Formation of [Cu 2 O 2 ] 2+ and [Cu 2 O] 2+ toward C–H …

 

 

Research Article

pubs.acs.org/acscatalysis

§

†,⊥

†,⊥,◆
§

Hacksung Kim,
∥,⊥

Formation of [Cu2O2]2+ and [Cu2O]2+ toward C−H Bond Activation in
Cu-SSZ-13 and Cu-SSZ-39
Matthew J. Wulfers,
Bahar Ipek,
Karl S. Booksh,
Joseph P. Smith,
†
Center for Catalytic Science and Technology,
Chemistry and Biochemistry, University of Delaware, Newark, Delaware 19716, United States
‡
Department of Chemistry, Center for Catalysis and Surface Science, Northwestern University, Evanston, Illinois 60208, United
States
#
Chemical Sciences and Engineering Division, Argonne National Laboratory, Argonne, Illinois 60439, United States
∇
Department of Chemistry and
Wisconsin 53706, United States
∥
Center for Neutron Research, National Institute of Standards and Technology, Gaithersburg, Maryland 20899, United States
*S Supporting Information

○
Department of Chemical and Biological Engineering, University of Wisconsin−Madison, Madison,

‡,#
Florian Göltl,
and Raul F. Lobo*,†,⊥

Department of Chemical and Biomolecular Engineering, and

Craig M. Brown,
⊥

Ive Hermans,

§
Department of

∇,○

∇

ABSTRACT: Cu-exchanged small-pore zeolites (CHA and
AEI) form methanol from methane (>95% selectivity) using a
3-step cyclic procedure (Wulfers et al. Chem. Commun. 2015,
51, 4447−4450) with methanol amounts higher than Cu-
ZSM-5 and Cu-mordenite on a per gram and per Cu basis.
Here, the CuxOy species formed on Cu-SSZ-13 and Cu-SSZ-
39 following O2 or He activation at 450 °C are identified as
trans-μ-1,2-peroxo dicopper(II) ([Cu2O2]2+) and mono-(μ-
oxo) dicopper(II) ([Cu2O]2+) using synchrotron X-ray
diffraction,
in situ UV−vis, and Raman spectroscopy and
theory. [Cu2O2]2+ and [Cu2O]2+ formed on Cu-SSZ-13
showed ligand-to-metal charge transfer (LMCT) energies
between 22,200 and 35,000 cm−1, Cu−O vibrations at 360, 510, 580, and 617 cm−1 and an O−O vibration at 837 cm−1. The
vibrations at 360, 510, 580, and 837 cm−1 are assigned to the trans-μ-1,2-peroxo dicopper(II) species, whereas the Cu−O
vibration at 617 cm−1 (Δ18O = 24 cm−1) is assigned to a stretching vibration of a thermodynamically favored mono-(μ-oxo)
dicopper(II) with a Cu−O−Cu angle of 95°. On the basis of the intensity loss of the broad LMCT band between 22,200 and
35,000 cm−1 and Raman intensity loss at 571 cm−1 upon reaction, both the trans-μ-1,2-peroxo dicopper(II) and mono-(μ-oxo)
dicopper(II) species are suggested to take part in methane activation at 200 °C with the trans-μ-1,2-peroxo dicopper(II) core
playing a dominant role. A relationship between the [Cu2Oy]2+ concentration and Cu(II) at the eight-membered ring is observed
and related to the concentration of [CuOH]+ suggested as an intermediate in [Cu2Oy]2+ formation.
KEYWORDS: methane activation, Cu-zeolite, Cu-SSZ-13, Cu-SSZ-39, active site, dicopper core, Raman spectroscopy

■ INTRODUCTION

The forecasted abundant supply and low cost of natural gas
have greatly incentivized its transformation into more valuable
products. Although direct conversion of methane to liquid fuels
and valuable chemicals is attractive,1 converting methane
selectively to oxygenated species in one-step is extremely
difficult because of the high stability of C−H bonds in methane
and the higher reactivity of oxygenated products such as
methanol and formaldehyde (resulting in complete combustion
even at low conversion).2,3 There are two groups of agents that
can selectively oxidize methane to methanol. The first is the
methane monooxygenase (MMO) enzyme, which can produce
methanol at ambient conditions with the active species
containing copper-oxygen4−6 or iron-oxygen7,8 intermediates.
The second group includes transition metal (Fe,9−13 Co,14,15

and Cu16−22)-exchanged zeolites that can form methanol with
high selectivity at temperatures between 120 and 300 °C with
transition metal−oxygen reactive sites. Although the reactive
site for Fe-exchanged zeolites, FeO,23 can only be formed by
18,21,24−26 reactive sites formed on Cu-
N2O,
exchanged zeolites can be obtained by reaction with O2,
making Cu- exchanged zeolites more appealing from a practical
standpoint.

the CuxOy

O2 activation on Cu-exchanged zeolites requires temper-
atures higher than 300 °C17 (to increase the concentration of
the active species20,27), but the optimum temperature range for

Received: October 21, 2016
Revised: May 10, 2017
Published: May 18, 2017

© XXXX American Chemical Society

4291

DOI: 10.1021/acscatal.6b03005
ACS Catal. 2017, 7, 4291−4303

ACS Catalysis

methane activation is between 120 and 200 °C.16,20 Thus,
selective methanol formation on copper-exchanged zeolites is
often achieved following a cyclic process containing the
following three steps:

(1) Oxidation of Cu-zeolite at temperatures between 300 and

(2) Methane reaction at 150−200 °C resulting in a stable

surface intermediate.

(3) Methanol extraction from the surface using steam or a

450 °C.

solvent.

Isothermal stepwise methanol production at 15020 or 200
°C27 or catalytic methanol production28,29 have been reported
on Cu-exchanged zeolites, but the methanol formation amounts
and methanol selectivity were significantly lower than the ones
reported for O2 activation performed at 450 °C.16

The reactive site in the cyclic procedure described above
forms during the first oxidation step and has been suggested to
be a mono-(μ-oxo) dicopper(II) complex, [Cu−O−Cu]2+, for
Cu-exchanged ZSM-518,24 and mordenite (Si/Al = 5).19,25,26
On the other hand, a trinuclear copper oxygen cluster,
[Cu3O3]2+, has also been suggested for Cu-mordenite (Si/Al
= 11).21 The [Cu−O−Cu]2+ core on Cu-ZSM-5 has been
associated with characteristic ligand-to-metal charge transfer
(LMCT) electronic transition at 22,700 cm−1 and resonance-
enhanced Raman Cu−O vibrations at 456 and 870 cm−1.24
Similar electronic transitions at 21,900 and 23,100 cm−1 and
Cu−O vibrations at 450 and 465 cm−1 have also been reported
for assigned mono-(μ-oxo) dicopper(II) complexes forming on
Cu-mordenite (Si/Al = 5).25

Cu-SSZ-13 (CHA), Cu-SSZ-16 (AFX), and Cu-SSZ-39
(AEI) have recently been shown to have interesting properties
for
selective methanol production following the 3-step
process,30 producing methanol from methane at 200 °C with
higher yields (per gram and per Cu basis) than those of Cu-
ZSM-5 and Cu-mordenite. Evidence for an LMCT band at
22,700 cm−1 was not observed for any of Cu-SSZ-13, Cu-SSZ-
16, and Cu-SSZ-39 using UV−vis spectroscopy after activation
in O2,30 suggesting the presence of a different CuxOy active
species.

SSZ-13, SSZ-16, and SSZ-39 have similar characteristic UV−
vis bands upon O2 treatment,30 indicating the presence of
similar Cu-O moieties. This is not unexpected because all three
have double-6-membered rings (d6MR) and 8-membered ring
(8MR) openings with pore dimensions of 3.4 and 3.8 Å.31
Among these small-pore zeolites, Cu-SSZ-13 is well-known for
its activity and selectivity for the selective catalytic reduction
32−37 and for NO oxidation
(SCR) of NOx with NH3
reactions.38,39 Besides Cu-SSZ-13, Cu-SSZ-39 is also effective
for the SCR of NOx.40,41 Numerous reports, using in situ and
operando methods, have attempted to understand copper
coordination in SSZ-13 and its relation to SCR activity.38,42−53
Some of these have verified the presence of CuxOy species on
Cu-SSZ-13 by X-ray absorption spectroscopy,49 EPR,47 and
theoretical
treatment con-
ditions. At this point, the structures of these species remain
to be determined.

investigations38 under different

Identification of CuxOy species on SSZ-13 is not only
essential to elucidate the mechanism of methanol production
but is also important for designing improved materials that
yield higher methanol amounts. We have conducted an
experimental and theoretical
investigation of copper(II)
acetate-exchanged Cu-SSZ-13 (Si/Al = 12) to understand the
structure and properties of these species. The investigation

Research Article

included in situ powder X-ray diffraction (PXRD), UV−vis
spectroscopy, and Raman spectroscopy following sample
oxidation at 450 °C because CuxOy
reactive species for
methanol production are formed during the oxidation step of
the 3-step cyclic process mentioned above. The oxidation of
Cu-SSZ-13 is contrasted to the effect of high temperature (450
°C) He treatment, and an oxidation route starting from Cu(I)-
SSZ-13 is evaluated to provide insights into the mechanism of
CuxOy formation. Finally, similarities between Cu-SSZ-13 and
Cu-SSZ-39 are highlighted based on Raman and UV−vis
spectroscopy.

■ RESULTS

O2 Treatment of Cu-SSZ-13. Synchrotron Powder X-ray
Diffraction (PXRD). Details of the locations of Cu and CuxOy
clusters in the zeolite depend on pretreatment conditions (such
as temperature and gas composition).24,51 In this section, we
discuss Cu sitting on hydrated and high-temperature (450 °C)
O2-treated Cu-SSZ-13 (and kept in an O2 atmosphere during
cooling to 50 °C). Cu coordination at 50 °C is relevant because
it is the temperature at which a similar sample was exposed to
methane in our previous report.30

The evolution of Cu-SSZ-13/12/0.4 upon O2 treatment was
monitored with variable-temperature synchrotron PXRD.
(Samples are identified as SSZ-xx/y/z where xx indicates the
zeolite type, y indicates the Si/Al ratio, and z indicates the Cu/
Al ratio; see Table S2 for elemental analyses). Before O2
treatment, in its hydrated form, Rietveld refinement of Cu-SSZ-
13/12/0.4 PXRD led to the identification of 15% of the Cu(II)
cations located close to 6MR, whereas the majority (85%) were
close to 8MR. Hydrated copper cations were found away from
framework oxygen atoms and close to the center of
the
chabazite cage with short distances to the oxygen atoms of
water molecules (Table 1 and Table S3).

Loss of water upon heating to 450 °C in O2 resulted in a
decrease in unit cell parameter a and an increase in parameter c
(Figure S4), as also found by Kwak et al. for Cu-SSZ-13/6/
0.4.48 There is a hysteresis in the change of the unit cell
parameters upon heating and cooling to 50 °C, which can be
explained by migration of Cu cations from their position away
from the zeolite framework toward the 6MR and 8MR windows
(Figure 1) with Cu−Ofw distances decreasing from 3.05(2) to
2.09(4) Å (Table 1).

After cooling (in O2) to 50 °C, Cu cations in 6MR moved to
the center between the three oxygen atoms (O1) with a mean
Cu−Ofw distance of 2.293(4) Å (Table 1 and Table S4). This
refined distance is close to the value of 2.329 Å previously
found on calcined Cu-SSZ-13 (Si/Al = 18) using neutron
diffraction.44 Cu placed on the symmetry axis of 6MR has often
been reported for SSZ-13 with different Si/Al ratios using XRD
and neutron diffraction.42,44,54,55 However, using extended X-
ray absorption fine structure analysis (EXAFS) and electron
it has been
paramagnetic resonance (EPR) spectroscopy,
shown that Cu(II) coordinates to four framework oxygen
atoms in a distorted square planar fashion on the plane of
6MR.34,44,47,51 This site, with two Al atoms on 6MR,
is
favorable by 54−71 kJ mol−155,56 for dehydrated Cu(II) cations
when compared to the Cu sites at 8MR having two Al atoms.
The fraction of Cu cations observed on 6MR was refined to a
value of 0.23 in close agreement with the reported value of 0.2
for occupation at 6MR on Cu-SSZ-13/15.5/0.45 using powder
XRD and the Rietveld/maximum entropy method.55

4292

DOI: 10.1021/acscatal.6b03005
ACS Catal. 2017, 7, 4291−4303

ACS Catalysis

Table 1. Cu Distribution and Bond Distances Obtained by
Rietveld Analysis of PXRD Data of Cu-SSZ-13/12/0.4
(H0.3Cu1.2Al2.7Si33.3O72) and Cu-SSZ-13/5/0.39
(H1.2Cu2.4Al6Si30O72) (APS, 17-BM-B, λ = 0.75009 Å,
Trigonal, R3̅m)

a

Cu-SSZ-13/12/
0.40

30 °C, hydrated Cu at
6MR

bond
Cu1−O1(fw)

*3

Cu1−O6(H2O)
Cu2−O3(fw)

Cu2−O5(H2O)

distance
(Å)

2.86(5)

2.00(7)

3.05(2)

1.92(3)

bond
Cu1−O1(fw)

*3

distance
(Å)

2.293(4)

Cu2−O2(fw)
Cu2−O3(fw)
Cu2−O4(fw)
Cu3−O2(fw)
Cu3−O3(fw)
Cu3···Cu3

*2

2.20(4)
2.74(4)
2.09(4)
2.76(3)
2.22(6)
2.5(1)

bond

distance
(Å)

Cu1−O1(fw)
Cu3−O1(fw)
Cu2−O3(fw)

*3
*3

Cu2−O2(fw)

2.195(5)
2.229(5)
2.186(6)

2.697(3)

Cu %

15

85

13.5804(2)
14.6747(3)
2343.83(9)
8.163
1.65
2.25

Cu %

47 (total)

23

26

21

31
38
9

13.5494(2)
14.7720(2)
2348.59(8)
2.610
1.51
1.99

Cu %

69 (total)

13.5028(2)
15.1105(6)
2385.95(9)
3.544
2.40
3.32

Cu at
8MR

a/Å
c/Å
V/Å3
χ2

Rp/%
wRp/ %

Cu at
6MR
Cu at
8MR

a/Å
c/Å
V/Å3
χ2

Rp/%
wRp/ %

Cu at
6MR

Cu at
8MR

a/Å
c/Å
V/Å3
χ2

Rp/%
wRp/ %

Cu-SSZ-13/
12/0.40
50 °C, O2
activated

Cu-SSZ-13/5/
0.39
50 °C, O2
activated

aValues in parentheses indicate one standard deviation in the prior
digit. See Figure 1 to identify different Cu positions on Cu-SSZ-13/12
and Figure S5 for Cu-SSZ-13/5.

Two Cu(II) sites were found near 8MR: Cu2 (26%) with
Cu−Ofw distances of 2.09(4) and 2.20(4) Å and Cu3 (21%)
further away from the framework oxygen atoms (Table 1 and
Figure 1). Cu−Ofw distances for Cu cations at 8MR are larger
than the optimized Cu−Ofw distances for bare Cu(II) with 2 Al
on 8MR (Cu−Ofw between 1.93 and 2.03 Å),51 indicating the
existence of CuxOyHz species coordinated to 8MR.

Research Article

O2 treatment of Cu-SSZ-13/5/0.39 at 450 °C followed by
cooling to 50 °C led to 69% of the Cu(II) cations near 6MR
with Cu−Ofw bond distances of 2.195(5) Å and 2.229(5) Å.
Only 9% of copper was near 8MR (Cu2) with Cu−Ofw
distances of 2.186(6) and 2.697(3) Å (Table 1). Other
structural parameters for Cu-SSZ-13/5/0.39 at 50 °C can be
found in Table S5. Like Cu-SSZ-3/12/0.4, not all Cu (22%) at
8MR could be distinctly located by the refinement because of
the predicted broad distribution of the Cu cations at 8MR.
Similarly, a refinement of the positions of extra-framework
oxygen atoms was not feasible.

The difference in Cu occupation at 6MR on the two SSZ-13
samples can be attributed to higher Al content in Cu-SSZ-13
with Si/Al = 5. Higher Al results in a higher proportion of
6MRs having two Al atoms.49 With more 6MRs having two Al
atoms, more Cu(II) cations (bare) occupy this energetically
favorable site as observed by PXRD here. Bates et al.49
calculated the maximum achievable Cu/Al for Cu associated
with two Al atoms on 6MR. Calculated Cu/Al ratios for
samples with Si/Al = 5 and 12 were found to be 0.23 and
0.11;49 from our Rietveld refinements, Cu occupation at 6MR
resulted in Cu/Al = 0.27 for Si/Al = 5 and 0.09 for Si/Al = 12,
which is in very good agreement with the estimates by Bates et
al. Because it was not possible to refine extra-framework oxygen
atom positions from the PXRD data, we used diffuse reflectance
(DR) UV−vis and Raman spectroscopy to obtain insight into
the structures and properties of CuxOy species.

Diffuse Reflectance UV−Vis Spectroscopy. DR UV−vis
spectra of Cu-SSZ-13 (Figure 2a) following O2 treatment at
450 °C displayed absorption features at 34,700, 30,000, 19,700,
16,500, and 13,600 cm−1, which were not present in the spectra
of hydrated Cu-SSZ-13/12. Bands at 19,700, 16,500, and
13,600 cm−1 were reported in the spectra of Cu-SSZ-13/13 and
Cu-SSZ-13/6 by Giordanino et al.45 and Wulfers et al.30 The
shoulders at 30,000 and 34,700 cm−1 on Cu-SSZ-13/12/0.4
were found more pronounced when compared to those of SSZ-
13/5/0.39, which had more than twice the Cu concentration of
Si/Al = 12 (Table S2), indicating the importance of the Si/Al
ratio and Cu distribution on CuxOy formation. Increasing the
Cu/Al ratio on SSZ-13 (Si/Al = 12) resulted in increased
absorption intensity between 30,000 and 34,700 cm−1,
indicating a higher concentration of CuxOy sites. The bands
at 13,600, 16,500, and 19,700 cm−1 also showed increased
intensity with increasing Cu/Al ratio. Godiksen et al.47 assigned
d ← d transition energies at 13,600, 16,600, 20,000 and 10,900,
13,600, and 16,600 cm−1 (dx
2 ← dzx, dyz, and
2 ← dxy, dx
2−y
dx
2 transition energies) to the distorted tetragonal
planar geometry of Cu(II) with 4-coordination to oxygen
atoms near two Al atoms on 6MR, where 13,600, 16,600, and
20,000 cm−1 are assigned to be the d ← d transition energies
for two Al atoms being the third-nearest neighbors and 10,900,
13,600, and 16,600 cm−1 for two Al atoms being the second-
nearest neighbors. With this assignment, the increased intensity
of the bands at 16,500 and 19,700 cm−1 with increased Cu/Al
ratio (Figure 2a) can be explained by higher Cu(II) occupation
of 6MRs with two Al atoms being the third-nearest neighbors.
When the temperature was increased gradually up to the
activation temperature of 450 °C, bands at 19,500 and 16,750
cm−1 were observed to evolve synchronously with the broad
band at 30,000 cm−1 (Figure 2b), which could indicate either
the contribution of absorption features of CuxOy sites or
evolution of bare Cu(II) sites during CuxOy
formation at
elevated temperatures.

2 ← dz

2−y

2−y

4293

DOI: 10.1021/acscatal.6b03005
ACS Catal. 2017, 7, 4291−4303

ACS Catalysis

Research Article

Figure 1. Schematic of d6MR and 8MR of hydrated and dehydrated H-Cu-SSZ-13/12/0.4 refined in space group R3̅m (No. 166). For illustration
purposes, all extra-framework atoms are represented in the same cage on the left (red spheres inside the cage represent O atoms of H2O). The three
refined Cu sites found after O2 treatment at 450 °C are given on the right.

The larger amount of methanol formation on Cu-SSZ-13/
12/0.35 (0.06 methanol/Cu) when compared to Cu-SSZ-13/
6/0.35 (0.03 methanol/Cu)30 suggests a higher concentration
of reactive CuxOy species on Cu-SSZ-13/12. Therefore, O2-
treated Cu-SSZ-13/12/0.4 was exposed to methane at 200 °C
to identify characteristic absorption energies of reactive CuxOy
species based on spectral changes. Reduction of absorption
intensities from a broad band centered at 29,500 cm−1 (and
shoulders at 35,000 and 22,200 cm−1) upon methane admission
(Figure 2c) indicates that the reactive species have absorption
features at approximately 29,500 cm−1.
Interestingly, no
absorption losses of the bands at 13,600, 16,500, and 19,700
cm−1 were observed.

In Situ Raman Spectroscopy of Cu-SSZ-13. Raman spectra
of O2-treated Cu-SSZ-13/12/0.4 were obtained using an
excitation wavelength of 458 nm (Figure 3). The bands at
467 and 487 cm−1 are assigned to the T−O−T bending
vibrations in the zeolite 4MR.57 The O2-activated sample
showed five new bands at 360, 510, 580, 617, and 837 cm−1
when compared to the hydrated sample. The vibrations at 360,
510, 580, and 837 cm−1 disappeared after 4 days at room
temperature in O2, showing the instability of the corresponding
CuxOy species, whereas the band at 617 cm−1 maintained a
constant intensity, suggesting a different (stable) species. The
decrease in intensity of these four bands over time could be due
to a small amount of H2O leaking into the system despite the
flow or due to intrinsic instability of the species.
continuous O2
In either case, the simultaneous disappearance of the bands
suggests that these vibrations belong to the same species. This
group of four vibrations is in excellent agreement with trans-μ-
1,2-peroxo dicopper(II)58 (Table 2), where 360 cm−1
is
assigned to the bending vibration of the Cu−O bond (δ
Cu‑O),
510 and 580 cm−1 are assigned to the stretching vibrations of
Cu‑O), and 837 cm−1 is assigned to the O−O
the Cu−O bond (ν
stretching vibration (ν

O‑O) of peroxide.

18O isotope experiments performed by treating the
flow for 13
dehydrated Cu-SSZ-13/12/0.4 at 450 °C in 18O2
h and then keeping it under 18O2 loop flow for nearly 4 days
reveal a 24 cm−1 shift for the band at 617 cm−1 (Figure S7).
The observed frequency at 617 cm−1 and the degree of oxygen
isotope shift of Δ(18O2) = 24 cm−1 are consistent with the
totally symmetric breathing mode (Ag with D2h and C2h
symmetry in the [Cu2O2]2+ diamond core)59 for bis(μ-oxo)
dicopper(III) species with a Cu−O−Cu angle of approximately
101° (Table 2).60 For example, a band at 610 cm−1 was

previously observed on Cu-SSZ-13/4.3/0.36 and assigned to a
bis(μ-oxo) dicopper(III) complex by Guo et al.57 However,
mono-(μ-oxo) dicopper species at a specific Cu−O−Cu angle
such as 100° also shows a vibrational band around 609 cm−1
and an almost identical degree of oxygen isotope shift (Table
2).24 Because the latter structure has been found to be much
more stable in a chabazite cage by our DFT calculations (see
below), we assign the Raman band at 617 cm−1 to the Cu−O−
Cu symmetric stretching vibration of mono-(μ-oxo) dicopper
species.

Giordanino et al.45 have previously suggested a side-on μ-
(η2:η2) peroxo dicopper(II) species, which has a characteristic
UV−vis absorption and Raman band61 at 29,000 and 763 cm−1,
respectively. Our UV−vis spectra show a characteristic side-on
peroxide absorption band at 30,000 cm−1; however, the Raman
spectral position at 837 cm−1 indicates an end-on peroxide
rather than side-on.

Theoretical Calculations. There are several

reports of
including trans-μ-1,2-peroxo dicopper(II),
CuxOyHz species,
bis(μ-oxo) dicopper(III), [CuOCu]2+, and [CuOH]+, forming
on Cu-exchanged zeolites.38,51,62,63 However, these features and
the relative stability are likely to vary with their
local
environment and therefore the zeolite support. To elucidate
the stability and spectroscopic features of CuxOyHz species in
SSZ-13, we modeled a series of mononuclear,64 dinuclear, and
trinuclear copper species within the framework of the material
(see Supporting Information section 4). The different
possibilities included the stoichiometries displayed in Table 2,
different positions within the framework as well as different
anchoring points consisting of activated O atoms (O atoms
closest to framework Al), silanol defects, and framework O
atoms. We also included a [CuOH]+ site in 8MR, which is the
most common site at high Si/Al ratios.51,63

To compare the stabilities of the sites, we calculated the
chemical potential of the Cu atoms (μCu) defined as the energy
of a single atom in the bonding environment as

Cu

μ

( ,
T P

O
2

P
/ )
0

=

Cu O H zeo

+

x

y

z

+
2H zeo

−

G

⎛
⎜
⎝

1
x

G

−

z

x

−
2

H O
2

μ

−

2
y

− +
z
4

2
x

O
2

μ

⎞
⎟
⎠

(1)

where G denotes the temperature- and pressure-corrected
Gibbs free energies of the species in the superscript. Using this

4294

DOI: 10.1021/acscatal.6b03005
ACS Catal. 2017, 7, 4291−4303

ACS Catalysis

Research Article

Figure 2. (a) DR UV−vis absorption spectra of Cu-exchanged zeolites after O2 treatment at 450 °C. The feature at 28,500 cm−1 is an artifact caused
by a lamp switch in the spectrometer. The pictures of the samples at 25 °C after 450 °C O2 treatment are given next to their spectrum. (b) DR UV−
vis absorption spectra of Cu-SSZ-13/12/0.4 after O2 treatment at 250, 300, 350, 400, and 450 °C. (c) DR UV−vis absorption spectra of Cu-SSZ-13/
12/0.4 after O2 treatment at 450 °C (black) followed by 2 h CH4 treatment at 200 °C (red). The difference spectrum is shown in green. Before the
spectra were collected, the temperature was reduced to room temperature. Note that small changes in sample position after each treatment render
comparison of spectral intensities only qualitative.

approach for the characterization, we find the [CuOH]+ site to
be the most stable under the characterization conditions (i.e.,
room temperature and vacuum). However, three other sites
lead to a μCu close enough to be present in the system. Starting
with the most stable one, these three sites are a [Cu2O2]2+ site
bridging an eight- (8MR) and a six-ring (6MR) (A), a
[Cu2O]2+ site located in 8MR both anchored on Al atoms (B),
and a [Cu2O2]2+ site located in the 8MR anchored on one
silanol defect and on activated framework Os (C). The relative
μCus and structures are displayed in Figure 4. Other stable
species are shown in the Supporting Information.

The species designated as A is trans-μ-1,2-peroxo dicopper-
(II) with a Cu−Cu distance of 3.98 Å and Cu−O (extra-
framework) distances of 1.82 and 1.83 Å. These are close to

trans-μ-1,2-peroxo

4.359(1) and 1.852(5) Å found for
dicopper(II) in biological systems.66

Site B is mono-(μ-oxo) dicopper(II) with a Cu−O−Cu angle
of 95°, a Cu−O distance of 1.76 Å, and a Cu···Cu distance of
2.60 Å. These optimized Cu−Cu distances and Cu−O−Cu
angles are typical for bis(μ-oxo) dicopper(III) with Cu···Cu
distances between 2.73 and 2.88 Å60 and Cu−O−Cu angle of
101°.60 Nevertheless, it is not simple to differentiate mono-(μ-
oxo) dicopper(II) and bis(μ-oxo) dicopper(III) using Raman
or X-ray absorption spectroscopy26 because of their similar
configuration. DFT calculations show that, within the zeolite
confinement, bis(μ-oxo) dicopper(III) is in fact not the most
stable species as it is in the homogeneous phase (see also
Figure S9).26,67

4295

DOI: 10.1021/acscatal.6b03005
ACS Catal. 2017, 7, 4291−4303

ACS Catalysis

Research Article

considered all possible Al configurations, and the most stable
species are given in Figure 4.

We also calculated vibrational spectra for the sites given in
Figure 4. Unfortunately, identifying Raman active vibrations in
this case is again not straightforward. Raman activity is
determined by a change in polarizability of the system with
respect to the given vibration, but doing so for a system
containing as many vibrations as this one would be very time
consuming. We therefore focus on identifying the stretching
vibrations of the Cu−O bonds, which are Raman active. We
then assign the Raman intensity of
this vibration as the
amplitude of the stretch. Furthermore, we calculated optical
absorption spectra for the three sites using a combination of
GW0 calculations to determine the accurate energetic position-
the electronic states and Bethe−Salpeter equation
ing of
calculations to accurately calculate the electron−hole inter-
actions upon excitation. Both spectroscopies are displayed in
Figure 5.

On the basis of the results from the stretching vibration
predictions, it can be concluded that all three species are likely
present on Cu-SSZ-13 following O2 treatment at 450 °C.
According to the calculations,
the experimentally observed
vibrations at 476, 510, and 580 cm−1 belong to two trans-μ-1,2-
peroxo dicopper(II) species (A and C in Figure 4), whereas the
vibration at 617 cm−1 could be assigned to the mono-(μ-oxo)
dicopper(II) species with a Cu−O−Cu angle of 95° (B in
Figure 4).

Figure 3. In situ Raman spectra of O2-activated (at 450 °C for 2 h) H-
Cu-SSZ-13/12/0.4. Blue and green spectra are taken after
the
indicated time spent in a pure O2 environment. Hydrated H-Cu-
SSZ-13/12 is given for comparison (black).

Vilella et al.62 have suggested that

trans-μ-1,2-peroxo
dicopper(II) is the most stable configuration for zeolites with
10MR and 12MR, such as MFI and MOR, having Al···Al
distances of ∼8 Å, and suggested a cis-μ-1,2-peroxo dicopper-
(II) core to be the most stable for the 8MR of CHA. However,
for SSZ-13/12, it is not reasonable to limit the calculations to
8MR due to the high Si/Al ratio. In our calculations, we

Table 2. Mononuclear and Dinuclear CuxOy Complexes and Their Spectroscopic Features

aCalculated in Woertink et al.24 for a Cu−O−Cu angle of 100°. bCalculated in this work for a Cu−O−Cu angle of 95°.

4296

DOI: 10.1021/acscatal.6b03005
ACS Catal. 2017, 7, 4291−4303

ACS Catalysis

Research Article

Figure 4. Three different structures with μCu within 5 kJ mol−1 of a [CuOH]+ site in the zeolite. Si atoms are displayed in yellow, O in red, Cu atoms
in blue, Al atoms in silver, and H atoms in white. The bond distances are given in units of Å.

Cu-SSZ-13 (Figure 2c), suggesting possible involvement of A−
C species in methane activation.

To gain further insight into species activating CH4, Raman
data collected after O2
treatment (using dry air) and
subsequent CH4 treatment were compared (Figure S12). The
decrease in the relative intensity of the band at 571 cm−1 was
found to be more dominant after CH4 reaction, indicating a
prevailing role of the assigned trans-μ-1,2-peroxo dicopper(II)
species in C−H bond activation.

He Activation of Na-Cu-SSZ-13 and Na-Cu-SSZ-39. He
activation of Na-Cu-SSZ-13/12/0.47 and Na-Cu-SSZ-39/10/
0.48 showed similar spectroscopic features when compared to
O2 activation, indicating formation of the same CuxOy species
in He. (The samples denoted as Na-Cu-SSZ-xx contain very
little sodium with Na/Al ratios between 0.025 and 0.029, see
Table S2.) The similarity of the spectroscopic features of Cu-
SSZ-13 and Cu-SSZ-39 is not surprising due to the structural
similarity between the two frameworks; however,
the
resemblance of the spectra of the O2- or He-treated samples
(with slightly pronounced band intensity in the 25,000−35,000
cm−1 region, which was related to the reactive species in Figure
2c following O2 treatment) is surprising when compared to Cu-
ZSM-5, on which mono-(μ-oxo) dicopper(II) decomposes
when treated at 450−500 °C in He.17,24

The UV−vis spectra of O2- and He-treated Na-Cu-SSZ-13
and Na-Cu-SSZ-39 (Figure 6a) displayed three bands in the
13,000−20,000 cm−1 region on both samples, indicating the
presence of Cu(II) regardless of
the treatment gas. He
treatment of Cu-SSZ-13 is known to result in autoreduction
the sample but only up to certain extent.51 The
of
autoreduction has been reported to proceed by the homolytic
loss of the OH ligands of Cu(II) cations coordinated near
single (AlO4/2)− units at temperatures higher than 300 °C in an
inert atmosphere.51,68 The bare Cu(II) cations on 6MR, on the
other hand, cannot be readily autoreduced. Calculations
performed by Paolucci et al. showed favorability of bare Cu(II)
cations on 6MR in a He environment for temperatures lower
than 500 °C.68 Therefore, at temperatures lower than 500 °C,
Cu(II) cations are expected to be on 6MR.

Observed Raman bands following O2 treatment of Na-Cu-
SSZ-13/12 (Figure 6b) were very similar to H-Cu-SSZ-13/12/
0.4 (Figure 3), confirming the negligible effect of the sodium
content on Na-Cu-SSZ-13 and Na-Cu-SSZ-39. Bands at 511,
574, and 616 cm−1 were observed following both O2 and He
treatment for Na-Cu-SSZ-13/12/0.47 with bands at 511 and
574 cm−1 having lower intensities after He treatment (Figure
6b).

4297

DOI: 10.1021/acscatal.6b03005
ACS Catal. 2017, 7, 4291−4303

Figure 5. Comparison of experimental and calculated (a) Raman and
(b) UV−vis spectra belonging to A, B, and C species (notations are
given in Figure 4). The experimental Raman spectrum of Cu-SSZ-13/
12/0.4 is given as a black line on the left figure. On the right, the
difference between the spectra of O2-activated Cu-SSZ-13/12, Cu/Al
= 0.4 and Cu/Al = 0.18, is given in black and denoted as ‘exp’.

The predicted UV−vis absorption spectra suggest that the
trans-μ-1,2-peroxo dicopper(II) species (A) have absorptions
between 16,000 and 20,000 cm−1 that could contribute to the
observed bands at 16,500 and 19,700 cm−1 (Figure 2a). On the
basis of the low methanol/Cu ratios observed on the Cu-SSZ-
13 sample (the highest methanol/Cu ratio on H-Cu-SSZ-13/12
is found to be 0.09 following 1 h of CH4 activation29), the
concentration of CuxOy sites is suggested to be lower than that
of bare Cu(II) sites and possible [CuOH]+ sites that are
suggested to be at 8MR near single Al atoms in CHA.51,63
Therefore, the intensity of absorptions belonging to trans-μ-1,2-
peroxo and mono-(μ-oxo) dicopper(II) species are suggested
to be lower than the intensity of absorptions for bare Cu(II)
and [CuOH]+ sites. Recently, Kulkarni et al.63 suggested that
[CuOH]+ is the active species for methanol formation in 8MR
zeolites. Even though methane activation by [CuOH]+ has
been theoretically suggested by Kulkarni et al,63 the UV−vis
bands between 22,200 and 35,000 cm−1 (Figure 2.a) are not
assigned to [CuOH]+ based on the explanation by Godiksen et
al.47 for the indiscernibility of [CuOH]+ in the UV−vis spectra.
Even if absorption features of [CuOH]+ were discernible in
UV−vis spectra, they would be related to the observed band at
13,600 cm−1 rather than the broad band around 29,500 cm−1
the
based on our calculations (see Figure S11).
predicted LMCT electronic transitions between 22,000 and
37,000 cm−1 (Figure 5b) of the A−C species are in good
agreement with the observed LMCT energies in this region
(Figure 2b) and with the broad absorption loss observed
between 22,200 and 35,000 cm−1 upon methane contact with

Instead,

ACS Catalysis

Research Article

Na-Cu-SSZ-39 also showed three major bands at 511, 572,
and 607 cm−1, where the first two were also less intense after
He treatment (Figure 6c). Considering the assignment of these
bands to trans-μ-1,2-peroxo dicopper(II), one could infer that
both He and O2 treatment result in the formation of trans-μ-
1,2-peroxo dicopper(II) and mono-(μ-oxo) dicopper(II)
species but that O2 treatment results in a higher concentration
of trans-μ-1,2-peroxo dicopper(II) species.

When the amounts of methanol produc