See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/7448186

High-pressure reactivity of propene

Article in The Journal of Chemical Physics · December 2005

DOI: 10.1063/1.2109947 · Source: PubMed

CITATIONS READS
27 5,523

4 authors, including:

Margherita Citroni Vincenzo Schettino

University of Florence University of Florence
53 PUBLICATIONS 1,062 CITATIONS 214 PUBLICATIONS 6,037 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Vincenzo Schettino on 01 June 2014.

The user has requested enhancement of the downloaded file.

 THE JOURNAL OF CHEMICAL PHYSICS 123, 194510 共2005兲

High-pressure reactivity of propene

Margherita Citroni, Matteo Ceppatelli, Roberto Bini,a兲 and Vincenzo Schettino
European Laboratory for Non-linear Spectroscopy (LENS) and Instituto Nazionale per la Fisica
della Materia (INFM), Via Carrara 1, I-50019 Sesto Fiorentino, Firenze, Italy and Dipartimento
di Chimica dell’Università di Firenze, Via della Lastruccia 3, I-50019 Sesto Fiorentino, Firenze, Italy
共Received 12 July 2005; accepted 9 September 2005; published online 16 November 2005兲

The phase diagram of propene has been investigated at high pressure by using the diamond anvil
cell technique and Fourier transform infrared spectroscopy. The pressure conditions necessary to
induce a spontaneous reaction of the sample have been found at different temperatures, allowing the
stability boundary of propene to be drawn. The reaction is diffusion controlled and seems to occur
only in the fluid phase, implying a slope inversion of the stability boundary at about 250 K. The
product of the reaction is a mixture of linear oligomers independently of the P-T conditions. The
activation volume and energy of the process have been obtained from the kinetic data. Also the
activation of the reaction by laser absorption has been carefully studied. A high proton mobility has
been identified as the likely reason that limits the lengthening of the chain up to six to eight
monomeric units preventing the polymer formation. © 2005 American Institute of Physics.
关DOI: 10.1063/1.2109947兴

I. INTRODUCTION high-pressure laser-assisted reaction the quantitative and se-

Among the thermodynamic variables pressure is the
most effective in increasing the density of matter. From a
microscopic point of view this corresponds to strengthen the
interactions among the atoms that can originate the structural
and electronic changes. This process can be pushed until the
redistribution of the electronic densities gives rise to a trans-
formation that can also produce new materials recoverable at
ambient conditions. Such effects are extremely common in
molecular systems which are highly compressible and the
electronic rearrangements can result in the reversible or irre-
versible breaking and reforming of chemical bonds. The
most spectacular examples of reversible transformations are
those concerning the insulator-to-metal transitions 关as in the
case of I2 共Refs. 1 and 2兲 and O2 共Ref. 3兲兴, and the
molecular-nonmolecular transformations observed in N2
共Refs. 4–6兲 and CO2.7,8 While these structural changes re-
quire extreme pressure and temperature conditions, it is pos-
sible to induce irreversible transformations with milder con-
ditions in simple unsaturated hydrocarbons.9,10 From a
chemical point of view, pressure not only acts as a regulatory
tool for the chemical equilibria and kinetics, but also triggers
the reaction mechanisms that are uniquely related to high-
pressure conditions. Because of the geometrical constraints
and of the changes in the electronic structure new reaction
pathways can be observed, and a higher selectivity intro-
duced, comparatively to those occurring at normal densities.
A very important role in high-pressure reactions can be
played by photochemical effects.10 The general effect played
by the optical catalysis is to lower the threshold pressure and
to increase the extent of the reaction.11,12 Moreover, irradia-
tion may select a specific process when alternative reactions
are possible, as in the case of liquid butadiene where in the
a兲
Electronic mail: bini@chim.unifi.it
lective formation of a linear polymer is observed.13
Propene 共CH2 v CH – CH3兲 is the starting monomer for
the production of polypropene, one of the most largely pro-
duced polymers, and is an important component in many
copolymers used as fibers and plastics. The polymerization
of propene takes place through the rupture of the double
bond, while the position of the methyl group with respect to
the molecular backbone determines the stereochemistry of
the resulting polypropenes: all on the same side 共isotactic兲,
alternated 共syndiotactic兲, and randomly positioned 共atactic兲.
The different regiochemistries, tacticities and lengths of the
polymer chains, arising from different reaction and catalysis
conditions give rise to polymeric samples ranging from very
ordered to completely disordered having very different me-
chanical properties.14 The prospect of inducing with high
pressure a topochemical reaction producing a stereochemi-
cally ordered polypropene chain, as obtained in the case of
butadiene,13 or to an ordered crystalline structure, as success-
fully performed in the case of ethylene,15 led us to investi-
gate the pressure-induced reactivity of propene. On the other
hand, the possibility that, at the pressure necessary to induce
the reaction, the methyl group may constitute a too strong
hindrance to the polymerization in favor of different pro-
cesses would provide interesting clues to understand the con-
densed phase reactive processes.
In this work we have probed the phase diagram of pro-
pene in order to find out the pressure and temperature con-
ditions which trigger the chemical reaction. Despite the dif-
ferent reaction conditions, also including the optical catalysis
of the chemical transformation, the polymer formation is not
observed at high pressure, being the product always com-
posed by a mixture of oligomers. Kinetic experiments per-
formed at constant pressure and temperature as a function of
temperature and pressure, respectively, allowed the thermo-
dynamic parameters of this reaction to be obtained.

0021-9606/2005/123共19兲/194510/9/$22.50 123, 194510-1 © 2005 American Institute of Physics

Downloaded 22 Nov 2005 to 150.217.156.167. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp
194510-2 Citroni et al.
FIG. 1. Phase diagram of propene. Empty and full circles are experimental
points relative to the melting from Ref. 18. The melting line, extrapolated
according to the Simon-Glatzel equation,18 is drawn as a dotted line above
2.5 GPa to indicate that it is not reliable above this pressure. The isobaric
and isothermal studies performed in this work are indicated by dotted ar-
rows. Empty square indicates the spontaneous polymerization observed in
Ref. 21 at 473 K. Full squares are the threshold conditions for the onset of
the reaction observed in this work. A tentative stability boundary for the
monomer is identified by the full line passing through these points and by
the dashed line that overlaps the melting above 250 K. Stars are the P-T
conditions at which the reaction kinetic was studied.
II. EXPERIMENT
A membrane diamond anvil cell 共MDAC兲 equipped with
IIa-type diamonds was employed to compress pure propene
samples. The sample was confined by a stainless-steel gasket
and its initial dimensions were generally 150 ␮m in diameter
and 45 ␮m in thickness. Loading of propene 共purity ⬎99.8%
from Rivoira兲 into the cell was achieved by condensing the
FIG. 2. IR spectra of propene in the C–H stretching region recorded during
an isothermal compression 共panel a兲 and decompression 共panel b兲 at 220 K.
Downloaded 22 Nov 2005 to 150.217.156.167. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp
J. Chem. Phys. 123, 194510 共2005兲
FIG. 3. Pressure shift of the ␯17 band during the three isothermal runs at
110, 170, and 220 K both on compression 共solid circles兲 and decompression
共empty circles兲. Dotted lines represent the expected transition pressures ac-
cording to Ref. 18.
gas close to the melting point 共88 K兲 in a nitrogen atmo-
sphere. After closing the cell, care was taken to maintain the
liquid propene to an initial pressure not exceeding 0.7 GPa.
The pressure was measured by the ruby fluorescence method
using the second-harmonic emission line 共532 nm兲 of a di-
ode pumped Nd:YAG 共yttrium aluminum garnet兲 laser as
excitation source. The laser power was reduced to 0.2 mW in
order to avoid unintended laser effects on the reaction pro-
cess. Fourier transform infrared 共FTIR兲 absorption measure-
ments, with an instrumental resolution of 1 cm−1, were per-
formed with a Bruker-IFS 120 HR spectrometer specifically
modified for high-pressure measurements.16,17 Experiments
as a function of temperature 共15– 400 K兲 were performed by
using a close-cycle cryostat coupled with the spectrometer
and a resistive heating of the MDAC. The temperature was
measured by a Si diode below 300 K, and by a J thermo-
couple above 300 K.
III. PHASE DIAGRAM OF PROPENE
Propene is a gas at ambient conditions with Tb = 226 K
and Tm = 88 K. The triple point is at Ttr = 88 K and Ptr
= 9.5 bars, while the critical point is at Tc = 365 K and Pc
= 46 bars. A melting line, obtained from volume measure-
ments in the 96– 142 K and 0.1– 1.0 GPa temperature and
pressure ranges, is available in literature.18 In the same work
the existence of two solid phases, I and II, with the high-
temperature phase II having a glassy structure, was sug-
194510-3 High-pressure reactivity of propene J. Chem. Phys. 123, 194510 共2005兲

TABLE I. Assignment of the vibrational modes of propene. Frequencies are expressed in cm−1, st= stretching,
w = wagging, r = rocking, t = twisting, b = bending, s = symmetric, a = antisymmetric, ip= in plane, oop= out of
plane.

Sym. Mode Description Calc.a Exp.a This workb

A⬘ 1 CH2 a-st 3092.2 3091.0 3088

2 CH st+ CH2 s-st 3017.2 3008c 3018
3 CH2 s-st+ CH st 2996.4 2991.0 2985
4 CH3 a-st 2972.3 2973.0 2972
5 CH3 s-st 2930.0 2931.9 2922
6 C v C st+ CH2 ip-b 1655.5 1652.8 1648
7 CH3 a-ip-b + CH2 ip-b 1460.3 1458.5 1452
8 CH2 ip-b + CH3 a-ip-b 1418.9 1414c 1416
9 CH3 s-ip-b 1380.7 1378.0 1372
10 CH ip-b + CH2 r + C v C st 1304.5 1298.0 1297
11 CH2 r + C – C st+ C – C v C ip-b 1179.2 1170 1170
12 CH3 r + CH2 r + CH ip-b 944.6 934.5c 936
13 C–C st+ CH2 r 914.2 919c
14 C v C – C ip-b 428.4 428
A⬙ 15 CH3 a-st 2930 2931.9 2949
16 CH3 a-ip-b 1448.7 1442.5 1436
17 CH3 r + CH oop-b 1047.3 1044.7 1041
18 CH2 t + CH oop-b 994.2 990.0 991
19 CH2 w 913.1 912 912
20 CH oop-b + CH2 t + CH3 r 575.2 575.2 584
21 CH3 t 189.2 188
a
Gas phase at ambient conditions 共Ref. 22兲.
b
T = 298 K and P = 0.4 GPa.
c
Liquid phase at low temperature 共Ref. 22兲.

gested. The melting line was extrapolated by the authors to along this isobar and in compression-decompression cycles
higher pressures and temperatures according to the Simon- at 110, 170, and 220 K, but in all cases the only spectral
Glatzel equation,19 variations were a line narrowing on lowering temperature,

冋冉 冊 册
c
T creasing pressure.
P − Ptr = a −1 ,
Ttr
where Ptr and Ttr are the coordinates of the triple point and a
and c are constants which depend on the substance under
study. The melting line rises indefinitely with increasing
pressure if a and c remain constant 共see Fig. 1兲, and actually
the extrapolation with this curve seems to be valid only up to
pressures about one order of magnitude greater than the
value of a obtained at low pressures.20 Therefore, being a
= 3.064 kbars and c = 3.871, the curve is not expected to re-
produce the melting line above 2 – 3 GPa. Finally, during
volumetric studies up to 1.0 GPa, a spontaneous polymeriza-
tion was detected at 473 K with an increasing efficiency on
rising pressure.21
In the attempt to extend the melting curve to higher pres-
sures and temperatures, and to collect information about the
structural properties of the solid phases, we probed the phase
diagram of propene between 40 and 370 K by means of the
FTIR spectroscopy. The melting curve reported in Ref. 18
was crossed on the increasing and decreasing pressure along
several isotherms, and on cooling or heating along one isobar
共P ⬃ 0.1 GPa兲 as shown in Fig. 1. No changes were detected
in the IR spectrum on cooling the sample from ambient tem-
perature down to 44 K at 0.1– 0.2 GPa but a line narrowing
that allowed the observation of unresolved absorptions at
ambient temperature. The melting line was likely crossed
and a continuous frequency shift and line broadening on in-
In spite of the high pressures reached, especially in the
170 K 共5.8 GPa兲 and 220 K 共8.2 GPa兲 cycles, the spectrum
of the monomer was completely recovered when the pressure
was released, thus excluding any chemical reaction in these
P-T conditions 共see Fig. 2兲. In order to detect any subtle
spectral change associated with a phase transition, the ab-
sorption profiles of the modes ␯17, ␯9, and ␯6 共the assignment
of the propene bands is reported in Table I兲, were fitted with
pseudo- Voigt functions and the pressure evolution of their
frequencies, bandwidths and integrated areas along each iso-
therm was analyzed. However, no discontinuity or slope
change was detected, so that the melting line could not be
identified. As an example, the pressure shift of the ␯17 band
along the 110, 170, and 220 K isotherms is reported in Fig.
3. This behavior likely indicates the glassy nature of the solid
phase but x-ray-diffraction measurements are mandatory to
confirm it as well as to detect the melting curve at higher
pressures.
IV. PRESSURE-INDUCED REACTION
The chemical reactivity of propene was studied at differ-
ent temperature and pressure conditions on freshly loaded
samples. In the experiments directed to identify the reaction
pressure threshold the pressure was slowly raised along an
isotherm by steps of 0.2– 0.3 GPa, monitoring the FTIR

Downloaded 22 Nov 2005 to 150.217.156.167. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp
194510-4 Citroni et al.
FIG. 4. Time evolution of the IR absorption spectra of propene recorded
during the reaction at 298 K and 3.2 GPa. The propene absorptions bands
共␯6, ␯9, and ␯18兲 and the product band indicated by the dotted line have been
used for the kinetic analysis.
spectrum for several hours at each step to detect the forma-
tion of new chemical entities. The appearance of new IR
bands was interpreted as the beginning of the reaction 共full
squares in Fig. 1兲. In other experiments, the desired P-T
conditions were rapidly achieved after the loading and main-
tained until an equilibrium for the transformation was
reached. The reaction evolution was monitored by the FTIR
spectrum and care was taken in keeping constant pressure
and temperature during all the reactive processes in order to
extract reliable kinetic information. Several experiments
were performed in the range of 295– 380 K and
1.1– 3.8 GPa, at the P-T conditions evidenced by the stars in
Downloaded 22 Nov 2005 to 150.217.156.167. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp
J. Chem. Phys. 123, 194510 共2005兲
Fig. 1. The onset of the reaction is revealed by the growth of
new absorption bands at 1384 and 1470 cm−1, as shoulders
of the ␯9 and ␯7 bands of propene, respectively, and by a
broad absorption in the C–H stretching region
共2820– 3020 cm−1兲, overlapped to the monomer peaks. This
behavior, together with the decrease of the propene bands,
can be appreciated in Fig. 4 where the evolution with time of
the FTIR spectrum at 298 K and 3.2 GPa is reported. In
general, a large amount of propene was still present in the
cell when the reaction equilibrium was reached 共i.e., when
the FTIR spectrum did not show any change after several
hours兲. In all the experiments, a further temperature or pres-
sure increase forced the propene to react to a greater extent
but its complete transformation was never obtained. A very
important result is that no reaction was observed in all the
compression experiments performed at T 艋 220 K. In par-
ticular, in a 220 K isotherm the sample was compressed up
to 11.3 GPa but, as it can be seen in Fig. 1, the reaction was
observed at this pressure only after rising the temperature up
to 290 K. This temperature was reached in steps of 5 K wait-
ing 12 h at each step to monitor the occurrence of the reac-
tion. This result is extremely interesting because along the
270 K isotherm the reaction was observed to occur at 4 GPa,
a temperature lower than that found along the 11.3 GPa iso-
bar.
A. The reaction product
The product recovered from all the reactions is a trans-
parent viscous gel-like material which, by visual inspection
at a microscope, appears as a film smeared on the diamond
surfaces. A consistent amount of this material is lost when
the cell is opened to allow the evaporation of the unreacted
propene. Only a weak IR spectrum of this sample could be
recorded 共trace a in Fig. 5兲. Beside the scantily informative
CH stretching region, two weak peaks at 973 and 1157 cm−1
and two more intense bands at 1379 and 1460 cm−1 are ob-
served. The state of the product rules out the possibility that
a polymerization occurred since polypropylene, regardless of
its tacticity, is solid. Therefore, according to its viscous na-
ture, the reaction product is likely a mixture of liquid and
solid oligomers, even though the quite narrow spectral band-
FIG. 5. Comparison of the IR spectra of the sample
recovered from the purely pressure-induced process
共trace a兲 and of several dimerization and trimerization
products 共left panel兲 and of commercial polypropylenes
共right panel兲. Specifically, left panel: 共b兲 1,3-
dimethylcyclobutane, 共c兲 2-methylpentane, 共d兲 2,4-
dimethylheptane, 共e兲 4-methyloctane, 共f兲 1,2,4-
trimethylcyclohexane, 共g兲 1,3,5-trimethylcyclohexane,
and 共h兲 1,2,3-trimethylcyclohexane; right panel: 共b兲
syndiotactic Mw 174 000, 共c兲 amorphous Mw 19 600,
共d兲 isotactic Mw 120 000 共130-␮m-thick in KBr兲, and
共e兲 isotactic Mw 580 000 共243 ␮m thick in KBr兲. The
displayed spectrum of the recovered product is enlarged
ten times.
194510-5 High-pressure reactivity of propene J. Chem. Phys. 123, 194510 共2005兲

widths could indicate a relatively sharp size distribution. It

should be considered that propene derivatives formed by up
to six monomeric units are still liquid.23 In the left panel of
Fig. 5 we compare the IR spectrum of the recovered sample
with those of some possible dimers and trimers, that form
from the monomer either from linear propagation of from
cyclization mechanisms. Similarities with linear branched
oligomers 共traces c–e兲, rather than with cyclic compounds,
are noticed. This observation is also supported by the close
resemblance of our IR spectra to those of longer linear oli-
gomers formed by five monomeric units.24 This comparison
also suggests the atactic nature of our sample since the rela-
tive intensities of the two bands at 972 and 993 cm−1 has
been used as a criterion to judge long oligomers and polymer
stereoregularity, being the latter as higher as stronger is the
band at 992 cm−1 that is hardly detectable in the spectrum of
our recovered compound. Also the comparison with the IR
spectra of several commercial polypropene isomers having
different tacticity and chain length, reported in the right
panel of Fig. 5, confirms this finding: only the spectrum of
the amorphous polymer 共trace c兲 shows a remarkable simi-
larity in the fingerprint region with that of the recovered
sample.

B. Kinetic analysis
FIG. 6. Time evolution of the integrated intensity of the product absorption
The time evolution of the product absorption band grow- band growing as a shoulder of the ␯9 mode of propene. The two sets of data
ing at 1384 cm−1 and of the ␯6 and ␯18 isolated bands of points have been fitted by the Avrami equation 关Eq. 共1兲兴.
propene, were used to perform the kinetic analysis of the
reactions. Their spectral profiles were fitted with pseudo- pene loaded and thus they are not indicative of the extent of
Voigt functions and their integrated intensities were reported the reaction. The time evolution of the absorption intensities
as a function of time to be analyzed with the Avrami of the ␯6 and ␯18 propene modes were also fitted to a decay
model.25 This model was originally developed for the crystal law built according to the Avrami model,
growth from a liquid phase but it has been later extended to n
At = A⬁ + 共A0 − A⬁兲e−关k共t − t0兲兴 , 共2兲
the study of solid-state reactions.26 The general formulation
is where A0, At, and A⬁ are the propene concentration at time 0,
−关k共t − t0兲兴n t, and at the equilibrium, respectively. Two examples are
At = A⬁关1 − e 兴, 共1兲
reported in Fig. 7. The results obtained by the fit of the
where At and A⬁ are the concentration of the product, the product and of the monomer bands are fully consistent. The
band absorbance in the present case, at time t and at the n values found for the different experiments range from 0.48
equilibrium, respectively; k is the rate constant; t0 is the start- to 0.70, while the average over all the seven experiments is
ing time of the growth process; and n is a parameter which 0.61. This result suggests the same reaction mechanism in all
depends on the nucleation law and on the dimensionality of the cases and indicates, within the Avrami model, a linear
the growth process. In some cases, such as for diffusion con- growth with all the nuclei active at the beginning of the
trolled polymerizations, the mechanism of the reaction can process.
be revealed by the value of the n parameter.27–29 The kinetic
curves built on the absorbance values of the product band TABLE II. Parameters obtained by the fit of the time evolutions of the
have been consistently reproduced by Eq. 共1兲 for all the dif- product absorption band, shoulder of the ␯9 mode of propene, at different
ferent experiments. The data measured in two high- temperatures and pressures by using the Avrami equation. The relative un-
certainty is 10% for k and n, and 1% for t0 and A⬁.
temperature experiments are reported in Fig. 6 together with
the best fit. In these examples it is evident how fast the T 共K兲 P 共GPa兲 n k 共h−1兲 t0 共h兲 A⬁
reaction develops at high temperature needing only 2 – 3 h to
reach the completion. This makes the determination of the 296 3.2 0.55 0.18 1.50 13.5
kinetic curve really challenging because several spectra must 297 3.8 0.65 0.25 1.06 12.7
370 0.95 0.64 0.64 −0.30 6.38
be measured in the early stages of the reaction in order to
378 3.4 0.64 14.6 0.18 16.0
have reliable kinetic curves. The parameters obtained by all
335 3.3 0.48 1.0 0.22 12.6
the kinetic studies are reported in Table II. The relative un- 380 1.1 0.70 2.0 0.26 10.9
certainty does not exceed 10% for k and n, and 1% for t0 and 373 2.1 0.60 15.4 0.06 16.5
A⬁. The values of A⬁ depend on the initial amount of pro-

Downloaded 22 Nov 2005 to 150.217.156.167. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp
194510-6 Citroni et al.
FIG. 7. Time evolution of the integrated intensity of the ␯6 mode absorption
band of propene in two different kinetic studies. The two sets of data point
have been fitted according to the Avrami model by Eq. 共2兲.
The dependence of the rate constant k upon temperature
at constant pressure is given by
冉 冊
⳵ ln k Ea
= 2,
⳵T P RT
while its dependence upon pressure at constant temperature
is
冉 冊
⳵ ln k
⳵P T
=−
⌬V⫽
RT
. 共4兲
In the transition state theory, Ea and ⌬V⫽ are the activation
energy and the activation volume of the process, respec-
tively. In the present case, the good agreement with the
Avrami model indicates the diffusion of the reagent species
to the nucleation centers as the rate-determining elementary
step, therefore the activation energy and volume refer to this
process. In the upper panel of Fig. 8 we have reported the
evolution of the ln k values collected at 3.4± 0.3 GPa as a
function of 1 / T 共295艋 T 艋 378 K兲; while the evolution of
ln k with pressure 共0.95艋 P 艋 3.4 GPa兲 at 375± 5 K is
shown in the lower panel. An activation energy of
46 KJ mol−1, with an uncertainty of 14%, was obtained for
the oligomerization process at 3.4 GPa by the linear fit of the
experimental points reported in the upper panel of Fig. 8
according to Eq. 共3兲. As the activation volume is concerned,
the nonlinear evolution of the ln k values at 375 K as a func-
tion of P was reproduced by a parabolic law30
Downloaded 22 Nov 2005 to 150.217.156.167. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp
J. Chem. Phys. 123, 194510 共2005兲
FIG. 8. Upper panel: linear fit of the observed values of ln k as a function of
1 / T at 3.4 GPa. Lower panel: polynomial fit 共second order兲 of the observed
values of ln k as a function of P at 375 K.
ln k = a + bP + cP2 ,
with a = −5共1兲, b = 6共2兲 GPa−1, and c = −1.1共3兲 GPa−2. By us-
ing this quadratic dependence on P of ln k the activation
共3兲
volume can be obtained
⌬V⫽ = − 共b + 2cP兲RT.
The parabolic fit of the data reported in the lower panel of
Fig. 8 provides a ⌬V⫽ equal to −19 cm3 mol−1 at atmo-
spheric pressure, −12 cm3 mol−1 at 1 GPa, −5.1 cm3 mol−1 at
2 GPa, zero at 2.7 GPa, and positive at higher pressures.
This means that the reaction is strongly accelerated by com-
pression at relatively low pressures, but this effect reduces on
increasing pressure until the process is retarded by further
compression above 2.7 GPa.
C. Laser-assisted high-pressure reaction
Since the polymerization of the sample is not achievable
in the purely pressure induced reaction we tried to induce the
reaction under the combined effect of pressure and laser ir-
radiation, in analogy with the butadiene and ethylene
cases.13,15 A propene sample was irradiated at room tempera-
ture and different pressures starting from 0.9 GPa. Each
cycle, of variable duration, was performed by focusing
20 mW of the 458 nm line of an Ar ion laser on the sample
with a 200 mm focal length. FTIR spectra were measured
before and after each cycle. A transformation, similar to that
observed at higher pressures or temperatures in the purely
pressure-induced reaction, was observed at 1.8 GPa 共see up-
194510-7 High-pressure reactivity of propene
FIG. 9. Evolution of the IR spectrum in the ␯9-␯7 region during the laser-
assisted reaction at 298 K. Upper panel: IR spectra measured at 1.8 GPa
after the indicated total irradiation duration with 20 mW of the 458 nm Ar+
line. Lower panel: IR spectra recorded during the reaction induced at
0.75 GPa by the UV multiline emission 共337.5 and 356.4 nm兲 of a Kr+ laser.
The total irradiation time with a power of 500 mW is reported for each
spectrum, the first 1 h cycle 共star兲 was performed with 100 mW.
per panel of Fig. 9兲. The occurrence of the reaction is re-
vealed after the first irradiation cycle by the growth of the
characteristic absorption bands of the product at 1380 and
1468 cm−1, by the intensity decrease of the absorption bands
assigned to propene, and by the intensification of the C–H
stretching modes involving sp3 C. Four further irradiation
cycles were performed at this pressure under the same con-
ditions, and the FTIR spectrum was monitored after each
cycle. The spectra show the reaction to proceed under irra-
diation 共see Fig. 9兲, while it proceeds very slowly when the
sample is not irradiated stopping after 24– 48 h and restart-
ing only after a new irradiation cycle. After 16 irradiation
hours, and a total time of 145 h, the intensity of the product
bands did not increase further.
A similar experiment was repeated on a new sample of
propene compressed at 0.75 GPa by using the UV multiline
共UVML兲 at 337.5– 356.4 nm of a Kr ion laser. A longer focal
length 共300 mm兲 was employed to focus the laser onto the
sample. In the first cycle the sample was irradiated for 1 h
with a power of 100 mW. As reported in the lower panel of
Fig. 9, the sample reacted to a very low extent. A new irra-
diation cycle of 1 h with 500 mW at the same wavelength
was performed. In this case, the reaction was clearly de-
Downloaded 22 Nov 2005 to 150.217.156.167. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp
J. Chem. Phys. 123, 194510 共2005兲
FIG. 10. Evolution of the absorption band of the ␯6 mode at room tempera-
ture during: 共a兲 the reaction activated at 0.75 GPa by the UV multiline
共500 mW兲; 共b兲 the reaction activated at 1.8 GPa by the 458 nm line
共20 mW兲; 共c兲 the purely pressure-induced reaction at 3.2 GPa. The time
values reported for the laser-assisted reactions 共a兲 and 共b兲 indicate the effec-
tive duration of laser irradiation since no appreciable changes are observed
at these pressures when the sample is not irradiated.
tected. Three more cycles of 7, 7 and 8 h were performed
with 500 mW at the same wavelength, until no more spectral
changes were observed.
At the end of both irradiation experiments a larger
amount of product was recovered than in the purely pressure
induced experiments. This result can be readily appreciated
in Fig. 10 where the band profile of the ␯6 propene mode is
compared during the three reactions. Nevertheless, a dense
liquid product was obtained also in the photoassisted reac-
tions, and the IR spectrum measured before opening the cell
reveals the reaction product to be the same as in the purely
pressure-induced reaction.
V. DISCUSSION
In spite of the wide P-T region in which propene was
investigated, the impossibility to detect the melting through
the IR spectra prevented the extension of the melting line
reported in the literature 共Fig. 1兲. The lack of spectral modi-
fications during the melting as well as the observed overcom-
pression and supercooling, suggest a glassy solid also in
analogy with other similar systems.18 An important informa-
tion about the melting line can be gained by the reactivity
results. During the isothermal compression at 220 K no sign
of reaction was observed up to 11.3 GPa, while in the fol-
lowing isobaric heating cycle the reaction threshold was
found at 290 K. This temperature is higher than the threshold
temperature of 270 K observed at 4.0 GPa. If we look at the
isothermal compression cycles reported in Fig. 1, it is evi-
dent that above 250 K the reaction occurs from the fluid
phase, while below this temperature the melting line should
be crossed to trigger the reaction, according to the extrapo-
lated melting and reaction lines. The missed observation of
the reaction in the glassy solid can be correlated to a dra-
matic reduction of the diffusivity with respect to the fluid.
The reactivity of propene at 11.3 GPa and 290 K could
therefore be related to the occurrence of the melting at these
P-T conditions. As a matter of fact, the agreement with the
194510-8 Citroni et al.
FIG. 11. Absorption spectra measured in the empty cell and with propene at
the pressure values where the reaction was activated by the laser light. The
two arrows indicate the laser lines employed in the two cases.
extrapolated melting line of Ref. 18 is quite good 共see Fig.
1兲. The importance of the diffusivity in the high-pressure
reactivity of propene is also confirmed by the fit of the ki-
netic data with the Avrami model. The low value of n 共0.6 on
the average兲 indicates that the activation of the reactive cen-
ters 共nucleation兲 is extremely rapid and does not contribute
to the reaction kinetics being the diffusion the rate-
determining step of the mechanism.
The activation energy and volumes obtained from con-
stant pressure and temperature experiments, respectively, are
therefore only related to the diffusion step. As obtained from
Fig. 8, the activation volume at room pressure results
⌬V⫽ = −19 cm3 mol−1 in nice agreement with the values
共⬃−20 cm3 mol−1兲 reported for radical polymerization
reactions.31,32 The negative value means that the reorienta-
tions and rearrangements leading to the formation of new
bonds occur in the sense of decreasing the volume of the
system. A negative contribution to the activation volume
arises from electrostriction if the transition state is more po-
lar than the initial state. This contribution is more important
when the reaction is performed in an apolar medium, as in
this case is the propene matrix. Geometric rearrangements
have a positive structural contribution to ⌬V⫽ when they
transiently require an extravolume to be achieved. The ob-
served increase of the ⌬V⫽ values with pressure, ⌬V⫽ ⬎ 0
above 2.7 GPa, is attributed to a dominant structural contri-
bution due to a growing steric hindrance. The activation en-
ergy at 3.4 GPa Ea = 46 KJ mol−1, corresponding to about
3800 cm−1, indicates that the rate-determining step consists
of a rearrangement requiring an energy in the region of the
molecular vibrations. This value is not far from that found by
Yoo and Nicol in the pressure-induced polymerization of
cyanoacetylene.27
As already discussed, the gel-like state of our recovered
sample likely suggests a reaction product composed by a
mixture of solid and liquid oligomers, even though the spec-
tral bandwidths indicate a relatively sharp size distribution.
The formation of a polymeric chain can be excluded but the
comparison among the IR spectra of our recovered product,
long oligomers,24 and commercial polypropenes can be use-
ful 共Fig. 5兲 to gain insight on the structure of the resulting
chains. Syndiotactic and isotactic polymers and oligomers
present characteristic absorption bands between 800 and
Downloaded 22 Nov 2005 to 150.217.156.167. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp
J. Chem. Phys. 123, 194510 共2005兲
1100 cm−1 which allow the identification of these two geo-
metrical arrangements. On the other hand, the spectrum of
the amorphous polypropene is a kind of superposition of the
two-ordered polymers spectra but with a general intensity
redistribution that makes also this spectrum extremely char-
acteristic. The similarity of this last spectrum with that of our
recovered product is remarkable, indicating that long amor-
phous oligomers are likely present in our product. Also the
comparison between the IR spectra of shorter 共see Fig. 5兲
and longer oligomers24 suggests the formation of linear
methyl-substituted alkanes rather than the cyclic products
that should be preferred for the stoichiometry of the process.
Linear products require in fact either extra protons, or the
formation of a terminal unsaturated C v CH2 bonds not de-
tected in our spectra. However, the protons necessary for the
chain-end saturation can be pictured as coming from other
propene molecules, which are in large excess also when the
equilibrium of the reaction is reached.
A noticeable support to the interpretation of the reaction
dynamics is provided by a recent ab initio molecular-
dynamics simulation.33 In this study the propene molecules
were found to undergo a complex reaction scheme to pro-
duce two different dimers: a saturated linear chain made up
of five propene units, formed by head to tail interactions and
leading to a polypropene sequence, and methyl acetylene.
The process was found to be highly cooperative, and a con-
sistent proton mobility characterized the sample during the
reaction making the reaction mechanism completely ionic.
This picture nicely agrees with the hypotheses drawn on the
basis of the present experimental findings: a diffusion limit-
ing step, a polar transition state that give a negative contri-
bution to ⌬V⫽, the formation of relatively small oligomers
stabilized by the hydrogen migration that prevents a further
lengthening of the chain. Furthermore, also in the simulation
the reaction was not complete, involving only 40% of the
monomeric units present in the simulation box.
As observed in several other cases the effect of laser
irradiation is to lower the reaction threshold pressure.10 The
reaction was induced at 1.8 GPa by using laser light at
458 nm, and at 0.75 GPa by using a multiline emission be-
tween 337.5 and 356.4 nm, while for lower pressures these
wavelengths were not able to trigger the reaction. This pres-
sures values are considerably lower than the purely pressure
threshold occurring at 3.1 GPa at 298 K. From Fig. 11 it can
be seen that the absorption spectra measured at the same
pressures of the two irradiation experiments only presents
the diamond absorption edge starting at 230 nm, therefore
the employed wavelengths cannot be absorbed by propene
through a one-photon process. This is also reasonably ex-
pected according to the position of the first electronic excited
state which is located between 160 and 190 nm
共52 631– 62 500 cm−1兲 at ambient pressure.34 Therefore the
laser light is absorbed through a multiphoton process. This
is, once again, a common behavior with other systems react-
ing through an optical catalysis in the fluid phase.13,15 The
product of the photoassisted reaction has the same character-
istics of the product recovered from the purely pressure-
induced process. This comparison may imply that the same
reaction path is explored in the two processes. The decrease
194510-9 High-pressure reactivity of propene J. Chem. Phys. 123, 194510 共2005兲

of the threshold pressure under irradiation indicates that the ACKNOWLEDGMENTS

electronic excited states are also involved in the purely
pressure-induced process, as clearly evidenced in the This work has been supported by the European Union
calculations.33 From the electronic point of view, the situa- under Contract No. HPRI-CT1999-00111, by the Italian
tion in propene is very similar to that of ethylene.15 The Ministero dell’Università e della Ricerca Scientifica e Tecno-
lowest electronic transition corresponds to a ␲ → ␲* excita- logica 共MURST兲, and by “Firenze Hydrolab” through a grant
tion, whose consequence is to destroy the planarity constraint by Ente Cassa di Risparmio di Firenze.
imposed by the ␲ bond in the fundamental state and to
1
lengthen the molecule. A twofold effect is thus gained: the A. S. Balchan and H. G. Drickamer, J. Chem. Phys. 34, 1948 共1961兲.
2
reduction of the intermolecular distances, thus simulating a B. M. Riggleman and H. G. Drickamer, J. Chem. Phys. 38, 2721 共1963兲.
3
S. Desgreniers, Y. K. Vohra, and A. L. Ruoff, J. Phys. Chem. 94, 1117
higher pressure value, and the possibility for the terminal 共1990兲.
groups to rotate minimizing the steric hindrance between 4
A. Goncharov, E. A. Gregoryanz, H. K. Mao, Z. Liu, and R. J. Hemley,
nearest-neighbor reacting molecules. Phys. Rev. Lett. 85, 1262 共2000兲.
5
M. I. Eremets, R. J. Hemley, H. K. Mao, and E. A. Gregoryanz, Nature
共London兲 411, 170 共2001兲.
VI. SUMMARY AND CONCLUSIONS 6
E. A. Gregoryanz, A. F. Goncharov, R. J. Hemley, and H. K. Mao, Phys.
The pressure-induced reaction in fluid propene has been Rev. B 64, 052103 共2001兲.
7
V. Iota, C. S. Yoo, and H. Cynn, Science 283, 1510 共1999兲.
extensively investigated by modifying the pressure, tempera- 8
S. Serra, C. Cavazzoni, G. L. Chiarotti, S. Scandolo, and E. Tosatti,
ture, and using light to activate the reaction. In spite of the Science 284, 788 共1999兲.
9
different tools employed to induce the reactivity of the V. Schettino and R. Bini, Phys. Chem. Chem. Phys. 5, 1951 共2003兲.
10
monomer, we were not able to produce a polymeric chain R. Bini, Acc. Chem. Res. 37, 95 共2004兲.
11
L. Ciabini, M. Santoro, R. Bini, and V. Schettino, Phys. Rev. Lett. 88,
contrary to the butadiene and ethylene cases. The product
085505 共2002兲.
always consists of a mixture of solid and liquid linear oligo- 12
M. Santoro, M. Ceppatelli, R. Bini, and V. Schettino, J. Chem. Phys. 118,
mers. The stabilization of such oligomers, and of the inter- 8321 共2003兲.
13
mediate species, could be searched in an ionic mechanism M. Citroni, M. Ceppatelli, R. Bini, and V. Schettino, Science 295, 2058
characterized by high proton mobility as suggested by recent 共2002兲.
14
V. Busico and R. Cipullo, Prog. Polym. Sci. 26, 443 共2001兲.
ab initio molecular-dynamics studies.33 The kinetic analysis 15
D. Chelazzi, M. Ceppatelli, M. Santoro, R. Bini, and V. Schettino, Nat.
of the reaction, performed in several different P-T condi- Mater. 3, 470 共2004兲.
16
tions, indicates an oligomerization kinetically controlled by a R. Bini, R. Ballerini, G. Pratesi, and H. J. Jodl, Rev. Sci. Instrum. 68,
diffusion step, dominated by the relative motions below 3154 共1997兲.
17
F. A. Gorelli, L. Ulivi, M. Santoro, and R. Bini, Phys. Rev. Lett. 83, 4093
2.7 GPa, and by molecular rearrangements at higher pres- 共1999兲.
sures. The pressure behavior of the activation volume evi- 18
L. E. Reeves, G. J. Scott, and S. E. Babb, Jr., J. Chem. Phys. 40, 3662
dences how pressure drives the reactivity by the two opposite 共1964兲.
19
effects of decreased intermolecular distances and of a de- F. E. Simon and G. Glatzel, Z. Anorg. Allg. Chem. 178, 309 共1929兲.
20
S. E. Babb, Jr., Rev. Mod. Phys. 35, 400 共1962兲.
creased diffusivity. This latter aspect is emphasized by the 21
S. L. Robertson and S. E. Babb, Jr., J. Chem. Phys. 51, 1357 共1969兲.
observation in the phase diagram of a slope inversion of the 22
B. Silvi, P. Labarbe, and J. P. Perchard, Spectrochim. Acta, Part A 29,
stability boundary likely connected to a lack of reactivity in 263 共1973兲.
23
the glassy phase even at relatively high pressures. Here the CRC Handbook of Chemistry and Physics, 76th ed., edited by D. R. Lide
共CRC, Boca Raton, FL, 1995兲.
thermal motions are strongly reduced and the decrease of the 24
E. Benedetti, S. Pucci, P. Pino, and V. Schettino, Spectrochim. Acta, Part
intermolecular distances due to a greater density is not suf- A 29, 1313 共1973兲.
25
ficient for the reaction to occur, since the internal rearrange- M. Avrami, J. Chem. Phys. 7, 1103 共1939兲; 8, 212 共1940兲; 9, 177
ments of the molecules are also required. The reactivity in 共1940兲.
26
S. F. Hulbert, J. Br. Ceram. Soc. 6, 11 共1969兲.
the glassy phase may be also prevented by the lack of long- 27
C. S. Yoo and M. Nicol, J. Phys. Chem. 90, 6732 共1986兲.
range order, i.e., of collective motions 共phonons兲 that acti- 28
M. Ceppatelli, M. Santoro, R. Bini, and V. Schettino, J. Chem. Phys. 113,
vate the chemical reaction by changing the interaction con- 5991 共2000兲.
29
tacts among the molecules and also transmit the lattice D. Chelazzi, M. Ceppatelli, M. Santoro, R. Bini, and V. Schettino, J.
distortion from site to site allowing the reaction to Phys. Chem. B 共in press兲.
30
E. V. Boldyreva, V. V. Boldyrev, and T. P. Shakhtshneider, in Reactivity
propagate.28,35 Like in several other unsaturated compounds of Molecular Solids, edited by E. V. Boldyreva and V. V. Boldyrev
the photochemical activation of the reaction determines a 共Wiley, Chichester, 1999兲, pp. 134.
31
lowering of the reaction threshold pressure, but a higher se- T. Asano, W. J. Le Noble, Chem. Rev. 共Washington, D.C.兲 78, 407
lectivity of the reactive process, as observed for ethylene and 共1978兲.
32
N. S. Isaac, in High Pressure Techniques in Chemistry and Physics, ed-
butadiene, is not introduced in the present case. This is prob- ited by W. B. Holzapfel and N. S. Issac 共Oxford University Press, New
ably due to the involvement of the first excited state in the York, 1997兲, p. 307–351.
purely pressure-induced reaction 共as indicated by ab initio 33
M. Mugnai, G. Cardini, and V. Schettino, J. Chem. Phys. 120, 5327
molecular-dynamics studies33兲, i.e., the geometry of the 共2004兲.
34
G. Herzberg, Electronic Spectra and Electronic Structure of Polyatomic
ground state at a sufficiently high pressure matches the ge- Molecules, Molecular Spectra and Molecular Structure, Vol. III 共D. Van
ometry of the excited state, so that excitation to this state Nostrand, Princeton, 1967兲.
35
leads to a chemical evolution in the same direction. K. Dwarakanath and P. N. Prasad, J. Am. Chem. Soc. 102, 4254 共1980兲.

Downloaded 22 Nov 2005 to 150.217.156.167. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp
View publication stats