Selective Oxidative Cracking of n-Butane to Light Olefins over Hexagonal Boron Nitride with Limited Formation of COx

Article abstract of DOI:10.1002/cssc.201901663

In recent years, hexagonal boron nitride (hBN) has emerged as an unexpected catalyst for the oxidative dehydrogenation of alkanes. Here, the versatility of hBN was extended to alkane oxidative cracking chemistry by investigating the production of ethylene and propylene from n-butane. Cracking selectivity was primarily controlled by the ratio of n-butane to O₂ within the reactant feed. Under O₂-lean conditions, increasing temperature led to increased selectivity to ethylene and propylene and decreased selectivity to CO_x. In addition to surface-mediated chemistry, homogeneous gas-phase reactions likely contributed to the observed product distribution, and a reaction mechanism was proposed based on these observations. The catalyst showed good stability under oxidative cracking conditions for 100 h time-on-stream while maintaining high selectivity to ethylene and propylene.

Full text of DOI:10.1002/cssc.201901663

Accepted Article
Title: Selective Oxidative Cracking of n-Butane to Light Olefins over
Hexagonal Boron Nitride with Limited Formation of COx
Authors: William McDermott, Juan Venegas, and Ive Hermans
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemSusChem 10.1002/cssc.201901663
Link to VoR: http://dx.doi.org/10.1002/cssc.201901663
A Journal of
www.chemsuschem.org

10.1002/cssc.201901663
ChemSusChem
FULL PAPER
Selective Oxidative Cracking of n-Butane to Light Olefins over
Hexagonal Boron Nitride with Limited Formation of COx
William P. McDermott[a], Juan Venegas[b], and Ive Hermans*[a,b]
Abstract: In recent years, hexagonal boron nitride (hBN) emerged as
an unexpected catalyst for the oxidative dehydrogenation (ODH) of
alkanes. Here, we extend the versatility of hBN to alkane oxidative
cracking chemistry by investigating the production of ethylene and
propylene from n-butane. Cracking selectivity is primarily controlled
by the ratio of n-butane to O2 within the reactant feed. Under O2-lean
conditions, increasing temperature leads to increased selectivity to
ethylene and propylene and decreased selectivity to COx. In addition
to surface-mediated chemistry, homogeneous gas-phase reactions
likely contribute to the observed product distribution and a reaction
mechanism is proposed based upon our observations. The catalyst
shows good stability under oxidative cracking conditions for 100 hr
time-on-stream while maintaining high selectivity to ethylene and
propylene.
The oxidative cracking of higher hydrocarbons to light olefins has
the potential to be economically competitive with steam
cracking.[6] The exothermic oxidative reaction lends itself to lower
reaction temperatures than the endothermic nonoxidative steam
cracking. The heat generated during oxidative cracking may even
enable autothermal reactor operation, reducing heat input during
operation and thus decreasing fuel consumption.[7,8] The
incorporation of O2 within the feed also prevents the fast build-up
of coke deposition.
Many catalysts have been studied for the oxidative cracking of
hydrocarbons to light olefins such as Li-doped metal oxides,[9–12]
metal oxychlorides,[13] noble metal-coated monoliths,[7,8,14–16]
supported Co catalysts,[17–19] and supported Au catalysts.[20–22]
Many of these systems have substantial disadvantages, such as
poor catalyst stability or reaction conditions not largely different
from those employed in nonoxidative cracking. The most
significant drawback shared by all oxidative cracking catalysts
explored thus far has been the significant co-production of CO2,
which has been identified as main bottleneck in achieving success
in oxidative cracking technology.[6]
Introduction
Light olefins such as ethylene and propylene are two of the most
vital building blocks in the chemical industry as they are the
starting point for many (co-)polymers, oxygenates, and other bulk
chemicals. In 2017, the global market for ethylene and propylene
stood at 170 Mt and 130 Mt, respectively.[1] Historically, the
majority of light olefins have been produced through steam
cracking of petroleum-derived naphtha.[2] Despite the emergence
of light olefin production through fluid catalytic cracking (FCC) or
direct methods like propane dehydrogenation (PDH), steam
cracking remains the “workhorse” light olefin production
technology.[2–4] Steam cracker furnaces are typically operated in
the temperature range of 850–900˚C to provide the energy for the
pyrolysis of naphtha into lighter molecules. These operating
conditions result in process emissions of approximately 1.8–2.0 t
CO2/t ethylene from fuel consumption.[5] Moreover, high reaction
temperatures lead to coke deposition within the reactors which
must be periodically removed leading to significant reactor
downtimes. For these reasons, the development of processes
that require less energy and do not produce coke deposits could
lead to significant cost-savings in the production of light olefins
and is of interest to academic and industrial researchers.
Leveles et al. investigated the kinetics and mechanism of the
oxidative cracking of propane over a Li/MgO catalyst.[10] They
reported that the reactive oxygen species on the catalyst surface
enable the energetically-preferred secondary H-abstraction from
propane to form i-propyl radical and under nonoxidative
conditions, this radical can undergo unimolecular C-C cleavage to
form ethylene and CH3•. The CH3• radical acts as a radical chain
carrier and abstract H atoms from propane and form i-propyl and
n-propyl radicals. The n-propyl radical can decompose into
propylene and H•. Under oxidative conditions, the role of O2 is
two-fold. Beyond regenerating the active sites, the presence of O2
changes the radical pool in the gas phase. The reaction of O2 with
propyl radicals forms propylene and HOO•. H-abstraction by
HOO• from propane leads to propyl radicals and H2O2. The
decomposition of H2O2 to HO• (chain initiation) adds to the pool
of gas phase radical chain carriers. The production of COand CO2
(COx) mainly occurred through overoxidation pathways on the
surface of the catalyst.
In recent years, our group has shown that hexagonal boron nitride
(hBN), boron nitride nanotubes (BNNTs), and other boron-
containing catalysts exhibit high selectivity in the oxidative
dehydrogenation (ODH) of propane, n-butane, and isobutane[23–
25] In these studies, a variety of spectroscopic techniques such as
XPS, Raman, and IR revealed that after exposure to reaction
[a]
[b]
W. P. McDermott, Prof. Dr. I. Hermans
Department of Chemistry
University of Wisconsin-Madison
1101 University Avenue, Madison, WI 53706 (USA)
E-mail: hermans@chem.wisc.edu
Dr. J. M. Venegas, Prof. Dr. I. Hermans
Department of Chemical and Biological Engineering
University of Wisconsin-Madison
conditions,
these
boron-containing
materials
become
functionalized with oxygen at their surface. We recently employed
solid-state NMR to investigate boron nitride materials and
demonstrated that the oxidized phase on used boron nitride
catalysts can be denoted as B(OH)xO3-x (0£x£2), which contains
the active site for ODH.[26] During reaction kinetic studies of hBN-
1415 Engineering Drive, Madison, WI 53706 (USA)
Supporting information for this article is given via a link at the end of
the document
This article is protected by copyright. All rights reserved.

10.1002/cssc.201901663
ChemSusChem
FULL PAPER
catalyzed ODH, it was revealed that the O2 partial pressure (PO2
)
greatly influenced the observed product distribution.[24]
Performing the oxidation of n-butane at dilute PO2 results in a
product distribution that consists mainly of cracking products (i.e.
ethylene, propylene). Increasing PO2 resulted in increased
selectivity to the primary ODH products, 1-butene, cis-2-butene
and trans-2-butene.
Thus far, reports from our group and others have primarily
discussed the role of the surface of hBN in the mechanism of
alkane ODH.[23,24,27–29] These contributions demonstrated the
importance of O2 activation on the hBN surface and that the
resulting reactive surface oxygen species likely initiate
dehydrogenation through H-abstraction from the gas phase
alkane. In a recent report evaluating mass and heat transfer
phenomena in the hBN-catalyzed ODH of propane,[30] we found
that increasing catalyst dilution led to a significant increase in the
rate of alkane consumption at a constant mass of catalyst,
indicating that reactions in the gas phase contribute substantially
to the reactivity of boron-containing catalysts. As our findings are
not consistent with a reaction in which the surface is mainly
responsible for catalytic activity, we put forward the hypothesis
that the oxidation of alkanes over hBN proceeds through a
surface-initiated radical propagation reaction. This hypothesis is
plausible given the evidence for such a mechanism has been put
forward for other catalyst systems such as Li/MgO.
In this contribution, we investigate hBN for the oxidation of n-
butane to produce ethylene and propylene. We have chosen n-
butane to serve as a model hydrocarbon substrate in these
studies so that the experimental results can inform the future
investigation of more complex hydrocarbon feeds. Our main
objective in this study is to probe the role of various reaction
parameters such as reactant partial pressures and temperature
and their effect on activity and product distribution. We also
demonstrate that hBN does not exhibit any deactivation over the
course of 100 hr time-on-stream (TOS) despite low O2 availability
to oxidize potential coke deposits. By altering PO2 and observing
how the rate of formation of each product evolves, we show that
the mechanism of oxidative cracking over hBN is likely similar to
that of a previously studied cracking catalyst which can be
attributed to homogeneous gas-phase chemistry present in these
Figure 1. (a) Conversion of reactants in the oxidative cracking of n-butane as a
function of reaction temperature. (b) Rate of formation of products as a function
of reaction temperature. Balance gas is N2 to a Ptot of 1 atm.
After catalyst pre-treatment, we initially sought to determine if the
systems. By comparison, hBN produces less COx, especially CO2. selectivity trends observed at low PO2 held at higher Xn-C4H10. Our
initial oxidative cracking experiments were performed mostly in
the differential regime (Xn-C4H10 < 10%) to obtain valuable kinetic
data.[24] Because of the apparent O2 adsorption dependence, it is
understood that O2-lean feeds lower alkane consumption levels.
Results and Discussion
To overcome lower alkane conversions observed under O2-lean
conditions, we employ reaction temperatures higher than those
typically employed for hBN-catalyzed n-butane ODH under O2-
rich conditions. A range of temperatures were scanned at a Pn-
C4H10:PO2 ratio of 16 and an inverse weight-hourly-space-velocity
(WHSV-1) of 27.5 kgcat s moln-C4H10-1. As temperature increases
from 475˚C to 575˚C, Xn-C4H10 increases linearly from 4.3% to
19.8% while XO2 increases from 36.2% to 76.6% where it begins
to level off (Figure 1a). The rate of C4 olefin production increases
slightly at 500˚ but then begins to decrease with increasing
Fresh hBN exhibits an induction period attributed to the oxidation
of surface boron species.[31] Prior to O2-lean oxidative cracking
experiments, fresh hBN was activated under an O2-rich feed of
0.20 atm n-C4H10, 0.10 atm O2, and 0.70 atm N2 for 24 hr at 550
°C. Figure S1 demonstrates the growth of the reactivity of hBN
with time on stream, with n-butane conversion (Xn-C4H10) 23.7%
before decreasing mildly and reaching a steady state value of
20.3%.
This article is protected by copyright. All rights reserved.

10.1002/cssc.201901663
ChemSusChem
FULL PAPER
Figure 2. Product distributions at isoconversion in the oxidation of n-butane with different ratios of Pn-C4H10:PO2 at T =
550˚C. Reaction conditions are further tabulated in Table S1.
temperature (Figure 1b). In this temperature range, the rates of
methane and propylene increase and maintain close to a 1:1 ratio,
suggesting that they arise from the same C4 intermediate that
decomposes into C3 and C1 fragments. The rate of ethylene
formation ethylene is significantly higher than that of ethane,
suggesting that under these conditions, the cracking of n-butane
into C2 fragments results in primarily ethylene. The production of
COx only shows slight increases.
temperatures under O2-rich conditions but with undesirable COx
formation. (Figure S4).
The influence of contact time was investigated at a Pn-C4H10:PO2 of
16 at a temperature of 550˚C. Although Xn-C4H10 increases from
7.8% to 17.2% as WHSV-1 increases from 15.7 to 55.0 kgcat s moln-
-1
(Figure S5), the product selectivity is not altered
C4H10
significantly. Thus, temperature and contact time can be
increased to overcome lower alkane conversions under O2-lean
conditions.
Under O2-lean conditions, the substantial difference between the
rates of formation of cracking products and C4 with increasing
temperature has a significant impact on selectivity. Over the
temperature range examined, selectivity to ethylene and
propylene increases from 18.5% to 24.5% and from 25.4% to
45.2%, respectively (Figure S2). As the rate of cracking product
formation outpaced that of C4 olefins, the combined selectivity to
C4 olefins (1-butene, trans-2-butene, cis-2-butene, isobutene, and
1,3-butadiene) decreases substantially from 45.7% to 7.8%. The
increase in reaction temperature also decreases COx selectivity,
which remained under 2%. This change in selectivity is not solely
conversion-dependent as the product distribution at
isoconversion is signficantly shifted toward cracking products at
higher temperature (Figure S3). We originally reported that under
O2-rich conditions at lower temperatures the product selectivity in
the oxidation of n-butane was independent of temperature in the
range of 440 – 480 °C.[24] The reaction conditions employed here
contribute to temperature-dependent reactivity trends.
We then examined the influence of Pn-C4H10:PO2 on the product
distribution to see if the trends observed at low alkane conversion
are consistent with those at elevated conversion (Figure 2). For a
valuable comparison, selectivity data were obtained at
isoconversion by altering the feed conditions to the desired partial
pressures and adjustment of contact time. As the ratio is
increases, the selectivity to propylene and methane increase
while ethylene remains relatively constant. Additionally, the
selectivity to ethane becomes significant, suggesting the
production of ethane is directly related to the O2 content in the
feed. Most importantly, the selectivity to COx decreases with
increasing Pn-C4H10:PO2. As oxidative cracking over hBN is affected
by PO2, we hypothesized that the undesirable selectivity to COx
from previously studied catalysts may have stemmed from non-
optimal reaction conditions. Many of the catalysts in the literature
are tested under O2-rich feed (Pn-C4H10:PO2 < 1) and it is unclear
how these catalysts would perform under a more optimized set of
reaction
conditions.
Therefore,
we
synthesized
an
Li2O/Dy2O3/MgO catalyst based upon a literature procedure[9] with
a nominal loading of 7.7% Li2O and 7.3% Dy2O3 and evaluated it
under O2-lean conditions. Investigating this catalyst at Pn-C4H10:PO2
of 16, we observed over a 30-fold increase in the CO2 selectivity
at comparable Xn-C4H10 relative to hBN. This difference suggests
that low COx selectivity is inherent to hBN and not only a
consequence of decreased O2 availability in the feed. Apart from
COx selectivity, the pattern in selectivity between hBN and
Li2O/Dy2O3/MgO under O2-lean conditions is nearly identical. This
similarity suggests that there exist commonalities in the
Indeed, the results under O2-lean conditions contrast with those
obtained under O2-rich conditions. With a Pn-C4H10:PO2 of 2:1 and
the previously employed WHSV-1 of 27.5 kgcat s moln-C4H10-1, C4
olefin production increases linearly with temperature from 475˚C
until 575˚C where the rate begins to decrease. Over this range,
the rate of formation of cracking product increases and the
production of CO follows the same trend. Therefore, enhanced
selectivity to cracking products can be obtained at high
This article is protected by copyright. All rights reserved.

10.1002/cssc.201901663
ChemSusChem
FULL PAPER
mechanism of oxidative cracking between these materials, which
we hypothesize to be the result of similar gas phase contributions
in their mechanism.
previously reported value of approximately 1.3–1.4 under slightly
different reaction conditions.[24] These values are closer to a
statistical distribution, signifying that the second H-abstraction is
most likely entropically controlled. As the formation of C4 olefins
increases substantially with PO2, we believe that the second H-
abstraction is performed in the gas phase with very reactive
abstracting agents with no energetic preference for primary or
secondary H atoms, hence why we observe a statistical
distribution.
Figure 3. Rate of product formation in the oxidation of n-butane over hBN as a
function of the partial pressure of O2. Xn-C4H10 = 0.5 – 5.1%. Balance gas is N2
to a Ptot of 1 atm.
In order to gain further insight into the reaction mechanism in
hBN-catalyzed oxidative cracking, we investigated the rate of
formation of each product at varying PO2 while maintaining
constant n-butane contact time and ensuring that the reaction is
under differential conditions (Xn-C4H10 < 10%). Under nonoxidative
conditions (PO2 = 0), only cracking products are observed, i.e.
methane, ethane, ethylene and propylene (Figure 3), consistent
with unimolecular decomposition of radicals in the gas phase.
Only trace amounts of C4 olefins were detected, suggesting that
O2 is essential for the production of ODH products under these
reaction conditions. The rate of formation of C4 olefins increases
sharply with increasing PO2 over the entire range investigated.
The rate of formation of methane decreases with a concurrent
increase in the rate of formation of CO. This relationship suggests
that CH3• species formed during cracking can abstract H• from
another gas phase species to form methane or be oxidized to CO
in the presence of O2. The rate of formation of ethylene and
propylene reach their maximum at approximately PO2 = 0.02 atm.
Above this point, any increases in PO2 only influences the
production of ODH and COx products.
Figure 4. Proposed gas phase mechanism in the oxidation of n-butane over
hBN. The reactions are numbered in black or red for pathways not dependent
or dependent on gas phase O2, respectively (see text).
Similarities in the product distribution between hBN and
Li2O/Dy2O3/MgO (vide supra) led us to investigate the influence
of PO2 in the rate of formation of products in the latter system. This
study reveals highly similar trends in the rate of product formation
under differential conditions (Figure S7) between these catalysts.
The most significant difference is that the production of CO2 is
more substantial over Li2O/Dy2O3/MgO than it is with hBN. These
similar trends in product formation rates allow us to surmise that
the cracking of alkanes over these materials have closely related
mechanisms. As the surfaces of these materials are presumably
quite different in nature, identical gas phase reactions occurring
in both systems could explain the analogies in product
distributions.
Figure 4 presents a gas phase reaction network that accounts for
our experimental observations. The fate of the radical species
formed during the reaction depends on the presence of O2. The
pathways independent of O2 (numbered in black in Figure 4)
To provide more insight into the formation of butenes in the
oxidative cracking of n-butane, the rate of formation of each C4
olefin in addition to the ratio of 1-butene:2-butenes was
investigated as a function of PO2 (Figure S6). The ratio of butene
isomers, 1-butene:trans-2-butene:cis-2-butene, gives information
on the nature of the second H-abstraction. A statistical distribution,
3:1:1, implies that the second H-abstraction to form the olefin is
entropically controlled and a thermodynamic distribution, 1:1:1,
implies that it is energetically controlled.[32,33] In our experiment,
the ratio 1-butene:2-butenes falls between 1.14–1.23, close to our
predominate under O2 starvation conditions. Regardless of PO2
,
we propose that the reactive B-O moieties on the surface abstract
the energetically-preferred secondary H-atom to form secondary
n-butyl radicals (1). Under O2-lean conditions, the majority of
these radicals will decompose to form propylene and CH3• (2),
according to the b-scission rule.[34] CH3• acts as a chain carrier
which will form CH4 upon H• abstraction from n-butane to form
primary and secondary n-butyl radicals (3).[35] Primary radicals
This article is protected by copyright. All rights reserved.

10.1002/cssc.201901663
ChemSusChem
FULL PAPER
undergo b-scission to form ethyl radical and ethylene (4). The
ethyl radical forms ethane upon H abstraction of n-butane (5) or
can eject H• to form ethylene (6), the latter reaction providing the
reasoning for ethylene comprising the large majority of the C2
selectivity under all tested conditions.[36]
suffer from deactivation due to coke deposition and oxidative
cracking catalysts deactivate from degradation of the active
phase through evaporation or phase change. hBN has been
shown to be highly stable for up to 300 hr in the ODH of ethane
at 530°C.[41] Since O2-lean feeds are being used here to favor
selectivity to light olefins, coke deposition may play a more
significant role as O2 availability to oxidize carbon species is
decreased. After the initial O2-rich treatment of hBN described
above, the stability of hBN was investigated for an extended time-
on-stream (TOS) under O2-lean conditions. Figure 5 shows Xn-
With addition of O2, new reaction pathways are accessible
(numbered in red in Figure 4). The reaction of alkyl radicals with
O2 to form alkylperoxy adducts is well established in the
literature.[37–39] The decomposition of these adducts result in the
formation of an olefin and HOO• . Thus, reaction of secondary n-
butyl radicals formed at the catalyst surface with O2 forms 1-
butene, trans-2-butene, and cis-2-butene (7). The HOO• radical,
like CH3•, will further react with n-butane molecules to form
additional alkyl radicals (4). The H2O2 formed in this step
generates two HO• radicals upon decomposition (8); this HO•
radical becomes the main chain carrier under oxidative
conditions.[10] In addition to the nonoxidative processes described
above, the primary radical can react with O2 to form 1-butene (9).
The presence of O2 also affects the fate of the alkyl radicals
formed from C-C cleavage. CH3• can oxidize to COx (10) and ethyl
radical forms ethylene upon H-abstraction (11), explaining the
loss of ethane production with increasing PO2. Further studies are
required to fully characterize radical species present in high
temperature hydrocarbon oxidations. Recently, Luo and
coworkers employed a synchrotron VUV photoionization mass
spectrometry (SVUV-PIMS) technique that allowed for the
detection and characterization of radical species present in the
oxidation of methane and ethane over Li/MgO.[40] Species
analogous to those proposed here such as alkylperoxy radicals
were detected and this may prove to be a valuable technique in
the study of oxidative cracking mechanisms.
C4H10 and product selectivity of hBN with for 100 hr TOS at 550ºC
-1
with a Pn-C4H10:PO2 of 16:1 and WHSV-1 of 55.0 kgcat s moln-C4H10
.
No loss of activity was observed in this period; moreover, Xn-C4H10
even increases slightly with TOS from 11.5% to 13.7%. High
selectivity to ethylene (24.2%) and propylene (43.9%) is
maintained over the course of the experiment and total COx
selectivity remained at 1.2%.
Despite the lack of deactivation in the TOS experiment, coke
deposition was still assessed through Raman spectroscopy and
TGA. Fresh hBN catalysts were treated for 48 hr at 550°C with a
feed of 0.20 atm n-butane with either PO2 of 0.0125 or 0.10 atm.
Additionally, contact time was altered to achieve different levels
of XO2. Raman spectroscopy reveals that catalysts treated under
conditions of limited O2 availability (i.e. low PO2 and/or high XO2
)
exhibited two D bands (disordered carbon) at 1180 and 1280 cm-
in addition to a G band (graphitic carbon) at 1600 cm-1,
1
respectively (Figure S8).[42] These bands were less pronounced
under O2-richer conditions, suggesting that coke deposition is
hindered by high O2 availability. TGA of the catalysts under
flowing air was then used to quantitatively determine the amount
of coke present after each treatment (Figure S9). After the initial
loss in mass on spent catalysts due to dehydration from 100 to
300ºC, the loss of mass due to the oxidation of coke deposits at
500 to 700ºC shows that no more than 0.25 wt% of the catalysts
is comprised of coke deposits. Raman spectroscopy was used to
evaluate the catalyst that underwent the largest mass decrease
in the oxidative TGA analysis and previously exhibited the largest
evidence for coke deposition in the Raman experiment (PO2
=
0.0125 atm, XO2 = 82%). The post-TGA spectrum indicates that
these bands significantly decrease in intensity after treatment in
flowing air, confirming that these bands can be correctly assigned
to coke deposition and that the coke can be oxidatively removed.
Although we have no indication that these minor coke deposits
affect the catalyst performance at lab scale, they may become an
issue with longer TOS at pilot scale.
Conclusions
Figure 5. Conversion and selectivity of the oxidative cracking of n-butane over
hBN as a function of time-on-stream. Pn-C4H10 = 0.20 atm, PO2 = 0.0125 atm,
In this contribution, hBN is established as a selective catalyst for
the oxidative cracking of n-butane to lighter olefins. The
adjustment of reactant feed from O2-rich to O2-lean alters the
product distribution in the oxidation of alkanes over hBN from
primarily butenes (ODH products) to ethylene and propylene
(cracking products), respectively. Lowering the O2 concentration
in the feed also reduces the selectivity to COx to less than 2% in
-
balance gas is N2 to a Ptot of 1 atm. T = 550 °C, WHSV-1 = 55.0 kgcat s moln-C4H10
1. Conversion and selectivity labels reflect their value when the experiment was
discontinued.
Catalyst stability is an important concern when considering a new
catalyst for oxidative cracking. Nonoxidative cracking catalysts
This article is protected by copyright. All rights reserved.

10.1002/cssc.201901663
ChemSusChem
FULL PAPER
each catalyst was pretreated under identical conditions prior to the
collection of reported data. Fresh hBN was heated to 550ºC under flow of
20 mL min-1 O2 and 20 mL min-1 N2. After the temperature stabilized, the
gas flow was changed to a mixture of 2:1:7 n-C4H10:O2:N2 at a flow rate of
80 mL min-1 and the catalyst was treated under this condition for 24 hr.
The temperature and gas flow rates were then adjusted to give the desired
reaction conditions.
most cases. Increasing reaction temperature allows for enhanced
rates of production of cracking products. The rate of product
formation as a function of PO2 reflects what we hypothesize to be
a
surface-initiated gas phase radical mechanism, further
supported by observations that the previously studied
Li2O/Dy2O3/MgO catalyst shares highly similar trends in product
formation rates. We propose a reaction mechanism for oxidative
cracking over hBN that accounts for our observations in selectivity
and rates of formation. There is evidence for slight coke
deposition in used samples but there is no observed catalyst
deactivation.
The H2O produced from the reaction was condensed using
a
thermoelectrically-cooled condenser held at -5oC. The dried reactor
effluent was monitored using an on-line µGC (Inficon Micro GC Fusion Gas
Analyzer) equipped with three GC modules and three micro thermal
conductivity detectors. O2, N2, CH4, and CO were analyzed using a Rt-
Molsieve 5A column, CO2 was analyzed using an Rt-U-Bond column, and
all hydrocarbons with the exception of CH4 were analyzed using an Rt-
Alumina Bond/Na2SO4 column. Reported experiments showed the carbon
balance closed within ±5%.
As other boron-containing catalysts (B4C, WB, NiB, etc.) with
oxidized surface boron species catalyze the oxidative
dehydrogenation (ODH) of alkanes with the same selectivity as
hBN,[25] we envision that these materials will follow the same
trends in reactivity observed for hBN when evaluated for their
performance in oxidative cracking. Therefore, boron-containing
materials are a highly versatile platform for selective hydrocarbon
oxidations and new materials can be engineered to meet the
needs of new oxidative technology.
Material Synthesis
The Li2O/Dy2O3/MgO catalyst was synthesized based upon a literature
procedure.[9] To achieve a nominal loading of 7.7 wt% Li2O and 7.3 wt%
Dy2O3 on MgO, 1.8 g LiNO3 (Sigma-Aldrich) , 0.37g Dy2O3 (Strem
Chemicals), and 4.2 g MgO (Aldrich) were added to 250 mL 18 MW H2O
and the resultant slurry was stirred for 2 hr. The mixture was then
rotovapped to dryness and dried at 100ºC overnight before being ground
and treated under flow of air for 2 hr at 200ºC. The material was ground
again and then heated under for flow of air to 750ºC at a rate of 4 C min-1
and held at 750ºC for 3 hr. The powder was then prepared and tested as
described for hBN, with the pretreatment performed at 575ºC.
Although current hydrocarbon cracking processes are well
established, their high emission of CO2 necessitates the
development of environmentally benign technology. Here, we
demonstrate that hBN is able to produce light olefins selectively
at milder temperatures than steam cracking with an order of
magnitude less CO2 selectivity than previously studied oxidative
cracking catalysts. Moreover, the tunability in product distribution
with simple adjustment of process parameters enables the flexible
production of olefins in a given market. The understanding of the
parameters involved in the production of ethylene and propylene
from n-butane over hBN can inform future studies investigating
the oxidation of more complex hydrocarbon feeds as we look
toward more efficient routes to valuable chemicals.
Thermogravimetric Analysis (TGA)
TGA was performed using
a
Mettler Toledo DSC/TGA
1
Thermogravemetric Analyzer. About 70-100 mg of used catalysts were
deposited into a ceramic 150 μL crucible. The change in mass was
monitored over time as the sample was heated 10°C min-1 from ambient
temperature to 900°C under air flow of 40 mL min-1.
Raman Spectroscopy
Experimental Section
Raman spectra were recorded by using an Invia Reflex Raman
Microscope (Renishaw) equipped with a λ = 785 nm laser. All of the
measurements were carried out at ambient conditions using pelletized
samples without dilution. Typically, 60 scans of 1 s in duration were
averaged. The spectrum was recorded between 600–1700 cm-1 using the
WiRE software package (Renishaw).
Catalytic Testing
Fresh hBN was obtained from Sigma-Aldrich and not further chemically
treated prior to catalytic testing. The powder was pelletized, crushed and
sieved to collect particles of 425–600 μm in diameter in order to limit any
potential mass transfer effects. About 300 mg of hBN was loaded into a
quartz tube reactor (9 mm ID) undiluted and supported on a bed of quartz
wool approximately 1 in. in length in the middle of the tube. A quartz rod
was placed beneath the catalyst bed to reduce dead volume and minimize
the potential for homogeneous gas-phase reactions past the catalyst bed.
Flowrates of n-C4H10 (instrument grade, Matheson), O2 (UHP, Airgas) and
N2 (UHP, Airgas) were controlled using three mass flow controllers
(Bronkhorst) calibrated to each individual gas to allow total flowrates of 40–
Acknowledgements
This research was supported by the U.S. Department of Energy,
Office of Science, Office of Basic Energy Sciences under Award
DE-SC0017918.
140 mL min-1 and WHSV-1 values ranging from 15.7–55.0 kgcat s moln-C4H10
-
1. All reactions were performed at a Ptot of 1 atm. The temperature was
controlled using a thermocouple inserted into the middle of the catalyst
bed.
Keywords: hexagonal boron nitride • oxidative cracking •
oxidation • radicals • hydrocarbon upgrading
[1]
[2]
LyondellBasell, 2017 Data Book, 2017.
Exposure of fresh hBN to reaction conditions results in an induction period
of approximately 18 hr before steady state activity is achieved; therefore,
S. M. Sadrameli, Fuel 2015, 140, 102–115.
This article is protected by copyright. All rights reserved.

10.1002/cssc.201901663
ChemSusChem
FULL PAPER
[3]
[4]
S. M. Sadrameli, Fuel 2016, 173, 285–297.
[25]
[26]
J. T. Grant, W. P. McDermott, J. M. Venegas, S. P. Burt, J. Micka,
S. P. Phivilay, C. A. Carerro, I. Hermans, ChemCatChem 2017, 9,
3623–3626.
J. J. H. B. Sattler, J. Ruiz-Martinez, E. Santillan-Jimenez, B. M.
Weckhuysen, Chem. Rev. 2014, 114, 10613–10653.
T. Ren, M. K. Patel, K. Blok, Energy 2008, 33, 817–833.
C. Boyadjian, L. Lefferts, K. Seshan, Appl. Catal. A Gen. 2010, 372,
167–174.
[5]
[6]
A. M. Love, B. Thomas, S. E. Specht, M. P. Hanrahan, J. M.
Venegas, S. P. Burt, J. T. Grant, M. C. Cendejas, W. P. McDermott,
A. J. Rossini, et al., J. Am. Chem. Soc. 2019, 141, 182–190.
L. Shi, B. Yan, D. Shao, F. Jiang, D. Wang, A. H. Lu, Cuihua
Xuebao/Chinese J. Catal. 2017, 38, 389–395.
L. Shi, D. Wang, W. Song, D. Shao, W. P. Zhang, A. H. Lu,
ChemCatChem 2017, 9, 1720.
[7]
[8]
[9]
L. D. Schmidt, M. Huff, Catal. Today 1994, 21, 443–454.
M. Huff, L. D. Schmidt, J. Catal. 1994, 149, 127–141.
L. Leveles, S. Fuchs, K. Seshan, J. A. Lercher, L. Lefferts, Appl.
Catal. A Gen. 2002, 227, 287–297.
[27]
[28]
[29]
[10]
[11]
[12]
[13]
[14]
L. Leveles, K. Seshan, J. A. Lercher, L. Lefferts, J. Catal. 2003, 218,
296–306.
R. Huang, B. Zhang, J. Wang, K. H. Wu, W. Shi, Y. Zhang, Y. Liu,
A. Zheng, R. Schlögl, D. S. Su, ChemCatChem 2017, 9, 3293–
3297.
L. Leveles, K. Seshan, J. A. Lercher, L. Lefferts, J. Catal. 2003, 218,
307–314.
[30]
[31]
J. M. Venegas, I. Hermans, Org. Process Res. Dev. 2018, 22,
1644–1652.
C. Trionfetti, I. V. Babich, K. Seshan, L. Lefferts, Appl. Catal. A Gen.
2006, 310, 105–113.
Y. Zhou, J. Lin, L. Li, X. Pan, X. Sun, X. Wang, J. Catal. 2018, 365,
14–23.
T. Sato, M. Saito, K. Matano, Y. Yoshimura, T. Oikawa, T. Konishi,
H. Yasuda, N. Kijima, Appl. Catal. A Gen. 2002, 206, 237–244.
A. G. Dietz, A. F. Carlsson, L. D. Schmidt, J. Catal. 1998, 176, 459–
473.
[32]
[33]
[34]
M. Chaar, D. Patel, M. Kung, H. Kung, J. Catal. 1987, 105.
T. Blasco, J. M. L. Nieto, A. Dejoz, M. I. Vazquez, 1995, 282.
I. Glassman, R. A. Yetter, N. G. Glumac, Combustion, Academic
Press, Burlington, MA, 2008.
[15]
[16]
S. S. Bharadwaj, L. D. Schmidt, J. Catal. 1995, 155, 403–413.
S. S. Bharadwaj, C. Yokoyama, L. D. Schmidt, Appl. Catal. A Gen.
1996, 140, 73–97.
[35]
[36]
[37]
[38]
[39]
[40]
[41]
[42]
J. A. Manion, D. A. Sheen, I. A. Awan, J. Phys. Chem. A 2015, 119,
7637–7658.
[17]
M. L. Kaliya, S. B. Kogan, N. Froumin, M. Herskowitz, Catal.
Commun. 2002, 3, 327–333.
S. L. Peukert, N. J. Labbe, R. Sivaramakrishnan, J. V. Michael, J.
Phys. Chem. A 2013, 117, 10228–10238.
[18]
[19]
M. L. Kaliya, S. B. Kogan, Catal. Today 2005, 106, 95–98.
M. Narayanappa, V. D. B. C. Dasireddy, H. B. Friedrich, Appl. Catal.
A Gen. 2012, 447–448, 135–143.
J. D. Desain, C. A. Taatjes, J. A. Miller, S. J. Klippensteina, D. K.
Hahn, R. March, Faraday Discuss 2001, 119, 101–120.
A. J. Eskola, T. T. Pekkanen, S. P. Joshi, R. S. Timonen, S. J.
Klippenstein, Proc. Combust. Inst. 2019, 37, 291–298.
A. J. Eskola, O. Welz, J. D. Savee, D. L. Osborn, C. A. Taatjes, J.
Phys. Chem. A 2013, 117, 12216–12235.
[20]
J. Sá, M. Ace, J. J. Delgado, A. Goguet, C. Hardacre, K. Morgan,
ChemCatChem 2011, 3, 394–398.
[21]
[22]
K. Narasimharao, T. T. Ali, Catal. Letters 2013, 143, 1074–1084.
F. S. Al-Sultan, S. N. Basahel, K. Narasimharao, Fuel 2018, 233,
796–804.
L. Luo, R. You, Y. Liu, J. Yang, Y. Zhu, W. Wen, Y. Pan, F. Qi, W.
Huang, ACS Catal. 2019, 9, 2514–2520.
[23]
[24]
J. Grant, C. A. Carrero, F. Goeltl, J. Venegas, P. Mueller, S. P. Burt,
S. E. Specht, W. P. McDermott, A. Chieregato, I. Hermans, Science
2016, 354, 1570–1573.
L. Shi, D. Wang, W. Song, D. Shao, W. P. Zhang, A. H. Lu,
ChemCatChem 2017, 9, 1720.
J. J. H. B. Sattler, A. M. Beale, B. M. Weckhuysen, Phys. Chem.
Chem. Phys. 2013, 15, 12095–12103.
J. M. Venegas, J. T. Grant, W. P. McDermott, S. P. Burt, J. Micka,
C. A. Carrero, I. Hermans, ChemCatChem 2017, 9, 2118–2127.
This article is protected by copyright. All rights reserved.

10.1002/cssc.201901663
ChemSusChem
FULL PAPER
Entry for the Table of Contents
FULL PAPER
Move over steam cracking! The
oxidative cracking of n-butane over
hexagonal boron nitride (hBN) under
O2-lean conditions gives high
William P. McDermott, Juan Venegas,
Ive Hermans
Page No. – Page No.
selectivity to ethylene and propylene
with sparing amounts of CO and CO2.
Oxidative cracking of hydrocarbons
may provide an efficient route to the
production of valuable building-block
chemicals.
Selective Oxidative Cracking of n-
Butane to Light Olefins over
Hexagonal Boron Nitride with
Limited Formation of COx
This article is protected by copyright. All rights reserved.