Supplementary information for

Steering CO2 Hydrogenation Towards C-C Coupling to Hydrocarbons

Using Porous Organic Polymer/Metal Interfaces

Chengshuang Zhou,1 Arun S. Asundi,1,2 Emmett D. Goodman,1,2 Jiyun Hong,3 Baraa

Werghi,1,2,3 Adam S. Hoffman,3 Sindhu Nathan,1,2 Stacey F. Bent,1,2 Simon R. Bare2,3 and

Matteo Cargnello1,2,*

1Department of Chemical Engineering, Stanford University, Stanford, CA 94305, USA

2SUNCAT Center for Interface Science and Catalysis, Stanford University, Stanford, CA

94305, USA

3Stanford Synchrotron Radiation Lightsource, SLAC National Accelerator Laboratory,

Menlo Park, CA 94025, USA

*Corresponding author email: mcargnello@stanford.edu

This file contains:

1. Additional Materials and Methods
2. Supplementary Figures (Fig. S1-S25)
3. Supplementary Tables (Table S1-S3)
4. Supplementary References
1. Additional Materials and Methods

Materials
Triruthenium dodecacarbonyl (Ru3(CO)12, 99 %), trioctylamine (98 %), oleylamine (OLAM,
70%), benzyl ether (98%), 1,4-dioxane (99%), acetic acid (99%), o-phthalaldehyde (o-PA,
97%), m-phthalaldehyde (m-PA, 99%), p-phthalaldehyde (p-PA, 99%), WO3 (99%) and ZrO2
(99%) were purchased from Sigma Aldrich. 1,3,5-tris(4-aminophenyl)benzene (TAPB, 97%)
was purchased from TCI. All solvents were of reagent grade and all reagents were used as-
received. Al2O3 Pluralox TH100/150 was obtained from Sasol and calcined at 900 °C for 24 h.
The calcined material was predominantly γ-Al2O3. TiO2 (Aeroxide, P25) was obtained from
Acros and used directly as support without additional treatments. SiO2 spheres were
synthesized using a Stöber method and functionalized with amino groups following a recently
reported procedure.1 All calcined supports and samples were ground and sieved below 180 μm
grain size.

Synthesis of 3 nm Ru nanoparticles
Ru nanoparticles were prepared by thermal decomposition of Ru3(CO)12 via colloidal synthesis
using standard Schlenk techniques and a previously reported procedure.1 40.8 mL of benzyl
ether (98%, Sigma-Aldrich) were added to 50 mg of Ru3(CO)12 (99%, Sigma-Aldrich) in a
three-neck flask. 0.1 mL of oleylamine (70%, Sigma-Aldrich) and 0.1 mL of trioctylamine
(99%, Sigma-Aldrich) were added as surfactants. The reaction content was degassed (<2 Torr)
for 30 min at room temperature. The flask was then flushed with nitrogen, heated to 270 °C at
a rate of ~20 °C min-1 and kept at this temperature for 30 min. The particles were purified by
precipitation with ethanol (total volume 30 ml) followed by centrifugation (8000 rpm, 3 min)
and redissolution in hexanes for three times, and finally dispersed in hexanes.

Inorganic Ru/oxide catalysts preparation

For the preparation of Ru/TiO2, an appropriate amount of 3 nm Ru nanoparticles was added to
TiO2 (Aeroxide P25, Acros) support dispersed in hexanes under vigorous stirring to achieve a
total metal loading of 0.5 wt. %. The mixture was stirred for 20 min to allow the nanoparticles
to adsorb on the support and the catalysts were then separated by centrifugation (8000 rpm, 3
min). Colorless supernatants were observed for each solution indicating complete adsorption
of the particles. The powders were dried at 80 °C for 3 h and sieved below 180 μm grain size.
Organic ligands were removed by fast calcination of the catalysts at 700 °C for 30 s.2
Other oxide-supported Ru catalysts were prepared by replacing TiO2 with appropriate oxides.
Specifically, WO3, ZrO2, Al2O3 and -NH2 functionalized3 SiO2 were used. It should be noted
that the weight loading of Ru in all of these catalysts was kept at 0.5 wt.% except Ru/WO3,
which was 0.25 wt.% due to limited adsorption of particles on this low surface are support.

Catalyst characterization
Thermogravimetric analysis (TGA) was performed on a TA Instruments TGA 5500. Samples
were heated at a ramp rate of 5 °C min−1 under a flow of 20 mL min−1 of air.
Bright-field transmission electron microscopy (TEM) was performed on a FEI 80-300 Titan
Environmental Transmission Electron Microscope operating at 300 kV. HAADF images were
collected with a Fischione Model 3000 annular detector using a camera length that set to the
inner collection angle to ~71 mrad operating at 200 kV.
X-ray photoelectron spectroscopy (XPS) was performed using a PHI VersaProbe 3 Scanning
XPS Microprobe equipped with a hemispherical electron analyser using Al(Kα) radiation
(1486.3 eV). For all samples, the incident X-ray spot size was 100 um, and an excitation of
100 W at 20 kV was applied. An Ar+ neutralizer and electron flood gun were used to
compensate for sample charging. Binding energies were referenced to the C 1s peak (284.8 eV)
or Si 2p peak (103.5 eV) from SiO2 as internal standard.
Gas adsorption measurements were carried out in a Micromeritics 3-Flex instrument, which
uses high-accuracy pressure transducers to measure gas adsorption. For N2 physisorption
experiments, samples were degassed in the Micromeritics SmartVacPrep unit below 0.1 Torr
for 19 h at 150 °C, with masses measured after degassing. N2 isotherm was then measured in
12 mm borosilicate tubes containing ~100 mg of powder samples, in a bath of liquid nitrogen.
To calculate pore size distributions, isotherms were fitted with the NLDFT carbon-slit model.
For CO chemisorption experiments, ~150 mg of catalyst powder was loaded in a quartz tube
and subjected to pretreatments consisting evacuation at 110 °C for 30 min, oxidation in O2(5
vol %)/Ar at 250 °C for 30 min, further evacuation at 250 °C for 30 min, and reduction in H2(5
vol %)/Ar at 250 °C for 30 min followed by evacuation at 250 °C for 2 h and 180 °C for 6 h.
The CO adsorption experiments were conducted at −70 °C to avoid the formation of carbonates
on TiO2 using an ethanol/dry ice bath and in the pressure range from 2 to 20 Torr.
Fourier Transform Infrared Spectroscopy (FTIR) measurements of the samples were performed
on a Thermo-Fisher Nicolet is-50 FTIR instrument with an attenuated total reflectance
attachment using a deuterated triglycine sulfate detector with 2 cm−1 precision.
Diffuse Reflectance Infrared Fourier Transform Spectroscopy (DRIFTS) was performed in a
Praying Mantis DRIFTS system (Harrick). Samples were loaded into a high temperature and
pressure reaction cell with ZnSe windows. Gas flow rates were controlled using EL-Flow series
mass flow controllers (Bronkhorst). Prior to measurement, catalysts were activated by
oxidation at 250 °C in 20 mL min-1 5% O2/N2 for 30 minutes and subsequent reduction at
250 °C in 20 mL min-1 5% H2/N2 atmosphere for another 30 minutes. For CO adsorption-
desorption experiments, sample cell was allowed to cool to room temperature and background
spectra were measured in 20 mL min-1 N2 at 1 atm. The feed gas was then switched to 20 mL
min-1 10% CO and the sample was allowed to adsorb CO until saturation. When the acquired
spectra stopped evolving (~20 minutes), the feed gas was switched back to 20 mL min-1 N2 and
reaction cell was vacuumed to ~1 torr to remove excess gas phase CO. The temperature of the
sample cell was then gradually increased to 250 °C (10 °C min-1). For operando CO2
hydrogenation measurements, feed gas was switched to 20 mL min-1 75% H2 after sample
activation at 250 °C and the sample cell was gradually pressurized to 6 bar. After measuring
the background spectra at 250 °C, 6 bar in 20 mL min-1 75% H2, the feed gas was switched to
20 mL min-1 of 25% CO2 + 75% H2. IR spectra were collected throughout the sample treatment
process using a Bruker Vertex 70 spectrometer with a liquid nitrogen-cooled mercury-
cadmium-telluride (MCT) detector. Each spectrum was measured by averaging 200 scans over
the range 600 - 4400 cm−1.
In-situ X-ray absorption measurements at Ru K-edge (11.564 keV) were performed at beamline
9-3 of the Stanford Synchrotron Radiation Lightsource at the SLAC National Laboratory.
Beamline 9–3 is a 16-pole, 2-tesla wiggler beamline with a vertically collimating mirror for
harmonic rejection and a cylindrically bent mirror for focusing. Incident photon energy was
selected by the liquid-nitrogen cooled, double-crystal Si(220) monochromator at crystal
orientation of = 90. The catalyst sample was pressed, crushed, and sized between 80 and 120
mesh sieves. Approximately 70 mg of the sieved sample was loaded to make a 9-mm bed in a
3-mm Kapton tube, held in place by two plugs of quartz wool. The capillary was loaded into a
custom-built flow-reactor reactor described previously3 . The reactor was mounted on the
sample stage at a 45° angle relative to the X-ray beam. Spectra were collected in fluorescence
detection mode using a PIPS detector orthogonal (90°) to the beam path with a 10-cm Soller
slit. The beam size of 1 mm [v] by 3 mm [h] was used. Ar-filled ion chambers were used to
measure the incident X-ray intensity and the Ru foil reference, which was scanned
simultaneously with the sample for energy calibration. For each steady state of interest, 4
EXAFS scans were collected and merged for analysis. (maybe add details about the treatment
steps as well). Gas flow rates were controlled using mass flow controllers (Brooks) and
temperature was controlled with a Eurotherm PID controller. To monitor the gas flows, a mass-
spectrometer (Hiden QGA) was used throughout the experiment. The XAS data was processed
using the Athena software of the Demeter package.4 The EXAFS spectra were energy
calibrated, merged, and normalized. Linear combination fitting of the XANES region in each
steady state was performed using bulk Ru metal and RuO2 as standards. The normalized
EXAFS spectra were modeled using the Artemis software of the Demeter package. S02 was
calculated to be 0.71 ± 0.06 by fitting Ru foil. For modelling EXAFS of different states,
scattering paths generated from metal Ru, and RuO2 were used. Ru_mp-33.cif and RuO2_mp-
825.cif from the materials project5, together with Ru-O, Ru-N, Ru-C first shell model from
Artemis software were used to construct models for EXAFS fitting.

Calculation of surface atoms in 2.7 nm Ru nanoparticles

Ruthenium (Ru) is a rare transition metal with 44 electrons. According to experimental data,
the preferred crystal structure of Ru is hexagonal closed pack (hcp) with a=270.59pm,
b=270.59pm, c=428.15pm.
As such, the volume of a unit cell containing 6 atoms is:
3 √3
= = 0.081446 nm
2
Total number of unit cells in 2.7 nm Ru sphere particles:
4 4
3 (1.35 ) 10.306 nm
= =3 = = 126.54 ( !)
0.081446 nm 0.081446 nm

"# = 6 × 126.54 = 759

Assuming that shell thickness = bond distances of Ru atoms = 0.2759 nm

Number of core atoms:
4 4
( − 0.2759) (1.0741 ) 5.1907 nm
3 =6×3
' =6× =6×
0.081446 nm 0.081446 nm

= 382 )* !
 +,-. /
In total, +,-
= 50 % atoms are on surface
Calculation of C-C coupling probability (α)
Generally, hydrocarbon production from CO2 hydrogenation on Ru catalysts share similar
mechanism with Fischer-Tropsch synthesis and thus follows Anderson–Schulz–Flory (ASF)
distribution because of the polymerization nature of the process. Specifically, the molar
percentage of a hydrocarbon product with carbon number n (Mn) is dependent on the C-C
coupling probability α (0 < α < 1) by the following equation:

12 = (1 − 3)3 2.4

In order to obtain the value of α from experimental data, the above equation is often
expressed in the logarithmic form and fitted linearly (ln(Mn) vs n) to get the slope, which
equals to ln(α), as described below:
1−3
ln(12 ) = (3) + ln( )
3

3
Ru/TiO2
ln(Mn) ∝ nln(α
α) o-IPOP/Ru/TiO2
2
m-IPOP/Ru/TiO2
p-IPOP/Ru/TiO2
1

0
ln(Mn)

-1

-2

-3

-4
2 3 4
Carbon Number (n)

Sample ln(α) α
Ru/TiO2 -2.53 0.08
o-IPOP/Ru/TiO2 -1.83 0.16
m-IPOP/Ru/TiO2 -1.34 0.26
p-IPOP/Ru/TiO2 -0.93 0.39

Calculation of transition state energies from Eyring plot

The rate determining step of CO2 hydrogenation reaction is often considered to be the initial
step of the reaction – the reduction of adsorbed *CO2 to *COxHy species where x and y depend
on the subsequent pathway of the reaction (e.g., *CO for RWGS pathway or *HCOO for
formate mediated pathway). To gain more insight on the reaction mechanism at a molecular
level, we invoke Transition State Theory (TST) to describe this initial step by the Eyring
equation, where ∆9: and ∆; : are enthalpy and entropy change of the transition:

<= > .∆@ A ∆D A

<( ) )= BC B
ℎ
In order to obtain the value of ∆9: and ∆; : , the Eyring equation is expressed in the
logarithmic form:

< −∆9: 1 ∆; : <=

ln E F = + + ln( )
> G > G ℎ

I 4
The values of ∆9 : and ∆; : can be obtained by the linear fit of ln HCJ vs C, where the slope

.∆@ A ∆D A I
equals to B
and y-intercept equals to B
+ ln( LK ).
Calculation of Catalyst Stability
Following the condensation reaction of the two precursors:
3
M N9 4 + M/ 9O P → M O2 9 N2 2 + 3 9 P
2
It can be calculated that the average C:N:H atomic ratio is 12:1:8, leading to 86.7% of carbon
mass in the polymer. Meanwhile, TGA analysis revealed that polymer accounts for 50% mass
of the sample. Combined, it can be derived that 86.7% × 60% = 52.0% mass of the p-
IPOP/Ru/TiO2 sample is carbon.
Hypothetically, if all carbon in the sample were converted into hydrocarbon products (including
CO, CH4 and higher hydrocarbons), the “equivalent” CO2 consumption would be:
520 R 'S 2 R
÷ 12 = 43.3 * /<R
1 <R ! T *
~43 mol/kg sample. Given that the actual detected CO2 reaction rate is 2.5 mol/(h*kg), in 77
hrs of testing, the catalyst had converted:
*
2.5 × 77 ℎ = 193 * /<R
ℎ ∙ <R
193 mol CO2 per kg sample far exceeds the actual content of organic polymer within the sample.
Thus, this serves as complementary evidence that the product detected cannot be simply from
the degradation of IPOP.

Calculation of Uncertainties
A. values that can be directly measured
For values than can be directly measured through experiments, such as reaction rates,
kinetic measurements were carried out for three times. Average values µ were
reported in the main text and figures, while standard deviations were calculated by
4
X = YZ ∑Z
]_4(\] − ^ ) and plotted as error bars.

B. values that are fitted (slopes, intercepts, etc)

To obtain values that were fitted from experiment data, such as activation energies,
transition state energies, rate orders and C-C coupling probabilities, experiments were
firstly performed for three times, data points were then plotted as average with error
bars, calculated in the same method above. The built-in function in OriginPro 2021b,
weighted least squares linear fitting, was then applied to the dataset with instrumental
4
weightings `b. Specifically:
a

\] c \c 1
E∑ F E∑ ] F − (∑ ] ] )(∑ )
X] X] X] X]
! *T =
\ \ 1
E∑ ] F − E∑ ] F (∑ )
X] X] X]
1
∑
X]
* # d =e
\ 1 \
E∑ ] F E∑ F − E∑ ] F
X] X] X]
 \] c] \
∑ − ! *T ∗ (∑ ] )
X] X]
c−f ) T) = \
∑ ]
X]

\]
∑
X]
* ]2 ' d =e
\ 1 \
E∑ ] F (∑ ) − E∑ ] F
X] X] X]
Calculation of Transport and Diffusion Criteria
We have calculated the Mears and Weisz-Prater criteria at 10% conversion (despite most of the
experiments having been run at 5% conversion or less) as follows and the external and internal
diffusion barriers don't limit our reactions.

• Calculation of Mears Criterion for p-IPOP/Ru/TiO2

− ′i = )f* ) 7.0 × 10.N * kl /(<R ∙ !) (experimentally measured)
300 R
jS = yz < { !f)c *| ) c!) y { = 127 <R/
( 0.5 ) × 3
G = ) c!) T )f {fz! 9 × 10., (catalysts sieved under 180 um)
= )f* * { 1* R
T 1.5 ∗ 101325 €
MiS = yz < R ! * ) )f* *| • (MP ) 25% ∗ 6 y = 1.5 y = = = 35.0 */
G × > 8.314 × 523 • ∙ * .4
Ф = T* *!f)c 0.4
.N
ƒi= = R ! − Tℎ ! {f||z!f f)c ) 523† 10 /! (Fogler, table 14-2)
{d = )* {f ) 10.

1G 1G
− ′i jS G = 4
= ‹ = | *Œ * f)c Reˆ
< MiS U{d =5! )6y × 10.,
< ˆ
Re = = 3.2
Ф ƒi= Sh′ (1 − Ф)ν = 1.9 × 10. /!
= < = 5.5 × 10. /!
1 − Ф {d
;ℎ′
;ℎ′ = (Re′)4/ (Sc)4/
= 0.83
ν = kinematic viscosity at
523 K, 1.5 bar
~10., /!
ν (https://www.engineeringtoolbox.com/carbon-dioxide-
; =
ƒi= dynamic-kinematic-viscosity-temperature-pressure-
;
d_2074.html)
= 0.1

MR=4*10-5<0.15, not external diffusion limited

• Calculation of Weisz-Prater Criterion

− ′i = )f* ) 7.0 × 10.N * kl /(<R ∙ !)
− ′i j G j = !* f{ ) c!) { !f)c 2120 <R/
M˜™
M˜™ = G = ) c!) T )f {fz! 9 × 10.,
ƒ Mi# = 3 × 10.,
ƒ = || )f R ! − Tℎ ! {f||z!f f)c 10., /!
Mi# =R ! * ) )f* *| • (MP2) ) ) c!) !z | 35.0 * /

CWP=3*10-5<1, not external diffusion limited

2. Supplementary Figures (Fig. S1-S25)

35 35

30 30

25 Ru/TiO2 25 o-IPOP/Ru/TiO2
2.8 ± 0.4 nm 2.7 ± 0.5 nm
20 20
Count

Count
15 15

10 10

5 5

0 0
1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
Ru Particle Diameter (nm) Ru Particle Diameter (nm)

35 30

30 25

25 m-IPOP/Ru/TiO2 p-IPOP/Ru/TiO2
20
2.7 ± 0.4 nm 2.8 ± 0.4 nm
20
Count

Count 15
15
10
10

5 5

0 0
1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
Ru Particle Diameter (nm) Ru Particle Diameter (nm)

Fig. S1 Ru particle size distributions for different samples (from no less than 100
particles for each distribution)
Fig. S2 Representative TEM images of a) o-IPOP/Ru/TiO2 and b) m-IPOP/Ru/TiO2
Fig. S3 Representative STEM images (a, e) and corresponding EDS mappings (b-d, f-h
of a-d) o-IPOP/Ru/TiO2 and e-h) m-IPOP/Ru/TiO2
Fig. S4 C1s and Ti 2p XPS spectra of o-IPOP/Ru/TiO2 and m-IPOP/Ru/TiO2
Fig. S5 Investigation of the effect of titania thermal pretreatment on polymer growth. a)
FT-IR spectra of TiO2 after different treatments. TEM images of p-IPOP growth results
on b) calcined TiO2 and c) re-hydrolyzed TiO2
 p-phthalaldehyde

m-phthalaldehyde
Normalized Absorbance (a.u.)

o-phthalaldehyde

1,3,5-Tris(4-aminophenyl)benzene

p-IPOP/Ru/TiO2

m-IPOP/Ru/TiO2

o-IPOP/Ru/TiO2

Ru/TiO2

3750 3500 3250 3000 2750 2500 2250 2000 1750 1500 1250 1000 750
Wavenumber (cm-1)

Fig. S6 FT-IR spectra of o-, m-, p-IPOP/Ru/TiO2 and their corresponding monomers (o-,
m-, p-phthalaldehyde and 1,3,5-Tris(4-aminophenyl)benzene)
 H2 consumption and NH3 production (a.u.)
polymer degradation

NH3 signal baseline

RuO2 reduction

H2 consumption baseline

0 100 200 300 400 500

Temperature (°C)

Fig. S7 Temperature programmed reduction results for p-IPOP/Ru/TiO2 in terms of H2

consumption (black squares) and ammonia production (red circles). The two sets of signals
were offset for clarity.
Fig. S8 Thermogravimetric analysis of o-IPOP/Ru/TiO2 and m-IPOP/Ru/TiO2
 250
Ru/TiO2
o-IPOP/Ru/TiO2

Quantity Adsorbed (cm³/g STP)

200 m-IPOP/Ru/TiO2
p-IPOP/Ru/TiO2
Specific Surface Area (m2/g)
Ru/TiO2 55.8
150 o-IPOP/Ru/TiO2 88.3
m-IPOP/Ru/TiO2 53.8
p-IPOP/Ru/TiO2 222.2

100

50

0
0.0 0.2 0.4 0.6 0.8 1.0
Relative Pressure (P/P0)

0.006
0.1 p-IPOP/Ru/TiO2 0.004

Incremental Volume (cm³⋅g-1⋅Å-1)

0.002
Cumulative Volume (cm³⋅g-1)

0.0 0.000
0.006
m-IPOP/Ru/TiO2
0.1
0.004

0.002

0.0 0.000
0.006
o-IPOP/Ru/TiO2
0.1
0.004

0.002

0.0 0.000
0.006
Ru/TiO2
0.1
0.004

0.002

0.0 0.000
0 20 40 60 80 100
Pore Width (Å)

Fig. S9 a) N2 physisorption isotherms and b) NLDFT pore size distributions using the
carbon slit model calculated from the isotherm of TiO2, o-,m-,p-IPOP/Ru/TiO2
 CO2 Conversion and Product Distribution (%)
5
CO CH4
100

CO Production TOF (103 s-1)

4
80

3
60

2
40

20 1

0 0
iO 2 /TiO 2 /TiO 2 /TiO 2
Ru/T IPO P/Ru IPO P/Ru IP OP/Ru
o- m- p-

Fig. S10 CO2 conversion and product selectivity for CO2 hydrogenation at atmospheric
pressure with different catalysts
 100 1000

90
CO2 H2 900

CO2 and H2 concentration (%) 80 800

C1-4 Product (GC area)

70 700

60 600

50 No CO2 No H2 500

40 400

30 300

20 200

10 100

0 0
0 2 4 6 8 10 12 14 16
Time on Stream (h)

Fig. S11 CO2 hydrogenation (6 bar total pressure, 75% H2 + 25% CO2, 30 sccm, 250 °
C) with p-IPOP/Ru/TiO2 catalyst. Reactant gas (CO2 and H2) were sequentially removed
to correlate C1-4 production with the presence of both reactants.
Fig. S12 Arrhenius plots for different samples. Conversions were controlled to be below
5% and apparent activation energies were calculated for individual pathways.
Fig. S13 Eyring plots for different samples. Conversions were controlled to be below 5%
and transition state energies were calculated for individual pathways.
Fig. S14 H2 rate orders for different samples. H2 concentration was varied between 45% and
80% while CO2 concentration was kept constant at 15% with conversions below 10%. Rate
orders were calculated for individual pathways.
Fig. S15 CO2 rate orders for different samples. CO2 concentration was varied between 15%
and 25% while H2 concentration remained at 75% with conversions below 10%. Rate orders
were calculated for individual pathways.
 0.36
o-IPOP/Ru/TiO2
1 atm
CO gas
labile CO 5% CO
0.24
25°°C

0.12
vac.
Absorbance (K-M unit)

25°°C
0.00

vac.
m-IPOP/Ru/TiO2 100°°C
0.28 CO gas

labile CO

0.14 vac.
180°°C

0.00
vac.
250°°C
-0.14

2300 2200 2100 2000 1900 1800

-1
Wavenumber (cm )

Fig. S16 DRIFT spectra of CO adsorption on o-IPOP/Ru/TiO2 and m-IPOP/Ru/TiO2

 Total Pressure
0.15

Absorbance (K-M unit)

6 bar
p-IPOP/Ru/TiO2
0.10
250°°C

0.05

0.00

1 bar
-0.05
2300 2200 2100 2000 1900 1800

Wavenumber (cm-1)
Fig. S17 DRIFT spectra of CO adsorption on p-IPOP/Ru/TiO2 with increasing pressure at
250 °C. Note that CO concentration is kept at 10% while total pressure increases from 1
bar to 6 bar.
Absorbance (K-M unit)
Time
0.02 o-IPOF/Ru/TiO2
Ru-CO
0.01

0.00

AB
-0.01

-0.02

-0.03
2200 2000 1800 1600 1400
Wavenumber (cm-1)
Absorbance (K-M unit)

0.02 Time
m-IPOF/Ru/TiO2
Ru-CO
0.01

Z
0.00

-0.01

2200 2000 1800 1600 1400

Wavenumber (cm-1)
Fig. S18 operando-DRIFT spectra of CO2 hydrogenation on o-IPOP/Ru/TiO2 and m-
IPOP/Ru/TiO2
 1.4
Ru foil reference
Normalized Absorbance (a.u.) p-IPOP/Ru/TiO2 - O2
1.2 p-IPOP/Ru/TiO2 - H2
p-IPOP/Ru/TiO2 - CO2 + H2
RuO2 reference
1.0

0.8

0.6

0.4

0.2

0.0
22080 22100 22120 22140 22160 22180

Energy (eV)

Fig. S19 Ru K-edge XANES spectra of the Ru foil (as reference), RuO2 powder (as
reference), p-IPOP/Ru/TiO2 after oxidative pretreatment (250°C, 5% O2, 30 mins),
reductive pretreatment (250°C, 5% H2, 30 mins) and under reaction conditions (250°C,
75% H2 + 25% CO2)
 RuO2
8

p-IPOP/Ru/TiO2 - O2
4
χ(k)·k2 (Å-2)

p-IPOP/Ru/TiO2 - H2

0
p-IPOP/Ru/TiO2 - CO2+H2

-4
Ru foil

-8

0 2 4 6 8 10 12 14

k (Å-1)

Fig. S20 k2-weighted Ru K-edge EXAFS spectra of the Ru foil (as reference), RuO2
powder (as reference), p-IPOP/Ru/TiO2 after oxidative pretreatment (250°C, 5% O2, 30
mins), reductive pretreatment (250°C, 5% H2, 30 mins) and under reaction conditions
(250°C, 75% H2 + 25% CO2)
Fig. S21 Representative STEM images (a, d) and corresponding EDS maps (b-d and f-h)
of a-d) o-IPOP/Ru/TiO2 and e-h) m-IPOP/Ru/TiO2 after catalytic tests (6 bar total
pressure, 75% H2 + 25% CO2, 30 sccm, 250 °C) for 72 hrs.
Fig. S22 N 1s and Ti 2p XPS spectra of o-IPOP/Ru/TiO2 and) m-IPOP/Ru/TiO2 after
catalytic tests (6 bar total pressure, 75% H2 + 25% CO2, 30 sccm, 250 °C) for 72 hrs.
 p-IPOP/Ru/TiO2-PT

m-IPOP/Ru/TiO2-PT
Normalized Absorbance (a.u.)

o-IPOP/Ru/TiO2-PT

p-IPOP/Ru/TiO2

m-IPOP/Ru/TiO2

o-IPOP/Ru/TiO2

Ru/TiO2

3750 3500 3250 3000 2750 2500 2250 2000 1750 1500 1250 1000 750
Wavenumber (cm-1)

Fig. S23 FT-IR spectra of samples after CO2 hydrogenation tests (6 bar total pressure,
75% H2 + 25% CO2, 30 sccm, 250 °C) for 72 hrs.
Fig. S24 Representative TEM images of encapsulated catalysts following the same
synthesis procedure used to prepare p-IPOP/Ru/TiO2
 0.20
TiO2 - CO2 + H2 *HCOOx

*CO
0.16
TiO2 - CO2

0.12
TiO2 - N2
Absorbance (K-M unit)

p-IPOP/TiO2 - CO2 + H2
0.08

0.04

0.00
p-IPOP/TiO2 - CO2

-0.04
Nδ+ ⋅⋅
⋅⋅⋅ CO2δ−

-0.08 p-IPOP/TiO2 - N2

2200 2100 2000 1900 1800 1700 1600 1500 1400

Wavenumber (cm-1)

Fig. S25 DRIFT spectra of bare TiO2 and p-IPOP/TiO2 under exposure of N2 or CO2 or
1:3 CO2+H2 at 250 °C , 6 bar total pressure. Specifically, IR spectrum was recorded
under 6 bar in N2 flow as background. 25% CO2 was then introduced also at 6 bar to
study the CO2-sample interaction. After ~1hr when spectra stopped changing, 25% CO2
+75% H2 was introduced at 6 bar to study the nature of surface intermediates during CO2
hydrogenation without Ru. Spectra regions with potential peaks corresponding to *CO,
*HCOOx and *CO2-N have been labeled.
 3. Supplementary Tables (Table S1-S3)

Ru/TiO2 o-IPOP/Ru/TiO2 m-IPOP/Ru/TiO2 p-IPOP/Ru/TiO2

Metallic Area
2.20 0.02 0.06 0.05
(m2·gtotal sample-1)

Table S1 Metallic (Ru) areas of different samples calculated from CO chemisorption experiments.
 reduced R
Sample Path R (Å) CN σ (Å)
2
ΔE0 (eV) chi2 factor
(%)
after Ru-O ±
2.00 0.01 5.33 ± 0.64 0.004 ± 0.001
2.55 ± 1.74 206.0 1.4
oxidation Ru-Ru (oxide) 3.11 ± 0.03 0.88 ± 0.63 0.004 ± 0.001
Ru-Ru (metal) 2.67 ± 0.01 8.48 ± 0.62 0.008 ± 0.001 -4.57 ± 0.72
after X=O 2.19 ± 0.05 0.65 ± 0.30 0.000 ± 0.004 14.52 ± 7.49 232.3
0.1
reduction Ru-X or X=N 2.20 ± 0.04 0.83 ± 0.36 0.000 ± 0.004 12.37 ± 7.34 207.7
or X=C 2.22 ± 0.04 1.15 ± 0.46 0.000 ± 0.004 8.38 ± 7.22 166.1
Ru-Ru (metal) 2.67 ± 0.01 8.62 ± 0.57 0.008 ± 0.001 -4.67 ± 0.61
under X=O 2.18 ± 0.05 0.50 ± 0.29 0.000 ± 0.005 12.06 ± 9.57 142.5
0.1
reaction Ru-X or X=N 2.19 ± 0.05 0.64 ± 0.34 0.000 ± 0.004 8.10 ± 9.09 123.5
or X=C 2.20 ± 0.04 0.84 ± 0.42 0.000 ± 0.004 3.23 ± 8.86 100.7

Table S2 Results of the models used to fit the Ru K-edge EXAFS data. R is the interatomic
distance and CN is coordination number, and σ2 is the mean square deviation in half-path-length. For
after reduction and under reaction data, multiple paths were tried and compared, and the overall best model
is highlighted in red.
 Ru/TiO2 o-IPOP/Ru/TiO2 m-IPOP/Ru/TiO2 p-IPOP/Ru/TiO2
CO rate
8.32 2.13 4.12 16.00
(10-5 molCO2·s-1·kgtotal sample-1)
CH4 rate
2903.46 122.11 114.93 46.39
(10-5 molCO2·s-1·kgtotal sample-1)
C2H6 rate
29.39 1.19 3.44 4.20
(10-5 molCO2·s-1·kgtotal sample-1)
C3H8 rate
3.51 0.30 1.44 2.70
(10-5 molCO2·s-1·kgtotal sample-1)
C4H10 rate
not detected not detected 0.31 0.94
(10-5 molCO2·s-1·kgtotal sample-1)
C2+ rate
32.90 1.50 5.18 7.84
(10 molCO2·s-1·kgtotal sample-1)
-5

Total CO2 conversion rate

2944.67 125.74 124.23 70.23
(10-5 molCO2·s-1·kgtotal sample-1)

Table S3 Mass-normalized reaction rates (CO2 based) of different samples, tested for CO2
hydrogenation at 6 bar total pressure, 75% H2 + 25% CO2, 30 sccm, 250 °C.

Discussion: are changes in reactivity primarily due to physical site blocking?

We can first assume that the polymer functions only as physical blocker with no effect on the intrinsic
TOF per site. Looking at the overall mass-based reaction rate (Table S3), we find that the reaction rate
of p-IPOP/Ru/TiO2 is 3% of unmodified Ru/TiO2. Then if we look at the number of exposed Ru sites
(Table S1) in p-IPOP/Ru/TiO2 (0.05 m2/g), we find that it is 2% of that in Ru/TiO2 (2.2 m2/g). The
close agreement between these two values (3% and 2%) suggests that the polymer is unlikely to
selectively block certain types of sites, as that would result in dramatic discrepancies between mass-
based reaction rate and chemisorption-measured exposed surface area.
However, when comparing the TOF for individual products, we see that CO and C2+ TOFs are notably
increased (50-80 fold) while CH4 TOF is less affected. This suggests that, instead of covering sites
selectively, the polymer is introducing beneficial effects to promote CO production, while suppressing
methanation on each site.
Overall, we wish to emphasize that the changes in selectivity could not simply originate from the
physical blockage by the polymer.
4. Supplementary References

1. A. Aitbekova, et al., Low-Temperature Restructuring of CeO2 Supported Ru

Nanoparticles Determines Selectivity in CO2 Catalytic Reduction. J. Am. Chem. Soc. 140,
13736–13745 (2018).
2. M. Cargnello, et al., Efficient Removal of Organic Ligands from Supported Nanocrystals
by Fast Thermal Annealing Enables Catalytic Studies on Well-Defined Active Phases. J.
Am. Chem. Soc. 137, 6906–6911 (2015).
3. K.-C. Kao, et al., M. A General Approach for Monolayer Adsorption of High Weight
Loadings of Uniform Nanocrystals on Oxide Supports. Angew. Chem. Int. Ed. 60, 7971
(2021).