The Electronic Structure or Charge Delocalization of Sulfated Zirconia (Supported on Multi-Walled Carbon Nanotubes): Acid Sites Probed by X-Ray Absorption Spectroscopy
Changchang Liu, Lisa Pfefferle and Gary L. Haller
Department of Chemical and Environmental Engineering, School of Engineering and Applied Science, Yale University, New Haven, CT 06520-8286, USA
Corresponding author:
Email:
Telephone: 203-432-4378
Fax: 203-432-4387
Supplementary Materials
In the supplementary materials, we address three issues that are mentioned but not elaborated in detail in the manuscript. First, in order to normalize the Zr LIII-edge XANES to the edge-jump, we selected three different post-edge ranges and provide reasons for the one chosen, as seen in Figure S1 and the following text. The resulting spectra of the normalized data were used in Figure 1(a) in the manuscript. Second, using the Zr and S XANES that cover all three Zr L-edges and S K-edge, collected at the beamline X15B at NSLS, BNL, we demonstrate the removal of S content by annealing the sample at 550°C in Figure S2 – adapted from Ref. 14, Fig.7a. Third, we used Scherrer equation fitting to calculate the particle size of bulk ZrO2 (monoclinic) and bulk commercial S-ZrO2 (tetragonal), hereby including the fitted XRD patterns in Figure S3. The fitting results (particle size) are mentioned in the manuscript.
(a)
(b)
(c)
Figure S1. Normalized Zr LIII edge XANES based on three different post-edge regions: (a) 2233 ~ 2300 eV, (b) 2233 ~ 2280 eV, and (c) 2280 ~ 2300 eV.
In order to gather per-atom information at the eg and t2g peak, we need to normalize the data to the edge-jump (to 1). Besides E0, the correct normalization method of the Zr LIII edge XANES requires proper selection of the pre-edge and post-edge, as determined by a straight line extrapolating through 2 data points chosen in the pre-edge and post-edge region, respectively. Athena calculates the subsequent background removal and normalization process automatically. Generally speaking, pre-edge selection is almost never ambiguous. Unlike EXAFS data, where the post-edge oscillation extends to > 800 eV till it is a “flat tail” (in other words, post-edge selection is easy), in XANES data collection, we are usually only interested in the information given within the post-edge range of < 100 eV. In this short range, fine features may still exist, which makes it difficult to make good post-edge selection for the normalization. Careless normalization results in complete or misinterpretation of data. Few study precedents to ours have explicitly expressed how to correctly choose a normalization range for XANES data. One more thing to note here is that, the software Athena automatically picks pre-edge and post-edge regions. While it is almost always correct for EXAFS data, it is almost always incorrect for XANES data (particularly the post-edge). That is why it is crucial to decide the post-edge selection based on the understanding of your own data. Here we use our data as an example to illustrate this important issue.
The pre-edge region is inarguably 2210 ~ 2215 eV. There are two fine features in the post-edge region: (1) 2233 eV ~ 2247 eV, and (2) 2247 eV ~ 2280 eV; and a flat tail region 2280 ~ 2300 eV. We have therefore selected 3 different post-edge regions based on which we normalized our data (Table 1). The results are given in Figure S1.
Table 1.
Post-edge Regions / 2 Fine Features / Flat Tail / Where in Figure 22233 ~ 2300 eV / Inclusive / Inclusive / (a)
2233 ~ 2280 eV / Inclusive / Exclusive / (b)
2280 ~ 2300 eV / Exclusive / Inclusive / (c)
In Figure S1, a dashed line is drawn at y=1 to indicate the edge-jump after normalization. Different post-edge selection gives completely different results at the Zr LIII peak when comparing the unannealed sample with the samples annealed at 4 different temperatures (250°C, 350°C, 450°C, 550°C). Yet, neither (a) nor (b) gives an edge-jump of 1. We therefore believe that only the selection in (c) is the correct post-edge region, i.e., 2280 ~ 2300 eV.
Figure S2. Zr L-edges and S K-edge XANES of S-ZrO2/MWCNT before annealing and after annealing at 550°C in He. (Data adapted from Ref. 14, Fig.7a)
(a)
(b)
Figure S3. Multi-peak fitting with Lorentzian on (a) monoclinic bulk ZrO2 and (b) tetragonal bulk S-ZrO2 to obtain particle size calculation using Scherrer equation.
In Figure S3(a), we used the peak at 2θ=28.27° and obtained the particle size of 30.4 nm; in (b), we used the peak at 2θ=30.24° and obtained the particle size of 9.2 nm.