The Role of Lewis Acid Sites in γ‐Al₂O₃ Oligomerization

Abstract

Olefin oligomerization by γ‐Al₂O₃ has recently been reported, and it was suggested that Lewis acid sites are catalytic. The goal of this study is to determine the number of active sites per gram of alumina to confirm that Lewis acid sites are indeed catalytic. Addition of an inorganic Sr oxide base resulted in a linear decrease in the propylene oligomerization conversion at loadings up to 0.3 wt %; while, there is a >95 % loss in conversion above 1 wt % Sr. Additionally, there was a linear decrease in the intensity of the Lewis acid peaks of absorbed pyridine in the IR spectra with an increase in Sr loading, which correlates with the loss in propylene conversion, suggesting that Lewis acid sites are catalytic. Characterization of the Sr structure by XAS and STEM indicates that single Sr²⁺ ions are bound to the γ‐Al₂O₃ surface and poison one catalytic site per Sr ion. The maximum loading needed to poison all catalytic sites, assuming uniform surface coverage, was ∼0.4 wt % Sr, giving an acid site density of ∼0.2 sites per nm² of γ‐Al₂O₃, or approximately 3 % of the alumina surface.

Olefin oligomerization to liquid hydrocarbons by γ‐Al₂O₃ has recently been reported, and it was suggested that Lewis acid sites are catalytic. In this paper, single Sr²⁺ ions poison the oligomerization activity and Lewis acid sites confirm these are catalytic, and also allow for determination of the number of active sites.

1. Introduction

Since the original report of olefin oligomerization by Day et al. in 1886 where ethylene was thermally converted to liquid hydrocarbon products at 400 °C, ^[1] there has been a continued interest in developing oligomerization technology for converting light olefins into value‐added products. Early studies focused on understanding thermal reactions at high temperatures. These reactions underwent radical chemistry[ ² , ³ , ⁴ , ⁵ , ⁶ , ⁷ , ⁸ ] and produced a wide range of products due to oligomerization and thermal cracking.[ ⁹ , ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ , ²¹ ] Later, heterogeneous catalysts such as metallic Pt, ^[1] metal oxides, ^[10] and metal chlorides were reported for olefin oligomerization; however, ^[22] the first commercial breakthrough used solid phosphoric acid (SPA), a Brønsted acid, for the production of gasoline range hydrocarbons. SPA produced primarily C₆, C₉, and C₁₂ olefins from propylene although small amounts of other olefin products were produced by cracking.[ ²³ , ²⁴ , ²⁵ ] In the 1980’s, Mobil developed a zeolite process using ZSM‐5. ^[26] At elevated pressure and at reaction temperatures from 250–400 °C, isomerization, and cracking produces a mixture of gasoline and diesel range hydrocarbon products; ^[28] while near‐atmospheric pressure and temperatures above 400 °C alkane and aromatic hydrocarbons are produced. ^[29]

Olefin oligomerization has also been developed using homogenous catalysts. For example, the Shell higher olefin process developed in 1968 using nickel(II) complexes. ^[27] Homogenous metal alkyl complexes produce primarily linear alpha olefins (LAOs) by the Cosse‐Arlman mechanism at reaction temperature below about 100 °C and pressures >20 bar. This is a highly effective catalyst; however, rapid deactivation occurs above 250 °C due to reduction of the Ni complex to metallic nickel nanoparticles. ^[30] These catalysts are also readily poisoned by traces of water and oxygen in the feeds.

Recent work by Conrad et al. reported that γ‐Al₂O₃ catalyzes olefin oligomerization at 250–450 °C and 1–40 bar with rates that are significantly higher than the thermal reactions. ^[31] In addition to olefin oligomerization, γ‐Al₂O₃ catalyzed additional reactions of double bond isomerization and H‐transfer, the latter resulting in a small amount of alkane products. While olefin oligomerization dominates at 20–40 bar resulting in oligomer products (C₆₊), at near atmospheric pressure, olefin disproportionation results in non‐oligomer products such as odd carbon number products from ethylene. ^[32] Although γ‐Al₂O₃ is catalytic for olefin oligomerization, it gives products that differ significantly from those typical for Brønsted acid and metal alkyl catalysts. Since γ‐Al₂O₃ is well known to only have Lewis acid sites, these were proposed as the catalytic site for olefin oligomerization.[ ³³ , ³⁴ , ³⁵ , ³⁶ ]

In the present work, the goal is to confirm that Lewis acid sites are the catalytic sites on γ‐Al₂O₃ and to also estimate the active site density. The propylene oligomerization conversion was determined for alumina at a constant set of conditions. The alumina was then poisoned with increasing amounts of Sr²⁺ base to determine the fractional loss in conversion. In addition, the structure of the Sr oxide was determined by X‐ray absorption spectroscopy (XAS) at the Sr K edge, STEM, and pyridine infrared (IR) spectroscopy. The fractional loss in catalytic conversion and Lewis acid sites by pyridine IR suggest that Lewis acid sites are the catalytic site. In addition, Sr on alumina is present as isolated ions suggesting that one Sr poisons one catalytic site, which allows for the determination of the number of catalytic sites.

Experimental Section

Materials

CATALOX SBa‐200 γ‐Al₂O₃ (98 %) was purchased from Sasol, strontium nitrate (99 %) and strontium oxide (97 %) were purchased from Thermo Fisher Scientific, and citric acid (99 %) was purchased from Sigma‐Aldrich. All gases used for catalytic testing were purchased from Indiana Oxygen Company. The ultra‐high purity He (>99.999 %) for XAS experiments was purchased from AirGas, Illinois.

Catalyst preparation

The γ‐Al₂O₃ support used was CATALOX SBa‐200 (pore size=4–10 nm, surface area=200 m² g⁻¹, pore volume ∼0.7 mL/g). XSr/Al₂O₃ catalysts (X wt % Sr) were synthesized using incipient wetness impregnation (IWI). Sr was chosen since is a basic oxide and additionally has a high enough molecular weight for structural characterization by STEM and XAS. Varied amounts of strontium nitrate (99 %, Thermo Fischer Scientific) corresponding to 0.05–10 wt % Sr were dissolved in ultrapure deionized water (18.2 MΩ‐cm). Citric acid (99 %, Sigma‐Aldrich) was added in a 1 : 1 molar ratio with strontium nitrate. Ultrapure deionized water was added until fully dissolved. All catalysts were added dropwise to 5 g CATALOX SBa‐200 γ‐Al₂O₃. The XSr/Al₂O₃ precursors were dried overnight at 125 °C and calcined at a ramp rate of 10 °C/min and held at 550 °C for 3 hours in a convection oven.

Low pressure propylene oligomerization

The propylene oligomerization conversion, product selectivity, and C₄ product distribution were used to compare the poisoning effect of the Sr base. All catalytic tests were performed in a 12.7 mm ID quartz tube, fixed‐bed reactor with a K‐type thermocouple placed in the center of the bed to monitor the reaction temperature. An electrically heated furnace connected to a temperature controller was used to manually adjust the bed temperature. Approximately 2 g of catalyst were loaded into the reactor (supported by quartz wool packed at a constriction in the middle of the reactor). The reactions were performed with high purity C₃H₆ (>99.8 %) at 425 °C, 1.5 bar, and 10 ccm flowrate. The pressure was maintained using a back‐pressure regulator in addition to a pressure gauge at the reactor inlet. The products were analyzed after 15 minutes using a Hewlett Packard 6890 GC with a flame ionization detector (FID). There was little deactivation for the duration of the test up to about 1 hour. Peaks were integrated and compared to a bypass run of pure propylene to determine the conversion. Product selectivity and C₄ distributions were determined by subtracting the impurities present in the high purity C₃H₆ tank (∼0.2 % C₃H₈). The background thermal conversion was <0.3 % for all flow rates.

X‐ray absorption spectroscopy (XAS)

XAS measurements were conducted at the 10‐BM‐B beamline at the Advanced Photon Source (APS) of Argonne National Lab (ANL) and at the 8‐ID beamline at the National Synchrotron Light Source II (NSLS‐II) of Brookhaven National Lab (BNL) ^[37] at the Sr K edge (16.105 keV) to determine the Sr coordination geometry for the 0.05–10 wt % Sr on alumina catalysts. At ANL all samples were ground into a fine powder (∼20 mg), pressed into a stainless‐steel sample holder with 6 sample wells, sealed in a sample cell (1′′ OD quartz tube) with Kapton end caps, and scanned as received in 50 ccm flowing ultra‐high purity He (>99.999 %). At BNL all samples were ground into a fine powder, pressed into wafers, and sealed between Kapton tape. Measurements had a Pb foil (LI edge=15.861 keV) and a third ion chamber that were used for energy calibrations. XAS data were fit using WinXAS 3.1 software. Least‐squares regression fits of the k²‐weighted Fourier transform data from 2.5 to 10.5 Å⁻¹ in k‐space were used to obtain the extended x‐ray absorption fine structure (EXAFS) coordination parameters. The first shell was used to fit all spectra. All samples were fit using theoretical scattering paths using FEFF software. An S_o ² value of 0.80, which was determined by fitting a SrO reference, was used for all samples. All samples were analyzed using the same value of σ².

Scanning transmission electron microscopy (STEM)

Samples were dispersed in ethanol (>99.9 %, VWR) by grinding with a mortar and pestle. The samples were collected by dipping the TEM grid (part# 3620 C‐MB, SPI Supplies) under the ethanol mixture and letting the grid dry in air. The samples were analyzed using a JEOL NEOARM 200CF transmission electron microscope equipped with spherical aberration correction to allow atomic resolution imaging, and an Oxford Aztec Energy Dispersive System (EDS) for elemental analysis. The microscope was equipped with two large area JEOL EDS detectors for higher throughput in acquisition of X‐ray fluorescence signals. Images were recorded in annular dark field (ADF) mode.

Pyridine infrared (IR) transmission spectroscopy

The acid sites of alumina were investigated using transmission IR spectroscopy of adsorbed pyridine. The spectra were collected using a Thermo Fisher Nicolet 4700 spectrometer equipped with a mercury cadmium telluride detector (MCT‐A). Each spectrum was recorded by collecting 32 scans from 4000 to 650 cm⁻¹ at a resolution of 4 cm⁻¹. The samples were pressed into wafers (∼0.1 g) and added to the cell (1” OD quartz tube) with CaF₂ windows. The samples were dehydrated at 350 °C (10 °C/min) in 50 mL/min pure He for 30 minutes then cooled to 150 °C and the unabsorbed spectra were collected. The samples were then introduced to vacuum (<0.1 torr) prior to dosing pyridine at ∼2 torr intervals until surface equilibrium was achieved. Physically adsorbed pyridine was removed under vacuum for 30 minutes followed by purging with pure He at 50 mL/min at 150 °C prior to data collection. The unabsorbed spectrum was subtracted from the saturated spectrum for each sample. The spectra were normalized by mass of catalyst.

2. Results

2.1. Catalyst synthesis

The γ‐Al₂O₃ reference catalyst, CATALOX SBa‐200 was used as received. The XSr/Al₂O₃ catalysts (X wt % Sr) were prepared using incipient wetness impregnation (IWI). Equimolar amounts of strontium nitrate and citric acid were added to deionized water. Citric acid acts as a chelating agent which helps prevent agglomeration of metal ions. ^[38] The resulting solutions were added dropwise to γ‐Al₂O₃. The Sr poison (base) precursors were dried and calcined at 550 °C to remove the organic ligands. The catalysts contained ∼0.05–10 wt % Sr.

2.2. Low pressure propylene oligomerization with γ‐Al₂O₃

The propylene oligomerization conversion of γ‐Al₂O₃ and Sr oxide poisoned γ‐Al₂O₃ catalysts were determined at near atmospheric pressure (Figure 1). Reaction products were collected after 15 minutes at 425 °C and 1.5 bar with 2 g Al₂O₃ at flow rates of 10–100 ccm C₃H₆. In the absence of a catalyst, the thermal conversion was <0.3 % at all flowrates. The space velocity was varied to give conversions from 6–31 % to determine the product selectivity (Figure 2) and C₄ distribution (Figure 3).

Figure 1.

Propylene oligomerization conversion of γ‐Al₂O₃ at 425 °C, 1.5 bar, and propylene.

Figure 2.

Product selectivity at 6.2 % (black), 8.5 % (red), 11 % (grey), 17 % (blue), 24 % (orange), and 31 % (green) conversion using γ‐Al₂O₃ at 425 °C, 1.5 bar, and propylene.

Figure 3.

C₄ product distribution of 1‐butene (red), trans‐2‐butene (green), cis‐2‐butene (purple), isobutene (yellow), and other C₄ species (blue) at different conversion levels using γ‐Al₂O₃ at 425 °C, 1.5 bar, and propylene.

The product selectivity was analyzed at each conversion (Figure 2). At all flow rates, there was ∼80 % selectivity to C_4‐6 hydrocarbons, ∼10–20 % selectivity to ethylene, and ∼5 % selectivity to propane. There was <2% selectivity to methane and only trace amounts of ethane and C₇₊ hydrocarbons. The propane selectivity decreased with increasing conversion. The C₄ selectivity was higher and the C₆ selectivity was lower at 31 % conversion. All conversions had similar C₅ selectivity.

The effect of increasing conversion on the C₄ product distribution was determined (Figure 3). Isobutene was the predominant C₄ product at all conversions. As conversion increased, the fraction of isobutene decreased. Generally, all other olefin products increased slightly. There were additionally small amounts of other hydrocarbons, e. g., n‐butane, isobutane, and butadiene. The reaction products from Figures 1–3 indicate that γ‐Al₂O₃ converts propylene with a significant selectivity to non‐oligomer C₄ and C₅ products consistent with the previous study. ^[31]

2.3. Poisoning of γ‐Al₂O₃ by Sr oxide base

10 wt % Sr oxide base was added to γ‐Al₂O₃ using incipient wetness impregnation (IWI) to poison the catalytic sites. The propylene oligomerization conversion was determined at the same reaction conditions as γ‐Al₂O₃ (2 g, 425 °C, 1.5 bar, 10 ccm flowrate) and the product selectivity and C₄ distribution were analyzed. With 10 % Sr, the conversion was 0.5 %, which corresponds to 99 % loss of γ‐Al₂O₃ catalytic activity (Equation 1). (The thermal background conversion was 0.3 %)

$${\displaystyle \%Poison={\left(1-\frac{X_{{Sr/Al}_2{O}_3}-X_{Thermal}}{X_{{\gamma -Al}_2{O}_3}-X_{Thermal}}\right)}^{*}100\%}$$

Loadings of 0.05–5 wt % Sr were used to partially poison the γ‐Al₂O₃ catalytic activity (Table 1, Figure 4). While the conversion of γ‐Al₂O₃ was 31 %, addition of Sr leads to a linear decrease in conversion up to 0.3 wt % Sr. At higher Sr loadings, there was a non‐linear loss in conversion (Figure 4a). Equation 1 was used to determine the fractional loss in conversion. >95 % of conversion was poisoned above 2 wt % Sr.

Table 1.

Propylene conversion and percent poisoning of XSr/Al₂O₃ catalysts with increasing Sr loading at 425 °C, 1.5 bar and 10 ccm.

Sr Loading (wt %)	Conversion (%)	Poisoning (%)
Thermal	0.3	–
0	31	0
0.05	28	13
0.1	22	30
0.2	16	49
0.3	11	66
0.4	8.5	74
0.5	5.1	85
1	5.0	85
2	1.2	97
5	1.6	96
10	0.5	99

Figure 4.

(a) Propylene conversion and (b) percent poisoning of the catalytic activity of XSr/Al₂O₃ catalysts at 425 °C, 1.5 bar, and 10 ccm propylene.

The product selectivities and C₄ hydrocarbon distribution of 0.2Sr (51 % sites remaining), 0.3Sr (34 % sites remaining) and 0.4Sr/Al₂O₃ (26 % of active sites remaining) were compared to γ‐Al₂O₃ at the same conversion in Table 2 and Figures 5–6. The overall product selectivities were similar when compared at equivalent conversion. The C₄ product distribution of both catalysts were also very similar. Since the selectivity is not changed up to the loss of about 75 % of the catalytic sites, this suggests that all active sites are catalytically identical.

Table 2.

Comparison of the reaction products at equivalent propylene conversion for γ‐Al₂O₃ and partially Sr poisoned γ‐Al₂O₃ at 425 °C and 1.5 bar.

Catalyst	Conversion (%)	CH4	C2H6	C2H4	C3H8	C4	C5	C6
Al2O3	8.5	2	<1	12	6	51	17	12
0.4Sr/Al2O3	8.5	3	<1	12	7	56	15	7
Al2O3	11	1	<1	20	5	50	16	9
0.3Sr/Al2O3	11	2	<1	14	4	51	19	10
Al2O3	17	2	<1	12	5	52	18	11
0.2Sr/Al2O3	16	2	<1	11	4	52	18	14

Figure 5.

Product selectivity of γ‐Al₂O₃ (black) and 0.4Sr/Al₂O₃ (red) at 425 °C, 1.5 bar at ∼8.5 % conversion.

Figure 6.

C₄ product distribution of 1‐butene (orange), trans‐2‐butene (green), cis‐2‐butene (purple), isobutene (yellow), and other C₄ hydrocarbons (blue) of γ‐Al₂O₃ and 0.4Sr/Al₂O₃ at 425 °C, 1.5 bar at ∼8.5 % conversion.

2.4. Local structure of the Sr ions

The local structure of the Sr ions was determined using Sr K‐edge XAS and STEM imaging. Sr K edge X‐ray absorption spectroscopy was used to determine the Sr oxidation state and coordination geometry. The XANES edge energy was used to determine the oxidation state of the Sr samples. Figure 7a shows the XANES of the Sr(NO₃)₂ standard and 0.1Sr/Al₂O₃ from 16.06 to 16.16 keV. The edge energies were 16.1150 and 16.1155 keV, respectively, indicating Sr²⁺ on alumina. The XANES energy of all Sr/Al₂O₃ catalysts were identical, see Table 3.

Figure 7.

(a) Sr K‐edge XANES from 16.06 to 16.16 keV of Sr(NO₃)₂ (black) and 0.1Sr/A1₂O₃ (red) and (b) Magnitude of Fourier transform of the k²‐weighted EXAFS from Δk=2.5 to 10.5 Å⁻¹ of SrO (black) and 10Sr/Al₂O₃ (red) scanned in flowing He at 30 °C.

Table 3.

EXAFS fitting parameters for SrO and all XSr/Al₂O₃ catalysts.

Sample / Sr Loading (wt %)	XANES Energy, keV	CNSr−O (±10 %)	R (Å) (±0.01)	σ2*103 (Å2)	Eo Shift (eV)
SrO	–	6	2.60	–	–
Sr(NO3)2	16.1150	–	–	–	–
0.1	16.1155	4.4	2.58	4.0	−4.5
0.2	16.1154	4.4	2.57	4.0	−4.7
0.3	16.1155	4.3	2.57	4.0	−4.4
0.4	16.1155	4.5	2.58	4.0	−4.5
0.5	16.1155	4.3	2.58	4.0	−4.7
1	16.1155	4.2	2.58	4.0	−3.8
2	16.1155	3.8	2.57	4.0	−3.7
5	16.1154	3.7	2.57	4.0	−3.3
10	16.1154	3.9	2.58	4.0	−3.7

The EXAFS was used to determine the coordination environment, for example the number of Sr−O bonds and bond distance, of the Sr atoms. There was one large peak between 1.5–2.4 Å (phase uncorrected distance) for SrO which corresponds to Sr−O first shell scattering. There was a second, larger peak between 2.7–3.7 Å which corresponds to Sr−O−Sr second shell scattering, Figure 7b (black spectrum). For 10Sr/Al₂O₃ (Figure 7b red spectrum), there was one large peak between 1.2–2.6 Å corresponding to Sr−O first shell scattering. The smaller magnitude of the Fourier transform indicates that there are fewer Sr−O bonds in the Sr/Al₂O₃ catalyst. In the latter, there is also a small peak between 2.6–3.3 Å which has a different magnitude and imaginary part of the Fourier transform when compared to SrO indicating there are no Sr−O−Sr bonds. The EXAFS of all Sr/Al₂O₃ samples were modeled using FEFF with an S_o ² of 0.80 and a σ² of 4.0*10⁻³. All Sr/Al₂O₃ catalysts had ∼4 Sr−O bonds at 2.57–2.58 Å (Figure 8, Table 3). This is slightly shorter than the 2.60 Å observed for SrO. This bond contraction is often observed in single ion species bound to the support surface.[ ³⁹ , ⁴⁰ , ⁴¹ ] The higher shell peak in the catalysts is likely Sr−O−Al, but was too weak to reliable fit; however, the magnitude and imaginary parts of the Fourier transform do not match that of SrO. The XAS data is consistent with isolated Sr²⁺ single ions bound to the oxygen atoms of the γ‐Al₂O₃ support.

Figure 8.

Fourier transform of the k²‐weighted EXAFS from Δk=2.5 to 10.5 Å⁻¹ of 0.1Sr/Al₂O₃ scanned in flowing He at 30 °C; solid line‐magnitude FT and dotted line‐imaginary part FT; red (data) and black (First shell Sr−O fit from ΔR=1.4 to 2.5 Å).

EDS and AC‐STEM images were collected on the 10Sr/Al₂O₃ catalyst (Figure 9–10). At lower loading the Sr could not be imaged. The EDS elemental compositions (Figure 9) had a range of 6–15 wt % Sr on different γ‐Al₂O₃ particles indicating that Sr is not evenly dispersed across the support. High resolution AC‐STEM images (Figure 10) indicate that Sr was atomically dispersed on γ‐Al₂O₃. No small SrO nanoparticles were detected in any image of this sample, which is consistent with the XAS analysis.

Figure 9.

EDS elemental compositions of 10Sr/Al₂O₃ catalyst particles.

Figure 10.

High resolution AC‐ STEM images of 10Sr/Al₂O₃ catalyst with EDS compositions.

2.4.1. Pyridine infrared spectroscopy

Pyridine transmission, infrared spectroscopy (IR) was used to determine the change in the fraction of Lewis acid sites present on γ‐Al₂O₃ with increasing amounts of added Sr oxide base. The IR spectra of adsorbed pyridine on γ‐Al₂O₃ and 10Sr/Al₂O₃ is shown in Figure 11. For γ‐Al₂O₃, there were peaks at 1450, 1492, 1575, and 1614 cm⁻¹, which have previously been assigned to Lewis acid sites.[ ³³ , ³⁴ , ³⁵ , ³⁶ ] Brønsted acid sites which occur at 1540 cm⁻¹ were absent. ^[33] On the 10Sr/Al₂O₃ catalyst, there were no IR peaks indicating that Sr addition had poisoned all Lewis acid sites.

Figure 11.

Pyridine IR spectra of γ‐Al₂O₃ (black) and 10Sr/Al₂O₃ at 150 °C.

The IR spectra of adsorbed pyridine were obtained for alumina with intermediate Sr loadings (0.05–0.2 wt %) to determine if partial Sr coverage leads to partial loss of Lewis acid sites (Figure 12, Table 4). The pyridine IR spectra were similar at all Sr loadings although the peak areas decreased as the Sr loading increased indicating the loss of Lewis acid sites with increasing Sr loading. Equation 2 was used to determine the fractional poisoning of the Lewis acid sites on γ‐Al₂O₃. The percent of poisoning for each Sr loading by pyridine IR were also similar to the percent loss in propylene oligomerization conversion indicating that Lewis acid sites are the catalytically active site on γ‐Al₂O₃ (Table 4).

$${\displaystyle \%Poison={\left(1-\frac{{Area}_{XSr/{Al}_2{O}_3}}{{Area}_{{Al}_2{O}_3}}\right)}^{*}100\%}$$

Figure 12.

Pyridine IR spectra of 0.05Sr/Al₂O₃ (orange), 0.1Sr/Al₂O₃ (blue), and 0.2Sr/Al₂O₃ (green) at 150 °C.

Table 4.

Poisoning results and normalized pyridine IR peak areas for select XSr/Al₂O₃ catalysts.

Sr Loading (wt %)	Oligo Poison (%)	IR Poison (%)
0	0	0
0.05	13	11
0.1	30	38
0.2	49	56
10	99	100

3. Discussion

3.1. Identification of the catalytic site on γ‐Al₂O₃

Previously, γ‐Al₂O₃ was shown to be catalytically active for ethylene oligomerization at reaction temperature from 250–450 °C, high pressure (10–40 atm), and propylene oligomerization is possible at near atmospheric pressure. Since γ‐Al₂O₃ is well known to have Lewis acid sites, these were proposed to be the catalytic sites. ^[31] In this study, the effect of Sr base partial poisoning of γ‐Al₂O₃ on the propylene conversion and IR spectra of adsorbed pyridine was determined to confirm that Lewis acid sites are catalytic and to determine the number of active sites. Sr was an effective poison of the catalytic activity (Figure 4) with loadings as low as 2 wt % resulting in near complete loss in conversion. To achieve complete conversion loss, a loading of 10 wt % Sr was required. The poisoning scales linearly up to ∼0.3 wt % Sr (66 % loss of conversion) suggesting that Sr effectively poisons the active sites. At higher loadings, however, Sr was not evenly dispersed across the support as observed by EDS on different γ‐Al₂O₃ particles, for example, 10 wt % Sr/ Al₂O₃ (Figures 9 and 10). At 10 wt % Sr there is also complete loss of Lewis acid sites. The loss of Lewis acid sites correlates with the loss of catalytic activity suggesting that these are the catalytic sites, as previously proposed. ^[31]

3.2. The local structure of the Sr²⁺ ions

The local structure of the Sr ions was determined using both XAS and STEM. By XANES, the Sr oxidation state in the XSr/Al₂O₃ catalysts is +2. While the number of Sr−O bonds in SrO is six; for Sr ions on γ‐Al₂O₃ there are four Sr−O bonds. The absence of Sr−O−Sr scatting from the small higher shell peak in the XSr/Al₂O₃ catalysts indicates that SrO nanoparticles are not formed and suggests that Sr is present as isolated ions bounded the O ions of the γ‐Al₂O₃ surface. This was confirmed by atomic resolution STEM for the 10 wt % Sr/Al₂O₃. Both the XAS and STEM indicate that isolated Sr²⁺ ions can be added to γ‐Al₂O₃ up to a loading of 10 wt %.

3.3. Determination of the number of catalytic sites on γ‐Al₂O₃

The pyridine IR spectra indicate that there is a decrease in Lewis acid sites with increasing Sr loading. Increasing the Sr loading results in a fractional loss of the number of Lewis acid sites. Table 4 also give the fractional loss in propylene oligomerization conversion with closely matches the loss in Lewis acid sites again indicating that Lewis acid sites are the catalytic site on γ‐Al₂O₃. The site density can be calculated by extrapolating the linear range of Sr poisoning to a fully covered γ‐Al₂O₃ surface. It is estimated that ∼0.4 wt % Sr is necessary to poison all catalytic activity. Since XAS and STEM indicate that Sr is present as a single ion, it is likely that one Sr²⁺ ion poisons one catalytic site. Alternatively, one can calculate the loss in the number of active sites in the range where there is a linear loss in conversion with increasing Sr. For example, at 0.3 wt % Sr there is a 66 % conversion loss (Table 1). Both methods correspond to ∼4*10⁻⁵ moles Sr per gram of catalyst or ∼0.2 Lewis acid sites per nm². The Lewis site coverage is estimated to be about 2–3 % of the alumina surface.

4. Conclusions

Sr/Al₂O₃ catalysts with 0.05–10 wt % Sr were synthesized using IWI and calcination at 550 °C. These synthesis conditions resulted in isolated Sr²⁺ ions bound to the γ‐Al₂O₃ support as determined by XAS and STEM. Increasing the Sr loading decreased the propylene oligomerization conversion. Partially poisoned γ‐Al₂O₃ had a similar product selectivity and C₄ distribution as γ‐Al₂O₃ at the same conversion indicating that all catalytic sites on γ‐Al₂O₃ are identical. Sr oxide base poisoned the active sites linearly up to ∼0.3 wt %; while at loadings of >2 wt % Sr, there is a loss in >95 % of the olefin conversion. STEM results indicated that the Sr was not evenly dispersed across the γ‐Al₂O₃ support. Poisoning of γ‐Al₂O₃ by Sr²⁺ ions also lead to a loss in the number of Lewis acid sites by pyridine IR. The loss in oligomerization activity directly correlated with the loss in Lewis acid sites indicating that Lewis acid sites are the catalytic sites on γ‐Al₂O₃. Extrapolation of the linear poisoning region suggests that ∼0.4 wt % Sr is required to fully poison the Lewis acid sites if Sr was evenly dispersed. This corresponds to ∼0.2 Lewis acid sites per nm² of γ‐Al₂O₃, or about 3 % of the alumina surface.

Conflict of interest

There are no conflicts to declare.

Breckner C. J., Pham H. N., Dempsey M. G., Perez-Ahuatl M. A., Kohl A. C., Lytle C. N., Datye A. K., Miller J. T., ChemPhysChem 2023, 24, e202300244.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.