An empirical study of the effect of the variables in a flash vacuum pyrolysis (FVP) experiment

Article abstract of DOI:10.1039/b410786c

The effect of the variation of the experimental parameters on the conversion of precursor to products in a typical flash vacuum pyrolysis (FVP) experiment was investigated empirically. Temperature-conversion plots can be used to optimise FVP conditions and their mechanistic significance is exemplified. At a given temperature, the conversion can be increased by an increase in the background pressure, or by packing a section of the furnace tube with inert material (particularly when placed at the trap end of the furnace tube) or by employing a catalyst. Despite the prevailing view that only intramolecular reactions take place by FVP, it has been shown by a 'dual-FVP' cross-over experiment that the dimerisation of benzyl radicals occurs in the gas-phase, before the cold trap, under standard conditions. However, reduction in through-put rate, increase in furnace temperature and reduction in background pressure all reduce the amount of gas-phase coupling.

Full text of DOI:10.1039/b410786c

View Article Online / Journal Homepage / Table of Contents for this issue
Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
A R T I C L E
An empirical study of the effect of the variables in a flash vacuum
pyrolysis (FVP) experiment
Emma F. Duffy, Jonathan S. Foot, Hamish McNab* and Andrew A. Milligan
School of Chemistry, The University of Edinburgh, West Mains Road, Edinburgh, UK EH9 3JJ
Received 14th July 2004, Accepted 23rd July 2004
First published as an Advance Article on the web 24th August 2004
The effect of the variation of the experimental parameters on the conversion of precursor to products in a typical flash
vacuum pyrolysis (FVP) experiment was investigated empirically. Temperature-conversion plots can be used to optimise
FVP conditions and their mechanistic significance is exemplified. At a given temperature, the conversion can be increased
by an increase in the background pressure, or by packing a section of the furnace tube with inert material (particularly
when placed at the trap end of the furnace tube) or by employing a catalyst. Despite the prevailing view that only
intramolecular reactions take place by FVP, it has been shown by a ‘dual-FVP’ cross-over experiment that the dimerisation
of benzyl radicals occurs in the gas-phase, before the cold trap, under standard conditions. However, reduction in through-
put rate, increase in furnace temperature and reduction in background pressure all reduce the amount of gas-phase
coupling.
The technique of flash vacuum pyrolysis (FVP) experimentally
involves the vacuum sublimation or distillation of a substrate
through a hot tube (generally at temperatures between 300–1000 °C)
to induce chemical change. After passing through the tube, the
products are quenched at low temperatures in a trap placed at the
exit point of the furnace tube.1,2 The pyrolysis is usually considered
to take place under dilute, short contact time conditions in the gas
phase such that individual molecules react intramolecularly in
the effective absence of other molecules of substrate, products or
substantial amounts of oxygen. The technique therefore produces
much cleaner results than other forms of pyrolysis. FVP is a robust,
Fig. 1 FVP apparatus.
highly reproducible method and has found widespread use in
applications ranging from matrix isolation and mechanistic studies
to preparative organic chemistry.3
1. Effect of furnace temperature: temperature–
conversion plots
As is the case with any method which depends on gas flow,
the residence time of substrate molecules in the hot (reacting)
zone is a key parameter,1 and this in turn depends on a number of
variables including the pressure and temperature, the dimensions
of the reactor tube, throughput rate and the presence of (inert)
packing material (if any).l However, in any one laboratory where
the furnace dimensions, temperature gradient and packing are
fixed and the vacuum system delivers a fairly constant pressure,
the main variables are likely to be throughput rate and furnace
temperature. It should be noted that atmospheric pressure flow
pyrolysis (carried out by injecting the substrate into a gas stream
at the top of a vertical heated tube), has many more variables
in practice, including choice of solvent (if any) and carrier gas,
nature of preheater (if any), heat exchange medium and mode of
product trapping.
Although FVP experiments are highly reproducible using one
set of apparatus (see below), the variables are often not specified
in literature reports of FVP applications, so it can be difficult to
reproduce such conditions in another laboratory without carrying
out a series of trial experiments. In the work described here, we have
therefore systematically varied the FVP parameters and monitored
the effect on conversion to products to provide empirical guidelines
on their relative importance and on the design of the preparative
FVP experiment. In all the experiments, the substrate was distilled
through a horizontal silica tube (overall dimensions 35 × 2.5 cm)
heated by an electrical tube furnace. The temperature was mea-
sured at the centre of the furnace; it is relatively constant for much
of the central region of the tube but becomes cooler towards the
ends, in particular the open (trap) end. The product trap (U-tube)
was situated at the exit point of the furnace and the pressure was
recorded between the product trap and a high capacity rotary oil
pump (Fig. 1).
Furnace temperature is intuitively the most important variable in
an FVP experiment. As a model reaction, the temperature depen-
dence of the cyclisation of the malonate 1 to the 3-oxopyrrolizine
carboxylic ester 2 (Scheme 1) was studied.4 This reaction was
chosen because it takes place over a typical FVP temperature range
[100% conversion at 650 °C (0.01 mbar) using an empty furnace
tube] and its sensitivity to experimental variables is readily moni-
tored because a temperature range of some 150 °C is required to
increase the conversion from 10% to 90%. In addition, there are
no significant side-reactions. The variation of substrate conver-
sion with temperature (Table 1, entries 1–5) follows a sigmoid
curve (Fig. 2) reflecting the Boltzmann distribution of molecules
in the tube. Thus at low temperatures only a few molecules possess
enough energy to react. At higher temperatures the steep region of
the curve reflects the greater number of molecules in the central
region of the Boltzmann distribution. At the highest temperatures
the curve again flattens out since only a few molecules do not
possess enough energy to react.
Scheme 1
T h i s j o u r n a l i s
©
T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 4
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2 , 2 6 7 7 – 2 6 8 3
2 6 7 7

View Article Online
Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Table 1 FVP of 1 to give 2 under a variety of FVP conditions
Entry
Quantity/mg
Ti/°Ca
Tf/°Cb
P/mbarc
Time/mind
Packing
% Product
Notes
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
30
32
31
30
32
30
25
30
30
30
25
29
36
30
32
31
31
32
26
28
26
26
28
26
32
31
32
29
120
120
120
120
120
120
120
120
80
120
200
120
120
120
120
120
120
120
120
120
120
120
120
120
120
120
120
120
450
500
550
600
650
500
500
500
500
500
500
400
450
475
500
450
450
450
450
450
450
450
450
450
450
300
350
450
0.01
0.01
0.01
0.01
0.01
0.01
0.1
5
5
5
5
5
5
6
15
4
5
5
7
5
5
7
5
5
5
4
5
15
5
4
5
3
5
5
none
none
none
none
none
none
none
none
none
none
none
SiO2e
SiO2e
SiO2e
SiO2e
SiO2e
SiO2e
SiO2f
SiO2f
SiO2f
SiO2f
porg
9.3
44
72
95
100
44
90
100
41
43
32
0
9.5
78
100
15
5
1
2
1
1
0.005
0.005
0.005
0.005
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.005
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
3
3
3
3
4
5
3
4
5
6
3
4
5
31
80
7
99
39
46
23
79
48
79
100
porg
porg
porg
SiO2h
SiO2h
SiO2h
4
4
4
5
a Inlet temperature. b Furnace temperature. c Background pressure (measured after the product trap). d Time required for complete distillation of the substrate
e
f
through the furnace (unless otherwise stated). A stack of eight SiO2 tubes (5 × 0.6 cm) was placed in the furnace. Sections of the furnace were packed
throughout the tube diameter with portions of silica wool (typically 5 cm length within the furnace tube, ca. 1.3 g silica wool). For entry 21, the whole length
of the tube was packed with silica wool. g Sections of the furnace were packed throughout the tube diameter with porcelain chips [typically 20 g, held in place
by a small amount of silica wool (3–4 cm)]. For entry 25, the whole length of the tube was packed with porcelain chips. h Eight silica tubes (5 × 0.6 cm) were
coated with TLC grade silica gel, as follows.14 Methanol (40 cm3) was added to silica gel (20 g) to produce a thick slurry. The silica tubes were dipped in the
slurry until thoroughly coated. Excess coating was removed from the outside of the tubes, which were then dried at 120 °C for 4 h before being stacked in the
furnace (trap end).
Notes: 1. Repeat of entry 2. 2. Only small amount of pyrolysate recovered from product trap. 3. Packing located in the centre of the furnace. 4. Packing located
at the trap end of the furnace. 5. Packing located at the inlet end of the furnace. 6. The whole length of the tube was packed with silica wool.
a 1,5-sigmatropic hydrogen atom shift takes place as a key step
at an intermediate stage which is not observed in radical pro-
cesses (Scheme 2).10 However, there is no substantive evidence
to support the radical cyclisation mechanism for the hydrazone 4,
since in substituted examples, the spirodienyl radical 6, if formed,
is transformed to products exclusively by C–N bond migration
(Scheme 2).9 Because a number of products are formed in these
reactions, the amount of remaining starting materials 3, 4 and 5 at
various furnace temperatures is shown on the graph (Fig. 3) and in
Table 2. As can be seen from the data, conversion of the hydrazones
3 and 4 to products begins below 500 °C and both show similar,
steep, temperature dependence. In contrast, the vinamidine 5 is
stable at above 600 °C and a greater temperature range is required
for complete conversion to products. Hence the similarity in mecha-
nism for the reactions of 3 and 4 is confirmed, which establishes that
the cyclisation of 4 almost certainly proceeds via an iminyl radical
intermediate rather than by a concerted mechanism.
Fig. 2 Temperature–conversion plot for 1→2 using an empty furnace
tube and using a furnace tube containing a central packing of silica tubes.
The results shown in Table 1, entries 2, 6 and 10, demonstrate the
reproducibility of a typical FVP experiment (44 ± 1% conversion),
even at a point in the sigmoid curve where there is a relatively steep
temperature dependence.
2. Effect of pressure variation
Temperature–conversion plots have been used in the past to
identify the optimum pyrolysis temperature5,6 and to monitor
the progress of consecutive reactions.7 However, the position
and shape of the sigmoid curve with respect to the temperature
axis is related in principle to the activation energy of the process
and therefore contains additional mechanistic information.
As an illustration of how this can be used in practice, we have
reinvestigated the pyrolyses of the hydrazones 38 and 49 and the
related vinamidine 5.10 Hydrazones such as 3 are known to cyclise
by a free radical mechanism8 because of the observed formation
of rearranged products via a spirodienyl intermediate (Scheme 2).
In contrast, vinamidines such as 5 are known to cyclise by a
concerted electrocyclisation-with-elimination process in which
The background pressure in the system was easily varied by 2
orders of magnitude by introduction of nitrogen gas via a needle
valve located between the product trap and the pumps. An increase
in pressure is expected to increase the contact time because of a
reduced pressure gradient between inlet and trap. Indeed, at 500 °C
the 45% conversion of 1 to 2 at 10−2 mbar is increased to 90% at
10−1 mbar and is almost quantitative at 1 mbar (Table 1, entries
6–8). These results show that conversion is strongly dependent on
the pressure in the system. It is therefore important that the pres-
sure is kept constant for each run used for the construction of tem-
perature–conversion plots. In practice this is not usually a problem
since, in the absence of leaks in the vacuum line, the pump usually
delivers a steady pressure. However, if a gas is generated in the FVP
2 6 7 8
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2 , 2 6 7 7 – 2 6 8 3

View Article Online
Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Table 2 FVP of 3–5 at various temperatures
Precursor
Quantity/mg
Ti/°Ca
Tf/°Cb
P/mbarc
Time/mind
% precursore
3
3
3
3
3
3
4
4
4
4
5
5
5
5
5
25
25
25
25
25
25
35
35
35
35
25
25
25
25
25
120
120
120
120
120
120
150
150
150
150
100
100
100
100
100
475
500
550
575
600
650
450
500
550
600
650
700
750
800
850
0.007
0.007
0.007
0.007
0.007
0.007
0.01
0.01
0.01
0.01
0.014
0.014
0.014
0.014
0.014
35
35
35
35
35
35
25
25
25
25
20
20
20
20
20
78
64
17
6
5
5
77
16
0
0
89
60
30
9
3
a Inlet temperature. b Furnace temperature. c Background pressure (measured after the product trap). d Time required for complete volatilisation of the substrate
through the furnace (unless otherwise stated). e See Experimental section for explanation.
Fig. 3 Temperature–conversion plot for the pyrolyses of 3–5, showing
remaining starting materials.
by collisions. Under these conditions of ‘chemical activation’,11
the excess energy may be dissipated by collision i.e. subsequent
secondary reactions may be minimised by performing the reaction
in the presence of a carrier gas.
FVP reactions can be carried out at a lower pressure, e.g. by
use of an oil diffusion pump in series with the rotary pump. In our
experience there is little advantage in using these conditions for
preparative reactions, except where the substrate is poorly volatile.
3. Effect of substrate throughput rate
The throughput rate is governed by the volatilisation temperature
of the substrate from the inlet. Variation of inlet temperature from
80–200 °C has relatively little effect on the conversion of the model
precursor 1 (Table 1, entries 9–11; 38 ± 6% at 500 °C). This para-
meter is therefore not especially critical (in the absence of evolved
gases—see above) and can be varied to give a convenient through-
put rate with relatively little effect on contact time and conversion to
products (though see Section 6 below). This result explains the high
reproducibility of FVP experiments; at constant pressure and with
an empty furnace tube, the furnace temperature becomes effectively
the only experimental variable. On the negative side, if an FVP
reaction cannot be optimised by control of furnace temperature then
there is little scope for further optimisation unless furnace packings
or catalysts (see below) can be introduced to the hot tube.
Scheme 2
4. Effect of furnace packing materials
reaction, which contributes to the background pressure (e.g. N2 or
CO), substantially increased contact times should be anticipated.
This is likely to be particularly important when a reaction is scaled
up, since the inevitable increase in throughput rate will contribute to
an even greater increase in background pressure (see below).
It should be noted that in some highly exothermic reactions
such as those involving carbenes or nitrenes, intermediates may be
generated in vibrationally excited states which may be deactivated
It is well known that packing a section of the furnace tube with an
‘inert’ material (such as silica chips) can result in a reduction of
furnace temperature required for a conversion by ca. 50–100 °C.12
However, as pointed out by Brown12 it is not clear whether this is
due to contact time effects or to some specific surface effect of the
so-called ‘inert’ packing. We have therefore studied the effects of
various packing materials located at different parts of the furnace
tube. Thus, if some specific surface phenomenon is involved, the
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2 , 2 6 7 7 – 2 6 8 3
2 6 7 9

View Article Online
Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
effect should be greatest when the packing is located at the centre
of the tube—i.e. where the temperature is greatest. Three common
packing materials were studied for their effect on the 1→2 conver-
sion, viz. silica tubes, silica wool and porcelain chips and in all
cases the whole section of the tube was packed.
22–25; 450 °C pyrolysis temperature; 7–9% conversion in the
absence of packing). Again the effect was greatest when the pack-
ing was at the trap end of the furnace tube (46% conversion) but
there was rather less distinction compared with the run in which the
packing was placed in the centre of the tube (39% conversion). In
this case packing at inlet end of furnace tube increased the conver-
sion significantly (to 23%) which suggests that the porcelain may
be a more efficient ‘pre-heater’ than the silica used in the previous
experiments.
These results show that relatively short sections of inert pack-
ing material within the furnace tube of an FVP experiment can
dramatically influence the conversion to products, particularly
when the packing is located away from the centre of the tube at the
trap end of the furnace. In the first place, the furnace temperature
can be reduced, but because the major effect seems to be an increase
in the contact time, this does not mean that ‘milder’ conditions are
experienced by substrate molecules. Perhaps more important, the
reduction in the temperature range required for complete conver-
sion to products may give the opportunity for much finer control,
particularly if two or more successive thermal processes can take
place.
As shown in Fig. 2, (and Table 1, entries 12–17) the major effect
of packing the centre of the furnace tube longitudinally with small
silica tubes (stack of 8 tubes, each 5 × 0.6 cm) is to narrow the
temperature range of the conversion. Thus a temperature increase
of only ca. 50 °C is required to increase the conversion from 10%
to >90% (c.f. the 150 °C range required in the absence of the pack-
ing). This implies that the packing has the effect of equalising the
energies of the precursor molecules (i.e. the half-height width of the
Boltzmann distribution is reduced). This may be due to disruption
of the flow in the centre of the tube so that all the molecules experi-
ence a similar (increased) number of surface collisions. Although
the temperature of onset of reaction is not significantly affected by
this packing, complete conversion to products is observed at 500 °C
rather than at 650 °C in the absence of the packing. These results
suggest that such packing might be used to minimise the effect of
secondary, subsequent pyrolysis processes on an initial reaction.
At low conversions, the position of the packing in the tube has
relatively little influence (Table 1 entries 13, 16 and 17), though
it appears that the major effect is not found when the packing is
located at the centre of the furnace tube.
This trend was even more apparent when silica wool (1.3 g,
spread along a ca. 5 cm length of tube) was used as the packing
material (Table 1, entries 18–21 and Fig. 4). For these experiments,
full temperature–conversion plots were not carried out, but the
effects of various packing configurations on the conversions at
450 °C were monitored. The silica wool has a much larger surface
area than the tubes, but if a surface reaction is involved, the effect
would be expected to be greatest where the surface is hottest—i.e.
when the wool is placed in the centre of the tube. In the event, the
conversion from 1 to 2 at 450 °C was increased from 7–9% (in the
absence of packing, or with packing at the inlet end of the furnace
tube) to 31% (with packing in the centre of the tube) but was further
increased to 80% when the packing was placed at the trap end of
the furnace tube. It is clear from these results that (i) silica wool has
a greater effect on the conversion than silica tubes; (ii) an increase
in the number of molecular collisions at the inlet end of the tube
(i.e. in a pre-heating zone) has little influence on the conversion;
(iii) the major effect is seen when the packing is at the trap end
and not when it is in the centre of the furnace tube. This suggests
that there is no ‘surface’phenomenon involved with ‘inert’packing
materials such as silica, and instead the major influence must be a
reduction in flow rate causing an increase in contact time. The effect
of this increase in contact time is greatest when the molecules are
forced to spend this period in the hottest zone of the furnace.
5. Effect of ‘catalysts’ in the furnace
Surprisingly little systematic work has been carried out on the use
of solid-phase catalysts in the short contact time gas-phase FVP
experiment,13 though solid-phase reagents have been employed
in the VGSR (Vacuum Gas-Solid Reaction) technique.2 In one
interesting application, Storr and co-workers used silica tubes
coated with TLC-grade silica-gel, and observed substantial
reduction in furnace temperatures for dehydration and related
reactions.14 The model reaction 1→2 involves consecutive elimina-
tion, (thought to be via an intramolecular concerted process) and
electrocyclisation and it is not clear that surface catalysis should
accelerate either step of this process. In the event, a very significant
effect was observed when Storr’s conditions were used [Table 1,
entries 26–28, (silica tubes at trap end of furnace) and Fig. 4]
such that the temperature for complete conversion was reduced
to 400 °C, some 250 °C lower than with an empty tube. In this set
of experiments, the steep temperature dependence found for silica
tubes centrally located in the furnace was not found (Fig. 2), but it
is anticipated that a greater effect may be observed if the ‘catalyst’
is placed in the hottest (centre) zone of the furnace tube. It is clear
that a more systematic study of catalytic conditions may provide an
added dimension to the FVP experiment and significantly increase
the utility of the method in synthesis. Further examples will be
reported in future publications.
6. Intermolecular reactions—the coupling of
benzyl radicals
FVP reactions are dominated by intramolecular processes. Of
the few intermolecular reactions which are preparatively useful,
most are dimerisations of reactive species such as carbenes or
radicals. Yet not all radicals dimerise. Phenyl radicals generally
undergo hydrogen atom capture whereas benzyl radicals couple
to give excellent yields of bibenzyls.15 It was therefore of interest
to understand the mechanism of such dimerisations more precisely
and in particular to define the region(s) of furnace and trap where
the actual coupling reaction takes place. In very early work, Hedaya
and McNeil have shown that the level of the cross-coupling reaction
of cyclopentadienyl and allyl radicals could be reduced by locating
the trap at a greater distance from the furnace exit, which suggests
that—in their system—the coupling reaction takes place in the cold
trap under standard conditions.16 In addition, benzyl radicals (gene-
rated by FVP of benzyl bromide) can survive as far as a neon matrix
where they can be detected by EPR spectroscopy.17 However, the
apparatus used for these experiments may be very different from
that in use today for ‘preparative’ FVP experiments.
Fig. 4 Temperature–conversion plot for 1→2 using an empty furnace
tube, using silica wool packing at various points in the furnace tube and
using a furnace tube containing a packing of silica gel ‘catalyst’ (trap end).
In one experiment in which the whole furnace tube was packed
with silica wool (Table 1, entry 21), the packing proved to be too
dense for effective gaseous flow through the furnace and only a
small amount of product (with high conversion) was obtained.
Broadly similar results were obtained when porcelain chips
(20 g, filling a ca. 5 cm segment of the furnace tube, held in place
with silica wool) were used as the packing material (Table 1, entries
The dimerisation of benzyl radicals 8, generated by pyrolysis of
oxalate esters18 (e.g. 7) (Scheme 3) to give bibenzyl 9 was therefore
2 6 8 0
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2 , 2 6 7 7 – 2 6 8 3