.
Interestingly, initial model chemical system for our experimental study has previously
shown ‘on water’ enhancement effects.6 Furthermore, an analogue of 2,5-dichloro-1,4-
benzoquinone has also shown accelerated reaction rates on ferric hydroxide nano particles
surfaces for a reaction with 2-methly indole.9 The proposed mechanism for the reaction is
outlined in Figure 4.18 which indicates participation of hydrogen bonds on the surface for
enhancing the chemical reaction.
Figure 4.18 Mechanism for reaction of benzoquinone with methyl indole on a surface bearing
hydrogen bonds.9
73
A hydrogen bond is an attractive interaction of a hydrogen atom with an electronegative
atom.16 If we take a closer look at polymeric structures of the material of construction for
microreactors, we see that the Radel R has a sulfone bond in every repeating-unit of their
polymer. Similarly, PEEK has one carbonyl group in its repeating-unit. Structures of the
materials are shown in table 4.4.
Table 4.4 Chemical structure and repeated units in the polymeric material
Material Chemical Structure
Radel R SO
OO
O
n
PEEK OO
O
n
FEP F
F
F
F
F
F
F
CF3n
Sulfone group has high polarity and has slightly higher donor properties than carbonyl
group due to less effective sulfur-oxygen p-bonding.17 At the same time, a benzoquinone
shows tautomeric isomerization (scheme 4.1).
O
O OH
O
a b
OH
OH
c
Scheme 4.1
74
It is fair to say that tautomers of benzoquinone (b, c) can participate in hydrogen bonding
with sulfone group present in the backbone of Radel R.18 The hydrogen bonding between
the sulfone group and the tautomers will be higher than the one between the carbonyl group
and the tautomers as sulfone has slightly higher electronegativity than carbonyl group.17
Thus, conversion of 2,5-(dichloro) 1,4-benzoquinone should be highest for Radel R,
followed by PEEK and FEP.
As the experimental data has substantial deviations, it is hard to correlate results of the
study with any of the discussed rate enhancement theories. We do see enhancement of
organic reactions in microreactor however the trends could not be explained.
4.9 Summary
In this chapter, we discussed development of the experimental protocol and several issues
and problems encountered during experimentation. The experimental observations were
reported, analyzed and discussed to arrive at an intermediate conclusion regarding effects of
surfaces on a homogeneous organic reaction.
The study revealed that surfaces do affect homogeneous organic reaction in a microreactor
although a clear picture could not be derived due to error associated with the experimental
data. Cnversion of 2,5-dichloro-1,4-benzoquinone in the model chemical reaction was
highest for Radel R microcapillary with internal diameter of 508 mm (surface-to-volume
ratio 7874 m2/m3). Plausible cause of the observed effect and rate enhancements effects
observed was discussed in light of studies on ‘on water’ reaction and ‘reactions on surfaces’
by other researchers. ‘On water’ model proposed by Jung et al seems most convincing and
applicable to our system, and was used as a base model to explain trends seen in the current
research work.10 It was proposed that the tautomers of 2,5-(dichloro) 1,4-benzoquinone
interacts with the surface of the microreactors (Radel R) and give higher conversion.
75
4.9 References
1. Zhang, H. et al. Synthesis of Aryl-Substituted 1,4-Benzoquinone via Water-Promoted and In(OTf)3-Catalyzed in situ Conjugate Addition-Dehydrogenation of Aromatic Compounds to 1,4-Benzoquinone in Water. Advanced Synthesis & Catalysis 348, 229-235(2006).
2. Jackman, R.J. et al. Microfluidic systems with on-line UV detection fabricated in photodefinable epoxy. Journal of Micromechanics and Microengineering 11, 263(2001).
3. Harris, D.C. Quantitative chemical analysis. (WH Freeman: 2003).
4. Mcwilliam, I.G. & Dewar, R.A. Flame Ionization Detector for Gas Chromatography. Nature 181, 760-760(1958).
5. Narayan, S. et al. "On water": unique reactivity of organic compounds in aqueous suspension. Angewandte Chemie (International ed. in English) 44, 3275-3279(2005).
6. Zhang, H. et al. “On Water”-Promoted Direct Coupling of Indoles with 1,4-Benzoquinones without Catalyst. European Journal of Organic Chemistry, 869-873(2006).
7. Pirrung, M.C. & Sarma, K.D. Multicomponent Reactions Are Accelerated in Water. Journal of the American Chemical Society 126, 444-445(2004).
8. Meltzer, P.S. & Branch, C.G. Fast reactions ‘on water’. Nature 435, 3-4(2005).
9. Niu, F. et al. Promotion of organic reactions by interfacial hydrogen bonds on hydroxyl group rich nano-solids. Chemical Communications, 2803–2805(2008).
10. Jung, Y. & Marcus, R.A. On the theory of organic catalysis "on water." Journal of the American Chemical Society 129, 5492-502(2007).
11. Nyberg, A.I., Usano, A. & Pihko, P.M. Proline-Catalyzed Ketone-Aldehyde Aldol Reactions are Accelerated by Water. Synlett, 1891-1896(2004).
12. Blackmond, D.G. et al. Water in organocatalytic processes: debunking the myths. Angewandte Chemie (International ed. in English) 46, 3798-3800(2007).
13. Basiuk, V.A. Organic reactions on the surface of silicon dioxide: synthetic applications. Russian Chemical Reviews 64, 1003-1019(1995).
14. Regen, S.L. & Koteel, C. Activation through impregnation. Permanganate-coated solid supports. Journal of the American Chemical Society 99, 3837-3838(1977).
15. Bladé-Font, A. Facile synthesis of γ-, α-, and e-lactams by cyclodehydration of ω-amino acids on alumina or silica gel. Tetrahedron Letters 21, 2443-2446(1980).
76
16. Jeffrey, G.A. An introduction to hydrogen bonding. (Oxford University Press, USA: 1997).
17. Drago, R.S., Wayland, B. & Carlson, R.L. Donor Properties of Sulfoxides, Alkyl Sulfites, and Sulfones. Journal of the American Chemical Society 85, 3125-3128(1963).
18. Oznobikhina, L. et al. Orientation of hydrogen bond in H-complexes of sulfones and sulfonamides. Russian Journal of General Chemistry 79, 1674-1682(2009).
{Bibliography}
77
5. Summary and Outlook
This thesis focused on single phase organic reactions in microreactors and analyzed surface
effects on homogeneous organic reactions in microreactors. The discussion started with a
general overview on microreactor types, fabrication methodologies, their unique physical
properties, and chemical properties. Organic reactions that have been carried out in
microreactors, and reaction rates and yield enhancement cases were highlighted. The
motivation to understanding the enhancement effects was discussed which laid the basis of
this thesis.
Enhancement effects were investigated by analyzing possible physical and chemical factors
(temperature, pressure and surfaces) which could affect reaction rates and yield of an
organic reaction in a microreactor. A conclusion was derived that effect of surfaces is a
strong factor which can influence chemical kinetics. A thorough analysis was planned by
designing systematic experimental studies with different materials and surface-to-volume
ratio for microreactors respectively. A model homogeneous organic reaction was chosen to
suit the engineering and analytical requirements of the study. This reaction chemistry was
optimized for reaction speed and compatible solvents, and reaction between 2,5-Dichloro-
1,4-benzoquinone with 2-methyl indole in acetonitrile.
In order to study our premise, two types of microreactors were chosen--polymer
microcapillaries and custom made silicon microreactors. Polymer microcapillaries are
becoming popular within scientific community as they are relatively cheaper and readily
available. Thus a study which could give a better insight about their efficacy will be
beneficial to the community. Microcapillaries made out of Radel R, PEEK and FEP with
internal diameter of 508 mm, 254 mm, 229 mm and 175 mm were purchased, and were
grouped together on the basis of similar internal surface-to-volume ratio but different
material and same material but different surface-to-volume ratio.
78
Silicon microreactors were fabricated at A*STAR Institute of Microelectronics, Singapore.
The silicon microreactors were designed such that they have same internal volume but
different surface-to-volume ratio.
In chapter 4 we discussed the setup designed for carrying out experiments. The problems
associated with experiments in silicon microreactors and UV-vis analysis were discussed
and analyzed. The experiments were performed such that the residence time in all the
microcapillaries was 18 min. Conversion of 2,5-(dichloro) 1,4-benzoquinone in different
microreactors were calculated and plotted both for similar surface-to-volume ratio and for
same material with different surface-to-volume ratio. The study revealed that surfaces do
affect homogeneous organic reaction in a microreactor with highest conversion obtained in
Radel R microcapillary (internal dia.-508 mm), however the trend was irregular. The
deviation associated with quantification by GC-FID was calculated to be 6.2%, which could
be the prime reason for such a high deviation.
The observed effects were explained in light of heterogeneity generated in ‘on water’
reactions and reactions occurring at surfaces of materials. The conversion trend observed
for Radel R, PEEK and FEP were analyzed in light of possible van der waal interactions
and hydrogen bonding occurring between transition state and microreactor surface.
However, some trends observed in the experimental data are still not completely
understood.
5.1 Principal Thesis Contributions
A systematic study was planned and executed to analyze the enhancement in reaction rates
and yield for homogeneous organic reactions in microreactors. Several physical and
chemical factors were analyzed which could result in change in kinetics of the reaction.
Microfabrication of silicon microreactor was reported and engineering problems and
challenges encountered while performing experiments were discussed. This will be of
immense use for someone fabricating and working with microreactors (especially silicon).
79
The thesis work also revealed that surfaces do affect the course of reaction and is indeed a
factor which can influence chemical kinetics of a reaction.
80
Appendix A
81
Appendix A
82
Appendix A
83
Appendix A
AutoCAD drawing of the patterned mask – wafer #1 - frontside
84
Appendix A
AutoCAD drawing of the patterned mask – wafer #1 - backside
85
Appendix A
AutoCAD drawing of the patterned mask – wafer #2 - frontside
86
Appendix A
AutoCAD drawing of the patterned mask – wafer #2 - backside
87
Appendix B
Experimental data obtained by GC-FID analysis of the products
88
Appendix B
89
Appendix B
