Selective oxygenation of α-olefins by means of metalloporphyrin catalysts mimicking cytochrome P-450

Article abstract of DOI:10.1135/cccc2010138

A significant problem associated with synthetic metalloporphyrin catalysts used to mimic oxygenation function of cytochrome P-450 is their deactivation due to the formation of inactive μ-oxo dimer. We prepared a series of porphyrin-based catalysts mimicking cytochrome P-450 potentially resistant to deactivation by varying the central metal ion, porphyrin ring substitution, and catalyst support. The influence of the nature of a solid support, the type of central cation and the chemical modification of the porphyrin ring on the catalytic efficiency was studied. The oxygenation ability of the synthesized catalysts was tested using 1-hexene as a substrate and tert-butyl hydroperoxide as an oxidant under various reaction conditions. Identification of the oxygenation products was performed with gas chromatography-mass spectrometry (GC-MS). Quantification of the products formed was based on GC with flame ionization detection (FID). It was found that the application of low GC injection temperature (150 °C) was necessary to detect primary products of the oxygenation, peroxo alkenes, from which secondary oxygenation products, mainly epoxides, alcohols, aldehydes and ketones, were formed slowly. The efficiency of the individual catalysts depended on their structure and varied significantly. The most active catalysts were Fe(II)- tetraphenylporphyrin (TPP) immobilized on aminopolystyrene (aminoPS) and Mn(II)-TPP immobilized on aminoPS.

Full text of DOI:10.1135/cccc2010138

Oxygenation of α-Olefins by Metalloporphyrin Catalyst
1163
SELECTIVE OXYGENATION OF α-OLEFINS BY MEANS OF
METALLOPORPHYRIN CATALYSTS MIMICKING CYTOCHROME P-450
1
2
3
4,
Pavel ZACHAŘ , Galina PETKOVA , David SÝKORA and Vladimír KRÁL *
Institute of Chemical Technology Prague, Department of Analytical Chemistry,
1
Technická 5, 166 28 Prague 6, Czech Republic; e-mail: pavel.zachar@vscht.cz,
2
3
4
galina.petkova@vscht.cz, david.sykora@vscht.cz, vladimir.kral@vscht.cz
Received November 22, 2010
Accepted May 6, 2011
Published online September 5, 2011
A significant problem associated with synthetic metalloporphyrin catalysts used to mimic
oxygenation function of cytochrome P-450 is their deactivation due to the formation of
inactive µ-oxo dimer. We prepared a series of porphyrin-based catalysts mimicking cyto-
chrome P-450 potentially resistant to deactivation by varying the central metal ion,
porphyrin ring substitution, and catalyst support. The influence of the nature of a solid sup-
port, the type of central cation and the chemical modification of the porphyrin ring on the
catalytic efficiency was studied. The oxygenation ability of the synthesized catalysts was
tested using 1-hexene as a substrate and tert-butyl hydroperoxide as an oxidant under vari-
ous reaction conditions. Identification of the oxygenation products was performed with gas
chromatography-mass spectrometry (GC-MS). Quantification of the products formed was
based on GC with flame ionization detection (FID). It was found that the application of low
GC injection temperature (150 °C) was necessary to detect primary products of the oxygena-
tion, peroxo alkenes, from which secondary oxygenation products, mainly epoxides, alco-
hols, aldehydes and ketones, were formed slowly. The efficiency of the individual catalysts
depended on their structure and varied significantly. The most active catalysts were Fe(II)-
tetraphenylporphyrin (TPP) immobilized on aminopolystyrene (aminoPS) and Mn(II)-TPP
immobilized on aminoPS.
Keywords: Cytochromes; Oxygenations; Porphyrinoids; Cytochrome P-450; Enzyme mimick-
ing; Oxygenation; Metalloporphyrin; Product distribution.
Cytochrome P-450 is a member of the monooxygenase family of the heme
enzymes that play an important role in metabolism of biomolecules.
Cytochrome P-450 catalyzes the oxidation processes taking place mainly in
the liver where numerous substrates of endogeneous and exogeneous origin
are degraded. Because of its importance, cytochrome P-450 is intensively
studied by many biologists, chemists and physicists. Despite of this effort a
detailed mechanism of its action still remains unknown.
Collect. Czech. Chem. Commun. 2011, Vol. 76, No. 10, pp. 1163–1175
© 2011 Institute of Organic Chemistry and Biochemistry
doi:10.1135/cccc2010138

1164
Zachař, Petkova, Sýkora, Král:
A significant problem associated with synthetic metalloporphyrin cata-
lysts is their deactivation due to the formation of inactive µ-oxo dimers
(Metal–O–Metal). There are two ways to prevent the deactivation: first, the
application of a homogeneous catalysis – modification of the porphyrin
skeleton with various bulky electronegative substituents and an axial bond
with an appropriate ligand of the central atom1–9; second, the use of a het-
erogeneous catalysis – an attachment of the catalysts to a suitable support.
As a support cyclodextrin, cyclofan, molecular sieves, Merrifield resin or
synthetic polymers can be used10–12
.
In our previous work13 we have used the homogeneous catalysis with dif-
ferent metallocomplexes than porphyrin. The synthesis of such type of
metallocomplexes based on novel ligands, specifically on arylmethyl substi-
tuted bis-β-aminoketones, has been described for instance in ref.14. The effi-
cacy of the catalysts for 1-hexene oxygenation has been tested as function
of Mn+ and ligand type and reaction conditions. Our results have revealed
that the best performance for conversion of 1-hexene to oxygenation prod-
ucts has been obtained for Co(II) metallocomplexes of ligand N,N′-
ethylenebis[4-amino-1,5-di-(2-naphthyl)]pent-3-en-2-one. The analysis of
the product distribution has been performed by GC-MS. Mild conditions of
the analysis allowed for the detection of primary products of the oxygena-
tion reaction.
Here, we present novel metalloporphyrins immobilized on a solid sup-
port (heterogeneous catalysis). We investigated the influence of the nature
of a solid support, the type of a central cation and the chemical modifica-
tion of the porphyrin ring on the catalytic efficiency.
EXPERIMENTAL
Catalytic activity of a group of catalysts mimicking cytochrome P-450 structure was studied.
Experiments were carried out as heterogeneous catalysis where given complexes were fixed
on the following supports: aminopropyl-silica gel (aminoSil), aminopolystyrene (aminoPS),
and benzhydrylamine polystyrene (BHA-PS). As the catalysts we used a group of porphyrin
metallocomplexes with various central cations (Fe2+, Mn2+, Co2+) and porphyrin rings with
two substituents, phenyl (P), tolyl (Tol), and several types of linkage to the support (Fig. 1).
Binding of the porphyrins to the appropriate support was realized by several different
procedures.
4-Carboxyphenyl-triphenyl porphyrin or 4-carboxyphenyl-tritolyl porphyrin was con-
verted to the acid chloride and reacted with the amino-substituted support by a standard
procedure. The reaction product was filtrated out the reaction mixture and washed with di-
chloromethane, methanol, water, and methanol (Fig. 1; catalysts 1–5, 9–11)15
.
The application of the pyridyl-triphenyl porphyrin was based on the following reaction
sequence: the pyridine ring on the porphyrin skeleton was quaternized with bromoacetic
Collect. Czech. Chem. Commun. 2011, Vol. 76, No. 10, pp. 1163–1175

Oxygenation of α-Olefins by Metalloporphyrin Catalyst
1165
acid and after conversion to chloride reacted with the amino support. The resulting product
was filtered off, washed with methanol, dichloromethane and water (Fig. 1; catalyst 12).
Amine linkage between porphyrin derivative and the support was obtained by alkylation
of the amino-substituted support with bromomethylphenyl-triphenyl porphyrin (catalysts
Fig. 1; 6–8).
Metallation (incorporation of the central metal into the porphyrin skeleton) was carried
out by a standard Macac strategy (heating of the porphyrin covalently immobilized on the
support in the presence of the appropriate acetylacetone metallocomplex (Fe, Mn, Co) in
DMF (80 °C, 8 h)). The resulting catalysts were washed with dichloromethane and methanol
and after drying used in the subsequent experiments.
The experiments were carried out with tert-butyl hydroperoxide (t-BuOOH) as the oxygen
donor and pyridine as the metallo-axial ligand, where pyridine was also used as an internal
standard for the evaluation of the conversion and the product distribution measurement. In
the present work 1-hexene was used as a substrate for α-olefin oxygenation.
The following aspects were studied and evaluated: (i) the efficiency of the individual cata-
lysts, (ii) the reaction kinetics of the catalysis and (iii) the influence of the amount of the
oxidant (t-BuOOH) on the quantity of the products.
A typical experimental setup:
The reaction mixture consisted of 30 mg of a catalyst, 500 µl of 1-hexene (substrate),
100 µl of t-BuOOH (oxidant) in dichloromethane (commercial aqueous 70% solution of
t-BuOOH was extracted (ratio 1:1, v/v) with dichloromethane), 500 µl of dichlormethane
(solvent) and 10 µl of pyridine (axial ligand and internal standard).
Catalyst-support
M
Catalyst No.
M-TrPP-aminoSil
R = C6H5; Y1
Fe2+
Mn2+
Co2+
1
2
3
M-TrPP-aminoPS
R = C6H5; Y1
Fe2+
Mn2+
4
5
M-TrPP-aminoPS
R = C6H5; Y2
Fe2+
Mn2+
Co2+
6
7
8
M-TrPP-BHA-PS
R = C6H5; Y1
Fe2+
Co2+
9
10
M-TrTolP-aminoSil
Co2+
Co2+
11
12
R = p-CH3-C6H4; Y1
M-TrPP-aminoSil
R = C6H5; Y3
FIG. 1
Structures and description of the studied catalysts (TrPP, triphenyl porphyrin; TrTolP, tritolyl
porphyrin)
Collect. Czech. Chem. Commun. 2011, Vol. 76, No. 10, pp. 1163–1175

1166
Zachař, Petkova, Sýkora, Král:
The mixture was left to react for 7 days at the laboratory temperature in a closed vial.
Then 1 µl of the supernatant (free of the catalyst) was injected into the gas chromatograph
for the analysis.
In the case of the reaction kinetic study the supernatant in the reaction vial was sampled
at specific times for a given time period.
The identification of the products in the reaction mixture was carried out by means of
GC-MS with a Hewlett–Packard 5890 gas chromatograph equipped with a quadrupole mass
detector 5971A (electron ionization at 70 eV), a capillary column used was BPX5 (25 m,
220 µm, film thickness 0.25 µm) and carrier gas, helium.
a
b
FIG. 2
Product distribution for oxygenation of 1-hexene with catalyst No. 12. GC-MS analysis, injec-
tion at: 250 °C (a), 150 °C (b)
Collect. Czech. Chem. Commun. 2011, Vol. 76, No. 10, pp. 1163–1175

Oxygenation of α-Olefins by Metalloporphyrin Catalyst
1167
Quantitative measurements were accomplished with a GC instrument Varian 3350 with
a flame ionization detector, a capillary column was DB5 (20 m, 320 µm, film thickness
0.22 µm) and carrier gas, hydrogen.
At the beginning of our work a suitable chromatographic method allowing for a reliable
monitoring of the progress of the reactions studied had to be developed. We found previ-
ously that the GC injection temperature had an essential influence on the appearance of the
chromatogram and that this effect was independent on the presence/absence of the oxidant
(t-BuOOH) in the injected sample13. The injection was carried out in the temperature range
from 150 to 250 °C. The GC chromatograms obtained at 150 and 250 °C, respectively, are
shown in Fig. 2.
As can be clearly seen, two dominant peaks found at 150 °C and marked X1 and X2a+X2b
nearly vanished at 250 °C (Fig. 2). In agreement with our previously published data13 we
found that three compounds X1, X2a and X2b are thermally labile primary reaction products
of the oxygenation and as such they represented precursors for consecutively forming sec-
ondary products identified in Fig. 3. Thus, it was determined experimentally that the injec-
tion temperatures above 150 °C led to a significant degradation of the primary reaction
products X1, X2a and X2b, and for that reason in all following experiments the injection
temperature 150 °C was employed. It should be also mentioned that in some cases depend-
ing on chromatographic conditions and the composition of the injected sample the com-
pounds X2a and X2b could be partially resolved (Fig. 3).
Our effort to identify unknowns X1, X2a and X2b led into the proposal of their tentative
structures (Fig. 4).
For the interpretation of our measured MS spectra for X1 and X2a+X2b we inspired by
ref.17, where a similar structure was proposed and verified by NMR spectra. Based on that,
we proposed structures for X1 and X2a+X2b. We did not have any corresponding standards
FIG. 3
Product distribution for oxygenation of 1-hexene (catalyst No. 5, after 7 days of the reaction).
GC-MS analysis, injection at 150 °C:
1 dichlormethane, 2 1-hexene, 3 pyridine,
4 1-hexen-3-one, 5 1-hexen-3-ol, 6 butyloxirane, 7 2-hexenal, 8 X1, 9 X2a, 10 X2b
Collect. Czech. Chem. Commun. 2011, Vol. 76, No. 10, pp. 1163–1175

1168
Zachař, Petkova, Sýkora, Král:
or tabulated MS. Thus, to further support our proposal we took advantage of the mass frag-
ment prediction software MassFrontier, ver. 4 (HighChem, USA), for fragment modeling.
The thorough comparison of the software generated and the measured MS spectra (Fig. 5)
showed a good agreement with the proposed structures for the compound X1, X2a and X2b
.
+
The spectra of all three compounds contain characteristic ions m/z 57 C4H9 (tert-butyl),
•+
+
67 C5H7+, 73 C3H5O2+, 82 C6H10 (the loss of t-BuOOH from the molecule), and 83 C6H11
(the loss of t-BuOO• from the molecule). In the spectra of X2a and X2b is present ion m/z
69 C5H9+, which, in agreement with the MassFrontier’s prediction, is missing in the MS
spectrum of X1 (Fig. 5).
RESULTS AND DISCUSSION
The majority of our catalysts were based on TrPP skeleton. The activity of
these catalysts was expressed as a sum of peak areas of the primary reaction
products X1 and X2 in % with respect to the peak area of pyridine used as
the internal standard (Fig. 6). The efficiency of the catalysts was not inten-
tionally expressed in % of the substrate (1-hexene) conversion because of
its high volatility leading to losses during the manipulation with the sam-
ples and, consequently to inaccuracy in data evaluation.
It follows from Fig. 6 that the studied M-TrPP-based catalysts can be clas-
sified into three groups according to their catalytic activity (based on the
H2C CH CH2 CH2 CH2 CH3
X1
H2C CH HC
CH2 CH2 CH3
OOtBu
tBuOO
CH2 CH2 CH2 CH3
C
C
H
H
X2a
or
tBuOO
CH2
CH2 CH2 CH3
C
C
H
H
tBuOO
H
C
C
H
CH2 CH2 CH2 CH3
X2b
or
H
tBuOO CH2
C
C
H
CH2 CH2 CH3
FIG. 4
Conversion scheme indicating the tentative structures of the major primary products
Collect. Czech. Chem. Commun. 2011, Vol. 76, No. 10, pp. 1163–1175

Oxygenation of α-Olefins by Metalloporphyrin Catalyst
1169
a
b
c
FIG. 5
Measured mass spectra of compounds X1 (a), X2a (b) and X2b (c); M = 172
Collect. Czech. Chem. Commun. 2011, Vol. 76, No. 10, pp. 1163–1175

1170
Zachař, Petkova, Sýkora, Král:
results obtained after 7 days of the reaction): (i) active catalysts (5, 6; the
sum of the primary product peak areas is above 200% of the internal stan-
dard peak area), (ii) medium active catalysts (1, 3, 8–10; the sum of the pri-
mary product peak areas is 50–120% of the internal standard peak area),
and (iii) inactive catalysts (2, 4, 7; the sum of the primary product peak
areas is less than 50% of the internal standard peak area, which is roughly
equivalent to the result obtained in blank experiment with no catalyst pres-
ent in the system).
Generally, the catalysts with Fe2+ as a central ion are usually slightly
more active than the catalysts with Mn2+ or Co2+ as a central ion with one
exception for the Mn2+-TrPP-Y1-aminoPS catalyst (5), which demonstrated
the highest activity followed by the Fe2+-TrPP-Y2-aminoPS catalyst (6).
Cobalt as a central atom provided relatively low efficiencies independ-
ently on the structure of the catalyst (Table I).
Next experiments with selected catalysts were focused on the reaction ki-
netics, i.e. the rate of X1, X2 and 2-hexenal formation and the rate of the
decrease of t-BuOOH concentration in 1, 5, 9, and 12 days of the catalysis
at the laboratory temperature (Table II).
Table II shows that after 12 days even in the blank experiment with no
catalyst present in the system some decrease in t-BuOOH takes place and
a low amount of the products is formed.
FIG. 6
The efficiency of the studied M-TrPP-based catalysts immobilized on aminopropyl-silica gel
(aminoSil), aminopolystyrene (aminoPS), and benzhydrylamine polystyrene (BHA-PS) used for
1-hexene oxygenation with t-BuOOH expressed as a peak areas of compounds X1+X2 in % of the
peak area of the internal standard (pyridine) obtained after 7 days of the reaction, data not
measured
Collect. Czech. Chem. Commun. 2011, Vol. 76, No. 10, pp. 1163–1175

Oxygenation of α-Olefins by Metalloporphyrin Catalyst
1171
The decrease in t-BuOOH amount is partially caused by its spontaneous
decomposition to 1-butanol, which was proven in the reaction mixture af-
ter long standing. Limited and slow decrease in the oxidant amount and at
the same time the creation of low amount of the products was observed in
experiments with the catalysts Mn2+-TrPP-Y1-aminoSil (2) and Fe2+-TrPP-
Y1-aminoPS (4), the inactive catalysts. Slightly better results were obtained
for the catalysts Fe2+-TrPP-Y1-BHA-PS (9) and Co2+-TrPP-Y3-aminoSil (12),
where a moderate decrease of t-BuOOH and the production of certain
amount of the primary products X1 and X2 were reached immediately in
the first day of the experiment. The best results were obtained with the
catalysts Mn2+-TrPP-Y1-aminoPS (5) and Fe2+-TrPP-Y2-aminoPS (6), which
proved their activity in previous experiments (vide supra). In the case of the
catalysts (5) and (6) a fast and pronounced decrease of the oxidant concen-
tration in the first day of the reaction was registered. Correspondingly, the
primary products X1 and X2 were quickly formed in the first day of the re-
action and then their concentration decreased in the following time, which
was in contrast to the secondary product 2-hexenal, which amount was
slowly and continuously increasing during 12 days.
Finally, the influence of the amount of the oxidant t-BuOOH on the
progress of the reaction for two catalysts, one active (Mn2+-TrPP-Y1-
aminoPS (5)) and the other inactive (Fe2+-TrPP-Y1-aminoPS (4)), differing
by the central atom only was studied (Table III).
As can be seen from Table III in the case of the inactive catalyst
Fe2+-TrPP-Y1-aminoPS (4) the increasing amount of the oxidant in the reac-
tion mixture had practically no influence on the amount of the formed
products X1 and X2 even for a long reaction time. A significant amount of
the oxidant remained in the system unreaceted even after 11 days. On the
other hand, for the active catalyst Mn2+-TrPP-Y1-aminoPS (5) the situation
was quite different. The amount of the primary reaction products X1 and
X2 was strongly dependent on the amount of the oxidant in the system.
TABLE I
Efficienciesa of the Co2+-metalloporphyrin catalysts
TrPP-Y1-aminoSil
(catalyst No. 3)
TrPP-Y2-aminoPS
(catalyst No. 8)
TrPP-Y1-BHA-PS
(catalyst No. 10)
TrTolP-Y1-aminoSil TrPP-Y2-aminoSil
(catalyst No. 11)
(catalyst No. 12)
50
80
50
80
80
a
The efficiency of the catalysts is expressed as a sum of peak areas of X1+X2 in % of peak
area of the internal standard (pyridine). Data obtained in the first day of the experiment.
The measurements were carried out after 7 days of the reaction.
Collect. Czech. Chem. Commun. 2011, Vol. 76, No. 10, pp. 1163–1175

1172
Zachař, Petkova, Sýkora, Král:
Collect. Czech. Chem. Commun. 2011, Vol. 76, No. 10, pp. 1163–1175

Oxygenation of α-Olefins by Metalloporphyrin Catalyst
1173
Collect. Czech. Chem. Commun. 2011, Vol. 76, No. 10, pp. 1163–1175

1174
Zachař, Petkova, Sýkora, Král:
Doubling the amount of t-BuOOH provided about twice as much of the oxi-
dation products X1 and X2. However, a further increase in t-BuOOH con-
centration led to a moderate increase of the oxidation products only and
this trend was independent on the reaction time. This finding was in a
good agreement with a fast decrease in t-BuOOH concentration in the first
day in the case that its amount in the system was doubled, and on the con-
trary with its only moderate concentration decrease when its amount in the
system was ten times higher.
CONCLUSIONS
The oxygenation of α-olefins using t-BuOOH was studied with the set of
twelve porphyrin-based catalysts with various central metal ions, porphyrin
ring substitution, type of linkage to the support, and the support itself,
all of which mimicking cytochrome P-450. It was demonstrated that effi-
ciency of the individual catalysts depended on a combination of all the
above mentioned parameters and differed significantly even for the cata-
lysts of the similar structures. The most active catalysts were Mn2+-TrPP-Y1-
aminoPS (5) and Fe2+-TrPP-Y2-aminoPS (6).
The primary reaction products were identified as the compounds contain-
ing peroxidic –OOt-Bu group, which slowly converted to common oxygena-
tion products i.e., epoxides, alcohols, aldehydes and ketones.
The conversion of the primary to the secondary products was slow at the
laboratory temperature. An increase of the reaction temperature would
accelerate the conversion process but then polymerization reactions could
take place as it was demonstrated when the GC injection temperature was
set above 150 °C.
Financial support from the Ministry of Education, Youth and Sports of the Czech Republic (grant
No. MSM 6046137307) is gratefully acknowledged.
REFERENCES
1. de Sousa A. N., de Carvalho M. E. M. D., Idemori Y. M.: J. Mol. Catal. A: Chem. 2001,
169, 1.
2. Nam W., Kim I., Lim M. H., Choi H. J., Lee J. S., Jang H. G.: Chem. Eur. J. 2002, 8, 2067.
3. Rose E., Andrioletti B., Zrig S., Quelquejeu-Etheve M.: Chem. Soc. Rev. 2005, 34, 573.
4. Liu Y., Zhang H.-J., Lu Y., Cai Y.-Q., Liu X.-L.: Green Chem. 2007, 9, 1114.
5. Stephenson N. A., Bell A. T.: J. Mol. Catal. A: Chem. 2007, 275, 54.
6. Atkinson S. T., Brady S. P., James J. P., Nolan K. B.: Pure Appl. Chem. 1995, 67, 1009.
7. Yang J., Weinberg R., Breslow R.: Chem. Commmun. 2000, 531.
Collect. Czech. Chem. Commun. 2011, Vol. 76, No. 10, pp. 1163–1175

Oxygenation of α-Olefins by Metalloporphyrin Catalyst
1175
8. Breslow R., Gabriele B., Yang J.: Tetrahedron Lett. 1998, 39, 2887.
9. Cai Y., Liu Y., Lu Y., Gao G., He M.: Catal. Lett. 2008, 124, 334.
10. Yang J., Gabriele B., Belvedere S., Huang Y., Breslow R.: J. Org. Chem. 2002, 67, 5057.
11. Yu X. Q., Huang J. S., Yu W. Y., Che C. M.: J. Am. Chem. Soc. 2000, 122, 5337.
12. Nestler O., Severin K.: Org. Lett. 2001, 3, 3907.
13. Pulkkinen J., Zachar P., Borek V. A., Laatikainen R., Král V.: J. Mol. Catal. A: Chem. 2004,
219, 21.
14. Pulkkinen J. T., Laatikainen R., Ahlgren M. J., Perakyla M., Vepsalainen J. J.: J. Chem.
Soc., Perkin Trans. 2 2000, 777.
15. Mammana A., Asakawa T., Bitsch-Jensen K., Wolfe A., Chaturantabut S., Otani Y., Li X.,
Li Z., Nakanishi K., Balaz M., Ellestad G., Berova N.: Bioorg. Med. Chem. 2008, 16, 6544.
16. Kubát P., Lang K., Anzenbacher P., Jursíková K., Král V., Ehrenberg B.: J. Chem. Soc.,
Perkin Trans. 1 2000, 933.
17. Griesbeck A. G., Bartoschek A.: Chem. Commun. 2002, 1594.
Collect. Czech. Chem. Commun. 2011, Vol. 76, No. 10, pp. 1163–1175

Copyright of Collection of Czechoslovak Chemical Communications is the property of Institute of Organic
Chemistry & Biochemistry, Academy of Sciences of the Czech Republic, v.v.i. and its content may not be
copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written
permission. However, users may print, download, or email articles for individual use.