Organic Process Research & Development
ARTICLE
Table 1. Ozonolysis flow setup operating parameters
mixture was ethyl acetate/nitrogen. A syringe pump (Cole Parmer
74900À35) was used to deliver the liquid, while the gas was in-
troduced by a mass flow controller (Brooks 5850). Gas and liquid
were brought together by a PEEK Y-mixer with 1 mm ID. For
residence time determination, coils of 2, 5, and 10 mL internal
volume, as employed in the ozonolysis experiments, were used.
Ethyl acetate containing a dye (Parker blue) was used in tracer
step-input experiments. Signals from the reactor inlet and outlet
were detected by light absorption. The detection unit consisted
of a linear diode array detector (TSL, 1401R-LF) which had 128
diodes, each of dimensions 63.5 μm by 55.5 μm. Illumination was
provided by two square LEDs (Kingbright L-1553IDT). The line
of 128 diodes in each sensor was positioned perpendicular to the
flow direction so that only a certain number of diodes were under
the tube. Using a Labview program, every 100 ms readings from
the relevant diodes on each sensor were collected, averaged, and
converted to a single absorbance measurement.
parameter
range
reactor volume
reactor internal diameter
gas flow rate
0.07, 2, 5, 10 mL
1 mm
25À66 mL minÀ1
0.25À1.08 mL minÀ1
0.25À1.44 mL minÀ1
À10 °C
reactant flow rate
quench flow rate
temperature
pressure
1.1À1.3 bar (reactor inlet)
with Carus Carulite 200 (manganese dioxide/copper oxide). The
efficiency of the catalyst is checked on a regular basis by bubbling
the treated gases through 10% potassium iodide solution. The pres-
ence of ozone would be immediately detected by the formation
of dark brown iodine.
The concentration of ozone at the outlet of the ozone generator is
dependent on the electric power consumption, gas flow rate, and
temperature. The dielectric is plain ceramic, and the electrodes
are tungsten with a maximum operating pressure of 2 bar gauge.
The ozone concentration can reach ∼280 g/Nm3 (∼19 wt %),
depending on the power setting, with the maximum concentra-
tion being reached around setting 6 (scale 1À10); higher power
settings result in a lower yield of ozone. At flow rates below
25 mL minÀ1, the ozone concentration was found to be unaccep-
tably variable and therefore unusable. The ozone analyser mea-
sures the ozone content with a UV photometer at a wavelength of
254 nm, and the maximum operating pressure is 1 bar gauge.
The Vapourtec R4 cooled reactor module consists of three
input streams (one after the main reactor cartridge) and one output
stream (Figure 1b). All internal tubing is 1 mm ID PFA. The
module is cooled via a stream of dried air chilled over dry ice and
distributed internally via a fan. Temperature control is via a pinch
valve controlled by the system. The ozone-containing oxygen gas
stream, flowing through a nonreturn valve, is precooled on entry
to the cooled reactor module (Figure 1b) and then mixed via a
1 mm ID “T” with the precooled substrate solution. The gas/
liquid mixture then proceeds through a reactor cartridge. The
cartridge consists of a weaved coil of circular PFA tubing with
1 mm ID, which gives a reactor volume of about 8.1 μL/cm.
Volumes used were 0.07, 2, 5, and 10 mL. The cartridge ends at
another 1 mm ID “T” where the gas/liquid stream is mixed with
the precooled quench solution. A short, cooled tube of 500 mm
length (0.39 mL) follows, allowing dissipation of any exotherm. The
output gas/liquid mixture is separated by gravity in a PTFE-
capped 50 mL bottle, where the liquid is collected at the bottom
and the gas allowed to exit via a 2-mm ID tube to the four-way valve
system. Ozone is highly aggressive to most plastics and rubbers and
degrades these materials rapidly, while stainless steel (316), PTFE
and PFA are ozone resistant. Therefore, all parts of the experimental
setup that are in contact with either fluids or gas consist of either
PTFA, PFA, or stainless steel (316). The connection tubes used
in the setup are either 1/8 in. (3.18 mm) or 1/16 in. (1.15 mm)
ID. The range of operating conditions used are shown in Table 1.
Hydrodynamics Characterisation. In order to characterise
the gas/liquid behaviour, flow patterns, film thickness, and residence
time were investigated. The flow patterns and the liquid film
thickness of the gas/liquid mixture were analysed using a Phantom
Miro 4 high-speed camera; the images were acquired at 200 pps.
StraightPFAtubingwith1mmIDand10cmlengthwasusedforthe
flow patterns and the liquid film thickness analysis. The gas/liquid
Analysis. Reaction analysis was carried out on a HP5890A
gas chromatograph equipped with a Supelco SPB 1701 column
(30 m  0.32 mm, 0.25 μm), FID detector (injection port
250 °C, detector 250 °C, Oven initial 100 °C, final 180 °C hold
for 2 min, 20 °C/min ramp); all product peaks (2 and 3) were
referenced with authentic samples obtained from Sigma-Aldrich;
spiked runs were also carried out. The GC relative peak areas were
checked and found to be 1:1:1 when a 1:1:1 molar solution of 1, 2,
and 3 was injected into the GC. No other 1-decene-derived
products apart from 2 and 3 were observed in the GC trace, and
thus, the total product peaks areas were assumed to equal unity.
NMR spectra were run on a Bruker 400 MHz spectrometer in
CDCl3 and referenced to the residual solvent peak.
2
Decene 1: Rf (GC) 3.6 min; Rf (SiO , 10% EtOAc in
cyclohexane) 0.75.
Nonanal 2: Rf (GC) 4.7 min; Rf (SiO , 10% EtOAc in
2
cyclohexane) 0.5; δH (400 MHz, CDCl3) 0.86 (3 H, t, J = 6.8,
CH3), 1.20À1.37 (10 H, m, 5 Â CH2), 1.62 (2 H, quin, J = 7.3
Hz, CH2), 2.40 (2 H, td, J = 7.3, 1.9 Hz, CH2) and 9.75 (1 H, t, J =
1.9 Hz, CHO).
H
Nonanoic Acid 3: Rf (GC) 6.8 min; δ (400 MHz, CDCl3)
0.88 (3 H, t, J = 6.8, CH3), 1.20À1.76 (12 H, m, 6 Â CH2), 2.33
(2 H, t, J = 7.5, CH2) and 11.00 (1 H, br s, COOH).
3-Octyl-1,2,4-trioxolane (decene ozonide) 4: δH (400 MHz,
CDCl3) 0.81 (3 H, t, J = 6.6, CH3), 1.10À1.30 (10 H, m, 5 Â
CH2), 1.32À1.41 (2 H, m, CH2), 1.61À1.68 (2H, m, CH2), 4.95
(1H, s, OCHAHBO), 5.05 (1H, t, J = 4.9, CH) and 5.11 (1 H, s,
OCHAHBO).
’ RESULTS
Hydrodynamics Characterization. Figure 2 shows the ob-
served flow patterns for different liquid (ethyl acetate) to gas
(nitrogen) flow rate ratio:(a) slug-annular (1:17, Vliquid = 1.5 mL
minÀ1, Vgas = 25 mL minÀ1), (b) wavy annular (1:33, Vliquid
=
0.75 mL minÀ1, Vgas = 25 mL minÀ1), (c) wavy annular (1:100,
Vliquid = 0.25 mL minÀ1, Vgas = 25 mL minÀ1), and annular flow
(1:400, Vliquid = 0.25 mL minÀ1, Vgas = 100 mL minÀ1). For the
liquid to gas flow rate ratio of (1:17) slug annular flow is observed
as an intermediate flow pattern between slug and annular flow. In
this regime the gas slugs are large and are separated by a thin
bridge (Figure 2a). Decreasing the liquid to gas flow rate ratio
leads to the disappearance of the bridge and the appearance of
waves. This constitutes the wavy annular flow regime (Figure 2b
and c) which persists up to liquid to gas flow rate ratios of 1:300.
991
dx.doi.org/10.1021/op200036d |Org. Process Res. Dev. 2011, 15, 989–996