Designing alkaline-rich Ba/MnO₂ catalysts for efficient oxidative coupling of alcohols and amines at low temperatures

Abstract

In this work, a series of alkaline earth metal-modified manganese oxide catalysts were prepared via the impregnation method and applied to the oxidative coupling reaction of alcohols and amines to imine. The prepared catalysts were characterized by X-ray diffraction (XRD), N₂-adsorption, field emission-scanning electron microscopy (FE-SEM), H₂ temperature-programmed reduction (H₂-TPR), NH₃-temperature-programmed desorption (NH₃-TPD), CO₂-temperature-programmed desorption (CO₂-TPD) and X-ray photoelectron spectroscopy. Among the catalysts, Ba/MnO₂ showed the best catalytic performance in the reaction and exhibited the highest Mn³⁺/Mn⁴⁺ ratio, surface weak basic sites and a relatively small number of weak acidic sites. Lattice oxygen mobility and Mn³⁺/Mn⁴⁺ ratio were found to play important roles in the catalytic activity of aerobic reactions. The weak and medium basic sites on the catalyst surface can further promote the alcohol–amine oxidative coupling process, as they serve as active centers for the activation of benzyl alcohol, which is the rate-determining step for this oxidative coupling reaction.

---

1. Introduction

Imines, also known as Schiff bases, are an important class of nitrogen-containing organic compounds. Due to their unsaturated C=N double bonds, they serve as crucial nitrogen sources and are widely applied in biological, agricultural, pharmaceutical, and fine chemical industries.^1–4 Current industrial synthesis of imines primarily relies on the condensation of carbonyl compounds with primary amines under strong acidic conditions, which suffers from high energy consumption, environmental pollution, and low product selectivity.^5–9 In contrast, the oxidative coupling of alcohols and amines in the presence of air or oxygen—producing only water as a by-product—represents a greener chemical process. Particularly in the context of growing environmental and resource constraints, developing aerobic oxidation technologies under mild conditions has emerged as a major research focus in catalysis.^10–13

However, the relative stability of molecular oxygen under mild conditions poses a key challenge for its efficient activation. While supported noble metal catalysts (e.g., Au, Pt, and Ru) demonstrate promising catalytic activity for this process,^14,15 their limited natural abundance and high cost remain significant drawbacks. Transition metal oxide catalysts (Mn, Co, Ni, Cu, Ce, Zn, etc.) offer advantages such as abundant reserves, low cost, and good stability, and generally exhibit high reactivity in aerobic oxidation reactions. However, they typically require larger catalyst loadings and harsher reaction conditions.^16,17 Therefore, there is an urgent need to develop highly active and stable transition metal catalysts to meet the low-cost requirements for industrial applications.

Among transition metal catalysts, manganese dioxide (MnO₂) exhibits superior redox performance due to its diverse oxidation states, versatile crystal structures, and excellent oxygen reduction/evolution activities, making it one of the most promising candidate materials to replace noble metal catalysts.^18–22 Jun Bu et al.¹⁸ prepared a series of manganese oxide catalysts with different crystal phases and applied them to the low-temperature oxidative coupling reaction of alcohol and amines. The results revealed that γ-MnO₂ exhibited the highest catalytic activity with the yield of imine reaching up to 98.1% in 12 h under an air atmosphere. The high activity of γ-MnO₂ was attributed to its high concentration of low-valence manganese ions. Fushan Chen et al.²² adopted a template-free oxalate route to synthesize different mesoporous manganese oxides by controlling the calcination conditions of the precursor. The optimal catalyst showed excellent catalytic activity and regenerative stability in the coupling reaction of alcohol–amine oxidation. The high activity of M-350 was mainly attributed to lattice oxygen mobility and (Mn³⁺ + Mn⁴⁺)/Mn²⁺ ratio. Clearly, the migration of lattice oxygen and multivalent manganese species are key factors influencing the reaction activity. However, in addition to these factors, the acid–base environment on the catalyst surface usually affects the adsorption and activation of reactant molecules and the desorption process of products, which may also play an important role. Jinling Song et al.²³ used Mg–Al composite oxides with dual acidic and basic properties as catalysts in the alcohol–amine oxidation coupling reaction, investigating the impact of the catalyst's surface acidity and basicity on catalytic performance. The study results showed that under the synergistic effect of acidic and basic centers, the catalyst can effectively catalyze the oxidation coupling reaction between alcohol and amine. Among these, the weakly basic site may be the primary active center for alcohol activation, and the oxidation of alcohol is the rate-determining step of the reaction. Therefore, they believed that the number of weakly basic centers on the surface to some extent determines the performance of the catalyst in this reaction.

Considering that the surface acid–base environment and lattice/adsorbed oxygen of catalysts may be critical factors influencing the alcohol–amine oxidative coupling reaction, this work focuses on manganese oxide as the research subject. A series of alkaline earth metal-modified supported manganese oxide catalysts were designed and prepared via a simple impregnation method to modulate the surface acid–base environment of manganese oxide. The correlation between the physicochemical properties of the catalysts (particularly surface acidity/basicity and Mn valence states) and their catalytic activity was systematically investigated. Based on the experimental results, a reaction mechanism hypothesis was proposed. Additionally, this work optimized three key reaction parameters—reaction temperature, catalyst dosage, and feedstock ratio—using the response surface methodology (RSM), with a comprehensive analysis of the interactive effects among these factors.

2. Experimental

2.1. Catalyst preparation

Magnesium acetate tetrahydrate (≥99.0%), calcium acetate monohydrate (≥98.0%), strontium acetate (≥99.0%), barium acetate (≥99.0%), isopropyl alcohol (≥99.7%), manganese acetate (≥98.0%) and potassium permanganate (≥98.0%) of analytical grade were purchased from Aladdin®.

MnO₂ was prepared by a simple chemical precipitation method. In a typical synthesis procedure, 13.2 g Mn(CH₃COO)₂·4H₂O was dissolved in 80 mL distilled water and 50 mL isopropanol. The reaction mixture was heated to 80 °C for 30 min under continuous stirring, and 50 mL aqueous solution of KMnO₄ (0.36 M) was added drop-wise to the reaction mixture and kept for 30 min. After the reaction mixture was cooled to room temperature, the product was filtered with distilled water and ethanol and then dried at 80 °C for 24 h.

The alkaline earth metal (Mg²⁺, Ca²⁺, Sr²⁺, Ba²⁺) loaded MnO₂ catalysts were prepared using the impregnation method. Taking Mg(CH₃COO)₂·4H₂O as an example: a specified amount was placed in a beaker, mixed with 50 mL of distilled water, and stirred thoroughly to form a solution. Subsequently, 5 g of MnO₂ was added to this solution, followed by continuous heating and stirring for 2 h. After cessation of stirring, the beaker was transferred to a drying oven and dried at 100 °C for 12 h. Upon cooling to room temperature, the sample was calcined in a muffle furnace at 450 °C for 3 h. After natural cooling to ambient temperature, the final product was collected and designated as Mg/MnO₂. The preparation procedures for Ca/MnO₂, Sr/MnO₂, and Ba/MnO₂ followed an identical methodology.

2.2. Catalyst characterization

X-ray diffraction (XRD) patterns of the catalyst samples were measured with an Empyrean X-ray diffractometer using a nickel-filtered Cu Kα source at a wavelength of 0.154 nm. An accelerating voltage of 40 kV and a current of 40 mA were used. A slit width of 0.25° was used as the source. Scans were collected using a PIXcel3D detector. N₂ adsorption/desorption isotherms were measured using a Micromeritics ASAP 2010 N analyzer. Specific surface areas were calculated using the BET model. Pore size distributions were evaluated from desorption branches of nitrogen isotherms using the BJH model. Scanning electron microscopy (SEM) observations of the catalyst samples were performed using a Hitachi SU8020. X-ray photoelectron spectroscopy (XPS) measurements were conducted on an ESCALAB250 spectrophotometer. The electron binding energies were referenced to the C 1s (Eb = 284.6 eV) peak. TPR was performed by using Micromeritics AutoChem II 2920 apparatus equipped with a HIDEN QIC-20 mass spectrometer (MS). The catalyst (50 mg) was pretreated at 300 °C for 1 h under flowing Ar (50 mL min⁻¹). Upon cooling to 50 °C, a flow rate of 50 mL min⁻¹ of 5% H₂–N₂ was used for reduction and the temperature was increased linearly from 100 °C to 600 °C at a rate of 10 °C min⁻¹. Temperature programmed desorption (TPD) of NH₃ and CO₂ were used to characterize the acidic and basic sites over the studied catalysts, respectively. The TPD experiments were conducted using a ChemBET Pulsar TPR/TPD instrument (Quantachrome Instruments) with a built-in TCD detector. Typically, a 100 mg catalyst was used in each measurement. The catalyst was first purged with He (UHP grade, Airgas) at 450 °C for 1 h with a 10 °C min⁻¹ heating ramp rate, and then cooled down to 50 °C. A flow of CO₂ (research grade, Airgas) or NH₃ (electronic grade, Airgas) was introduced into the tubular catalyst bed for 30 min at 120 °C for CO₂ adsorption and 120 °C for ammonia adsorption. After purging the catalyst bed for approximately 30 min with He to evacuate the physisorbed NH₃ or CO₂, the catalyst was heated to 500 °C with a 10 °C min⁻¹ heating ramp rate. The change in thermal conductivity due to the concentration change of NH₃ or CO₂ in the effluent was recorded on the TCD.

2.3. Catalyst evaluation

The catalytic evaluation for alcohol–amine oxidative coupling was conducted using a 50 mL dual-neck reactor containing 10 mL toluene, 0.3 g catalyst, 1.0 mmol benzyl alcohol, and 2.0 mmol aniline. The system was maintained at 60 °C under atmospheric oxygen supplied through an inflated balloon. Reaction monitoring involved periodic sampling through a filtered extraction device, with subsequent analysis using HP-5 column gas chromatography coupled with flame ionization detection. Reference compounds including parent alcohols, amines, and imines served as analytical standards. The calculation formula, chromatographic conditions and other experimental operation information are detailed in the SI.

2.4. Response surface methodology (RSM)

The response surface methodology (RSM) was employed to design experiments for optimizing the conditions of Ba/MnO₂-catalyzed oxidative coupling of alcohols and amines for imine synthesis. Experiments were designed using Design-Expert Version 13 (Stat-Ease, Inc., Minneapolis, MN, USA).^24,25 The Box–Behnken design (BBD) was used to establish the relationship between imine yield (response) and three variables: reaction time (A), aniline/alcohol ratio (B), and toluene dosage (C). Based on the 33-point Box–Behnken design framework, 17 experiments were conducted, including 12 factorial points and 5 central points. See the SI for detailed design conditions and other contents.

3. Results and discussion

3.1. Catalyst characterization

Fig. 1 shows the XRD patterns of alkaline earth metal-loaded manganese dioxide catalysts. For all catalysts, nine characteristic diffraction peaks were observed at 2θ = 12.78°, 18.11°, 28.84°, 37.52°, 41.97°, 49.86°, 60.27°, 65.11° and 69.71°, which can be assigned to the (110), (200), (310), (211), (301), (411), (521), (002) and (541) crystal planes of α-MnO₂ (PDF#44-0141), respectively. In the case of loaded manganese oxide catalysts, no characteristic diffraction peaks corresponding to alkaline earth metal species were detected, indicating that the alkaline earth metal species either exist as small crystallites or are uniformly dispersed on the manganese dioxide surface.^26–28

Fig. 1 XRD patterns of alkaline earth metal-loaded manganese dioxide catalysts.

Textural properties of catalysts were investigated by nitrogen adsorption–desorption measurements (Fig. S1 and Table S2). All of them were type IV patterns according to the International Union of Pure and Applied Chemistry (IUPAC) classification, which confirms that these MnO₂ catalysts have a mesoporous structure. For loaded catalysts, a decrease in the BET surface area was observed due to the loading of alkaline earth metals. However, their pore volume and mesoporous diameter were minimal. Pore size distribution plots showed that the most integrable pore size of alkaline earth metal-loaded manganese dioxide catalysts was centered at around 3.6–3.7 nm (Fig. S1(b)). These results indicate that the loading of alkaline earth metals had limited impact on the pore structure of the catalyst.^29,30 To further investigate the morphological structure and surface elemental distribution of the catalysts, SEM-EDS tests were performed on the catalyst samples. Fig. 2 shows the SEM-EDS spectra of MnO₂. The SEM image (Fig. 2(a)) reveals that the catalyst exhibits an irregular spherical morphology. The EDS spectra (Fig. 2(f)) demonstrate that only Mn and O elements are present on the catalyst surface, indicating that the MnO₂ support is pure and free of impurities. Fig. 2(b–e) and (g–j) displays the microstructures of the alkaline earth metal-loaded catalysts. All catalysts exhibit irregular spherical morphologies, consistent with the results shown in Fig. 2(b–e), suggesting that the loading of alkaline earth metals (Mg, Ca, Sr, and Ba) does not alter the inherent morphological structure of MnO₂. The corresponding EDS spectra of the loaded catalysts (Fig. 2(g–j)) confirm that Mg-, Ca-, Sr-, and Ba-related species are successfully loaded and uniformly dispersed on the MnO₂ support surface.

Fig. 2 SEM-EDS spectra of alkaline earth metal-loaded manganese dioxide catalysts: SEM (a–e) and EDS mapping images (f–j).

The surface acidic and basic properties of the samples were investigated by TPD measurements using NH₃ and CO₂ as probe molecules, respectively. Fig. 3(a) shows the desorption profiles of CO₂ from alkaline earth metal-loaded manganese dioxide catalysts. From Fig. 3(a), we can see that all catalysts exhibited a prominent CO₂ desorption peak in the range of 150–350 °C and the basic sites on these catalysts were arbitrarily considered weak and medium.^18,21,31 Adding alkaline earth metal species slightly increased the total CO₂ uptake compared to the MnO₂ support, indicating some new stronger base sites formed on the catalyst surface. Notably, among the supported catalysts, the Ba/MnO₂ catalyst exhibited the largest desorption peak area and the highest CO₂ desorption peak center temperature (around 250 °C), indicating that the Ba/MnO₂ catalyst possessed the highest amount of surface base and relatively stronger surface basic sites. The order of the number of surface basic sites on the catalysts was as follows: Ba/MnO₂ > Sr/MnO₂ > Ca/MnO₂ > Mg/MnO₂ > MnO₂. Fig. 3(b) shows the desorption profiles of NH₃ from the studied catalysts. From Fig. 3(b), all catalysts exhibited a prominent NH₃ desorption peak in the range of 150–350 °C and the acid sites on these catalysts were arbitrarily considered weak and medium.^23,30 The loading of alkaline earth metals reduces the number of acid sites of various strengths on the catalyst surface. This could be due to an interaction between alkaline earth metal species and acid sites, or by the coverage of alkaline earth metal species. Notably, the trend in the change of surface acidity exhibits a negative correlation with the order of alkaline strength of alkaline earth metals, meaning the surface acid amount follows the sequence: Ba/MnO₂ < Sr/MnO₂ < Ca/MnO₂ < Mg/MnO₂ < MnO₂.

Fig. 3 CO₂-TPD (a) and NH₃-TPD (b) profiles of alkaline earth metal-loaded manganese dioxide catalysts at different temperatures.

To gain a deeper understanding of the chemical states on the catalyst surfaces, XPS analysis was performed on these five catalysts, and their spectra were deconvoluted through peak fitting. Fig. 4(a) shows the Mn 2p_3/2 spectra for MnO₂ and alkaline earth metal loaded MnO₂ catalysts. The peaks at 641.9 eV and 643.0 eV can be assigned to characteristic peaks of Mn³⁺ and Mn⁴⁺ species, respectively.^26,32–34 The calculation of the fitted peak areas revealed the surface concentrations and ratios of Mn³⁺ and Mn⁴⁺ (Table 1). The Mn³⁺/Mn⁴⁺ ratio progressively increased with the enhanced alkalinity of the alkaline earth metals (as follows): MnO₂ (1.04) < Mg/MnO₂ (1.17) < Ca/MnO₂ (1.33) < Sr/MnO₂ (1.38) < Ba/MnO₂ (1.63). According to established principles, surface Mn³⁺ species and their abundance, along with unstable lattice oxygen, play a crucial role in promoting the cross-coupling reaction of alcohols and amines.^{19,20,35–37} Oxygen molecules from air can adsorb onto oxygen vacancies on the catalyst surface and subsequently be reduced by two electrons from reduced Mn⁺ species. To further explore the oxygen vacancies and their quantity changes in these catalysts, as well as their relationship with the aforementioned Mn³⁺, O 1s spectra were fitted and analyzed, as shown in Fig. 4(b). In the O 1s spectrum of MnO₂, three fitted peaks were identified at binding energies of approximately 529.5 eV, 531.6 eV, and 533.3 eV, attributed to Mn–O lattice oxygen (denoted as O_α), defective oxygen species (denoted as O_β, also referred to as hydroxyl-like surface adsorbed oxygen), and adsorbed molecular water (O_γ), respectively.^32–34,38 The surface O_β (such as O atoms at bridge sites) tends to exhibit higher activities than O_α. Notably, Ba/MnO₂ exhibits the highest proportion of defective oxygen species (0.60 for O_β/O_α), followed by Sr/MnO₂ (0.58), Ca/MnO₂ (0.46), Mg/MnO₂ (0.41) and MnO₂ (0.23). It is not difficult to find that as the alkalinity of alkali earth metals increases, the concentration of defective oxygen species on the catalyst also increases correspondingly. The above comparison indicates that the prepared MnO₂ itself possesses relatively high concentrations of both Mn³⁺ species and defective oxygen species, but these concentrations were further enhanced upon loading the alkaline earth metals. To further verify this result, the Mn 3s spectra for these five catalysts were deconvoluted through peak fitting, as shown Fig. 4(c). It can be observed that the ΔE_3s (the binding energy difference between the double peaks in the Mn 3s spectrum) values for MnO₂, Mg/MnO₂, Ca/MnO₂, Sr/MnO₂, and Ba/MnO₂ were 4.76 eV, 4.80 eV, 4.84 eV, 4.89 eV, and 4.90 eV, respectively. Although the changes were relatively small, they exhibited a gradual increase. Additionally, the average oxidation state (AOS) of Mn in these five catalysts was further analyzed (Table 1) using the formula AOS = 9.956–1.126ΔE_3s.^18,26 As shown in Table 1, the AOS values followed a decreasing trend in the order: MnO₂ (3.60) > Mg/MnO₂ (3.55) > Ca/MnO₂ (3.51) > Sr/MnO₂ (3.45) > Ba/MnO₂ (3.37). This trend aligns with the behavior of Mn³⁺, as a higher Mn³⁺ content leads to a lower AOS.^26,32,33 In conclusion, with the increase in the basicity of the alkaline earth metals supported on the catalyst surface, both the concentration of Mn³⁺ and the number of oxygen vacancies on the catalyst increased.

Fig. 4 XPS spectra of (a) Mn 2p, (b) O 1s, and (c) Mn 3s of alkaline earth metal loaded MnO₂ catalysts.

Table 1 XPS results of the catalysts

Samples	Mn 2p3/2			Mn 3s	O 1s
	Mn^3+	Mn^4+	Mn^3+/Mn^4+	AOS	Oα	Oβ	Oγ	Oβ/Oα
MnO2	0.51	0.49	1.04	3.60	0.71	0.16	0.12	0.23
Mg/MnO2	0.54	0.46	1.17	3.55	0.68	0.28	0.04	0.41
Ca/MnO2	0.57	0.43	1.33	3.51	0.61	0.28	0.11	0.46
Sr/MnO2	0.58	0.42	1.38	3.45	0.57	0.33	0.1	0.58
Ba/MnO2	0.62	0.38	1.63	3.37	0.53	0.32	0.15	0.60

---

To further investigate the reducibility of the studied catalysts, H₂-TPR (temperature-programmed reduction with H₂) analysis was also employed. Fig. 5(a) shows the H₂-TPR profiles for MnO₂ and the alkaline earth metal loaded MnO₂ catalysts. On all profiles, two H₂ consumption peaks can be observed. The H₂ consumption peak at lower temperature is attributed to the consumption of surface adsorbed oxygen and the reduction processes: Mn⁴⁺ → Mn³⁺ and Mn³⁺ → Mn²⁺. The peak at higher temperature can be attributed to the reduction process Mn³⁺ → Mn²⁺.^18,26,37,39 Based on calculations of the temperature and intensity of the first H₂ consumption peak, the surface oxygen storage capacity (OSC) and the ease of reduction of the catalyst samples decreased in the order: Ba/MnO₂ > Sr/MnO₂ > Ca/MnO₂ > Mg/MnO₂ > MnO₂. This result is consistent with the XPS characterization results.

Fig. 5 H₂-TPR (a) and O₂-TPD (b) profiles of MnO₂ and the alkaline earth metal loaded MnO₂ catalysts.

The surface oxygen species of different catalysts were analyzed via O₂-TPD experiments, and the results are presented in Fig. 5(b). The desorption peaks in the medium-low temperature region can be assigned to surface-adsorbed oxygen species.^40,41 Compared with unmodified MnO₂, the surface-adsorbed oxygen content of X/MnO₂ (X = Mg, Ca, Sr, Ba) all increased, indicating that the loading of alkaline earth metals is conducive to enhancing the mobility of surface active oxygen species. In particular, Ba/MnO₂ has the highest content of surface-adsorbed oxygen species and exhibits the optimal oxygen species mobility, which is consistent with the results of XPS and H₂-TPR analyses.

3.2. Catalyst performance

A series of prepared supported manganese oxide catalysts were applied to the oxidative coupling reaction of benzyl alcohol and aniline under low-temperature conditions. Fig. 6 and Table S4 show the specific activity evaluation results and turnover frequencies of the catalysts: specifically, the variation of imine yield with reaction time (Fig. 6) and the moles of imine formed per unit catalyst amount per hour (Table S4). Regarding the support, manganese oxide itself exhibited some activity for the target reaction. After 4 hours of reaction, the imine yield was 50.2% (TOF = 0.42 mmol g⁻¹ h⁻¹), and after 10 hours, it reached 93.1%. For the supported catalysts, the loading of alkaline earth metals significantly enhanced the catalytic activity. At the same reaction time of 4 hours, the imine yields and turnover frequencies (TOF) reached 55.3%, 0.46 mmol g⁻¹ h⁻¹ (Mg/MnO₂), 60.5%, 0.50 mmol g⁻¹ h⁻¹ (Ca/MnO₂), 86.2%, 0.72 mmol g⁻¹ h⁻¹ (Sr/MnO₂), and 92.2%, 0.77 mmol g⁻¹ h⁻¹ (Ba/MnO₂). Clearly, the catalytic activity progressively increased with the strengthening alkalinity of the alkaline earth metal. Among them, the optimal catalyst was Ba/MnO₂. Under general reaction conditions (reaction temperature 60 °C, ambient pressure, benzyl alcohol to aniline molar ratio of 1 : 2, catalyst dosage 0.3 g), an imine yield of 98.9% was achieved after just 6 hours of reaction. This result is already at a considerably high level in this research field. Combining the catalyst activity evaluation results with the previous characterization analysis, we can readily see that the loading of alkaline earth metals improved the oxidative coupling activity of the catalysts. And, this enhancement in catalytic activity is likely closely related to the increased proportion of low-valent manganese species and the higher number of oxygen vacancies in the materials. However, the loading of alkaline earth metals also simultaneously altered the surface acid–base environment of the catalysts. Specifically, as the alkalinity of the loaded alkaline earth metal increased, the number of acidic sites on the catalyst surface decreased, while the number of basic sites increased. So, how exactly does the acidity/basicity affect the catalytic activity in the target reaction?

Fig. 6 Catalytic performance of manganese oxide modified with alkaline earth metals at different reaction times.

To further explore the influence of acidic and basic sites on the catalytic activity of alkaline earth metal-supported manganese-based catalysts, acid–base poisoning experiments were conducted using pyridine (basic molecule) and pyrrole (acidic molecule). Fig. 7 shows a comparison chart of the catalytic activity of Ba/MnO₂ before and after acid–base poisoning with pyridine and pyrrole. Compared with the untreated catalyst, the activity of the catalyst poisoned by both pyridine and pyrrole significantly decreased. The activity of the catalyst treated with pyrrole was significantly lower than that treated with pyridine, indicating that the basic sites on the catalyst played a crucial role in the alcohol–amine oxidative coupling reaction. This activity result is consistent with the characterization results of CO₂-TPD and NH₂-TPD. It can be seen that in addition to the number of oxygen vacancies, the alkaline sites on the surface of the catalyst are also important factors affecting the catalytic activity of the catalyst, which may further promote the occurrence of alcohol–amine oxidation coupling reaction.

Fig. 7 Catalytic performance of Ba/MnO₂ before and after treatment of pyridine and pyrrole poisoning.

The stability and reusability of a heterogeneous catalyst are important for potential practical applications. As for Ba/MnO₂, a hot filtration test was carried out to confirm the heterogeneous nature of the sample. The results showed that Ba/MnO₂ behaved as an operationally heterogeneous catalyst, since the imine concentration in the filtrate did not increase (Fig. S2). Almost no Ba and Mn species in the filtration could be detected by ICP measurement (Table S3). Furthermore, multiple reaction cycles were performed to investigate the recoverability of this sample. After five cycles, the catalyst in the reaction at 60 °C still retained a high activity (Fig. 8). It could be concluded that the Ba/MnO₂ catalyst really has good stability and reusability.

Fig. 8 Reusability of the Ba/MnO₂ catalyst for the reaction at 60 °C. Reaction conditions: benzyl alcohol (1 mmol), aniline (2 mmol), catalyst (0.3 g), toluene (10 mL), air (0.1 MPa).

Additionally, a universality investigation experiment was conducted on the Ba/MnO₂ catalyst. As shown in the Table 2, the catalyst efficiently yielded the corresponding imines in high yields from the reactions of benzyl alcohol derivatives (with electron-withdrawing and electron-donating groups) and aniline. Among them, p-methoxybenzyl alcohol exhibited the highest activity, with an imine yield of up to 99.4% after 10 hours of reaction. The imine yields for p-hydroxybenzyl alcohol, o-methylbenzyl alcohol, p-isopropylbenzyl alcohol, p-chlorobenzyl alcohol, and p-nitrobenzyl alcohol were 93.5%, 94.6%, 95.3%, 97.7%, and 98.2%, respectively. The imine yields decreased to a certain extent due to steric hindrance. The results of the universality experiment demonstrate that the Ba/MnO₂ catalyst has excellent applicability.

Table 2 Aerobic oxidative coupling of various alcohols and aniline over Ba/MnO₂ catalysts

Entry	Alcohol	Yield (%)
1		93.5%
2		94.6%
3		95.3%
4		97.7%
5		99.4%
6		98.2%

---

3.3. RSM analysis and model fitting

Combined with the above experimental results, to explore the effect of various reaction conditions on the yield of imine synthesized by oxidative coupling of alcohol–amine, the Box–Behnken experimental design (BBD) of the response surface methodology (RSM) was used to optimize the reaction conditions of the Ba/MnO₂ catalyst. The experimental design had three factors and three levels, with 17 experimental groups. The experimental points and results are shown in Table 3. Under different reaction conditions, the yield of imine increased from 32.69% to 97.44%.

Table 3 Box–Behnken design and response values

Std	Run	Factor 1	Factor 2	Factor 3	Response 1
		A: reaction time	B: amine/alcohol	C: toluene	Yield
		h		mL	%
1	14	2	1	10	42.92
2	9	6	1	10	97.44
3	4	2	3	10	32.69
4	1	6	3	10	89.01
5	2	2	2	5	46.48
6	6	6	2	5	96.8
7	5	2	2	15	38.01
8	13	6	2	15	97.06
9	17	4	1	5	70.51
10	3	4	3	5	66.08
11	15	4	1	15	64.49
12	8	4	3	15	55.76
13	16	4	2	10	92.21
14	11	4	2	10	92.21
15	7	4	2	10	92.21
16	12	4	2	10	92.21
17	10	4	2	10	92.21

---

The validity and accuracy of the quadratic model, as well as the effects of single factors and factor interactions on the response value, were verified through standard analysis of variance (ANOVA). The P-value was used to test the significance of variables, where a smaller P-value indicates a more significant variable; the F-value was estimated from the sum of squares, representing the ratio of the mean square to the mean square error effect. As shown in Table 4 for the ANOVA of the quadratic model, the F-value and P-value were 549.23 and <0.0001, respectively, indicating that the quadratic model equation can be used to study the relationship between reaction variables and response values with highly reliable results. The value “P < 0.0001” is far less than 0.05, demonstrating the significance of the model terms.

Table 4 The analysis of variance (ANOVA) for the response surface quadratic model

Response: yield source	Sum of squares	df	Mean square	F-Value	p-Value
Model	8695.20	9	966.13	549.23	<0.0001	Significant
A – reaction time	6061.56	1	6061.56	3445.89	<0.0001
B – amine/alcohol	126.56	1	126.56	71.95	<0.0001
C – toluene	75.34	1	75.34	42.83	0.0003
AB	0.8100	1	0.8100	0.4605	0.5192
AC	19.05	1	19.05	10.83	0.0133
BC	4.62	1	4.62	2.63	0.1490
A ^2	478.35	1	478.35	271.94	<0.0001
B ^2	1082.78	1	1082.78	615.54	<0.0001
C ^2	602.66	1	602.66	342.60	<0.0001
Residual	12.31	7	1.76
Lack of fit	12.31	3	4.10
Pure error	0.0000	4	0.0000
Cor total	8707.51	16

---

Table 5 shows the fitting results for imine yield, with an R² of 0.9986 indicating the accuracy of the polynomial model. The adjusted R² and predicted R² were 0.9968 and 0.9774, respectively, with an absolute difference less than 0.02, showing a high degree of consistency. The adeq precision was 63.35, far greater than 4, indicating that the model has a sufficient signal-to-noise ratio for measurement. Additionally, the lack-of-fit term had an F-value of 4.34, suggesting good fitting of the model to all data. Taken together, these results fully demonstrate the high accuracy of the model, which can be used to analyze and predict imine yields. Therefore, the equation form of the quadratic model obtained by fitting is as follows:

Coded variable form:

Y = 92.21 + 27.53 × A − 3.98 × B − 3.07 × C + 0.4500 × AB + 2.18 × AC − 1.07 × BC − 10.66 × A² − 16.04 × B² − 11.96 × C²

Actual variable form:

Y = −97.15500 + 32.44812 × A + 61.41750 × B + 8.51425 × C + 0.225000 × AB + 0.218250 × AC − 0.215000 × BC − 2.66469 × A² − 16.03625 × B² − 0.478550 × C²

where Y represents the imine yield; A is the reaction time; B is the aniline/benzyl alcohol ratio (amine/alcohol); C is the toluene dosage.

Table 5 The fitting results of the imine yield

Std. dev.	1.33
Mean	74.02
C.V.%	1.79
R ^2	0.9986
Adjusted R^2	0.9968
Predicted R^2	0.9774
Adeq precision	63.3476

---

In the equations, the positive sign (+) indicates a synergistic effect between specified variables, while the negative sign (−) indicates an antagonistic effect. Additionally, the equations show that the imine yield exhibits both linear and quadratic effects. The coefficient of reaction time (A) is 27.53, having the greatest impact on the response, followed by the amine–alcohol ratio and toluene dosage. The coefficient of parameter AC is 2.18, which is greater than the coefficients of AB (0.45) and BC (−1.07), indicating that the interaction of AC is the strongest, followed by AB and BC. The BC interaction shows an antagonistic effect, which is attributed to the cumulative effect of reaction parameters. The response actual values and simulated values calculated through the above two equations are shown in Table 6. The need to improve the model is determined by checking the ratio of the maximum response value to the minimum response value. If the ratio is greater than 10, model improvement is required. In this model, the ratio is 2.98, which is less than 10, indicating that the model does not need to be improved. Combined with the previous discussion, it can be used to analyze the interaction between imine yield and various reaction conditions. Furthermore, Table 6 shows that the simulated values and actual values are in good agreement, further demonstrating the high reliability of the model.

Table 6 The predicted values and actual values obtained from model fitting

Group	Predicted	Actual	Group	Predicted	Actual
1	42.42	42.92	10	64.38	66.08
2	96.57	97.44	11	66.19	64.49
3	33.56	32.69	12	56.09	55.76
4	89.51	89.01	13	92.21	92.21
5	47.31	46.48	14	92.21	92.21
6	98.00	96.8	15	92.21	92.21
7	36.81	38.01	16	92.21	92.21
8	96.23	97.06	17	92.21	92.21
9	70.18	70.51

---

Fig. 9(a) shows the normal distribution plot of residuals for imine yield in this model. The data points are uniformly distributed along the diagonal, indicating that the residuals approximately follow a normal distribution. This suggests that the selected experimental model can be used to predict the experimental process. Fig. 9(b) depicts the relationship between the actual and predicted values of imine yield. The data in the figure exhibit a near-linear distribution, with the predicted values closely approaching the experimental values. This demonstrates that the quadratic polynomial model is suitable for describing the correlation between experimental variables and imine yield, which is consistent with the results in Table 6 and further validates the effectiveness of the model.

Fig. 9 Residual normal distribution diagram of imine yield (a); relationship between experimental values and predicted values of imine yield (b).

Based on the prediction model, the interaction between variables has a significant impact on imine yield. Therefore, by fixing one variable, a sub-model of the interaction between the remaining two variables on the response value can be obtained, and the data are plotted as response surfaces and contour plots. As shown in Fig. 10, (a) and (b) are the 3D surface and contour plots of the interaction between reaction time and amine/alcohol ratio with imine yield; (c) and (d) are those between reaction time and toluene dosage; (e) and (f) are those between amine/alcohol ratio and toluene dosage. The shape of the contour lines reflects the magnitude of the interaction to a certain extent: an elliptical shape indicates a significant interaction between two variables, while a circular shape indicates the opposite.^24,25Fig. 10(b, d and f) shows elliptical contours, indicating non-negligible interactions between variables. Additionally, combined with the P-values in the analysis of variance (ANOVA) Table 4, the P-values of AB, AC, and BC terms are 0.52, 0.01, and 0.15, respectively, further confirming the existence of interactions between experimental variables. The interaction between reaction time and toluene dosage (AC) is the most significant, as the regression coefficient of the AC interaction term (2.18) is greater than those of AB (0.45) and BC (−1.07). From Fig. 10(a and b), it can be seen that imine yield increases with the extension of reaction time. With the increase of amine/alcohol ratio, the yield first increases and then decreases. Simultaneously increasing reaction time and amine/alcohol ratio promotes yield, but the steepness of the response surface shows that the growth rate of yield first accelerates and then slows down. The contour lines indicate that higher imine yields occur when the reaction time is 5–6 h and the amine/alcohol ratio is 1.5–2.5, confirming that this experimental condition range is more suitable for oxidative coupling of alcohol–amine to synthesize imine. Fig. 10(c and d) shows a similar trend: as toluene dosage increases, imine yield first increases and then decreases. Increasing reaction time and toluene dosage simultaneously enhances yield, but the response surface is relatively gentle, indicating a slow growth rate. The contour lines show that higher yields are achieved at a reaction time of approximately 5 h and toluene dosage of 7–13 mL, suggesting that this range is optimal for the synthesis. Fig. 10(e and f) demonstrates that increasing both amine/alcohol ratio and toluene dosage leads to a trend of yield first increasing and then decreasing. Although a simultaneous increase slightly improves yield, the overall response surface is flat, indicating that the interaction between toluene dosage and amine/alcohol ratio has the least significant impact on imine yield.

Fig. 10 (a) and (b) The 3D surface and contour plots of reaction time, amine–alcohol ratio and imine yield; (c) and (d) the 3D surface and contour plots of reaction time, toluene dosage and imine yield; (e) and (f) the 3D surface and contour plots of amine–alcohol ratio, toluene dosage and imine yield.

The experimental conditions were optimized using the numerical function of Design Expert software. Three groups of experimental conditions with a satisfaction degree of 1 were selected to predict the imine yield, and the prediction results were verified through experiments. The data are shown in Table 7. According to the results in Table 7, in the first group of experiments, when the reaction time was 6 h, the amine/alcohol ratio was 2, and the toluene dosage was 15 mL, the predicted maximum imine yield reached 96.23% with a satisfaction degree of 1. The actual yield obtained through the experiment was 97.06%. The error was calculated as the absolute value of the difference between the actual yield and the predicted yield divided by the actual yield, resulting in 0.86%. The absolute errors of the second and third groups were 0.89% and 1.24%, respectively. The absolute errors of all three groups of experiments were less than 5%, indicating that the model can effectively predict the imine yield.

Table 7 The predicted values and actual values after optimizing experimental conditions

No.	Reaction time (h)	Aniline/alcohol	Toluene (mL)	Predicted (%)	Desirability	Actual (%)	Error (%)
1	6	2	15	96.23	1.00	97.06	0.86
2	6	1	10	96.57	1.00	97.44	0.89
3	6	2	5	98.00	1.00	96.80	1.24

---

Based on the experimental results, a reaction mechanism hypothesis was proposed. Fig. S3 shows the experimental data of the related mechanism study, indicating that the presence or absence of a catalyst has a significant impact on the activity of the target reaction. Relevant studies have shown that the oxidative dehydrogenation of benzyl alcohol is the rate-determining step in the reaction.^42–44 On reducible oxide catalysts, this reaction proceeds through two consecutive steps: first, the oxidative dehydrogenation of benzyl alcohol to form benzaldehyde, and then the reaction of benzaldehyde with aniline to form an imine. In this case, when aniline is absent, the oxidative dehydrogenation of benzyl alcohol to form benzaldehyde hardly occurs on non-reducible Ba/MnO₂ (Fig. S3(b and c)). This implies that the oxidative coupling of benzyl alcohol and aniline on Ba/MnO₂ should follow a different reaction pathway. Molecular O₂ has been confirmed to be involved in this reaction, as almost no imine is observed in the absence of molecular O₂ (Fig. S3(a)). However, the activation of molecular oxygen and benzyl alcohol is a relatively complex process that requires the participation of aniline. Therefore, although the activation of benzyl alcohol is not directly achieved by forming benzaldehyde, it remains the rate-determining step of this oxidative coupling reaction. In this study, the weak and moderate basic sites on the catalyst surface can further promote the oxidative coupling reaction of alcohols and amines. This is likely because they act as active centers for the activation of benzyl alcohol, which is the rate-determining step of this oxidative coupling reaction (Scheme 1).²³

Scheme 1 The plausible reaction mechanism of oxidative coupling of benzyl alcohol and aniline using the Ba/MnO₂ catalyst.

4. Conclusions

A series of alkaline earth metal-modified supported manganese oxide catalysts were designed and prepared via a simple impregnation method to modulate the surface acid–base environment of manganese oxide. The correlation between the physicochemical properties of the catalysts (particularly surface acidity/basicity and Mn valence states) and their catalytic activity was systematically investigated. According to the experimental results and the characterization results, Ba/MnO₂ showed the best catalytic performance in the reaction, and exhibited the highest Mn³⁺/Mn⁴⁺ ratio, surface weak basic sites and a relatively small number of weak acidic sites. Lattice oxygen mobility and Mn³⁺/Mn⁴⁺ ratio were found to play important roles in the catalytic activity of aerobic reactions. The weak and medium basic sites on the catalyst surface can further promote the alcohol–amine oxidative coupling process. Additionally, this work optimized three key reaction parameters, reaction temperature, toluene dosage, and amine–alcohol ratio, using the response surface methodology (RSM), with a comprehensive analysis of the interactive effects among these factors. The results show that the reaction time has the greatest influence on the yield, followed by the amine–alcohol ratio, and finally the toluene dosage. The interaction between variables was also analyzed and compared, and it was concluded that the interaction between reaction time and toluene dosage is the most significant. This study on transition metal catalysts holds important theoretical significance for the process of catalytic oxidative coupling of alcohols and amines to synthesize imines.

Author contributions

Qiang Bao: methodology, investigation, data curation, writing – original draft, writing – review & editing, and supervision. Wenhui Feng: methodology, investigation, data analysis, writing – review & editing. Yunfeng Hu and Zhenlu Wang: conceptualization and supervision. Guoliang Wu and Zhirui Chen: investigation and data curation. Haocheng Li and Chenguang Shi: data curation.

Conflicts of interest

There are no conflicts to declare.

Data availability

All data included in this study are available upon request by contact with the corresponding author.

Acknowledgements

This work was supported by the Northeast Petroleum University Youth Science Fund (grant no. 15071202002) and the Northeast Petroleum University Outstanding Scientific Research Talents Supporting Program in Superior and Featured Subjects (grant no. 15041260349).

References

1. S. Wu, W. Sun, J. Chen, J. Zhao, Q. Cao, W. Fang and Q. Zhao, J. Catal., 2019, 377, 110–121.
2. M. Purohit, S. Kalla and R. Jangir, ChemistrySelect, 2023, 8, 202300386.
3. S. Murahashi, Angew. Chem., Int. Ed. Engl., 1995, 34, 2443–2465.
4. Y. Chen, H. Lim, Q. Tang, Y. Gao, T. Sun, Q. Yan and Y. Yang, Appl. Catal., A, 2010, 380, 55–65.
5. B. Chen, L. Wang and S. Gao, ACS Catal., 2015, 5, 5851–5876.
6. R. D. Patil and S. Adimurthy, Asian J. Org. Chem., 2013, 2, 726–744.
7. M. T. Schümperli, C. Hammond and I. Hermans, ACS Catal., 2012, 2, 1108–1117.
8. L. Jiang, L. Jin, H. Tian, X. Yuan, X. Yu and Q. Xu, Chem. Commun., 2011, 47, 10833.
9. M. Largeron, Eur. J. Org. Chem., 2013, 2013, 5225–5235.
10. S. Shi, H. Lan, C. Tao, Y. Zhang, H. Zhang, Q. Sun, X. Zhang, G. Guo, Q. He, X. He and H. Ji, Chem. Pap., 2023, 77, 3495–3500.
11. H. Schiff, Justus Liebigs Ann. Chem., 1864, 131, 118–119.
12. P. Anastas and N. Eghbali, Chem. Soc. Rev., 2010, 39, 310–312.
13. J. Płotka-Wasylka, H. M. Mohamed, A. Kurowska-Susdorf, R. Dewani, M. Y. Fares and V. Andruch, Curr. Opin. Green Sustainable Chem., 2021, 31, 100508.
14. A. Corma, P. Concepción, M. Boronat, M. J. Sabater, J. Navas, M. José Yacaman, E. Larios, A. Posadas, M. A. López-Quintela, D. Buceta, E. Mendoza, G. Guilera and A. Mayoral, Nat. Chem., 2013, 5, 775–781.
15. S. H. Joo, S. J. Choi, I. Oh, J. Kwak, Z. Liu, O. Terasaki and R. Ryoo, Nature, 2001, 412, 169–172.
16. C. Yuan, H. Wu, Y. Xie and X. W. Lou, Angew. Chem., Int. Ed., 2014, 53, 1488–1504.
17. H. M. T. Galvis, J. H. Bitter, C. B. Khare, M. Ruitenbeek, A. Iulian Dugulan and K. P. de Jong, Science, 2012, 335, 835–838.
18. J. Bu, W. Wang, H. Zhang, C. Wu, B. Zhang, Y. Cao, F. Zhou, Q. Zhang and X. Zhang, Mol. Catal., 2018, 459, 46–54.
19. C. Wu, J. Bu, W. Wang, H. Shen, Y. Cao and H. Zhang, Ind. Eng. Chem. Res., 2022, 61, 5442–5452.
20. T. Chutimasakul, W. Tirdtrakool, P. N. Nakhonpanom, H. Kreethatorn, P. Jaruwatee, T. Bunchuay and J. Tantirungrotechai, ChemistrySelect, 2022, 7, 202203028.
21. J. Shi, T. Qi, B. Sun, G. Chu and J. Chen, Chem. Eng. J., 2022, 440, 135802.
22. F. Chen, S. Zhao, T. Yang, T. Jiang, J. Ni, H. Xiong, Q. Zhang and X. Li, Chin. J. Chem. Eng., 2019, 27, 2438–2446.
23. J. Song, G. Yu, X. Li, X. Yang, W. Zhang, W. Yan and G. Liu, Chin. J. Catal., 2018, 39, 309–318.
24. G. Zhang, H. Zhang, D. Yang, C. Li, Z. Peng and S. Zhang, Catal. Sci. Technol., 2016, 6, 6417–6430.
25. D. R. Pinheiro, R. F. Neves and S. P. A. Paz, Microporous Mesoporous Mater., 2021, 323, 111250.
26. Z. Wang, H. Jia, T. Zheng, Y. Dai, C. Zhang, X. Guo, T. Wang and L. Zhu, Appl. Catal., B, 2020, 272, 119030.
27. M. M. Kadam, K. B. Dhopte, N. Jha, V. G. Gaikar and P. R. Nemade, New J. Chem., 2016, 40, 1436–1442.
28. L. Wu, Z. Li, Y. Xiang, W. Dong, H. Wu, Y. Xu, Z. Ling, M. S. Chae, D. Sharon, N. Shpigel and X. Zhang, ACS Energy Lett., 2024, 9, 5801–5809.
29. J. Waikar, Bull. Mater. Sci., 2021, 44, 268.
30. A. Wang, H. Wang, M. Ji and H. Yin, New J. Chem., 2024, 48, 4995–5007.
31. C. Ni, J. Hou, L. Li, Y. Li, M. Wang, H. Yin and W. Tan, Chemosphere, 2020, 250, 126211.
32. P. Mahamallik, S. Saha and A. Pal, Chem. Eng. J., 2015, 276, 155–165.
33. J. Wang, J. Li, C. Jiang, P. Zhou, P. Zhang and J. Yu, Appl. Catal., B, 2017, 204, 147–155.
34. T. Subramaniam, M. Balarabe Idris, G. Harshini Sai and S. Devaraj, Mater. Chem. Phys., 2023, 303, 127845.
35. Y. Wang, T. Hu, Q. Liu and L. Zhang, Chem. Commun., 2018, 54, 4005–4008.
36. Y. Wang, H. Wang, P. Meng, D. Song, Z. Qi and X. Zhang, Chin. J. Chem., 2021, 39, 2983–2989.
37. J. Qin, Y. Long, G. Gou, W. Wu, Y. Luo, X. Cao, S. Luo, K. Wang and J. Ma, Catal. Sci. Technol., 2020, 10, 5628–5640.
38. R. Rajesh, P. Balla, P. Rani Mishra, R. Gundeeboyina, S. Kim, V. Perupogu, N. Arumugam, A. I. Almansour, U. Lassi and P. K. Seelam, Appl. Surf. Sci., 2025, 680, 161404.
39. Y. Dong, M. Ran, X. Zhang, S. Lin, W. Li, Y. Yang, H. Song, W. Wu, S. Liu, C. Zheng and X. Gao, J. Environ. Chem. Eng., 2024, 12, 112823.
40. S. Wu, H. Zhang, Q. Cao, Q. Zhao and W. Fang, Catal. Sci. Technol., 2021, 11, 810–822.
41. R. Rajesh, P. Balla, P. R. Mishra, R. Gundeeboyina, S. Kim, V. Perupogu, N. Arumugam, A. I. Almansour, U. Lassi and P. K. Seelam, Appl. Surf. Sci., 2025, 680, 161404.
42. W. Cui, J. Wang, Sagala and M. Jia, Russ. J. Phys. Chem. A, 2021, 95, 958–967.
43. Y. Long, H. Zhang, Z. Gao, J. Qin, Y. Pan, J. Zhao, Y. Luo, Z. Ma, Y. Xiong and J. Ma, Inorg. Chem. Front., 2019, 6, 829–836.
44. Q. Kang and Y. Zhang, Green Chem., 2012, 14, 1016–1019.

---