Beruflich Dokumente
Kultur Dokumente
Corresponding author.
C
for 4h with a heating rate of 1
Cmin
1
. The bimetallic oxide
precursor was prepared following a procedure similar to one
reported by Korlann et al. [23] by adding an aqueous solution of
cobalt (II) nitrate hexahydrate (Acros Organics, Co(NO
3
)
2
6H
2
O,
99%, 16.472g, 100mL of H
2
O) dropwise to a second solution con-
sisting of AHM dissolved in deionized water (10g, 80mL of H
2
O)
under stirring. This was done to obtain a 1:1 Co/Mo mole ratio. The
solution was vaporized to dryness in a water bath, and a purple,
solid precipitate was collected by vacuum ltration. The prod-
uct was rinsed with two washings of water followed by a single
washing with ethanol. The solid was dried overnight at 120
C and
calcined in air at 500
Cmin
1
.
Molybdenum nitrides were prepared by loading a 10mm i.d.
quartz reactor tube with about 4g of the oxidic precursor, while
passing NH
3
(MathersonTri-Gas, NH
3
, 99.99%), or a N
2
/H
2
(Mather-
son Tri-Gas, H
2
, 99.99%) mixture over the sample [6,7]. The reactor
was purged with nitrogen (BOC Gases, N
2
, Grade 5) prior to nitrida-
tionfor 30min andswitchedtoNH
3
or H
2
/N
2
mixtures (H
2
/N
2
=1/5,
v/v). Thetemperaturewas linearlyincreasedfromambient temper-
ature to 300
Cmin
1
), then from300
C to
500
Cmin
1
, and from500
C to 700
C
by a heating rate of 2
Cmin
1
. Then, the temperature was main-
tained at 700
and a
counting time of 1s per step for bulk metal oxides and 10s per
step for bulk nitrides. Identication of the phases was achieved by
reference to JCPDS diffraction le data.
BET surface area: Nitrogen adsorption/desorption isotherms
were measured at 196
C for 2h.
Elemental analysis: Metal andnitrogencontents intheMonitride
samples were determined by the Analytical Laboratory of the
Department of Plant, Soil and Environmental Sciences at the Uni-
versity of Maine or by Galbraith Laboratory using ICP-AES for the
metal analysis and combustion method for the nitrogen analysis.
2.3. Catalyst activities
Guaiacol HDO activity measurements were carried out in a
300mL stainless steel batch autoclave (Parr Model 4841), at 300
C
and under a hydrogen pressure of 5MPa. Prior to catalytic test-
ing, the passivated samples were activated ex situ under a ow
of H
2
(AGA Chile, 99.99%), at a ow rate of 60mL min
1
and
a temperature of 450
C (heating rate of 10
Cmin
1
). Approxi-
mately 0.25g of pre-treated catalyst was added to the reactor
charged with 80mL decalin (Merck, 99.5%) as the solvent, 2.53mL
guaiacol (0.232mol L
1
, Merck, 99%), and 700L of hexadecane
(Merck, 99%) as internal standard for quantitative GC analysis. The
sealed reactor was ushed with nitrogen (AGA Chile, Grade 5) for
30min to evacuate air from the system. While continuously stir-
ring the mixture, the reactor was heated to 300
C under N
2
. Once
the reaction temperature was attained, N
2
was replaced with H
2
and then pressurized to 5MPa. This pressure was maintained for
the entire duration of the experiment. Samples were periodically
withdrawn from the reactor after the sampling line was purged
with a small amount of the reactant mixture. The liquid samples
were analyzed by a Perkin Elmer (Clarus 400) gas chromatograph
equipped with a Flame Ionization Detector (FID) and a CP-Sil 5
CB column (Agilent, 30m0.53mm1.0mlmthickness). The
product distributions were identied by their column retention
time in comparison with available standards. The catalytic activ-
ity was expressed by the initial reaction rate which was calculated
from the slopes of the conversion vs. reaction time plot, and by
the intrinsic activity (i.e. the reaction rates per unit area per unit
Mo content). The phenol/catechol ratios were determined at 10%
conversion of guaiacol to avoid internal and external mass transfer
limitations.
3. Results and discussion
3.1. Catalysts preparation and characterization
Wide-angle XRD patterns for bulk molybdates, pre-and post
nitridation, are shown in Fig. 1. The patterns obtained were com-
pared to JCPDS data les for phase identication (MoO
3
: 35-0609,
-Mo
2
N: 25-1366, -Mo
2
N
0.78
: 25-1368, MoO
2
: 32-0671, Mo: 42-
1120, -CoMoO
4
: 021-0868, Co
3
Mo
3
N: 04-008-1301). The MoO
3
phase was observed for the bulk Mo oxide precursor samples.
The result presented indicated a complete thermal decomposition
in air of AHM to MoO
3
. XRD patterns of the bulk nitrides indi-
cated that the nitridation procedure affected the phases present
80 I.T. Ghampson et al. / Applied Catalysis A: General 413414 (2012) 7884
Table 1
Crystallite sizes of bulk different phases present in bulk nitride catalysts.
Sample Space velocity (h
1
) Crystallite size (nm)
-Mo
2
N (200) -Mo
2
N
0.78
(200) MoO
2
(111) Mo (110) Co
3
Mo
3
N (411)
MoN-A-i 19 5.7 43.3
MoN-A-ii 29 9.3
MoN-NH-i 19 29.3
MoN-NH-ii 29 6.2 78.6 13.8
CoMoN-A 29 7.1 21.2
in the nal metal nitride catalyst. XRD patterns of the MoN-A-i
and MoN-A-ii catalysts revealed characteristic peaks of -Mo
2
N
(2 =37.25, 43.37, 63.11, 75.65, and 79.69). However, the diffrac-
tion patterns of the MoN-A-i catalyst also exhibited crystalline
MoO
2
peaks (2 =26.03, 53.45, and 66.63), indicating incomplete
nitridation. The XRD patterns of the MoN-NH-i catalyst revealed
peaks that corresponded to -Mo
2
N
0.78
(2 =37.57, 43.09, 45.03,
62.45, 64.05, 75.35, 78.17, and 80.23). The -Mo
2
N
0.78
phase is a
tetragonal Mo nitride phase which results fromthe transformation
of -Mo
2
N [24]. The formation of the -Mo
2
N
0.78
phase in MoN-
NH-i was consistent with work published by Gong et al. [25] who
reported the synthesis of -Mo
2
N
0.78
under similar synthesis con-
ditions. Nagai and co-workers [26] also reported the synthesis of
bulk -Mo
2
N
0.78
when they cooled NH
3
-nitrided products in He
to roomtemperature prior to passivation. However, the diffraction
pattern of the MoN-NH-ii catalyst revealed the additional presence
of MoO
2
(2 =26.03, 31.79, 37.39, 41.45, 41.97, 49.51, 53.11, 53.57,
53.99, 60.57, 66.67, 67.75, 72.53, and 78.71), Mo metal (2 =40.51
and 73.41), and -Mo
2
N(2 =37.09, 43.45, 63.23, 75.45, and 79.75)
Fig. 1. XRD of bulk oxides and nitrides.
phases. According to Choi et al. [27], MoO
2
and Mo metal are
potential reaction intermediates to produce low surface area Mo
nitride material. Thus, the XRD result of MoN-NH-ii suggests
incomplete conversion of MoO
3
to -Mo
2
N.
Thermal decomposition of the cobalt molybdate precursor in air
formedMoO
3
and-CoMoO
4
phases, as observedbyXRD. Although
the theoretical Co/Mo ratio for the CoMo oxide is 1, there was no
evidence by XRDof Co
3
O
4
. Peak locations and relative intensities of
the XRD pattern of the CoMoN-A catalyst revealed the presence of
Co
3
Mo
3
N(2 =32.29, 35.51, 40.05, 42.57, 46.53, 49.49, 55.21, 59.83,
64.93, 69.79, and 72.77) and -Mo
2
N (2 =37.37, 43.37, 63.09,
75.67, and 79.51). Since monometallic and bimetallic oxides are
precursors for monometallic and bimetallic nitrides respectively,
the presence of both -Mo
2
N and Co
3
Mo
3
N in the CoMoN-A cata-
lyst could have been due to the direct transformation of MoO
3
and
-CoMoO
4
during temperature-programmed reaction with NH
3
.
Crystallite sizes of the different molybdenum crystal phases of
bulk Mo nitrides calculated from peak widths using the Scherrer
equation are presented in Table 1. The full width at half-maximum
(FWHM) data was corrected for instrumental contribution to the
line broadening by using Warrens formula (
2
= B
2
b
2
0
), where
B is FWHM of the measured peak, b
0
is the FWHM of LaB
6
stan-
dard (NIST SRM 660) interpolated to the angle of interest, and
is the corrected FWHM. The calculated particle diameter of
Mo
2
N in the samples ranged from 5.7 to 9.3nm, while the MoO
2
present in MoN-A-i and MoN-NH-ii had much large crystallite sizes
of 43.3nm and 78.6nm, respectively. Crystallite sizes of 21.2nm
were calculated for the Co
3
Mo
3
N phase present in CoMoN-A
catalyst.
The elemental analysis of passivated Mo nitrides along with
phases identied by XRD is shown in Table 2. The catalyst contain-
ing the highest atomic N/Mo was also comprised of single phase
-Mo
2
N. The catalyst containingthe lowest atomic N/Mocontained
oxides and molybdenum metal in addition to the metal nitride.
While the theoretical N/Mo is 0.5 for -Mo
2
N, it is plausible that
excess N could reside in interstitial sites and defects like grain
boundaries.
The BET surface areas of the passivated nitride catalysts are
summarized in Table 3. Comparison of bulk oxidic precursors
and corresponding nitrides shows a signicant increase in tex-
tural properties after nitridation. The BET surface area of the
bulk Mo oxide precursor was 2m
2
g
1
, while the corresponding
nitrides ranged from 9 to 115m
2
g
1
. The BET surface area of the
bulk CoMo oxide precursor and the corresponding bulk CoMo
nitride were 17m
2
g
1
and 50m
2
g
1
, respectively. The signicant
Table 2
Chemical composition bulk Mo nitrides.
Catalyst Elemental analysis (wt %) N/Mo atomic ratio
Mo N
MoN-A-i 67.10 6.24 0.64
MoN-A-ii 76.20 9.62 0.86
MoN-NH-i 79.40 6.63 0.57
MoN-NH-ii 93.20 3.16 0.23
I.T. Ghampson et al. / Applied Catalysis A: General 413414 (2012) 7884 81
Table 3
Catalytic activity of Unsupported Mo nitrides.
Catalyst SBET (m
2
g
1
) Reaction rate Intrinsic activity
10
6
mol g
1
catalyst
s
1
Reaction rate/Mo atoms Reaction rate/Mo atomsSBET
10
4
molec.Moat
1
s
1
10
5
gmolec.Moat
1
m
2
s
MoN-A-i 115 4.98 7.12 0.62
MoN-A-ii 40 6.23 7.84 1.96
MoN-NH-i 9 1.54 1.85 2.09
MoN-NH-ii 10 1.81 1.86 1.86
CoMoN-A 50 5.21
increase in surface area upon nitridation may be due to an increase
in density during the transformation of the oxide to the nitride,
which caused the evolution of cracks [27]. Differences in surface
areas of individual catalysts can be attributed to differences in syn-
thesis parameters including nitridation gas, temperature prole,
and gas space velocity [27]. A model, based on reaction pathways
fromthe oxide tonitride, has beenproposedby Choi et al. toexplain
the effect of synthesis conditions on nitride surface area [27].
3.2. Catalyst activity measurements
The evolution of the transformation of guaiacol and the yield
of products for the unsupported nitride catalysts are shown in
Fig. 2. The products formed from the reaction over all the cata-
lysts were similar, with guaiacol yielding principally phenol. Minor
amounts of catechol, benzene, cyclohexene and cyclohexane were
also detected, with the deoxygenated products appearing at longer
reaction times. The variation of the reactant and products with
time can be explained by the schematic in Fig. 3. On the basis
of this reaction scheme, the majority of the reaction proceeded
through the direct demethoxylation route (DMO) to form phenol
via hydrogenolysis of the aromatic carbonoxygen bond. The sec-
ond reaction pathway involved hydrogenolysis of the OCH
3
bond
(demethylation route, DME) on guaiacol to form catechol, which
was then transformed to phenol. Continuous production of cate-
chol, as evident in Fig. 2, indicated that the conversion of catechol
to phenol was not prominent over the time scale used in this study,
consistent with our earlier work [4]. Therefore, the demethylation
pathway appeared to be the minor route to phenol production.
The reactionrate results of the nitride catalysts for guaiacol con-
version are summarized in Table 3. Intrinsic activities normalized
tomolybdenumcontent andS
BET
are alsoshowninTable 3. It canbe
seen fromTable 3 that the MoN-A-ii and MoN-NH-ii catalysts dis-
played higher intrinsic activities and total reaction rates than the
MoN-A-i and MoN-NH-i catalysts, respectively. Also, Table 3 shows
that the trend obtained for the intrinsic activities was similar to the
reaction rates, suggesting that the conversion of guaiacol is inde-
pendent of the BET surface area and Mo content of the catalysts.
In other words, the changes in conversion of guaiacol observed in
Table 3 are not due to an increase in the active sites produced by
an increase in the specic surface area or Mo content. These results
suggest that the conversion of guaiacol depends on the different
phases present in the unsupported catalysts.
The catalyst with the highest reaction rate for guaiacol con-
version was MoN-A-ii. This catalyst had the highest atomic N/Mo
and consisted of single-phase -Mo
2
N, as determined by XRD.
The MoN-A-i catalyst contained predominantly -Mo
2
N and some
unconverted molybdenum oxide, which resulted in a slightly
lower reaction rate. The MoN-NH-i catalyst contained -Mo
2
N
0.78
suggesting that this phase of molybdenum nitride is less active
than the gamma phase for hydrodeoxygenation. The MoN-NH-ii
catalyst contained signicant impurities of molybdenum oxide
and molybdenum metal which could have resulted in a relatively
lower reaction rate. Our data is consistent with Nagai et al. [26]
who demonstrated that the CNhydrogenolysis activity decreased
according to the order -Mo
2
N>-Mo
2
N
0.78
Mo metal.
The catalytic performance of the CoMo bimetallic nitride cat-
alyst relative to the Mo monometallic nitride indicated that the
addition of Co did not enhance the activity of the catalyst. Our
observation is inconsistent with the positive effects of Co pro-
moters on bulk Mo nitrides reported for HDS and HDN catalysis
[22,28]. This result may be attributed to incomplete formation of
the bimetallic nitride, Co
3
Mo
3
N, as evident by the presence of
Mo
2
N. Particle sizes calculated from the XRD data in Table 1 indi-
cated that Co
3
Mo
3
N particles were three times larger than Mo
2
N
particles in the CoMoN-A catalyst. It is possible that the Mo
2
N par-
ticles were embeddedwithinCo
3
Mo
3
Ncrystallites, andhence, only
a portion of the bimetallic nitride was accessible for catalysis.
However, the additionof Coincreasedthe yieldof deoxygenated
products such as benzene, cyclohexene and cyclohexane as illus-
tratedinFig. 4. Fig. 4Ashows that appreciable quantities of different
deoxygenated products such as benzene, cyclohexene and cyclo-
hexane were formed during reaction over the CoMoN-A catalyst.
Fig. 4B shows that cyclohexene was the principal deoxygenated
product formed during reactions involving the MoN-A-ii catalyst.
The result in Fig. 4B indicated that hydrogen is conserved and
double bond hydrogenation is minimized over MoN-A-ii catalyst
relative to data in Fig. 4A for CoMoN-A catalyst. A plot of the
total yield of deoxygenated products vs. conversion (Fig. 4C) clearly
shows a higher yieldfor reactions catalyzedby CoMoN-Acompared
to that catalyzed by MoN-A-ii. Thus, it appears that the addition of
cobalt creatednew, or modied, active sites relatedtothe Co
3
Mo
3
N
phase which enhanced the elimination of oxygen from phenol to
formbenzene, as well as increasethehydrogenationof thearomatic
ring of benzene to form cyclohexene and cyclohexane. Fig. 4 also
shows that the yields of the deoxygenated compounds produced
by both catalysts increased at high conversions. The results indi-
cate that further studies to optimize the Co composition in order to
produce single-phase Co
3
Mo
3
N in the CoMoN catalyst, as well
as studies focusing on the nature of the surface of the catalyst for
HDO catalysis are warranted.
The selectivity in the transformation of guaiacol is expressed
in terms of the phenol/catechol (Phe/Cat) ratio, calculated at 10%
guaiacol conversion, and shown in Fig. 5. The results in Fig. 5 show
that, for most of the experiments, except for reactions over the
MoN-A-ii catalyst, the Phe/Cat ratios calculated were similar, sug-
gesting that the active sites were the same. The yield of formation
of phenol at 10% guaiacol conversion for all the catalysts ranged
from 8.79.7%, while the production of catechol by all the cata-
lysts ranged from0.2 to 1%. For the MoN-A-ii catalyst, 9.7% phenol
and 0.2% catechol were produced at 10% guaiacol conversion which
corresponded to the highest Phe/Cat ratio. The low value of cate-
chol produced resulted in a very high Phe/Cat ratio of 499. The
lowest Phe/Cat ratio was obtained for the MoN-A-i catalyst which
produced 8.9% phenol and 1% catechol at 10% guaiacol conversion.
The reason for the surprisingly high Phe/Cat ratio of the MoN-A-ii
catalyst could be attributed to the presence of single-phase Mo
2
N,
with high activity for the direct demethoxylation route. The other
82 I.T. Ghampson et al. / Applied Catalysis A: General 413414 (2012) 7884
Fig. 2. Variation of the transformation of guaiacol and the yield of products with time.
monometallic catalysts contained other phases, such as MoO
2
,
-Mo
2
N
0.78
or Mo. The presence of the MoO
2
phase decreased
the ability of the active Mo
2
N to rapidly remove the oxygen atom,
while the presence of Mo metal increased the hydrogenation char-
acter of the catalysts and hence channeled some of the reaction
through the demethylation route to formcatechol. A closer look at
the Phe/Cat ratio differences between the catalysts agrees with this
assertion. The difference inthe Phe/Cat ratiobetweenthe MoN-NH-
i and the MoN-NH-ii catalysts indicated that the MoN-NH-i catalyst
containing only the -Mo
2
N
0.78
phase was more selective towards
Fig. 3. Hydrodeoxygenation pathway of guaiacol adapted fromFerrari et al. [30].
I.T. Ghampson et al. / Applied Catalysis A: General 413414 (2012) 7884 83
Fig. 4. Deoxygenated product distribution vs. conversion of guaiacol for (A) CoMoN-A, (B) MoN-A-ii, and (C) total yield of deoxygenated products for CoMoN-A and MoN-A-ii
catalysts.
phenol thanthe catalyst containing the nitride, Mo metal andMoO
2
phases. Despite these differences, the phenol/catechol ratios for all
the catalysts tested which ranged from 9 to 48 were substantially
higher than values reported for sulded catalysts [29]. The higher
amount of phenol produced at low conversion has implications in
terms of the prevention of excessive hydrogen consumption [3].
Additionally, the rapid transformation of guaiacol to phenol will
lead to less formation of coke due to the propensity of guaiacol
and catechol to formpolycondensation products and coke [3]. The
reduction in hydrogen consumption and negligible coke forma-
tion will be critical for the ultimate commercial success of an HDO
catalyst.
Fig. 5. Phenol/catechol ratio for bulk metal nitride catalysts.
4. Conclusion
All the metal nitride catalysts prepared and characterized in
this study demonstrated high activity and rapid production of
signicant amounts of phenol, indicating that the transformation
of guaiacol mostly proceeded through the direct demethoxyla-
tion route, bypassing the formation of catechol. The activity trends
indicated that the Mo nitride catalyst with the highest degree of
nitridation contained only the -Mo
2
N phase and was the most
active. The presence of another Mo nitride phase (-Mo
2
N
0.78
)
and other molybdenum compounds such as MoO
2
and Mo metal,
decreased the activities of the catalysts. Furthermore, comparison
of phenol/catechol ratios indicated that the catalysts possessing
only the -Mo
2
N phase displayed the highest selectivity towards
the production of phenol. The addition of Co did not enhance the
overall activity of the Mo nitride catalysts. This may have been
due to the lack of single phase Co
3
Mo
3
N. However, the bimetallic
nitrideproducedhigher yields of deoxygenatedproducts compared
to monometallic nitride prepared using the same synthesis condi-
tion, and this result was also attributed to the presence of Co
3
Mo
3
N
particles.
Acknowledgments
The authors acknowledge the nancial support of DOE Epscor
Grant #DE-FG02-07ER46373 and nancial support from CONICYT
Chile, projects PFB-27, PIA-ACT-130 and FONDECYT No 1100512
grants. I.T. Ghampson is indebted to NSF Career Award 0547103
for sponsoring a trip to the Universidad de Concepcin. The authors
alsogratefullyacknowledgethetechnical assistanceof NickHill and
Manuel Veliz.
84 I.T. Ghampson et al. / Applied Catalysis A: General 413414 (2012) 7884
References
[1] R.D. Perlack, L.L. Wright, A.F. Turhollow, R.L. Graham, B.J. Stokes, D.C. Erbach,
DOE/GO-102005-2135, Oak Ridge National Laboratory, Oak Ridge, Tennesse,
2005.
[2] http://ec.europa.eu/energy/energy policy/doc/03 renewable energy roadmap
en.pdf, 2007.
[3] J. Zakzeski, P.C.A. Bruijnincx, A.L. Jongerius, B.M. Weckhuysen, Chem. Rev. 110
(2010) 35523599.
[4] C. Seplveda, K. Leiva, R. Garca, L.R. Radovic, I.T. Ghampson, W.J. DeSisto, J.L.G.
Fierro, N. Escalona, Catal. Today 172 (2011) 232239.
[5] H.Y. Zhao, D. Li, P. Bui, S.T. Oyama, Appl. Catal. A: Gen. 391 (2011)
305310.
[6] L. Volpe, M. Boudart, J. Solid State Chem. 59 (1985) 332347.
[7] R.S. Wise, E.J. Markel, J. Catal. 145 (1994) 344355.
[8] D.J. Sajkowski, S.T. Oyama, Appl. Catal. A: Gen. 134 (1996) 339349.
[9] H. Abe, T.K. Cheung, A.T. Bell, Catal. Lett. 21 (1993) 1118.
[10] E.J. Markel, J.W. Van Zee, J. Catal. 126 (1990) 643657.
[11] S. Ramanathan, S.T. Oyama, J. Phys. Chem. 99 (1995) 1636516372.
[12] H. Abe, A.T. Bell, Catal. Lett. 18 (1993) 18.
[13] E. Furimsky, Appl. Catal. A: Gen. 199 (2000) 147190.
[14] V.N. Bui, D. Laurenti, P. Afanasiev, C. Geantet, Appl. Catal. B101(2011) 239245.
[15] V.N. Bui, D. Laurenti, P. Delichre, C. Geantet, Appl. Catal. B101 (2011) 246255.
[16] V.N. Bui, G. Toussaint, D. Laurenti, C. Mirodatos, C. Geantet, Catal. Today 143
(2009) 172178.
[17] A. Gutierrez, R.K. Kaila, M.L. Honkela, R. Slioor, A.O.I. Krause, Catal. Today 147
(2009) 239246.
[18] Y.-C. Lin, C.-L. Li, H.-P. Wan, H.-T. Lee, C.-F. Liu, Energy Fuels 25 (2011)
890896.
[19] T. Nimmanwudipong, R. Runnebaum, D. Block, B. Gates, Catal. Lett. 141 (2011)
779783.
[20] P.E. Ruiz, K. Leiva, R. Garcia, P. Reyes, J.L.G. Fierro, N. Escalona, Appl. Catal. A
384 (2010) 7883.
[21] Y.Q. Liu, C.G. Liu, G.H. Que, Energy Fuels 16 (2002) 531535.
[22] D.W. Kim, D.K. Lee, S.K. Ihm, Catal. Lett. 43 (1997) 9195.
[23] S. Korlann, B. Diaz, M.E. Bussell, Chem. Mater. 14 (2002) 40494058.
[24] Z.B. Wei, Q. Xin, P. Grange, B. Delmon, J. Catal. 168 (1997) 176182.
[25] S.W. Gong, H.K. Chen, W. Li, B.Q. Li, Appl. Catal. A: Gen. 279 (2005) 257261.
[26] M. Nagai, Y. Goto, O. Uchino, S. Omi, Catal. Today 43 (1998) 249259.
[27] J.G. Choi, R.L. Curl, L.T. Thompson, J. Catal. 146 (1994) 218227.
[28] Y. Li, Y. Zhang, R. Raval, C. Li, R. Zhai, Q. Xin, Catal. Lett. 48 (1997) 239245.
[29] A. Centeno, E. Laurent, B. Delmon, J. Catal. 154 (1995) 288298.
[30] M. Ferrari, B. Delmon, P. Grange, Carbon 40 (2002) 497511.