Cu/TiO2 is a well-known photocatalyst for the photocatalytic transformation of CO2 into methane. The formation of C2+ products such as ethane and ethanol rather than methane is more interesting due to their higher energy density and economic value, but the formation of C–C bonds is currently a major challenge in CO2 photoreduction. In this context, we report the dominant formation of a C2 product, namely, ethane, from the gas-phase photoreduction of CO2 using TiO2 nanotube arrays (TNTAs) decorated with large-sized (80–200 nm) Ag and Cu nanoparticles without the use of a sacrificial agent or hole scavenger. Isotope-labeled mass spectrometry was used to verify the origin and identity of the reaction products. Under 2 h AM1.5G 1-sun illumination, the total rate of hydrocarbon production (methane + ethane) was highest for AgCu-TNTA with a total CxH2x+2 rate of 23.88 μmol g–1 h–1. Under identical conditions, the CxH2x+2 production rates for Ag-TNTA and Cu-TNTA were 6.54 and 1.39 μmol g–1 h–1, respectively. The ethane selectivity was the highest for AgCu-TNTA with 60.7%, while the ethane selectivity was found to be 15.9 and 10% for the Ag-TNTA and Cu-TNTA, respectively. Adjacent adsorption sites in our photocatalyst develop an asymmetric charge distribution due to quadrupole resonances in large metal nanoparticles and multipole resonances in Ag–Cu heterodimers. Such an asymmetric charge distribution decreases adsorbate–adsorbate repulsion and facilitates C–C coupling of reaction intermediates, which otherwise occurs poorly in TNTAs decorated with small metal nanoparticles.

1. S1
SUPPORTING INFORMATION
Asymmetric Multipole Plasmon-mediated Catalysis Shifts the Product
Selectivity of CO2 Photoreduction Towards C2+ Products
Ehsan Vahidzadeh,a
Sheng Zeng,a
Ajay P. Manuel,a
Saralyn Riddell,a
Pawan Kumar,a
Kazi Alam,a, b
and Karthik Shankara
aDepartment of Electrical and Computer Engineering, University of Alberta, 9211 - 116 St, Edmonton, AB T6G
1H9, Canada.
bNational Research Council Nanotechnology Research Centre, 11421 Saskatchewan Dr NW, Edmonton, AB T6G
2M9, Canada
CONTENTS
1. Materials and Methods p. S1
2. Schematic illustration of fabrication process p. S7
3. Field emission scanning electron micrographs of photocatalyst p. S7
4. FDTD simulated electric field profiles p. S8-9
5. Gas chromatograms of CO2 photoreduction experiments collected using
a gas chromatograph equipped with a pulse discharge detector
p. S10-12
6. Ultraviolet photon spectra and band alignment at AgCu-TNTA interface p. S13
7. Ion chromatograms of isotope-labeled CO2 photoreduction experiments
collected using a gas chromatograph-mass spectrometric system (GC-
MS)
8. Elemental abundance ratio of Ag:Cu in AgCu-TNTA sample
p. S15
p. S16
9. Elemental chemical maps of Cu-TNTA sample p. S17
10.References p. S18
1.Materials and Methods
1.1. Materials
Titanium foils (99%, 0.89mm thickness) were purchased from Alfa Aesar. Ammonium fluoride, ethylene
glycol (EG) and methanol were purchased from Fisher Scientific. Silver acetylacetonate (Ag(acac)) was
supplied by Sigma Aldrich and copper acetylacetonate (Cu(acac)2) was purchased from Acros. The

2. S2
solvents used for sample rinsing and material synthesis were deionized water and HPLC grade methanol
purchased from Fisher Scientific. The CO2 gas used for photocatalytic studies had 99% purity and was
obtained from Praxair while the 13
CO2 isotope gas used for sanity tests had 99% purity and was obtained
from Cambridge Isotope Laboratories, Inc. (USA).
1.2. Synthesis of TiO2 nanotube arrays (TNTAs), Ag-TNTA, Cu-TNTA and Ag-Cu-TNTA
Figure S1 schematically depicts the fabrication process used to synthesize metal nanoparticle decorated
TiO2 nanotube arrays. Electrochemical anodization of Ti foils was used to obtain TNTAs. Titanium foils
were cut into smaller pieces (1 cm × 2.5 cm) and washed with methanol, water and acetone thoroughly
before the anodization process. An EG based electrolyte containing 4% v/v water and 0.3 wt% NH4F was
used for anodization. The anodization was performed at room temperature under 40 V applied bias
between the anode and the cathode (another Ti foil) for 3 hours, and the anodized foils were washed with
methanol and water and dried with nitrogen. The obtained TNTAs were subsequently annealed at 450 °C
for 3 hours to induce crystallinity.1
For depositing the metal nanoparticles, two standard 2 mM solutions of Ag(acac)2 and Cu(acac)2
in methanol were prepared. The prepared TNTAs were immersed in a 2 ml solution of Ag(acac)2 or
Cu(acac)2 (for preparing the Ag-TNTA and Cu-TNTA samples respectively) and subjected to UV (365
nm) irradiation of a 4 W lamp for 3 hours. The synthesis procedure for preparing the AgCu-TNTA was
similar except that the irradiated solution was 3 ml in volume consisted of a combination of 2 ml of
Ag(acac)2 and 1 ml of Cu(acac)2 standard solutions (the molar ratio between Ag to Cu was adjusted to
2/1). It is worthy to mention that for AgCu-TNTA sample the concentration of the photodeposition
solution and the molar ratio of Ag/Cu in this solution was optimised (Table 1) to reach the best
photocatalytic performance with highest selectivity toward production of ethane. Based on these results,
it is evident that Ag/Cu with molar ratio of 2/1 resulted in the highest ethane selectivity. Decreasing the

3. S3
concentration of the Ag and Cu solutions from 2mM to 1 and 0.5mM resulted in lower photocatalytic
activity of the sample, also increasing the concentration of the precursor solution to 4mM decreased the
photocatalytic performance, this is probably due to shielding effect of extremely large nanoparticles which
blocks the irradiation of light to the underlying TNTA scaffold.2
Table 1. Optimization of the amount of Ag and Cu in AgCu-TNTA photocatalyst
Sample
number
Concentration
of Ag or Cu in
precursor
solutions
Photodeposition
Procedure
Ag/Cu Molar ratio
in photodeposition
solution
CO2 photoreduction results
1 2 mM 2ml of Ag
precursor solution
+ 2ml of Cu
precursor solution
1/1 9.5 PPM of Ethane
15.2 PPM of Methane
2 2 mM 2ml of Ag
precursor solution
+ 1ml of Cu
precursor solution
2/1 16.9 PPM of Ethane
10.92 PPM of Methane
4 1 mM 2ml of Ag
precursor solution
+ 1ml of Cu
precursor solution
2/1 4.38 PPM of Ethane
3.6 PPM of Methane
5 0.5 mM 2ml of Ag
precursor solution
+ 1ml of Cu
precursor solution
2/1 0.86 PPM of Ethane
3.3 PPM of Methane
6 4 mM 2ml of Ag
precursor solution
+ 1ml of Cu
precursor solution
2/1 9.92 PPM of Ethane
8.12 PPM of Methane
1.3. Photocatalytic reduction of CO2
Experiments measuring the photoreduction of CO2 were conducted using a 32 ml cylindrical shaped
stainless steel flanged reactor equipped with a quartz window. The photocatalyst (on Titanium foil) and a
few droplets of water were placed at the bottom of the reactor following which the reactor was purged
with pure CO2 gas three times in order to remove residual air inside the reaction chamber and the tubing.

4. S4
The purged reactor was pressurised with pure CO2 gas to 50 psi and subsequently placed on top of a hot
plate set to 50 °C to evaporate the water droplets. To investigate the photocatalytic activity of the samples,
the reactor was subjected to solar irradiation. The light source used in these experiments was a standard
Class A solar simulator (Newport) with an intensity of 100 mW cm-2
(one sun). The duration of solar
irradiation was 2 hours. After the irradiation step, the products of the reaction were analyzed using a
Shimadzu gas chromatograph (GC-2014) equipped with both Porapak Q and a molecular sieve column
and a pulsed discharge detector (PDD). The carrier gas used was ultrahigh purity He. Each of the
measurements was repeated three times and the average of these three measurements was reported as the
final result. Since it is impossible to purge all of the air from the GC after each run, the gas chromatogram
always exhibit huge peaks for oxygen and nitrogen and a really small peak for carbon monoxide
(originating from the lab’s atmosphere). Hydrocarbon mixtures of known composition were used to
calibrate the output of the GC. The examples of chromatogram of the products obtained from the GC
measurements for each of the samples are shown in Figures S6 through S8. In addition to normal gas
chromatography tests, a GC-MS was used to identify the products in the isotope labelling test using 13
CO2
instead of normal 12
CO2 gas. These tests were performed under identical reaction and illumination
conditions and their purpose was to verify that the hydrocarbon products originated from the reactant
(CO2) instead of carbon contamination. The resulting ion chromatograms are shown in Figure S10.
1.4 Characterization
The morphologies of the samples were investigated using a field emission scanning electron microscope
(FESEM, Zeiss Sigma) operating at an accelerating voltage of 5 kV. The high-resolution morphological
and structural properties of the samples were investigated using a high-resolution transmission electron
microscope (HR-TEM, JEOL-2200 FS TEM) operating at an acceleration voltage of 200 kV. The
crystallinity of the materials was determined using x-ray diffraction (XRD, Bruker-D8 Discover

5. S5
instrument) equipped with a sealed Cu Kα X-ray source (40 kV 44 mA, λ = 0.15418 nm) in the 2θ range
of 10–70°. The vibrational properties of the samples were studied using a Raman spectrometer (Nd:YAG
laser Raman Microscope, Nicolet Omega XR). The Raman spectrometer was equipped with a 532 nm
laser source with an incident power of 10 mW, which was used to excite the samples. The optical
absorption properties of the materials were studied by collecting the UV-visible diffuse reflectance spectra
(UV-Vis DRS, Perkin Elymer-Lambda-1050 UV-Vis-NIR spectrophotometer). The surface binding
energies and the oxidation states of the elements were investigated using x-ray photoelectron spectroscopy
(XPS) with Kratos Analytical-Axis-Ultra photoelectron spectrometer which was quipped with a
monochromatic Al-Kα source (15 kV, 50 W) and a photon energy of 1486.7 eV. C1s with binding energy
of ≈ 284.8 eV was chosen as the reference peak and the binding energies of all the elements were corrected
based on C1s reference point. The valence band positions and work functions of the synthesized materials
were determined using ultraviolet photoelectron spectroscopy (UPS) equipped with an excitation source
consisting of a 21.21 eV He lamp.
1.5 Conversion of GC peak areas to moles
The GC was calibrated with a standard cylinder with concentrations of 100 ppm methane and 100 ppm
ethane. As a result, the reported chromatogram results are the areas correspond to the ppm levels of the
products. The ppm concentration of the product can be converted into mole fraction of product using the
following formula:
= eq.(s1)
Where Mproduct is the mole fraction of the product in the reactor and ppmproduct is the ppm value reported
by the GC. Considering that the total pressure (Ptotal) in the reactor was 50 psi (3.45 × 105
Pa), using

6. S6
Dalton’s law, the partial pressure of the product can be calculated using eq. (s2). After the partial
pressure is obtained, the moles of product can simply obtained using the ideal gas law (eq. (s3)).
= × eq.(s2)
=
.
.
eq.(s3)
Where V is the volume of the reactor (32 ml), R is the ideal gas constant (8.314 J mol-1
K-1
) and T is the
temperature of the reactor (323 K).
The selectivity toward production of methane were calculated using the following formula:
!ℎ# $ %$&$'!()(!* =
+,-. ! -/ $!ℎ# $ 01-2.'$2 (µ,-&)
6-!#& #,-. ! -/ 01-2.'!7 (µ,-&)
eq. (s4)
1.6 Finite difference time domain (FDTD) simulations
The near field and far field electromagnetic properties of the samples were simulated by FDTD method
using Lumerical software package. The modeling process involves defining the structures, meshing, and
solving the Maxwell’s equations to obtain the desired results. The simulations were performed for TNTs,
Ag-TNTA, Cu-TNTA, and AgCu-TNTA samples under vacuum condition. In all of the simulations, the
length of the TNTAs was set to 6 microns with an inner diameter of 60 nm and an outer diameter of 75nm.
A total field scatter field (TFSF) light source with a bandwidth of 250-800 nm was used to illuminate the
samples at normal incidence from above, and the boundary conditions were set to perfectly matched layer
(PML). Further details regarding the FDTD simulations may be found in our previous works.3, 4

7. S7
Figure S1. Schematic diagram of the fabrication process of AgCu-TNTA samples.
Figure S2. FESEM images of (a) TNTAs and (b) AgCu-TNTA samples

8. S8
Figure S3. Lumerical FDTD simulation results showing the local electric field profiles for (a) Bare
TNTA (b) Cu-TNTA (c) Ag-TNTA and (d) AgCu-TNTA. The Ag-TNTA and AgCu-TNTA samples
have the largest number of hot spots.
Figure S4. Lumerical FDTD simulation results showing the local electric field profiles for (a) Isolated
120 nm diameter Ag NP in vacuum (xy plane) (b) Isolated 120 nm diameter Ag NP on a single TNTA

9. S9
(xy plane) (c) Electric field profile in the substrate (xy) plane for a Ag-Ag homodimer with 3 nm edge-
to-edge spacing situated on adjacent TNTAs and (d) Electric field profile in the substrate (xz) plane for a
Ag-Ag homodimer with 3 nm edge-to-edge spacing situated on adjacent TNTAs. For the isolated Ag
NPs in (a) and (b) no field scaling was used while for the Ag-Ag homodimers in (c) and (d), the electric
field intensities were scaled to a jet colormap.
Figure S5. Lumerical FDTD simulation results showing the local electric field profiles for (a) Isolated
120 nm diameter Cu NP in vacuum (xy plane) (b) Isolated 120 nm diameter Cu NP on a single TNTA
(xy plane) (c) Electric field profile in the substrate (xy) plane for a Cu-Cu homodimer with 3 nm edge-
to-edge spacing situated on adjacent TNTAs and (d) Electric field profile in the substrate (xz) plane for a
Cu-Cu homodimer with 3 nm edge-to-edge spacing situated on adjacent TNTAs. For the isolated Cu
NPs in (a) and (b) no field scaling was used while for the Cu-Cu homodimers in (c) and (d), the electric
field intensities were scaled to a jet colormap.

10. S10
Figure S6. Gas chromatogram of gaseous products after illumination of using AgCu-TNTA sample with
AM 1.5G simulated sunlight for 2 h.

11. S11
Figure S7. Gas chromatogram of gaseous products after illumination of using Cu-TNTA sample with
AM 1.5G simulated sunlight for 2 h.

12. S12
Figure S8. Gas chromatogram of gaseous products after illumination of using Ag-TNTA sample with
AM 1.5G simulated sunlight for 2 h.
Table 2. Comparison between the photocatalystic performance of AgCu-TNTA sample with previous
literature
Sample Light Source Products (µmoles/gr hr) Reference
Ag-TiO2 Xe lamp (300 W) 1.4 Methane 5
Ag-TiO2 365nm -UVA lamp (8 W) 2.64 Methane 6
Ag-TiO2 Solar Simulator (AM 1.5) 4.93 Methane 7
Cu-TiO2 Xe lamp (200 W) 5.4 CO
8.7 Methane
8
Cu-TiO2 Xe lamp (300 W) 13.6 CO
1 Methane
9
Cu-TiO2 UV Light 2.35 Methane 10
AgCu-TNTA Solar Simulator (AM 1.5) 9.38 Methane
14.5 Ethane
This
Work

13. S13
Figure S9. (a) UPS work function spectra of TNTA and AgCu-TNTA. The inset shows cut-off energy
(Ecut-off) of secondary electrons. The value of work function (WF) was determined using the expression
WF (ϕ) = 21.21 – Ecut-off, where 21.21 eV is the energy of the incident He I line of He discharge lamp
(b) UPS valence band spectra showing the position of the valence band maxima (VBmax) below the
Fermi level (c) Schematic diagram of the proposed band diagram at the AgCu-TNTA interface showing
upward band bending due to the formation of a Schottky barrier and concomitant interfacial depletion
region and (d) Schematic diagram of an accumulation-type heterojunction at the metal/semiconductor
interface where metal nanoparticles play the role of an electron sink.
To gain insight into the electronic band structure and charge migration mechanism, work function
(WF) and valence band (VB) spectra of TNTAs and AgCu-TNTAs were investigated using ultraviolet
photoelectron spectroscopy (UPS) (Fig. S11a and S11b). The work function value was estimated from
work function spectra using the expression WF (ϕ) = 21.21−Ecut-off, where 21.21 eV is the energy of the
incident, He I line of a He discharge lamp, and Ecut-off is the cut-off energy of secondary electrons. The
intersecting points after extrapolation of linear region of graph on horizontal and vertical region give the

14. S14
cut-off energy of secondary electrons (Ecut-off). The Ecut-off values for compact TNTAs and AgCu-TNTAs
were calculated to be 17.06 and 16.41 eV and the associated WF values (ϕ) were determined to be 4.15
and 4.80 eV respectively (Fig. S11b and Inset of Fig. S11a). Since the work function represents the
position of the Fermi level with respect to vacuum, we found that the Fermi level of TiO2 was downshifted
following decoration of Cu, Ag nanoparticles on TiO2 nanotube arrays demonstrating the formation of a
Schottky junction. Further, valence band maxima calculated from the intersecting point of extrapolated
linear region of graph for of TNTAs and AgCu-TNTAs were found to be 3.20 and 1.43 eV below the
Fermi level (Fig. S11a).
To understand the mechanism of charge separation, the band diagram of AgCu-TNTAs sample is
schematically depicted in Fig. S11c. The formation of a Schottky junction at the interface of the TNTAs
and the AgCu nanoparticles creates a depletion region in TiO2 concomitant with an upward bending of
the bands. We estimate the magnitude of the band-bending to be at least 1.2 V, in line with prior reports.
Many reports have previously measured the electron density in anodically formed TiO2 nanotube arrays
to be 1018
-1020
cm-3
,11
indicating the titania to exhibit near-degenerate n-type semiconducting behavior.
This is confirmed by the valence band maximum in Fig. S10a to be 3.2 eV below the Fermi level, almost
equal to the bandgap of anatase TiO2 thus implying that the Fermi level in TiO2 coincides with the
conduction band edge (implying degenerate n-type doping). Taken together, the combination of a high
built-in potential and a large electron density TiO2 indicate a very strong interfacial electric field. The
formation of the Schottky junction at the interface along with a strong internal electric field is beneficial
for the photocatalytic activity of the sample as the internal electric field can increase the charge separation.
In the Schottky barrier band alignment, the electrons from metal nanoparticles can get injected into the
semiconductor leaving the holes behind inside the metal nanoparticles.12, 13
Even though the UPS
measurements suggests an upward band bending at the interface, there is another widely accepted

15. S15
hypothesis in the literature which suggests that the metallic nanoparticles can act as an electron sink (trap).
The conventional belief is that the metallic nanoparticles can accept the electrons from the TiO2
semiconductor and the CO2 photoreduction occurs on the reactants adsorbed on the metallic nanoparticle
surface, a phenomenon termed ‘Schottky Barrier Electron Trapping’ in the literature (Fig. S11b).12
However, the metallic electron sink explanation is unlikely to be true for our AgCu-TNTA photocatalyst,
since the presence of the internal electric field in the direction implied by Fig. S11c would direct
photogenerated electrons toward TiO2 and holes toward the bimetallic nanoparticles.
Figure S10. GC-MS ion chromatogram showing the results of 13
C isotope labeled CO2 photoreduction
experiment for the AgCu-TNTA sample (a) The peak at m/e = 17 confirms the presence of 13
CH4 and (b)
The peak m/e = 32 confirms the presence of 13
C2H6.

16. S16
Figure S11. Elemental abundance ratio of Ag:Cu on the surface of AgCu-TNTA sample estimated to be
43.4:1 from measurement of the areas under the Ag3d and Cu2p. The area ratio of [Ag3d]:[Cu2p] is
31.2:1. The photoionization cross-sections of Ag3d and Cu2p are 0.2474 and 0.3438 respectively.14

17. S17

18. S18
Figure S12. Energy dispersive x-ray spectroscopy (EDS) mapping of Cu-TNTA sample (a) Scanning
electron micrograph of sample spot selected for EDS mapping, and elemental maps of various elemnts
(b) Ti (c) O (d) N (e) Cu (f) S (g) Si (h) C (i) Al (j) F. It is evident from the elemental maps that while
Ti and O are present nearly everywhere on the sample excepting the agglomerated nanoparticles on the
surface, Cu is present only at locations corresponding to the agglomerated nanoparticles. The remaining
elements are absent from the surface of the sample.
REFERENCES
1. Varghese, O. K.; Gong, D.; Paulose, M.; Grimes, C. A.; Dickey, E. C., Crystallization and high-temperature
structural stability of titanium oxide nanotube arrays. J. Mater. Res. 2003, 18 (1), 156-165.
2. Jo, S.; Verma, P.; Kuwahara, Y.; Mori, K.; Choi, W.; Yamashita, H., Enhanced hydrogen production from
ammonia borane using controlled plasmonic performance of Au nanoparticles deposited on TiO 2. Journal of
Materials Chemistry A 2017, 5 (41), 21883-21892.
3. Manuel, A. P.; Barya, P.; Riddell, S.; Zeng, S.; Alam, K. M.; Shankar, K., Plasmonic photocatalysis and SERS
sensing using ellipsometrically modeled Ag nanoisland substrates. Nanotechnology 2020, 31 (36), article no
032134.
4. Kar, P.; Zeng, S.; Zhang, Y.; Vahidzadeh, E.; Manuel, A.; Kisslinger, R.; Alam, K. M.; Thakur, U. K.; Mahdi,
N.; Kumar, P., High rate CO2 photoreduction using flame annealed TiO2 nanotubes. Appl. Catal. B-Environ. 2019,
243, 522-536.
5. Yu, B.; Zhou, Y.; Li, P.; Tu, W.; Li, P.; Tang, L.; Ye, J.; Zou, Z., Photocatalytic reduction of CO2 over Ag/TiO2
nanocomposites prepared with a simple and rapid silver mirror method. Nanoscale 2016, 8 (23), 11870-11874.
6. Kong, D.; Tan, J. Z. Y.; Yang, F.; Zeng, J.; Zhang, X., Electrodeposited Ag nanoparticles on TiO2 nanorods
for enhanced UV visible light photoreduction CO2 to CH4. Applied Surface Science 2013, 277, 105-110.
7. Hong, D.; Lyu, L.-M.; Koga, K.; Shimoyama, Y.; Kon, Y., Plasmonic Ag@ TiO2 Core–Shell Nanoparticles for
Enhanced CO2 Photoconversion to CH4. ACS Sustainable Chemistry & Engineering 2019, 7 (23), 18955-18964.
8. Zhai, Q.; Xie, S.; Fan, W.; Zhang, Q.; Wang, Y.; Deng, W.; Wang, Y., Photocatalytic conversion of carbon
dioxide with water into methane: platinum and copper (I) oxide co-catalysts with a core–shell structure.
Angewandte Chemie 2013, 125 (22), 5888-5891.
9. Wang, Z.; Zhou, W.; Wang, X.; Zhang, X.; Chen, H.; Hu, H.; Liu, L.; Ye, J.; Wang, D., Enhanced
Photocatalytic CO2 Reduction over TiO2 Using Metalloporphyrin as the Cocatalyst. Catalysts 2020, 10 (6), article
no 654.
10. Mekasuwandumrong, O.; Jantarasorn, N.; Panpranot, J.; Ratova, M.; Kelly, P.; Praserthdam, P., Synthesis
of Cu/TiO2 catalysts by reactive magnetron sputtering deposition and its application for photocatalytic reduction
of CO2 and H2O to CH4. Ceramics International 2019, 45 (17), 22961-22971.
11. Mohammadpour, A.; Kar, P.; D Wiltshire, B.; M Askar, A.; Shankar, K., Electron transport, trapping and
recombination in anodic TiO2 nanotube arrays. Curr. Nanosci. 2015, 11 (5), 593-614.
12. Farsinezhad, S.; Sharma, H.; Shankar, K., Interfacial band alignment for photocatalytic charge separation
in TiO 2 nanotube arrays coated with CuPt nanoparticles. Phys. Chem. Chem. Phys. 2015, 17 (44), 29723-29733.
13. Kar, P.; Zhang, Y.; Mahdi, N.; Thakur, U. K.; Wiltshire, B. D.; Kisslinger, R.; Shankar, K., Heterojunctions of
mixed phase TiO2 nanotubes with Cu, CuPt, and Pt nanoparticles: interfacial band alignment and visible light
photoelectrochemical activity. Nanotechnology 2017, 29 (1), article no 014002.
14. Yeh, J. J.; Lindau, I., Atomic subshell photoionization cross sections and asymmetry parameters: 1 ⩽ Z ⩽
103. Atomic Data and Nuclear Data Tables 1985, 32 (1), 1-155.