View Article Online / Journal Homepage / Table of Contents for this issue

Polymer Dynamic Article Links < C

Chemistry
Cite this: Polym. Chem., 2012, 3, 2698
www.rsc.org/polymers REVIEW
Published on 03 July 2012. Downloaded by State University of New York at Stony Brook on 30/10/2014 21:16:45.

The spinning disk reactor: an example of a process intensification technology

for polymers and particles
Stephen D. Pask,*a Oskar Nuykenb and Zhizhong Caic
Received 18th April 2012, Accepted 29th June 2012
DOI: 10.1039/c2py20237a

This review introduces the spinning disk reactor and emphasises its suitability for all aspects of polymer
chemistry. The literature reviewed includes university theses, as well as key patent and openly published
literature; it is not an exhaustive collation of the literature to date. The spinning disk reactor (SDR) is
introduced in the context of process intensification (PI). The SDR is the apparatus which has the best
mass and heat transfer characteristics of all typical chemical reactor types, while being flexible and
economical (material and energy). It is inherently safer than traditional stirred tank reactors. The
theory essential for the practical use of SDR is briefly summarised to emphasise the ease with which
laboratory SDR experiments can be scaled up to produce commercial quantities. The use of SDR to
produce well defined filler particles with simple (e.g. BaSO4) and reactive (Ag) precipitation reactions is
discussed. In order to emphasise the general and facile utility of the SDR in all aspects of polymer
chemistry, the literature available for polymer analogue (modification, e.g. hydrogenation) reactions is
presented. Further, step-growth- (polyurethanes), chain-growth- (e.g. polystyrene) and condensation
(e.g. polyesters) polymerisations are examined. UV-activated bulk, emulsion and solution
polymerisations and co-polymerisations (acrylates) are also described. Some of the experiments
described were carried out in the senior author’s laboratory and as well as critically reviewing the
available literature, a possible direction for future work is indicated.

Introduction overlain today by increasingly sophisticated Information Tech-

nology (IT) and control systems.
This review is not intended to be an exhaustive summary of the In order to understand trends in the chemical industry, it is
literature to date but rather an introduction to a technology, the always necessary to examine at least two aspects: economics and
spinning disk reactor, which is slowly becoming accepted in both safety. Thus, the trend away from acetylene as a chemical base
industry and academia as a viable alternative to traditional raw material1 has been driven by the perceived difficulties with
reactors. The emphasis is on the suitability of these types of controlling acetylene chemistry (although considerable advances
reactors for use in polymer and polymer related chemistry. have been made e.g. with an understanding of copper acetylide
Over the last, say 200 years, the development of the chemical chemistry2) and the availability of cheap oil. In particular, the
industry has tended to move to ‘‘big is beautiful’’, driven by the development of a world economy but also the increasing cost of
economics of scale but also, e.g. by concepts of surface area to transport and the obvious limited supply of oil as a raw material
volume (allowing scale-up, despite less precise purification are leading to a greater emphasis on the total environmental
procedures). However, the increasing size of chemical equipment footprint of a product during its entire lifetime, while continuing
has led e.g. to long investment lead times, barriers to introducing pressure to maintain market leadership is emphasizing cost to
new materials, unnecessarily expensive supply chains and a market.3
considerably greater impact of failure. While much of In an attempt to overcome the shortcomings and address
manufacturing has seen a rapid change in the technologies it social and market pressures, considerable efforts are now being
employs, most of the process plants in place today would be focused on the area of process intensification (PI). In the chem-
recognized by engineers who worked 100 years ago, albeit ical industry intensification can be applied at all stages of the
process:
a
SteDaPa Consulting, Im Gerstenfeld 44, D-41542 Dormagen, Germany.
Raw material storage
E-mail: [email protected]
Reaction
b
Wacker-Lehrstuhl f€
ur Makromolekulare Chemie, Technische Universit€
at
Separation steps
M€ unchen, Lichtenbergstr. 4, D-85747 Garching, Germany
c
Product isolation and purification
BASF Construction Chemicals GmbH, Chemiepark, Dr.-Albert-Frank-
Str. 32, D-83308 Trostberg, Germany
Packaging

2698 | Polym. Chem., 2012, 3, 2698–2707 This journal is ª The Royal Society of Chemistry 2012
 View Article Online

Although probably only about 10% of the equipment inven- large; small scale is a batch process. However, even assuming
tory is caused by the reaction step,4 the reactor determines the that a reaction takes place in a 10% solution, a throughput of
amount of material inventory and the size of the subsequent 20 ml s1 will yield 50 t a1 product. The laboratory equipment
equipment; it is generally the process step involving the highest and the production equipment can be identical, there are simply
risk to health and safety. Furthermore, improvements in reactor more units in the production and/or, as in the case of the spinning
technology which also lead to improved reaction specificity and a disk, the scale up process is trivial. Thus, the introduction of new
Published on 03 July 2012. Downloaded by State University of New York at Stony Brook on 30/10/2014 21:16:45.

reduction in the amount or number of by-products help to reduce products can also be accelerated.18
the cost of subsequent product purification (which has been Although major effects have been achieved by developing
estimated at 40% for most commodity chemicals5). novel equipment for carrying out chemical reactions, the devel-
The spinning disk reactor (SDR) can be seen as one of the opment of appropriate auxiliary equipment (e.g. heat
leading building blocks of the PI trend in chemical production. exchangers, mixers or distillation units) and the use of novel
The origins of PI can be traced back at least thirty years and the techniques are just as important for the overall success of PI. PI
pioneering work of C. Ramshaw6,7 is now being increasingly can be seen as an essential aspect of sustainable development;
incorporated into the industrial world but the origin of the producing more with less is process intensification.
spinning disk can perhaps be attributed to Kenneth Hickman It is believed that the hesitant acceptance of PI in industry is
and his work to improve the distillation process during the 1930s not only due to the desire to exhaust the value of existing plants;
(see e.g. the US patent literature8–10). A major difference between it can also be ascribed to a lack of knowledge of the possibilities
large scale and micro-scale reactors, typical of PI, and particu- as well as a degree of antipathy to things new. This article is not
larly of spinning disk technology, is the considerably improved intended to be an exhaustive review of the literature concerning
heat transfer and the importance of surface interactions.11 This SDR but rather an overview of some aspects of polymer chem-
explains the increasing scientific interest in this technology.12–16 istry where using spinning disk technology can improve existing
Such process intensification (PI) considerably reduces the scale processes and open the door to novel products.
of the production process which in turn can lead to: Brief sections on some inorganic materials which are used as

reduced investment; smaller plant profile; simpler plant fillers for polymers are included because some of the techniques

reduced stocks; more rapid, easier grade changes and effects have relevance for polymer chemistry and point to

reduced waste; lower material costs; increased reaction possibilities for novel production processes (see especially the
specificity generation of silver nano-particles in a polyol as a precursor for

reduced energy costs; improved temperature control; excel- reactive polyurethanes). No attempt is made to compare other
lent heat transfer elements of the process intensification toolbox.

reduced time and cost to market for new products; simplified With a more widespread acceptance of spinning disk tech-
‘‘scale up’’ nology it may well be that some reactions, which are entrenched

reduced safety risks; and smaller volumes in the process in a traditional picture (for example cationic polymerisations,
Indeed, a change in a chemical production process in terms of low temperature, difficult to control, etc. See below), come to be
technique or equipment that leads to one or more of the above seen in a new light.
reductions has been seen as a definition of PI.17 An overview of
the so-called ‘‘PI toolbox’’ is given in Fig. 1.
It can be argued that the use of novel solvents, such as
Introducing the spinning disk reactor
supercritical fluids, or energy sources, such as ultrasound, also
need to be considered as PI. Indeed, the use of smaller reactors A schematic picture of a typical spinning disk is shown in
makes some techniques, which have been developed in the Fig. 2 19 which outlines the general operating set-up including
laboratory but have proven unmanageable on a large scale, able heating and cooling of the disk and reactor walls.
to be effectively employed with PI equipment due to the much The object of a spinning disk reactor is to generate a highly
smaller scale of typical PI equipment. Furthermore, with the sheared liquid film when a liquid is applied to the unit at or near
introduction of the novel technology each process step needs to its centre. The film is instantly accelerated tangentially by the
be analysed and adapted (Responsive Processing) rather than the shear stresses established at the disk/liquid interface. This causes
traditional approach, adapting the process to existing the liquid to approach the disk’s angular velocity and then move
equipment. outward as a thinning/diverging film under the prevailing
A typical feature of much of the chemical industry is its cyclical centrifugal acceleration. This force stretches and contorts the
nature which is exacerbated by the traditional desire for largest film.20 The detailed form of the film is complex and depends not
scale equipment. Thus, when developing new products or only on the viscosity of the film material and the speed of rota-
considering plant expansion, the first question usually addressed tion but also on the surface geometry of the disk.21 The thin
is: how big can we build? Additionally, the large scale of the liquid film generated allows for high rates of mass transfer such
investment generally requires a long decision time. that it favours absorption, stripping,22,23 mixing and reaction
When the new capacity does come on stream there is a period processes.
of oversupply and price pressure. Instead of one large unit, the PI Although the precise thickness of the film of material on the
concept envisages several smaller units and involves considerably disk requires a detailed and rather complicated analysis for most
shorter residence times. Another feature of PI equipment is that synthetic work simple formulae are adequate. Thus, the film
they are continuous reactors and in many branches of the thickness and the residence time can be calculated using eqn (1)
chemical industry, continuous tends to be synonymous with and (2):

This journal is ª The Royal Society of Chemistry 2012 Polym. Chem., 2012, 3, 2698–2707 | 2699
Published on 03 July 2012. Downloaded by State University of New York at Stony Brook on 30/10/2014 21:16:45. View Article Online

Fig. 1 The PI toolbox.17

 1=3  1=3
3nQ 81p2 n 4=3
Film thickness ðmÞ : f ¼ (1) Mean residence time ðsÞ : t ¼ r (2)
2prr2 u2 16rQ2 u2
where y ¼ liquid viscosity (Pa s), Q ¼ liquid flow rate (m3 s1), where r ¼ disk radius (m), y ¼ liquid viscosity (Pa s), r ¼ liquid
r ¼ liquid density (kg m3), r ¼ disk radius (m), u ¼ rotational density (kg m3), Q ¼ liquid flow rate (m3 s1), u ¼ rotational speed
speed of the disk (radian s1) and f ¼ liquid film thickness (m). of the disk (radian s1) and t ¼ mean residence time of the liquid (s).

Fig. 2 A schematic view of a spinning disk reactor.

2700 | Polym. Chem., 2012, 3, 2698–2707 This journal is ª The Royal Society of Chemistry 2012
 View Article Online

An SDR generally has a working disk with a diameter between hundreds of thousands of m2 m3 for low viscosity systems
10 cm and 1 m. The limit is determined by engineering rather typical of a wide range of chemical synthesis routes.
than theoretical constraints. Disks made of various metals, often The high surface to volume ratio can be used to allow both
with a base of copper (excellent thermal conductivity) with a thin rapid transport between gas and liquids either for simple oper-
chrome plating for chemical resistance, are employed but again, ations such as stripping liquids of volatiles, scrubbing gases or
the limits are practical rather than theoretical. The disk can be for more complex gas–liquid reactions.
Published on 03 July 2012. Downloaded by State University of New York at Stony Brook on 30/10/2014 21:16:45.

smooth, grooved or meshed depending on the application and For the special case of rapid volatiles removal the short resi-
the throughput requirement. The rotational speeds of an SDR dence time and rapid heating offered by spinning disk processing
can vary from 100 to 6000 rpm (typical is 1000 rpm). The can also provide the ability to use much higher liquid tempera-
spinning disk reactor has the following characteristics:24 tures, without degradation, producing higher vapour pressures

Intense shearing and mixing in thin liquid film and further increasing mass flux and allowing the possibility for

Short liquid residence time much lower residual concentrations in the product without the

High solid/liquid heat/mass transfer need for an extreme vacuum.

High liquid/vapour heat/mass transfer A general benefit of process intensification is the advantage
Some general observations concerning SDR are worth offered during industrial scale-up. Experience has shown that the
mentioning to develop an understanding of the technology, in scale up of batch reactors is fraught with fundamental problems
particular, with reference to polymer chemistry and that can result in the journey from a laboratory experiment to a
manufacture. manufacturing process being long, arduous and expensive.
Viscous liquids, including polymers, behave in an unexpected The characteristics of heat transfer in a batch reactor illustrate
way on a spinning disk. As the liquid moves over the disc shear the problem. Thus, the doubling of the dimensions of a reactor
forces will tend to shear thin the fluid resulting in an apparent results in the surface area being increased by a factor of 4 but the
reduction in viscosity (this leads to films being thinner than volume by a factor of 8. This relative loss of surface to volume
calculated according to eqn (1)). Also, as the liquid film has to will, at best, result in longer heat up and cool down rates and
cover more area as it moves towards the edge of the disk the exothermic reactions may have to be operated at lower temper-
liquid film is continuously drawn. This drawing phenomenon is atures to avoid runaway reactions. Such changes often result
characteristic of SDR and tends to untangle the long polymer in changes to the chemistry and can lead to yield losses or
chains.14,24,25 Polymeric products often contain residual mono- unexpected by-products.
mers and/or solvents and these small molecules are trapped Agitation in a batch reactor is also impossible to scale in a
within the polymer coils and are difficult to remove. The chain fundamental sense. Two classic methods involve keeping the
untangling effect means that any monomers or solvent in the power per volume constant or keeping the stirrer tip speed the
polymer are no longer trapped and removal by evaporation is same. It is, however, impossible to keep both the same.
much more effective. This effect may also lead to improved A spinning disk reactor has very different characteristics.
reactivity for polymer modification reactions, especially where Clearly there are a range of important variables for a process on a
one of the reactants is in the gas phase (see below). disc such as residence time, velocity, and film thickness that will
The heat transfer performance of an SDR is dependent on the affect the processing. However, by using the simple scale up
exact form of the film as it moves across the disc. Waves tend to strategy of keeping the rotational speed the same irrespective of
form in the film that significantly enhances heat transfer (as well scale and keeping the flow rate per area constant, then for
as mass transfer) giving film coefficients of around 20–50 kW varying sizes of disks the data shown in Table 1 can be calculated
m2 K1 for low viscosity fluids. However, the film coefficient for for a rotational speed of 600 rpm and for fluids similar to water.
the heating/cooling fluid and the thermal resistance of the disc Last but not least, the scale-up is trivial since an SDR process
itself are critical. can be scaled up without changing any of the key parameters; it is
Generally, an SDR has an overall heat transfer coefficient of determined solely by simple physical laws without recourse to the
approximately 7–10 kW m2 K1 even for organic liquids. This is various operating numbers such as Prantl and Reynolds coeffi-
typically 5 to 10 times that achieved by most heat transfer devices cients. Processes can be transferred from the laboratory to plant
and enables small discs with low process fluid inventory to handle scale without elaborate or particularly costly alterations and can
significant thermal duties. be used for a continuous production of tons of product with a
At high viscosities the diffusion control limitations imposed by very high degree of confidence that the resulting products and
the bulk viscous reaction mixture in a typical batch reactor yields will be similar, if not identical. It should be noted that,
system are easily overcome in the thin film formed on the rotating rather than is the case for some other PI tools (such as micro- or
disk surface, thus enabling the polymerisation to proceed on the
SDR at inherently faster rates. Table 1 Scale up effects for Spinning Disk Reactors14
The ability to generate thin liquid films through high centrif-
ugal acceleration over the spinning disc surface offers a wide Heat transfer
coefficient in
range of opportunities for processes conventionally limited by Disc Flow Avg. film Residence film
interfacial mass transfer. The generation of films of typically diameter (m) rate (ml s1) thickness (mm) time (s) (kW m2 K1)
fractions of a millimetre down to a few microns, through
controlled flow and disc speed, can deliver surface to volume 0.1 2 68.9 0.270 52
0.3 18 68.9 0.270 52
ratios tailored to the processing requirements from thousands of 1.0 200 68.9 0.270 52
m2 m3 for high viscosity materials, such as polymer melts, up to

This journal is ª The Royal Society of Chemistry 2012 Polym. Chem., 2012, 3, 2698–2707 | 2701
 View Article Online

tubular-reactors26), the SDR can be scaled-up and does not

necessarily require numbering-up.
In his PhD thesis Cai14 showed that it was possible to
synthesize both silylated and non-silylated polyurethanes of
similar quality using an SDR as those synthesized in typical
batch reactors but that the SDR technology can produce the
Published on 03 July 2012. Downloaded by State University of New York at Stony Brook on 30/10/2014 21:16:45.

same amount of product in a shorter (ca. 60%) time, using less

(ca. 60%) energy and using almost one hundred times less solvent
to clean the apparatus. His results are given in Table 2.
It should also be noted that within a single SDR it has been
shown possible to have either several disks operating in sequence
or parallel,27 allowing either a corresponding increase in resi-
dence time under identical conditions, which can also be achieved
by recycling the reaction solution, or simply increasing the
throughput without considerably increasing the size of the
apparatus.
In summary, ‘‘every molecule that passes over the disc has Fig. 3 A comparison of the particle size distribution of barium sulphate
exactly the same experience as every other molecule in terms of precipitated using an SDR with that produced in a stirred tank reactor.
its heat profile, mixing characteristics, mass transfer and the
concentration gradient’’.28
Perhaps the most detailed and easily available description of
the precipitation of barium sulphate using an SDR is given in GB
Barium sulphate and calcium carbonate Patent 2377661.30 As well as being unusual in the amount of
scientific detail given, this patent makes the point that employing
In the next two sections the use of SDR to produce small, well
an SDR to produce solid particles is not only practical but also
defined inorganic particles will be described but SDR can also be
uses considerably less energy (of the order of kW kg1) than is
used to prepare similar particles from organic materials with just
required for T- or Y-mixing equipment. It is also emphasised that
the same advantages.29
an SDR is more easily kept free of fouling than traditional
Barium sulphate occurs naturally as the mineral barite (heavy
equipment or even T-mixers. The flexibility of an SDR approach
spar) and finds its major use as an additive for drilling muds due
is also demonstrated by applying ultrasonic energy to the spin-
to its high density, lack of magnetism, chemical inertness and
ning film or by adding agar to the solutions used; both lead to a
high thermal stability. Finely ground barite is also used as a
change in the median of the narrow particle size distributions by
weighting filler for plastics and rubber. For better uniformity of
a factor of ca. 10. Typical size distribution curves (from GB
colour barite is employed as a filler for rubbers and plastics, as
2377661)30 are shown in Fig. 3. Importantly, for the use of SDR
well as in paints and coatings. Furthermore, due to its large
as the apparatus of choice for precipitation production of nano-
coefficient of absorption for both gamma- and X-rays, it is used
particles is the micro-mixing time which needs to be very short to
medicinally as a contrast agent. The naturally occurring mineral
ensure that homogeneous nucleation dominates31 and this patent
is first converted to the sulphide, which is partially (ca. 7%) water
demonstrates that on an SDR mixing times of less than 1 ms are
soluble, and the sulphate precipitated by adding either sulphuric
easily obtained, an aspect which should be remembered when
acid or sodium sulphate to the solution. This process leads to
very pure, white barium sulphate (Blanc fixe) and can be
controlled to yield material of varying particle size. Commer-
cially, precipitated barium sulphate is sold with particle sizes
ranging from ca. 40 mm down to ca. 2 mm. The advent of rela-
tively cheap methods to enable particle size and distribution
control offers the possibility for novel and perhaps interesting
new applications of traditional fillers.

Table 2 A comparison of the production of 3.6 t of an alkoxy silane

endcapped polyurethane prepolymer in a batch reactor and in an SDR

Batch-process SDR-process

Scale 4 m3 100 cm disk

Loading 1.5 h 100 ml s1
Heat-up period 2.5 h 27 min
Reaction 7h 3s
Cooling 1.5 h 20 min
Emptying 1.5 h 100 ml s1
Cleaning 2h 100 ml s1
Total time 16 h 10 h
Cleaning fluid 1000 kg 12.5 kg Fig. 4 SDR cover with sight glass and variable nozzles for the precipi-
tation of BaSO4.

2702 | Polym. Chem., 2012, 3, 2698–2707 This journal is ª The Royal Society of Chemistry 2012
Published on 03 July 2012. Downloaded by State University of New York at Stony Brook on 30/10/2014 21:16:45. View Article Online

Fig. 5 Cubic and spherical particles of calcium carbonate produced simply by changing the speed of the disk rotation.30

considering chemical reactions on an SDR and making short residence times associated with SDR allow homogeneous
comparisons with simple stirred reactors. nucleation to be achieved without exceptional precautions, thus
It can be assumed32 that nano-fillers will have other properties leading to narrow particle size distributions.
than fillers of larger particle size and control of such different Typical choices of polyols include ethylene glycol, 1,2-
properties will become an essential aspect of this development. propylene glycol, and 1,5-pentanediol. Because the solution
Thus, it is also interesting to consider, that addition of other viscosity is largely determined by the length of the polyol’s
substances to the product is particularly easy using an SDR. A hydrocarbon chain (for the solvents listed, h ¼ 16.1, 40.4, and
typical example of such an experimental set-up is shown in Fig.4 140 mPa s, respectively), the choice of solvent can greatly impact
taken from Cai’s PhD thesis,14 who ably demonstrates that diffusion and growth processes associated with metal nano-
simply by changing the site of the addition of the coagulant crystal formation. In addition, the polyol acts as the reducing
(sodium sulphate) towards the outer rim of the disk the average agent for metal reduction and, at the relatively high temperatures
particle size of barium sulphate particles could be reduced from employed, can lead to the polyol being oxidized to various
ca. 5 mm to 170 nm. aldehyde and ketone species. This reaction becomes very inter-
In GB Patent 237766130 a facile control of particle morphology esting if one considers that the resulting solution of nano-
is described using the example of the precipitation of calcium particular silver could form the basis for the production of
carbonate from calcium hydroxide solutions with gaseous polyurethane adhesives or coatings or indeed other poly-
carbon dioxide. Thus, simply by changing the speed of rotation condensates, especially since an SDR approach limits secondary
of the disk (from 1100 to 1650 rpm) it was possible to produce reactions due to the short times for which the dispersion expe-
either ‘‘cubic’’ or ‘‘spherical’’ particles (see Fig. 5). These pictures riences higher temperatures. Using an SDR it has been shown
also emphasise the very uniform nature of the particles produced possible to vary both the size and the shape of silver nano-
employing an SDR. particles by precipitating the silver from aqueous solutions with
ascorbic acid in aqueous solution and simply changing the disk
speed or the reagent concentration.38 More recent work14
Metal nano-particles
employing a polypropylene glycol with terminal alcohol moieties
The market for well defined (narrow particle distribution) metal as the solvent and an average molar mass of 8000 has shown that
nano-particles is increasing rapidly. Silver, for example, is being spherical silver particles with a narrow particle size distribution
used for conductive inks and adhesives in various electronic (see Fig. 6) can be produced. It is worth emphasizing that even
units.33,34 using the higher molar mass polyol the thickness of the film
It should be remembered that silver powder has long been used remains well under 0.5 mm so that heat transfer as well as mixing
medicinally, particularly in postnatal prevention of infection remain excellent.
and, provided the technology is economical, the use of fine Another example of the preparation of metallic nano-particles,
particulate silver as a general biocide could become more general. which may have special relevance to polymer analogue reactions
A very successful reaction for producing metal nano-particles considering some of the comments made in this discussion, is the
with controlled shape is the so-called ‘‘polyol process’’.35 Using work of Zou et al.39 In this work very uniform aggregates
surface-capping reagents, such as poly(vinyl pyrrolidone) (PVP), (160 nm) of 5 nm palladium spheres are described using an SDR.
short, nano-scale silver wires can be produced.36 An aqueous solution of H2PdCl4 was reduced with hydrogen gas
Two principal methods of generating nano-particles from and, by adding polyvinyl pyrrolidone (PVP), the palladium
metals involve either homogeneous or heterogeneous nucleation. spheres could be collected as made and, after freeze drying,
In order to control the shape and crystal form of the particles via stored for several months. The authors show that by varying the
homogeneous nucleation, it is critical to ensure that the level of molar mass of the PVP, the ratio of PVP/Pd and the rotational
nucleating species is controlled after initial nucleation.37 The speed of the disc, the size of the spheres and their uniformity can

This journal is ª The Royal Society of Chemistry 2012 Polym. Chem., 2012, 3, 2698–2707 | 2703
Published on 03 July 2012. Downloaded by State University of New York at Stony Brook on 30/10/2014 21:16:45. View Article Online

Fig. 6 Silver particles produced using an SDR, particle size distribution and a TEM of the same sample.14

be optimized. The palladium nano-particles could be recycled specialty rubber (formerly Hypalon from du Pont), using an
several times and were employed for a Mizoroki–Heck coupling SDR.
reaction of aryl halides with aliphatic double bonds. Despite the The use of immobilized catalysts on an SDR is described in the
interest of this research group in SDR they do not report trying work of Vicevic et al.45 Although this work involves small
the coupling reaction using an SDR. Nevertheless, this report molecule chemistry it presents an interesting approach and the
and a consideration of the catalytic cycle for this reaction, which advantages of using an SDR for reactions which, up to now, have
is relatively well studied,40 do suggest some interesting directions required heterogeneous or homogeneous catalysts. In particular,
for future work. this work emphasizes the additional information (selectivity)
about reactions which only become apparent when new kinetic
information becomes available.
Polymer hydrogenation
The hydrogenation of commercial polymers (e.g. polybutadiene Polyurethanes
(BR) or acrylonitrile-co-butadiene (NBR)) yields polymers that
are more resistant to ageing and can have other desirable prop- Since the discovery of polyurethanes by Otto Bayer in 1937,
erties. BR yields special forms of, formally, ethylene propylene polyurethanes have become one of the most important classes of
copolymers (EPM) and NBR yields products known commer- polymers with applications varying from coatings and adhesives,
cially as Therban or Zetpol† – well known speciality elasto- to plastics, thermally insulating foams, elastomers and thermo-
mers. The current process for hydrogenating NBR is carried out plastic elastomers. Some 10 million tons of polyurethanes are
batchwise using either homogeneous or heterogeneous catalysts. produced globally every year. The basic reaction, a nucleophilic
One disadvantage of the process, typical of polymer analogue attack of a hydroxyl group on an isocyanate group, gives poly-
reactions, is the diffusion control, both with respect to the mers via an addition polymerisation if the reactants are multi-
catalyst and the hydrogen. The latter leads to very high pressures functional. The variability of the isocyanate- and the hydroxyl-
being required to enable economic rates of hydrogenation.41 terminated molecules has allowed the myriad of different mate-
From ref. 41 it can be derived that using an SDR technology rials based on this simple chemistry. Industrially, polyurethanes
the hydrogen pressure necessary for an economical hydrogena- are generally produced in conventional stirred tank reactors
tion could be reduced by a factor of 10 with an accompanying either batchwise or continuously (thermoplastic elastomeric
reduction in the cost of equipment and safety measures. It can be polyurethanes are produced using twin screw extruders).
assumed that with an SDR the mass transfer from the gas to the In his PhD thesis Cai14 compared the synthesis of both sily-
liquid phase is considerably improved and probably that the lated and nonsilylated polyurethane prepolymer binding agents
stretched polymer molecules are perhaps more easily using an SDR and in batch. His results, in terms of the advan-
hydrogenated. tages of using an SDR, have been discussed above. More
Furthermore, it should be possible to use similar equipment to importantly he was able to show that not only is an SDR a more
produce the lower molar mass HNBR described in US 7381781 42 suitable (e.g. more economic) apparatus for making such mate-
and indeed the use of ultrasound to reduce molar masses rials than a batch reactor but that the materials so produced were
described in EP 180113443 could be coupled to give a complete of a similar quality. One interesting aspect of the work, which is
and new type of production. not further investigated, is the variation in catalyst activity
The use of an SDR has been described for the chlorination and between the batch process and the SDR process, which is perhaps
sulphonation of hydrocarbons44 so that it should also be possible a function of the much higher temperatures (up to 200  C as
to produce chlor-sulphonated polyethylene (CSM), an important compared with 50–80  C for the batch process) and short resi-
dence time associated with the SDR process. Cai found that
† Therban is a registered trademark of Lanxess Deutschland GmbH bismuth(III)-neodecanoate was the better catalyst for the SDR
and Zetpol is a registered trademark of the Zeon Corporation. process when compared with zirconium-2,4-pentandionate. He

2704 | Polym. Chem., 2012, 3, 2698–2707 This journal is ª The Royal Society of Chemistry 2012
 View Article Online

also showed that, at least at higher temperatures, products could monomer. It would be interesting to see if the same could be
be made with identical isocyanate functionality, similar MWD achieved for 2-methyl propene (isobutylene, butyl rubber), the
but lower residual starting isocyanate. largest volume, cationically polymerised polymer available
One interesting point that Cai makes is that the centrifugal commercially.
force operating on an SDR (which even for smaller discs is as
much as 1000 times gravity for rotational speeds of ca. 1500–
Published on 03 July 2012. Downloaded by State University of New York at Stony Brook on 30/10/2014 21:16:45.

Polycondensation
3000 rpm) leads to the polymer molecules being aligned along the
vector of the force (shear thinning) rather than in loose coils as Polycondensation reactions are also a source of commercially
they tend to be in a stirred tank reactor. This has been postulated important polymers such as polyester (e.g. PET and PBT) and
by Jachuck25 and may well be another reason for the catalysts polyamides. The reaction producing polyesters is an equilibrium
having different activities in a stirred tank and a spinning disk reaction
reactor. It may also be one of the reasons for the improved
½Polyester½Water
efficiency of the modification of polymers which has been dis- K¼
½Alcohol½Acid
cussed in the section concerned with the hydrogenation of
polymers. and the yield and molar mass of the products depend on the
removal of water from the system. Typically, these reactions are
carried out in batch and the temperature slowly increased as the
Styrene
reaction proceeds in order to maintain a workable viscosity,
Polystyrene is considered here as an example of olefin addition to control side reactions and eventually to reduce the amount
polymerisation. Polystyrene and copolymers of styrene with e.g. of residual monomer to the required level. Reaction times of
acrylonitrile are important industrial polymers and are produced 15–20 h are not uncommon. Using a prepolymer from a batch
either in bulk, solution, suspension or emulsion. The advantage reactor Boodhoo and Jachuk52 could show that the spinning disk
of bulk systems is the purity of the final product, but such allows much higher reaction temperatures (180–300  C) and thus
systems need to be transferred from stirred tank reactors to considerably shorter reaction times once the acid value (the
extruder type equipment at higher conversion due to the viscosity amount of potassium hydroxide in milligrams that is needed to
of the system. Probably the earliest examination of the solution neutralize the acids in one gram of the reaction mixture which
free radical polymerisation of styrene on a spinning disk is reflects the residual acid moieties and thus the conversion) is
described by Boodhoo and Jachuck46 who also noted that under below 300. The very thin film on the SDR and the excellent heat
appropriate conditions the reaction rate is faster on a spinning transfer allow optimal removal of the water while reducing the
disk and that the MWD is narrower. In their study of the free occurrence of side reactions due to the short residence times. The
radical solution (toluene) polymerisation of styrene Vicevic authors52 emphasise that for the initial stages of the reaction
et al.,47 carrying on work done by Leveson et al.,48 prepared there is no advantage in using an SDR rather than a batch
prepolymer solutions in batch and transferred these solutions to reactor. In the early stages of the reaction the mass transfer in the
a spinning disk for further conversion. A technique also used by batch reactor is adequate to ensure acceptable reaction rates and
Moghbeli et al.49 for a bulk polymerisation of styrene. Vicevic such reactor types require less energy to run than an SDR. The
et al.47 modeled the system and showed that the rate of poly- authors mention earlier work suggesting that it may be that an
merisation on the spinning disk is some ten times that prevailing SDR could be used at even higher temperatures (up to 400  C) if
in the batch reactor and that the rate increases with the speed of the results from work on static thin films can be transferred to an
rotation of the disk (500–1500 rpm) by almost 30%. They also SDR. Such a development could mean that the SDR could be a
estimate that the energy of activation for propagation is some viable alternative for the complete polycondensation reaction.
10% smaller on the spinning disk. These authors suggest that the
smaller molar masses of material produced on the spinning disk
UV initiated polymerisation
can be attributed to greater radical concentration (reflected in
higher rates of radical generation while termination rates are UV radiation cured coatings are not something new; the prin-
similar or smaller on the disk to those measured in batch). This ciple has been used for centuries to prepare scratch resistant, high
explanation seems more plausible than the proposal that longer gloss wood varnishes with the help of sunlight. Nevertheless,
chains start to break up at higher rotational speeds and over the last almost 100 years the intentional use of specially
conversions given by Moghbeli et al.49 An interesting paper generated UV light has become an important branch of the
describing the cationic polymerisation of styrene on a spinning commercial world of polymers, for inks, coatings (for both wood
disk50 utilizes the technique of fixing catalysts onto the surface of and plastics) and adhesives. The use of lasers has further
the disk developed in their earlier work on the rearrangement of enhanced this field, due to an improved ability to focus the light
a-pinene oxide.51 Their work, including a comparison with a and to generate well defined energy levels of very specific wave-
stirred tank type reactor, suggests that the SDR is capable of lengths.53,54 There are several advantages of this type of initiation
considerably increasing the temperature at which cationic poly- for polymerisation, in particular, the negligible temperature
merisations can be carried out while maintaining useful molar effect and the facile control via the intensity of the light source.
masses and their distribution. Thus, using a fixed BF3 catalyst the However, although some attempts have been made to develop
authors show that it is possible on an SDR at 40  C to produce batch reactors for polymerisation using UV initiation tech-
polystyrenes with molar masses similar to those produced in a niques55 the scale up of such equipment is difficult, mixing and
batch reactor at 60  C. and with much higher concentrations of heat removal causing the major problems to be solved, so that

This journal is ª The Royal Society of Chemistry 2012 Polym. Chem., 2012, 3, 2698–2707 | 2705
 View Article Online

such equipment has not been developed to an industrial scale and 9 K. C. D. Hickman, US Pat. 2199994, Distillation Products, Inc., New
the commercial use of UV as a source of initiation has remained York, 1940.
10 K. C. D. Hickman, US Pat. 2343667, Distillation Products, Inc., New
confined to thin films.56 Recently, the development of flow York, 1944.
reactors wrapped around a lamp may make this type of process 11 A. A. Lapkin and P. K. Plucinski, in, Chemical Reactions and
industrially viable.57 Processes under Flow Conditions RSC, ed. S. V. Luis and E. Garcia-
The advantages of a spinning disk for carrying out energy Verdugo, Cambridge, UK, 2010.
Published on 03 July 2012. Downloaded by State University of New York at Stony Brook on 30/10/2014 21:16:45.

12 K. V. K. Boodhoo, PhD thesis, U. of Newcastle upon Tyne, 1999.

initiated polymerisations are obvious from the prior discussion 13 J. C. Johnstone, PhD thesis, U of Newcastle upon Tyne, 2000.
and have been studied by Boodhoo et al.58 These authors 14 Z. Cai, PhD thesis, Technische Universit€at M€
unchen, 2008.
describe the influence of light intensity and the duration for the 15 N. C. Jacobsen, PhD thesis, Technische Universit€at M€unchen, 2011.
16 M. Meeuwse, J. van der Schaaf and J. C. Schouten, AIChE J., 2012,
bulk polymerisation of butyl acrylate. This system is also the 58, 247.
basis for a very detailed study as part of Johnstones’ PhD thesis13 17 A. I. Stankiewicz and J. A. Moulijn, Chem. Eng. Prog., 2000, 96, 22.
which also demonstrates that it is necessary to optimize the 18 C. O’Driscoll, Chem. World, 2004, 21.
photoinitiator used as well as the rate of material supply to the 19 C. Ramshaw, Chem. Eng., 1985, 416, 30.
20 K. V. K. Boodhoo and R. J. Jachuck, Appl. Therm. Eng., 2000, 20,
disk and the temperature. 1127.
In Cai’s thesis14 the UV-initiated emulsion polymerisation of 21 A. Aoune and C. Ramshaw, Int. J. Heat Mass Transfer, 1999, 42,
butyl acrylate is described. His work shows that the molar mass 2543.
22 S. Yanniotis and D. Kolokotsa, Int. Commun. Heat Mass Transfer,
of the products produced during a single pass on a 20 cm
1996, 23, 721.
diameter disk and the conversion were similar to that which can 23 S. Yanniotis and D. Kolokotsa, J. Food Process Eng., 1996, 30, 313.
be achieved in bulk55,56 or in batch:59 2.7  105 and 80% 24 R. J. Jachuck, Chem. Eng. Res. Des., 2002, 80(A), 233.
respectively. In this work,14 a graft copolymer of acrylic acid and 25 R. J. J. Jachuck, WO Pat. 200300846, Protensive Limited, England,
2003.
vinyl-oxybutyl-polyethylene glycol (molar mass 1100 Daltons) is 26 C. H. Hornung, C. Guerrero-Sanchez, M. Brasholz, S. Saubern,
described, prepared using an SDR and UV initiation. After two J. Ciefari, G. Moad, E. Rizzardo and S. H. Thang, Org. Process
passes on the disk a conversion of the acrylic acid of ca. 70% Res. Dev., 2011, 15, 593.
could be achieved. The molar mass of the acrylic acid chains 27 C. Ramshaw and R. J. J. Jachuck, US Pat. 6972113, Protensive Ltd,
Cambridge, UK, 2005.
(Mw) was ca. 6  104 Daltons and the polydispersity was 1.8. An 28 A Chemical Week Associates Custom Publication, 28 May 2003.
improvement in yield or an increase in the molar mass could not 29 N. Anantachoke, M. Makha, C. L. Raston, V. Reutrakul,
be achieved using alternative dosage strategies. N. G. Smith and M. Saunders, J. Am. Chem. Soc., 2006, 128, 1384.
30 R. J. J. Jachuck and S. Cook, GB Pat. 2377661, Protensive Ltd.,
Cambridge, UK, 2003.
Outlook 31 A. E. Nielson, Acta Chem. Scand., 1961, 15, 441.
32 G. A. Ozin and A. C. Arsenault, Nanochemistry: a Chemical Approach
It has been demonstrated that the SDR is a very flexible appa- to Nanomaterials, RSC Publishing, Cambridge, UK, 2006.
ratus which can be utilized to great advantage for a variety of 33 J. R. Larry, R. M. Rosenberg and R. G. Uhler, IEEE Trans. Compon.,
Hybrids, Manuf. Technol., 1980, 3, 211.
reactions of interest to the polymer chemist. It is believed that 34 G. Fisher, Ceram. Ind., 1983, 120, 80.
much work is being done in industry today which has not yet 35 F. Fievet, J. P. Lagier and M. Figlarz, MRS Bull., 1989, 14, 29.
become public knowledge and it will be interesting to see whether 36 B. Wiley, Y. Sun and Y. Xia, Acc. Chem. Res., 2007, 40, 1067.
large scale production using SDR technology will become 37 A. R. Tao, S. Habas and P. Yang, Small, 2008, 4, 310.
38 K. S. Iyer, C. L. Raston and M. Saunders, Lab Chip, 2007, 7, 1800.
commercially important in the not too distant future. 39 J. Zou, K. S. Iyer, S. G. Stewart and C. L. Raston, New J. Chem.,
In the meantime there is much to be examined and this area of 2011, 35, 854.
research should prove a fruitful one for academia. In particular, 40 F. Ozawa, A. Kubo and T. Ahayashi, Chem. Lett., 1992, 2177.
41 C. M. Ong, S. Pask, S. Guo and F. Mersmann, US Pat. 7666950,
it is recommended that SDR could be combined with some of the Lanxess Deutschland GmbH, 2010.
newer green technologies (such as ionic liquids60–62) to combine 42 C. M. Ong, S. X. Guo and F. Guerin, US Pat. 7381781, Lanxess Inc.,
the advantages. The facile scale-up of an SDR makes the value of Sarnia, Ont. Canada, 2008.
such academic research greater than the more traditional 43 L. Reif and S. Pask, EP Pat. 1801134, Lanxess Deutschland GmbH,
2007.
approaches. 44 J. Leyland, GB Pat. 1080863, Unilever Ltd, England, 1967.
45 M. Vicevic, R. J. J. Jachuck, K. Scott, J. H. Clark and K. Wilson,
Notes and references Green Chem., 2004, 6, 533.
46 K. V. K. Boodhoo and R. J. Jachuck, Appl. Therm. Eng., 2000, 20,
1 M. M. Green and H. A. Wittcoff, Organic Chemistry Principles and 1127.
Industrial Practice, Wiley-VCH, Weinheim, 2003, pp. 249–292. 47 M. Vicevic, K. Novakovic, K. V. K. Boodhoo and A. J. Morris,
2 D. J. Artrip, Safety in the Use of Copper-containing Catalysts in Conference on Process Intensification and Process Innovation,
Ethylene Plants, AIChE 2006 Spring National Meeting Ethylene Christchurch, New Zealand, September 2006, pp. 24–29.
Producers Conference, Safety Session, Orlando, Florida, 2006. 48 P. Leveson, W. A. E. Dunk and R. J. Jachuck, J. Appl. Polym. Sci.,
3 M. R. Simmons, Twilight in the Desert, J. Wiley & Sons, Inc, 2003, 90, 693.
Hoboken, NJ, 2005. 49 M. R. Moghbeli, S. Mohammadi and S. M. Alavi, J. Appl. Polym.
4 Cleaner Technology in the UK, DTI Report, HMSO, London, 1991, Sci., 2009, 113, 709.
ISBN 0115152938. 50 K. V. K. Boodhoo, W. A. E. Dunk, M. Vicevic, R. J. Jachuck,
5 C. Fell, CEA, Chem. Eng. Aust., 1987, 12, 25. V. Sage, D. J. Macquarrie and J. H. Clark, J. Appl. Polym. Sci.,
6 C. Ramshaw, Chem. Eng., 1983, 389, 13. 2006, 101, 8.
7 C. Ramshaw, The Incentive for Process Intensification, in 51 K. V. K. Boodhoo and R. J. Jachuck, Green Chem., 2004, 6, 533.
Proceedings, 1st Intl. Conf. Proc. Intensif. For Chem. Ind., 18, BHR 52 K. V. K. Boodhoo and R. J. Jachuck, Green Chem., 2000, 2, 235.
Group, London, 1995, pp. 167–171. 53 K. V. K. Boodhoo, W. A. E. Dunk and R. J. Jachuk, Photoinitiated
8 K. C. D. Hickman, US Pat. 2180053, Distillation Products, Inc., New Polymerisation, ed. K. D. Belfield and J. V. Crivello, ACS Symp.
York, 1939. Ser., 2003, vol. 847, p. 437.

2706 | Polym. Chem., 2012, 3, 2698–2707 This journal is ª The Royal Society of Chemistry 2012
 View Article Online

54 C. Decker, Laser Curing of Arylic Coatings, Proceedings of a 58 K. V. K. Boodhoo, W. A. E. Dunk. and M. S. Jassim, J. Appl. Polym.
symposium of N.W. Region, Ind. Div., Royal Society of Chemistry, Sci., 2004, 91, 2079.
vol. 16, 1986. 59 I. Capek, Macromol. Rep., 1996, A33, 209.
55 C. F. Ryan and J. J. Gormly, Macromol. Synth., 1963, 60 C. Guerrero-Sanchez, R. Hoogenboom and U. S. Schubert, Chem.
1, 30. Commun., 2006, 3797.
56 C. Decker, T. Bendaikha, D. Decker and K. Zahouily, Polym. Prepr. 61 C. Guerrero-Sanchez, M. Lobert, R. Hoogenboom and
(Am. Chem. Soc., Div. Polym. Chem.), 1997, 38(1), 487. U. S. Schubert, Macromol. Rapid Commun., 2007, 28, 456.
Published on 03 July 2012. Downloaded by State University of New York at Stony Brook on 30/10/2014 21:16:45.

57 F. Levesque and P. H. Seeberger, Angew. Chem., Int. Ed., 2012, 51, 62 T. Erdmenger, C. Guerrero-Sanchez, J. Vitz, R. Hoogenboom and
1706. U. S. Schubert, Chem. Soc. Rev., 2010, 39, 3317.

This journal is ª The Royal Society of Chemistry 2012 Polym. Chem., 2012, 3, 2698–2707 | 2707