Ultrasound-assisted synthesis and stereoselectivity analysis of para-menthane-3,8-diol from citronellal via the Prins reaction

Graphical abstract

Abstract

para-Menthane-3,8-diol (PMD) is a widely used bio-based insect repellent. However, its conventional synthesis is limited by protracted reaction times. Herein, we report an ultrasound-assisted Prins cyclization of citronellal, a monoterpenoid aldehyde, to PMD under 20 kHz irradiation. Using pure citronellal, ultrasound irradiation at 40% amplitude (∼48 µm) afforded PMD with 98% GC-FID purity within 2 h, compared with 87% PMD obtained after 10 h under conventional heating. Even at 100% amplitude (∼120 µm), the reaction proceeded within 10 min while maintaining a high PMD content of approximately 93%. For lemon eucalyptus (Eucalyptus citriodora) oil, ultrasound irradiation at 40–60% amplitude (∼48–72 µm) reduced reaction times from 15 h to 0.5–2.0 h while providing PMD contents comparable to those obtained by conventional heating. Notably, crystallization from the natural oil feedstock increased the cis-isomer content to 90%, compared with 65% obtained from pure citronellal. The observed cis stereoselectivity is proposed to arise from steric effects associated with the 3-methyl substituent of citronellal during intramolecular cyclization. Density functional theory calculations qualitatively support the proposed stereoselective mechanism. These findings provide an efficient approach for the synthesis and purification of PMD while offering mechanistic insights into stereochemical control in Prins cyclization reactions.

1. Introduction

para-Menthane-3,8-diol is a naturally derived monoterpenoid known for its potent repellent activity against mosquitoes, ticks, and other biting insects [1]. It is registered with the U.S. Environmental Protection Agency (EPA) as an active repellent ingredient, and the U.S. Centers for Disease Control and Prevention (CDC) lists PMD among the most efficacious agents for mosquito control [2]. PMD exists as eight stereoisomers ca, ca′, cb, cb′, ta, ta′, tb, and tb′ (Fig. 1; t: 3,4-trans-isomer, referred to as the trans-isomer; c: 3,4-cis-isomer, referred to as the cis-isomer; a: major product; b: minor product; ′: enantiomer), all of which exhibit repellent activity. Notably, the cis-PMD isomers demonstrate superior efficacy compared to their trans-PMD counterparts [3], [4], [5]. This performance is associated with differences among PMD isomers in mosquito olfactory receptor binding affinity and evaporation behavior [5].

Fig. 1.

The structure of the 8 PMD stereoisomers. (t: 3,4-trans-isomer, c: 3,4-cis-isomer, a: major product, b: minor product, ′: enantiomer).

The Prins reaction is a powerful method for constructing oxygen-containing heterocycles and carbocycles through the acid-catalyzed addition of alkenes to carbonyl compounds, forming new C—C and C—O bonds. Various catalytic systems have been developed to promote Prins cyclization, including rhenium oxides for spirocyclization, chiral transition-metal complexes for asymmetric synthesis, and supramolecular catalysts that suppress side reactions [6]. The Prins cyclization of citronellal, a monoterpenoid aldehyde, has been acknowledged for a long time as an effective route to PMD (Scheme 1). This reaction was carried out in aqueous medium, where water acts as the nucleophile attacking the cyclic carbocation intermediate generated from the acid-catalyzed reaction between the aldehyde group and the carbon–carbon double bond, leading to the formation of para-menthane-3,8-diol (Scheme 1). Since both citronellal and PMD are monoterpenoids with low water solubility [7], the reaction proceeds as a biphasic system. Yuasa et al. reported that treatment of (+)-citronellal with 0.25% H₂SO₄ at 55 °C for 11 h afforded 99% conversion and 92.3% PMD selectivity (ca/cb/ta/tb = 61.2/1.2/35.0/2.6), yet the process suffered from extended reaction times and generated acetal impurities (∼2.1%) [8]. Efforts towards greener processes, such as replacing H₂SO₄ with CO₂/H₂O mixtures as the acid medium, have not yet outperformed mineral acid catalysis [9].

Scheme 1.

The Prins cyclization mechanism of citronellal.

The leaf essential oil of lemon eucalyptus (Eucalyptus citriodora) is characterized by a high content of citronellal (62–79% [10], [11]), providing a viable precursor for the synthesis of botanical insect repellents. Through acid-catalyzed Prins cyclization, the citronellal constituent of the essential oil is converted into PMD, yielding a PMD-enriched essential oil. This modified botanical product can be integrated directly into topical formulations, such as sprays and lotions, bypassing the need for the isolation of pure PMD [12], [13]. Drapeau et al. previously demonstrated the feasibility of this approach by utilizing lemon eucalyptus oil (comprising 74% citronellal) to produce an enriched oil containing 64% PMD, with citric acid serving as the catalyst. However, this method requires a prolonged 15-hour reaction time [12].

Advancements have been made in the synthesis of PMD, particularly through the development of diverse catalytic systems. These include formic acid [14], acetic acid [14], zeolite ZSM-5 [9], alkali lignin-derived carbon acid [15], Mo(CO)₅(OTf)₂ [16], or biosourced ammonium salts [17]. These studies have largely relied on conventional stirring techniques. Meanwhile, the present work explores an alternative approach to enhance reaction efficiency and reduce reaction times.

With operational frequencies ranging from 20 kHz to 1 MHz, sonochemistry has emerged as a useful tool for promoting a variety of organic transformations, particularly those involving heterogeneous systems or organometallic reagents [18], [19]. Sonochemistry has found widespread applications in industrial, environmental, and biomedical fields. In organic synthesis, ultrasound has proven particularly valuable for the preparation of natural products and medicinally relevant heterocycles, offering a more efficient and sustainable alternative to conventional methods [20]. The effects of ultrasound are primarily attributed to acoustic cavitation, in which alternating cycles of compression and rarefaction generate, grow, and collapse microbubbles within a liquid medium [19], [21]. The rapid collapse of these bubbles generates localized “hotspots” with extremely high temperatures and pressures on very short timescales, creating conditions that may facilitate chemical transformations [21], [22]. Despite these transient extreme conditions, reactions can proceed under relatively mild bulk environments. In addition, cavitation induces microstreaming and shock waves that enhance mass transfer and mixing, thereby increasing the likelihood of productive interactions between reactants and facilitating rapid reaction rates even in viscous or biphasic media at ambient temperatures [19], [21]. Beyond increased speed, ultrasound has been shown in some cases to influence reaction pathways through so-called “sonochemical switching”, enabling access to radical or single-electron transfer (SET) processes that may differ from those observed under conventional thermal conditions [18]. Collectively, these effects make ultrasound a promising green approach for accelerating organic transformations [20], [21]. More recently, the integration of ultrasound with continuous-flow microreactors has been investigated as a means of improving cavitation control, process reproducibility, and scalability [18], [19].

While ultrasound has been applied to the improvement of numerous synthetic transformations, its application to the Prins reaction has been scarcely reported, in particular regarding the stereoselection bias [20], [21], [22]. Mahía et al. reported the asymmetric synthesis of 7-(alkyl or aryl)-6-oxa-2-azabicyclo[3.2.1]octane derivatives via an aza-Prins cyclization strategy with ultrasound-assisted processing. However, the ultrasound irradiation was utilized only in the subsequent detosylation step and not in the Prins cyclization step itself [23]. To date, the use of ultrasound-assisted Prins cyclization for the synthesis of para-menthane-3,8-diol has not been reported.

In a study on the cyclization mechanism of citronellal to form para-menthane-3,8-diol, Zimmerman et al. indicated that the observed (ca + ca′)/(ta + ta′) selectivity arises from the preferential formation of the more stable corresponding intermediates [14]. In complementary kinetic studies of the Prins cyclization of citronellal, Price et al. identified the generation of the cyclized carbocation intermediate as the rate-limiting step (Scheme 1) [24]. Building on these foundations, Clark et al. proposed a mechanistic framework in which PMD formation proceeds via two distinct pathways, leading separately to 3,4-cis-PMD and 3,4-trans-PMD [25]. Despite these advances, the stereochemical basis underlying the distribution of the four pairs of optical isomers (ca + ca′), (ta + ta′), (tb + tb′), and (cb + cb′) remains incompletely resolved in the literature.

Herein, we report the use of low-frequency (20 kHz) ultrasound irradiation to accelerate the Prins cyclization in the synthesis of PMD from both pure citronellal and Eucalyptus citriodora essential oil (Fig. 2). Furthermore, the mechanism underlying the preferential formation of cis-PMD was investigated with the support of density functional theory (DFT) calculations.

Fig. 2.

Synthesis and purification of PMD in our study.

2. Materials and methods

2.1. Reagents and instrumentation

Reagents: (±)-Citronellal (Sigma-Aldrich, denoted as A), Eucalyptus citriodora essential oil provided by Tram Thien Nhien Co., Ltd. (denoted as B), and Eucalyptus citriodora essential oil obtained from Notessen Vietnam Co., Ltd (denoted as C) were analyzed by GC prior to use. Material A was determined to be 96% pure citronellal by GC–FID. Material B contained 53% citronellal by GC–FID (55% by GC–MS), and material C contained 74% citronellal by GC–MS. Other chemicals, including sulfuric acid, sodium bicarbonate, sodium sulfate, and solvents (Xilong, China), were used without further purification.

Instrumentation: Ultrasound-assisted reactions were performed using a Q500 sonicator (Qsonica, USA) equipped with a ½-inch titanium probe operating at 20 kHz and 100% amplitude (corresponding to a 120 µm amplitude). Comparative reactions under conventional heated magnetic stirring employed a C-MAG HS 10 digital stirrer/hotplate (IKA, China). Analytical thin-layer chromatography (TLC) was performed on pre-coated silica gel 60F-254 plates and visualized by staining with vanillin/sulfuric acid solution (0.5% vanillin, 5% H₂SO₄ in ethanol). Product composition was determined by a SCION 456-GC gas chromatograph equipped with a flame ionization detector (FID) or mass spectrometric detector (MS).

2.2. Synthesis and purification of PMD

2.2.1. General procedures

The heated magnetic stirring method was adapted from procedures reported by Yuasa et al. for pure citronellal [8] and by Drapeau et al. for Eucalyptus citriodora essential oil [12]. Reactions were conducted under the conditions specified in Exps. A0, B0, and C0 (Table 1), and their progress was monitored by thin-layer chromatography (TLC).

Table 1.

Investigated conditions for PMD synthesis from citronellal/E. citriodora essential oil.

Method	Exp.	Citro (g)	EO (g)	m_acid (g)	Temp (°C)	Amp (%)	Pulse	Time (hours)	EI (kJ. g−1)
Heated magnetic stirring method	A0*	1	−	40	50–60	−	−	10.00
	B0**	−	1	22	50–60	−	−	15.00
	C0**	−	1	40	50–60	−	−	15.00
Ultrasound method	A1	1	−	40	−	20	6	1.00	41.3
	A2	1	−	40	−	40	6	2.00	212.4
	A3	1	−	40	−	20	2	2.00	61.3
	A4	1	−	40	−	40	2	1.00	90.1
	A5	1	−	40	−	60	6	0.50	99.0
	A6	1	−	20	−	40	2	1.00	35.8
	A7	1	−	40	−	70	6	0.25	61.0
	A8	1	−	60	−	100	6	0.17	65.8
	A9	1	−	40	−	40	6	0.50	53.1
	B1	−	1	40		60	6	0.50	98.0
	B2	−	1	40		40	6	2.00	219.0
	C1	−	1	40		60	6	0.50	93.9
	C2	−	1	40		60	6	0.50	93.9
	C3	−	1	40		40	6	0.50	53.4
	C4	−	1	40		40	6	2.00	213.6

*Performed based on conditions published by Yuasa et al. [8].

**Performed based on conditions published by Drapeau et al. [12].

(A: pure citronellal used; B: 55% citronellal oil used; C: 74% citronellal oil used; Citro: mass of pure citronellal; EO: mass of E. citriodora essential oil; m_acid: mass of 0.25% (weight/weight) sulfuric acid solution (^**: 7% (weight/weight) citric acid solution); temp: reaction temperature; amp: ultrasound amplitude; pulse: ultrasound pulse (30 s on:5 s off = 6; 30 s on:15 s off = 2); EI: energy input, kJ. g⁻¹ citronellal or essential oil − see Section S1 in the Supplementary Data).

The ultrasound method: 0.25% (w/w) aqueous H₂SO₄ was measured into a 50 mL cylindrical beaker (4.3 cm diameter × 6.2 cm height), placed in an ice bath, and cooled to 5–10 °C. Citronellal (1.00 g) or Eucalyptus citriodora essential oil (1.00 g) was then introduced. The ultrasound probe was immersed 1 cm into the reaction mixture. Three ultrasonic parameters were selected for investigation: ultrasonic amplitude (100% corresponding to 120 µm), ultrasound pulse mode (two pulse regimes were evaluated: 30 s on/5 s off, corresponding to a pulse ratio of 6, and 30 s on/15 s off, corresponding to a pulse ratio of 2), and ultrasound time. The energy input (EI, kJ·g⁻¹), calculated for 1.00 g of citronellal or essential oil, was determined using a simple calorimetric method based on the initial temperature rise of the liquid medium during the ultrasonic experiments (see Section S1 in the Supplementary Data). The reaction was subjected to ultrasonic irradiation under the conditions outlined in Table 1 (Exps. A1–A9, B1–B2, C1–C4). The beaker was maintained in the ice bath throughout the sonication process.

Post-reaction work-up. Upon completion, the mixture was neutralized with 8% (w/w) aqueous NaHCO₃. The organic fraction was extracted successively with n-hexane (10 mL × 2) and dichloromethane (10 mL × 2) to ensure complete extraction of the products. The combined organic phases were dried over anhydrous Na₂SO₄ (1.00 g) and concentrated in vacuo at 40 °C. The resulting crude oil was analyzed directly by gas chromatography (GC).

Crystallization of PMD. For products derived from pure citronellal, n-hexane (5 mL) was added to the crude oil before cooling to − 20 °C for 24 h to induce crystallization. For products obtained from Eucalyptus citriodora oil, the crude product was cooled directly under identical conditions. In both cases, the resulting solids were collected by filtration and washed with cold n-hexane (5 mL).

2.3. Gas chromatography conditions

GC–MS analysis. Gas chromatography–mass spectrometry (GC–MS) was carried out on a SCION 456-GC instrument coupled to a mass spectrometric detector, equipped with an Rxi®-5MS fused silica capillary column (Crossbond® 5% diphenyl/95% dimethylpolysiloxane, Restek, USA; 30 m × 0.25 mm × 0.25 μm). Electron ionization (EI) mode was employed. The injector temperature was maintained at 250 °C, and the column flow rate was set to 1.0 mL/min with helium as the carrier gas. Samples were prepared in n-hexane, and 1 μL aliquots were injected (For further details, see Section S2 in the Supplementary Data).

GC–FID analysis. Gas chromatography with flame ionization detection (GC–FID) was performed on a SCION 456-GC instrument fitted with an Rxi®-5MS fused silica capillary column (Crossbond® 5% diphenyl/95% dimethyl polysiloxane, Restek, USA; 30 m × 0.25 mm × 0.50 μm). The injector temperature was maintained at 250 °C, with helium as the carrier gas at a flow rate of 1.0 mL/min. Hydrogen served as the combustion gas. Samples were prepared in n-hexane, and 1 μL aliquots were injected (For further details, see Section S2 in the Supplementary Data).

2.4. Computational methods and data processing

Density Functional Theory (DFT). All calculations were performed with ORCA 6.0.1. Geometry optimizations and frequency analyses employed the PBEh-3c composite method. Solvent effects were described using the SMD implicit solvation model with water as the solvent. Transition state (TS) searches and optimizations were conducted using the nudged elastic band (NEB) method. Reactant and product structures were fully optimized prior to execution of the %NEB command [26], [27], [28], [29], [30], [31], [32], [33].

Unless otherwise stated, the energies reported herein are single-point values including thermodynamic corrections at standard conditions (1 atm, 298 K). Energies are expressed in Hartree (E_h) and converted to kcal/mol where appropriate. Additional computational details are provided in Section S9 in the Supplementary Data.

The conformational distribution is determined based on the following equation [34]:

$$K_i=\frac{[{\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{f}\mathrm{o}\mathrm{r}\mathrm{m}\mathrm{e}\mathrm{r}}_i]}{{[\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{f}\mathrm{o}\mathrm{r}\mathrm{m}\mathrm{e}\mathrm{r}}_{\mathrm{m}\mathrm{i}\mathrm{n}}]}=\exp (-\frac{{\mathrm{\Delta }\mathrm{G}}_i}{\mathit{RT}})$$

$${\%\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{f}\mathrm{o}\mathrm{r}\mathrm{m}\mathrm{e}\mathrm{r}}_i=\frac{{\mathrm{K}}_i}{{\sum }_{i=1}^n{K}_i}$$

conformer_min: the conformer with the lowest relative Gibbs free energy; n: the number of conformers

Other data processing and statistical analyses were carried out in RStudio, making use of the “FrF2,” “dplyr,” “alr4,” “pid,” and “ggpubr” packages.

3. Results and discussion

3.1. Synthesis of para-menthane-3,8-diol (PMD) via the heated magnetic stirring method

Using the heated magnetic stirring method with pure citronellal as the substrate (Exp. A0, Table 2), the reaction reached 99.6% conversion after 10 h, affording a crude product with 86.77% PMD. These values are in close agreement with those reported by Yuasa et al. (92.3% PMD, 99% conversion) [8].

Table 2.

Synthesis results of PMD from citronellal and E. citriodora oil.

Method	Exp.	m (g)		p_im1 (%)	p_citro (%)	p_PMD (%)	p_PMD_m (%)***	p_im2 (%)	p_cis (%)
Magnetic stirring heating method	A0*	0.95		7.47	0.47	83.88	86.77	8.18	62.55
	B0**	0.90		23.39	7.38	53.51	−	15.72	62.66
	C0**	0.90		18.17	8.92	71.23	−	1.68	64.70
Ultrasound method	A1	0.90		6.85	11.90	80.78	83.77	0.47	67.68
	A2	0.96		4.05	0.32	94.71	98.09	0.92	65.52
	A3	0.80		6.19	27.62	65.16	67.70	1.03	67.10
	A4	0.93		5.28	0.39	93.34	96.30	0.99	65.24
	A5	0.95		4.72	0.40	94.05	96.64	0.83	64.28
	A6	0.95		6.00	0.60	91.65	94.97	1.75	65.29
	A7	0.90		6.74	0.74	91.67	94.39	0.85	63.81
	A8	0.90		8.05	0.62	90.53	93.06	0.80	63.15
	A9	0.95		6.96	2.20	90.07	92.67	0.77	65.90
	B1	0.92		24.78	6.21	52.81	−	15.85	66.92
	B2	0.90		19.57	5.12	62.54	−	13.21	63.88
	C1	0.90		16.45	0.00	80.67	−	2.89	62.45
	C2	0.95		20.70	0.00	77.17	−	2.12	64.90
	C3	0.72		13.34	45.36	37.91	−	3.39	65.02
	C4	0.92		18.14	0.00	78.91	−	2.49	61.74

* Performed based on conditions published by Yuasa et al. [8].

** Performed based on conditions published by Drapeau et al. [12].

*** Calculations were exclusively performed on experiments utilizing (±)-citronellal as the starting material.

For further details, see Section 3 in the Supplementary Data.

(m: mass of product; p_im1: peak area percentage of volatile components eluting prior to PMD, excluding citronellal; p_citro: peak area percentage of residual citronellal; p_PMD: peak area percentage of PMD; p_PMD_m: peak area percentage of PMD, corrected for impurities present in the starting citronellal material; p_cis: peak area percentage of cis-PMD isomers relative to total PMD; p_imp2: peak area percentage of volatile components eluting after PMD).

When E. citriodora oil served as the substrate, the following outcomes were observed:

• For oil containing 53% citronellal (Exp. B0), the crude product obtained after 15 h contained 54.00% PMD and 7.38% residual citronellal.

• For oil containing 74% citronellal (Exp. C0), the crude product obtained after 15 h contained 71.23% PMD and 8.92% residual citronellal.

These findings align with those of Drapeau et al. [6], who reported 79.6% PMD content and ∼1% residual citronellal after 15 h using E. citriodora oil with 74% citronellal. The agreement across these experiments underscores the reproducibility of the heated magnetic stirring method and provides a robust reference point for comparison with the ultrasound-assisted approach described herein.

3.2. Synthesis of PMD via the ultrasound method

3.2.1. Ultrasound-assisted synthesis of PMD from pure citronellal

An initial series of experiments (Exps. A1–A4, Table 2) was conducted to assess the influence of key ultrasound parameters on PMD synthesis from citronellal, namely ultrasound amplitude (20% and 40%), pulse mode (2 and 6), and sonication time (1 and 2 h). Among these, ultrasound amplitude exerted the most pronounced effect (Fig. 3): increasing amplitude enhanced PMD content in the crude product while simultaneously reducing reaction time. Higher pulse modes likewise improved reaction efficiency. Guided by these results, a second series of experiments (Exps. A5–A9) was undertaken, extending the amplitude range to 40–100% and correspondingly reducing reaction times.

Fig. 3.

Effects of the sonication parameters on PMD content and cis-PMD proportion. (amp: ultrasound amplitude, pulse: ultrasound pulse mode, time: sonication time, p_cis: peak area percentage of cis-PMD isomers relative to total PMD). (a, b) Pareto plots illustrating the relative effects of sonication parameters on PMD content and cis-PMD proportion. Initial screening identified ultrasound amplitude as the most significant factor: while increasing amplitude enhanced PMD content, it simultaneously led to a reduction in the cis-PMD proportion. The experimental design comprised two levels for each variable: ultrasound amplitude (20% and 40%), pulse mode (30 s on/5 s off, ratio = 6; and 30 s on/15 s off, ratio = 2), and sonication time (1 h and 2 h). (c) Linear regression plot depicting the inverse relationship between ultrasound amplitude and cis-PMD proportion, where higher amplitudes correlate with lower cis-PMD proportions.

At ultrasound amplitudes above 40% (Exps. A5, A7, and A8), PMD content remained consistently high from 90 to 94%. Increasing the amplitude from 20% to 100% enabled a marked reduction in sonication time – from 2 h (Exp. A2) to approximately 0.5 h (Exp. A5), and as short as 0.17 h (10 min) (Exp. A8) while maintaining PMD content above 92% and achieving > 99% citronellal conversion. However, concurrent increases in amplitude and decreases in reaction time were accompanied by a decline in PMD content within the product mixture and an increase in volatile impurities (components eluting prior to PMD in GC) (Exps. A7 and A8). We propose that the increased energy input at higher amplitudes may promote certain side reactions (e.g., the formation of isopulegol, as shown in Scheme 1), thereby increasing the proportion of these volatile byproducts.

Variation in the acid solution volume had minimal influence on PMD yield: under otherwise identical conditions, Exp. A4 (40 mL) and Exp. A6 (20 mL) gave PMD contents of 96.30% and 94.97%, respectively.

The cis-PMD proportion was found to depend primarily on ultrasound amplitude, with higher amplitudes correlating with lower cis-PMD proportions (Fig. 3c). The negative correlation between ultrasound amplitude and cis-PMD proportion (R =  − 0.94) was confirmed by linear regression, which showed statistical significance (Fisher’s test, p = 0.0001657). We observed the interconversion between cis-PMD and trans-PMD in 0.25% sulfuric acid at 60 °C, as well as under 30-minute ultrasonication (see Section S8 in Supplementary Data). The conversion of cis-PMD to trans-PMD proceeded more rapidly, suggesting that the cis-isomer is thermodynamically less stable than its trans-counterpart. Based on these findings, it is postulated that higher ultrasound amplitudes provide greater energy input to the reaction system; these harsher conditions likely facilitate the isomerization of cis-PMD to trans-PMD, resulting in a reduced cis-proportion compared to lower amplitudes.

Based on the results and analysis, two optimal parameter sets were identified to balance reaction time and PMD yield. The first, employing an amplitude of 40%, a pulse mode of 30 s on/5 s off, and 2 h of sonication (Exp. A2, Fig. 4), afforded a PMD content of 98.09%. The second, utilizing an amplitude of 60%, identical pulse mode, and 0.5 h of sonication (Exp. A5), yielded 96.64% PMD.

Fig. 4.

GC–FID analysis of the crude product from Exp. A2. (t: trans-isomer, c: cis-isomer, a: major product, b: minor product, ′: enantiomer).

3.2.2. Ultrasound-assisted synthesis of PMD from E. citriodora essential oil

The optimized ultrasound conditions from Exp. A2 and Exp. A5 were applied to the synthesis of PMD from E. citriodora essential oil. These protocols were evaluated using essential oil samples with varying citronellal content: 54% (Exp. B) and 74% (Exp. C). In both cases, high PMD content was obtained: 52–63% for the 54% citronellal oil and 77–80% for the 74% citronellal oil. These results confirm that the ultrasound-assisted method is effective across a range of citronellal concentrations in E. citriodora essential oil.

Regarding the cis-PMD content, it should be noted that essential oils contain numerous constituents capable of affecting stereoisomer distribution. Clark et al. showed, for example, that the addition of surfactants can increase the cis-PMD ratio [25]. Consequently, the effect of ultrasound amplitude on cis-PMD content in essential oils may deviate from trends observed with pure citronellal. Nonetheless, the content of cis-PMD obtained in these experiments was generally higher than 60%.

3.3. Comparative analysis of heated magnetic stirring and ultrasound-assisted methods

Comparison of the PMD synthesis results obtained via the heated magnetic stirring method (Exps. A0, B0, and C0) and the ultrasound-assisted method revealed clear advantages for the latter. In ultrasound conditions, the citronellal feedstock was dispersed far more finely than with conventional stirring (Fig. 5). Acoustic waves emitted from the sonication probe dispersed citronellal into micron-scale droplets, markedly increasing the interfacial area between the oil and the aqueous acid phases. This enlarged interface directly enhanced mass transfer and, in turn, accelerated the reaction rate.

Fig. 5.

Comparison of reactions between the magnetic stirring heating method (a) and the ultrasound-assisted method (b). (Top-view images: (a) mixture heated in an oil bath; (b) the reaction captured at the moment sonication ceases, showing more finely dispersed oil particles).

In experiments with pure citronellal, ultrasound significantly decreased reaction times. At 100% amplitude (Exp. A8), the reaction was complete in just 10 min – a significant reduction from several hours – while maintaining a high PMD proportion (93%) in the crude product and achieving > 99% citronellal conversion. The proportion of less volatile impurities (including the citronellal acetal byproduct) in ultrasound experiments (∼1%) was substantially lower than in non-ultrasound controls (∼8% in the present study’s heated stirring experiments; 2.7% in the work of Yuasa et al. [8]; 3–10% in the study by Borrego et al. when performed at − 78 °C, with even higher levels observed at 70 °C [5]). Overall impurity levels in ultrasound-derived products were comparable to, and in some cases lower than, those obtained via conventional heated magnetic stirring (Table 2).

A similar trend was observed when E. citriodora essential oil was used as the feedstock. Ultrasound reduced reaction times from 11 h to 0.5–2 h, while delivering PMD content comparable to the heated magnetic stirring method. Chromatograms of the post-sonication products revealed impurities distinct from those present in the starting essential oil and in products from heated stirring; however, their total content remained low (∼0.8%).

We observed that longer reaction times led to increased impurity formation:

• −Upon heating at 60 °C or ultrasonic irradiation at 40% amplitude in 0.25% sulfuric acid, both cis-PMD and trans-PMD gradually underwent isomerization accompanied by the formation of less polar impurities. Although these changes were only faintly detectable by TLC after 30 min, they became more evident with prolonged reaction times (see Section S8 in the Supplementary Data).
• −Borrego et al. reported that Prins cyclization of citronellal under harsher conditions, including elevated temperature (70 °C), sulfuric acid with higher concentrations (0.73–11%), and prolonged reaction times (7 h), resulted in higher trans-PMD proportions relative to cis-PMD together with substantial amounts of acetal impurities (13–92%) [5]. These observations are associated with PMD isomerization, in which cis-PMD gradually converts to the thermodynamically more stable trans-PMD, as well as with PMD depletion caused by acetal formation between PMD and citronellal during prolonged reaction times. The longer citronellal remained in the reaction mixture, the more favorable the formation of acetal impurities became.

Ultrasound irradiation may provide a more favorable balance between reaction time, cis-PMD preservation, and impurity suppression. The shortened reaction time likely limits both cis-to-trans PMD isomerization and the formation of related impurities such as isopulegol, citronellal acetal, and other degradation products. These advantages may arise from efficient energy transfer through acoustic cavitation, which promotes the reaction within localized microenvironments generated during bubble collapse while simultaneously allowing effective bulk-temperature control through external ice-water cooling, a condition difficult to achieve using conventional heated magnetic stirring.

3.4. Product crystallization

In experiments with pure citronellal, the crystallized PMD obtained was of high purity (>98%, Fig. 6), with crystallization yields ranging from 47 to 75% depending on the impurity content of the crude product (see Section S5 in the Supplementary Data). The proportion of cis-PMD within the total PMD isomer mixture was 63–66%.

Fig. 6.

GC–FID analysis of crude products and crystals from reactions using citronellal (left*) and essential oil (right) as starting materials.

(Crystals obtained from the essential oil reaction exhibited a significantly higher cis-PMD content compared to those from the pure citronellal reaction, * GC chromatogram of the crystals obtained from the experiment using pure citronellal as starting material, as previously reported in our study [35]).

When E. citriodora essential oil served as the starting material, the crystallized PMD was slightly less pure (90–92%, Fig. 6), with crystallization yields of 20–40% (see Section S5 in the Supplementary Data). Notably, crude products from the essential oil experiments yielded crystals with a significantly higher cis-PMD proportion, increasing from 67% in the crude mixture to > 96% after crystallization, as confirmed by TLC and GC analysis (see Section S4.23 and S7 in the Supplementary Data). We propose that other constituents of the essential oil may facilitate selective cis-PMD crystallization by preferentially solvating or better solubilizing trans-PMD, thereby suppressing its co-crystallization.

Subsequent recrystallization of these solids from n-hexane, performed once or twice, afforded cis-PMD of exceptional purity (∼100%) with yields ranging from 75 to 90%.

3.5. Stereoselectivity in PMD synthesis

In our previous work, the retention times of all PMD isomers were assigned on both GC–MS and GC–FID chromatograms (See Section S3 in the Supplementary Data) [35]. These assignments established the foundation for calculating the relative proportions of the individual isomers (Table 3) and for developing the proposed reaction mechanism.

Table 3.

Ratios of PMD isomers within the PMD isomer mixture.

Exp.	% (ca + ca′)	% (cb + cb′)	% (ta + ta′)	% (tb + tb′)
A0	61.68%	0.87%	34.97%	2.48%
A1	66.95%	0.74%	30.62%	1.70%
A2	64.59%	0.93%	32.52%	1.96%
A3	66.40%	0.70%	31.30%	1.60%
A4	64.40%	0.84%	32.64%	2.12%
A5	63.40%	0.88%	33.41%	2.31%
A6	64.39%	0.90%	32.69%	2.02%
A7	62.90%	0.91%	33.97%	2.22%
A8	62.14%	1.01%	34.45%	2.40%
A9	65.15%	0.75%	32.36%	1.73%
Average	64.20%	0.85%	32.89%	2.05%

(t: trans-isomer, c: cis-isomer, a: major product, b: minor product, ′: enantiomer).

Under heating magnetic stirring conditions (Exp. A0, Table 3), the isomer ratio of (ca + ca′)/(cb + cb′)/(ta + ta′)/(tb + tb′) was 61.7/0.9/35.0/2.5, which was similar to the results reported by Yuasa et al. (ca/cb/ta/tb = 61.2/1.2/35.0/2.6) [8]. Regarding the proportion of the (ca + ca′) isomers, the conventional heating magnetic stirring method gave lower values than the ultrasound-assisted method (62.14–66.95%, Exp. A1–A9). Nevertheless, the overall isomer distribution followed the order (ca + ca′) > (ta + ta′) > (tb + tb′) > (cb + cb′), corresponding to approximately 64%, 33%, 2%, and 1%, respectively.

Hart-Cooper et al. reported that cyclization of the 3,3-dimethylated analogue of citronellal (3,3,7-trimethyloct-6-enal) afforded predominantly a product in which the hydroxyl substituent and the three-carbon substituent (hereafter referred to as the 3C substituent, encompassing the propan-2-ylidene moiety and related three-carbon structural variants) adopted a trans configuration [36]. Conversely, cyclization of citronellal was found to predominantly yield the (ca + ca′) isomers, which exhibit a cis relationship between these substituents. This divergence indicates a significant stereochemical influence of the 3-methyl substituent on the preferential formation of cis-isomers during citronellal cyclization. To rationalize this observation, a plausible mechanistic hypothesis is proposed based mainly on the experimental observations, while DFT calculations were used to support the proposed interpretation (illustrated using the (S)-citronellal enantiomer).

The formation of the cyclic carbocation intermediate (INT2) is the rate-limiting step that dictates the ratio of the PMD isomers in the final product mixture (Scheme 1). For cyclization to proceed, the oxocarbenium ion intermediates (INT1) must undergo a conformational rearrangement in space, where only geometrically favorable conformations permit the ring closure. Eight distinct chair conformers, designated as conformers (i)–(viii), may be formed (Fig. 7).

• −From the thermodynamic perspective − the chair conformation in which the 3C substituent occupies an equatorial position is energetically favoured.

Fig. 7.

Formation of cyclic carbocation intermediates (INT2) corresponding to the PMD isomers (illustrated using the (S)-citronellal enantiomer) (illustrated using the optical isomer (S)-citronellal, t: trans-isomer, c: cis-isomer, a: major product, b: minor product, ′: enantiomer, eq: equatorial, ax: axial, INT: intermediate, 3C: three-carbon substituent).

Among the three substituents attached to the cyclohexane ring (the 3C substituent, methyl, and hydroxyl), the 3C substituent is the most sterically demanding. Consequently, it preferentially adopts an equatorial orientation to minimize steric repulsion and maximize energetic stabilisation. DFT calculations further showed that conformers with the 3C substituent in the equatorial position (including conformers (i), (iii), (v), and (vii)) predominantly account for 99.98% (Table 4). These conformers therefore represent the predominant conformational isomers involved in PMD formation.

• −From a kinetic perspective − steric influence of the methyl substituent on isomer ratios.
• (1)Methyl–3C steric interactions: Conformations in which the methyl and 3C substituents are located on opposite sides of the ring plane are formed preferentially over those in which both substituents reside on the same side, (ca′) ≫ (cb′) and (ta′) ≫ (tb′)

Table 4.

DFT calculations of the carbocation ring-closure step in the PMD formation from citronellal (illustrated using the (S)-citronellal enantiomer).

Isomer	Conformer	%INT1	TS	ΔG‡INT1 (kcal/mol)
ca′	(i)	37.24%		0.76
	(ii)	4 × 10-10%		1.43
ta′	(iii)	58.12%		0.85
	(iv)	10-11%		1.48
tb′	(v)	3.71%		2.84
	(vi)	10-10%		2.38
cb′	(vii)	0.91%		4.26
	(viii)	0.02%		1.52

(t: trans-isomer; c: cis-isomer; a: major product; b: minor product; ′: enantiomer; TS: transition state; INT1: intermediate formed after protonation of citronellal, including eight conformers (i)–(viii); INT1 subsequently undergoes cyclization to form the cyclic intermediate INT2; ΔG^‡_INT1: activation energy of INT1).

The methyl substituent is the second most sterically demanding group among the three substituents. Therefore, its steric interaction with the bulkiest 3C substituent exerts a major influence on isomer selectivity.

When the 3C substituent is drawn towards the —C⁺(OH) group during cyclization, it encounters significant steric hindrance with the methyl substituent. Accordingly, conformations where the methyl and 3C substituents are on opposite sides of the ring plane experience less steric interference from the methyl substituent during the ring-closing step compared to those where they are on the same side. Consequently, the propensity for product formation is higher for conformers (i), (ii), (iii), and (iv) than for (v), (vi), (vii), and (viii).

DFT calculations revealed that conformers (i), (ii), (iii), and (iv) exhibit lower activation energies for the conversion of INT1 to INT2 than conformers (v), (vi), (vii), and (viii). The corresponding ΔG^‡ values ranged from 0.76 to 1.48 kcal/mol for the former group and from 1.52 to 4.26 kcal/mol for the latter group (Table 4). In addition to their lower activation barriers, conformers (i) and (iii) also accounted for overwhelmingly higher conformational populations than the other conformers. Consequently, the formation of ca′ and ta′ was strongly favored over that of cb′ and tb′.

Product formation ratios:	[ca′ = (i) + (ii)] ≫ [cb′ = (vii) + (viii)]
	[ta′ = (iii) + (iv)] ≫ [tb′ = (v) + (vi)]

• (2)Methyl–hydroxyl steric interactions: The formation of structures where the methyl and hydroxyl substituents are located on opposite sides of the ring plane is more favored than those where they reside on the same side, ca′ > ta′.

When the —C⁺(OH) group interacts with the double bond, the 3C substituent is drawn toward the —C⁺(OH) group, resulting in steric repulsion exerted onto the hydroxyl substituent. Subsequently, the hydroxyl substituent exerts steric influence on adjacent structures. Two notable steric interactions are the 1,3-diaxial interaction between hydroxyl^axial and methyl^axial substituents in conformer (vii) and the equatorial 3-substituent interaction from 3C^equatorial to hydroxyl^equatorial to methylene to methyl^equatorial in conformer (iii).

The equatorial 3-substituent interaction makes conformer (iii) less favorable for ring closure than conformer (i). DFT calculations also revealed that the activation energy of conformer (i) is lower than that of conformer (iii). Although conformer (iii) (58.12%) was thermodynamically more populated than conformer (i) (37.24%), corresponding to only a 1.56-fold difference, the lower activation energy of conformer (i) resulted in preferential product formation relative to conformer (iii) (Table 4).

• Product formation ratios: [ca′ = (i) + (ii)] > [ta′ = (iii) + (iv)]

The 1,3-diaxial interaction between hydroxyl^axial and methyl^axial substituents makes ring closure of conformer (vii) less favorable than that of conformer (v). In addition, conformer (vii) accounted for a lower population than conformer (v) because the hydroxyl substituent occupies an axial rather than an equatorial position. Consequently, product formation from conformer (vii) was less favored than from conformer (v). Comparison between conformers (v) and (viii) further showed that conformer (v) was overwhelmingly more populated than conformer (viii), by approximately 157-fold according to the DFT calculations, due to the preferential equatorial orientation of the 3C substituent. As a result, product formation from conformer (v) was strongly favored. Conformer (v) was also identified as the predominant conformer responsible for tb′ formation (Table 4).

• Product formation ratios: [tb′ = (v) + (vi)] > [cb′ = (vii) + (viii)]

Collectively, the product distribution following the order ca′ > ta′ ≫ tb′ > cb′ was successfully rationalized by the proposed mechanistic model and supported by the DFT calculations.

3.6. Stereoselectivity in the Prins cyclization of citronellal under different reaction conditions and in the Prins cyclization of citronellal analogues

In the study reported by Borrego et al., Prins cyclization of citronellal performed at low temperature (−78 °C) afforded remarkably high cis-PMD proportions, accounting for 84–88% of the isomeric mixture [5]. These results are consistent with our proposed stereoselective hypotheses, in which cis-PMD is predicted to be the major product formed during cyclization. In contrast, when the reaction was carried out at elevated temperature (70 °C), trans-PMD became the predominant isomer. We propose that this reversal may be attributed to PMD isomerization and side reactions, as discussed in Section 3.3.

In investigation of the stereoselectivity of (ca + ca′) and (ta + ta′) isomer formation, Clark et al. proposed that the pathway leading to (ta + ta′) involves the formation of the hydroxy^equatorial-3C^axial-methyl^axial conformation (ii) [25]. This proposal was used to rationalize the observed ratio increase of (ca + ca′)/(ta + ta′) from 2 to 5 times upon the addition of the surfactant sodium dodecyl sulfate (SDS). However, our computational and experimental data lead to a distinct conclusion compared to the hypothesis by Clark et al. Our findings suggest the hydroxy^axial-3C^equatorial-methyl^equatorial conformation (iii) is preferentially formed. We propose that the presence of a surfactant during the cyclization step induces the structural distribution of the acyclic carbocation intermediate between the oil and aqueous phases. The key differentiating factor is the hydrophobic interaction between the surfactant's non-polar chain and the methyl substituent. This interaction significantly increases the effective local steric hindrance imposed upon the hydroxyl substituent during the ring closure. Consequently, the transition state that requires the hydroxyl and methyl substituents to be on opposite faces of the ring is energetically disfavored.

Based on the aforementioned stereoselectivity analyses, we have demonstrated that the substituent at the third carbon position (C3) of citronellal plays a pivotal role in modulating the formation of the cis-PMD isomer. In the study reported by Hart-Cooper et al., when two methyl substituents were present at the C3 position, the trans-isomer was formed predominantly, accounting for approximately 75% of the stereoisomeric mixture [36]. Based on these observations, we propose the following stereochemical trends:

• When the two substituents/atoms attached at the C3 position are identical, the trans-isomer is preferentially formed. During the cyclization process, analogues bearing two identical substituents/atoms at C3 experience pronounced 1,3-diaxial interactions between the hydroxyl group and the axial C3 substituent, which disfavors formation of the cis-isomer. DFT calculations performed on the unsubstituted C3 analogue and the 3,3-dimethyl analogue further supported the preferential formation of the trans-isomer (Fig. 8).

• When the two substituents/atoms at C3 differ significantly in steric size, specifically in systems containing one hydrogen atom and one substituent at least as bulky as a methyl group, the cis-isomer is preferentially formed. In such cases, the equatorial 3-substituent interaction hinders trans-isomer formation during the ring-closing step. DFT calculations performed on a 3-tert-butyl-substituted analogue, which is substantially bulkier than the 3-methyl substituent of citronellal, also predicted preferential formation of the cis-isomer (Fig. 8).

Fig. 8.

The preferential formation of the cis-isomer in the presence of C3-substituents. (ΔG^‡ calculated by DFT).

4. Conclusion

The present study suggests that low-frequency (20 kHz) ultrasound irradiation has potential for the synthesis of para-menthane-3,8-diol through the Prins cyclization of citronellal. Under optimized ultrasonic conditions, product mixtures with high PMD contents were obtained with fewer byproducts and shorter reaction times compared with the prolonged conventional thermal method. The potential advantages of ultrasound may arise from localized energy generation and more effective dispersion through the collapse of acoustic cavitation bubbles, thereby allowing efficient reaction activation while the bulk reaction medium remains cooled. These laboratory-scale results establish a technical foundation for future studies on larger-scale ultrasound-assisted PMD synthesis or the use of ultrasound as a complementary tool that can be integrated with different catalytic systems to improve selectivity, reduce byproduct formation, and shorten reaction times in Prins cyclizations.

In addition, the crystallization of PMD-enriched Eucalyptus citriodora essential oil offered a convenient approach for enriching cis-PMD without chromatographic purification. This route is particularly valuable for resource-limited settings or essential oil feedstocks with relatively low citronellal contents (e.g., approximately 53%). However, the key compositional and physicochemical factors influencing the selective crystallization of cis-PMD remain unclear and require further study.

To support the experimental stereochemical interpretation, density functional theory calculations were performed using the PBEh-3c composite functional, which was selected to provide a balance between computational cost and accuracy. The calculated results supported the proposed mechanistic hypotheses, rationalizing the experimental observations not only in the present study on citronellal cyclization but also in the work of Hart-Cooper et al. on the cyclization of the 3,3-dimethyl citronellal analogue.

Overall, these findings contribute to a better understanding of ultrasound-assisted Prins cyclizations and the mechanistic factors governing stereoselectivity in citronellal cyclization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following are the Supplementary data to this article:

Supplementary Data 1

The Supplementary Data consist of energy data, GC results, TLC results, and DFT calculations.