DES/Extraction and Process Optimization of Callistemon citrinus (Curtis) Skeels Essential Oil

Abstract

The essential oil of Callistemon citrinus (Curtis) Skeels is a valuable natural product with a complex volatile composition. In this study, an ultrasonic-assisted hydrodistillation process was systematically optimized using a combination of the steepest ascent method and response surface methodology (RSM) to refine key extraction parameters and identify the central point of the experimental design. Heating time, heating power, and ultrasonic treatment time were selected as critical variables influencing extraction efficiency. The optimized conditions were determined to be a heating time of 57 min, a heating power of 563 W, and an ultrasonic treatment time of 23 min, under which an essential oil yield of 0.8 mL/kg (dry weight basis) was achieved. Gas chromatography–mass spectrometry (GC–MS) analysis indicated that pretreatment with different metal ions resulted in pronounced differences in the chemical profiles of the extracted essential oils, particularly in major constituents such as cineole and α-terpineol. Among the tested ions, Sr²⁺ and Ca²⁺ pretreatments were associated with a higher relative abundance of monoterpenoid compounds, which suggests a selective influence of divalent metal ions on the extraction behavior of volatile components. Principal component analysis (PCA) further revealed clear discrimination among the essential oils obtained under different ion pretreatment conditions, confirming the compositional variability induced by ion-assisted extraction.

1. Introduction

Callistemon citrinus is an aromatic species widely distributed in subtropical and temperate regions and has attracted increasing attention due to its rich phytochemical profile and broad biological potential. Extracts and essential oils derived from C. citrinus are known to be abundant in terpenoids and phenolic compounds, which have been associated with multiple bioactivities, including insecticidal, antioxidant, and neuroactive effects. In addition, Callistemon citrinus (Curtis) Skeels (C. citrinus) is widely cultivated as a landscape plant because of its distinctive bottlebrush-like inflorescences [1,2] (Abdelhady and Aly, 2012; Salem et al., 2013a). Previous studies have reported that its leaves contain a wide range of functional compounds [3,4,5]; however, these phytochemical resources are often underutilized and insufficiently explored. Despite increasing reports on its chemical composition and general bioactivities, most existing studies remain largely descriptive, with limited insight into the molecular interactions between individual bioactive constituents and their corresponding biological targets. Consequently, the mechanistic understanding of how specific components from C. citrinus exert their biological effects remains fragmented. Addressing this knowledge gap is, therefore, essential for providing a clearer rationale for species selection and for advancing the targeted functional or biocontrol applications of C. citrinus.

Essential oils extracted from C. citrinus have been reported to have antifungal and insecticidal activity [1,6], and most have been used as antimicrobial and antifungal agents [2,7]. However, although the raw plant material contains a relatively high proportion of essential oils, their direct utilization remains constrained by the lack of efficient and environmentally friendly separation strategies [8], so it is crucial to find an efficient, green extraction method that maximizes the yield and quality of essential oils. Methods commonly used for plant essential oil extraction include steam distillation, microwave-assisted extraction, supercritical fluid extraction, and ultrasound-assisted extraction, et al. [9,10]. Therefore, it is important to develop an efficient and environmentally friendly pretreatment technology to destroy cell walls and improve the extraction efficiency of essential oils. The previous research results of Khursheed et al. showed that metal salt solutions can effectively dissolve cellulose components [11] (Khursheed et al., 2020), and metal ions with different valence states may have different interaction mechanisms with cell wall components due to their differences in charge density and ion radius [12], thus selectively affecting the release of active ingredients [13]; this phenomenon provides a new idea for the directional extraction of plant essential oils through ion regulation.

Deep eutectic solvent (DES) is a new type of green solvent formed by a hydrogen bond acceptor and a hydrogen bond donor through a hydrogen bond interaction [14]. Compared to traditional ionic liquids, deep eutectic solvent has designable physical and chemical properties, low toxicity and good biocompatibility [15]. It is widely used in the extraction of various active ingredients and shows great potential to replace traditional solvents [16]. Based on the above background, an innovative pretreatment strategy was proposed to couple the solvent function of DES with its pH-regulating ability and the cell wall destruction function of metal salt solution. We choose choline chloride and dilute hydrochloric acid or sodium hydroxide to construct a deep eutectic solvent system with different pH values and directly configure different types of metal salt (NaCl, KCl, SrCl₂, CaCl₂) pretreatment solutions in this deep eutectic solvent. Based on this, we employed an ultrasound-assisted steam distillation method to extract essential oil from C. citrinus leaves.

The purpose of this study was to systematically optimize the extraction process through single-factor experiments and the response surface method. The chemical composition differences of essential oils obtained by different pretreatments were analyzed by gas chromatography–mass spectrometry (G–MS), and their chemical profile characteristics were revealed by principal component analysis. Finally, the potential inhibitory activity of acetylcholinesterase of essential oil was predicted by molecular docking technology. It provides a theoretical basis and technical path for the directional and efficient extraction of C. citrinus essential oil and its application in functional product development.

2. Results and Discussion

2.1. Single-Factor Analysis

As shown in Figure 1, the yields of essential oils obtained from the treatment with different metal salt solutions are presented. Different inorganic salt solutions were used to treat millephrine within a unit of time, and the relationship between the extraction rate and different inorganic salt solutions was discussed. Through comparison, it was found that the SrCl₂ pretreatment group had the highest essential oil yield, reaching 0.55 mL/kgDW, followed by CaCl₂ and KCl, while the NaCl group had the lowest yield. This difference mainly stems from the physicochemical properties of different metal ions and their mechanisms of interaction with plant cell walls: divalent metal ions (such as ${Sr}^{2+}$, ${Ca}^{2+}$), due to their high charge density and large ionic radius, can more effectively undergo cross-linking or displacement reactions with components such as pectin and cellulose in the cell wall [12] (Dudev and Lim, 2014), disrupting the integrity of the cell structure and thereby enhancing the release efficiency of essential oils. However, the effect of monovalent ions (${Na}^{+}$, $K^{+}$) is relatively weak. The results of this study provide a theoretical basis for further optimizing the essential oil extraction process, indicating that choosing an appropriate metal salt pretreatment agent can significantly improve extraction efficiency, which has particular application potential in the green extraction technology of natural products [17] (Modi et al., 2021).

2.2. Plackett–Burman Design (PBD) Experimental Analysis

PBD is an effective method to screen out key factors from all independent variables. A total of 12 sets of experiments were conducted according to the PBD method.

Table 1. Response surface PBD optimization of essential oil yield and variance (ANOVA) fitting.

Run	PBD Experiments a							ANOVA
	PH	Ion Type	Ion Mass (g)	Ultrasonic Time (min)	Ultrasonic Power (W)	Heating Power (W)	Heating Time (min)	${\mathbf{Y}}_{\mathbf{E}\mathbf{O}}$ (mL/kgDW)	Source	Sum of Squares b	Df c	Mean Square d	F-Value e	p-Value f
1	2	$\mathrm{Sr}{\mathrm{Cl}}_2$	1.0	60	75	140	10	0.10	Model	0.2115	7	0.0302	6.3000	0.0472 *
2	2	NaCl	1.0	10	600	700	10	0.25	${\mathrm{X}}_1$	0.0002	1	0.0002	0.0435	0.8450
3	2	$\mathrm{Sr}{\mathrm{Cl}}_2$	1.0	10	600	700	120	0.50	${\mathrm{X}}_2$	0.0052	1	0.0052	1.0900	0.3560
4	9	$\mathrm{Sr}{\mathrm{Cl}}_2$	1.0	10	75	140	120	0.30	${\mathrm{X}}_3$	0.0002	1	0.0002	0.0435	0.8450
5	9	NaCl	1.0	60	75	700	120	0.50	${\mathrm{X}}_4$	0.0169	1	0.0169	3.5200	0.1338
6	9	$\mathrm{Sr}{\mathrm{Cl}}_2$	0.1	60	600	700	10	0.10	${\mathrm{X}}_5$	0.0019	1	0.0019	0.3913	0.5655
7	9	NaCl	0.1	10	600	140	120	0.40	$X_6$	0.0352	1	0.0352	7.3500	0.0535
8	2	NaCl	0.1	60	75	700	120	0.40	${\mathrm{X}}_7$	0.1519	1	0.1519	31.7000	0.0049 **
9	2	NaCl	0.1	10	75	140	10	0.20	Residual	0.0192	4	0.0048
10	9	$\mathrm{Sr}{\mathrm{Cl}}_2$	0.1	10	75	700	10	0.30	Cor Total	0.2306	11
11	2	$\mathrm{Sr}{\mathrm{Cl}}_2$	0.1	60	600	140	120	0.30	${\mathrm{R}}^2$	0.9169
12	9	NaCl	1.0	60	600	140	10	0.1	Adj.${\mathrm{R}}^2$	0.7715

^a: The results were obtained using Design Expert 13.0.1. X₁: pH. X₂: Ion type. X₃: Ion mass (g). X₄: Heating time (min). X₅: Ultrasonic power (W). X₆: Heating power (W). X₇: Ultrasonic time (min). ^b The sum of squares between the average values and the overall mean. ^c Degrees of freedom. ^d Sum of the squares divided by degrees of freedom. ^e Test for comparing term variance with residual variance. ^f Probability of the observed F-value. “*” is significant and “**” is extremely significant.

shows the essential oil yield and analysis of variance (ANOVA) of the 12 groups of experiments. The experimental data were fitted by multiple regression with Design Expert 13.0 software, and the second-order polynomial modulus of essential oil yield ($Y_{essentialoil}$) and each coding factor was obtained as follows:

$$Y_{essentialoil}=0.2875-0.0042X_1-0.0202X_2+0.0042X_3-0.0375X_4-0.0125X_5+0.0542X_6+0.1125X_7$$

(1)

where $Y_{essentialoil}$ is the essential oil yield, X₁ is pH effect, X₂ is ion type, X₃ is ion mass, X₄ is ultrasonic time, X₅ is ultrasonic power, X₆ is heating power, and X₇ is heating time. From Table 1, it can be seen that the F-value of the model is 6.30 and the p-value is 0.0472, which indicate that the model is significant at the 5% level and that the probability that this F-value is caused by noise is only 4.72%. The coefficient of determination (R²) of the model was 0.9169, indicating that the model could explain 91.69% of the variance of the response values, and the fitting degree was good. Among the seven factors investigated, only heating time had a significant effect on the yield of essential oils (p = 0.0049); other factors were not significant (p > 0.05), which indicated that heating time was the key factor affecting the yield of essential oils. By comparing the p-values of the seven factors, we concluded that the importance of the seven factors for the yield of essential oils was heating time > heating power > ultrasonic time > ion type > ultrasonic power > ion type. After comprehensive consideration, the factors of heating time, heating power and ultrasonic time were selected for further optimization.

2.3. Optimization of Essential Oil Yields with Box–Behnken Design

Table 2. Response surface BBD optimization of essential oil yield and variance (ANOVA) fitting.

No.	BBD Experiments a				ANOVA
	Heating Time (min)	Heating Power (W)	Ultrasonic Time (min)	${\mathit{Y}}_{\mathbf{E}\mathbf{O}}$(mL/kgDW) (mL/kgDW)	Source	Sum of Squares b	Df c	Mean Square d	F-Value e	p-Value f
1	40	700	20	0.62	Model	0.0777	9	0.0086	10.1900	0.0029 **
2	40	566	10	0.60	${\mathrm{X}}_7$	0.0392	1	0.0392	46.2700	0.0003 **
3	60	566	10	0.75	${\mathrm{X}}_6$	0.0112	1	0.0112	13.2800	0.0082 **
4	50	432	30	0.82	${\mathrm{X}}_4$	0.0144	1	0.0144	17.0600	0.0044 **
5	40	566	30	0.72	${\mathrm{X}}_6{\mathrm{X}}_7$	0.0020	1	0.0020	2.3900	0.1660
6	50	566	20	0.69	${\mathrm{X}}_4{\mathrm{X}}_7$	0.0000	1	0.0000	0.0295	0.8685
7	60	432	20	0.82	${\mathrm{X}}_4{\mathrm{X}}_6$	0.0000	1	0.0000	0.0295	0.8685
8	60	700	20	0.70	${\mathrm{X}}_7^2$	0.0013	1	0.0013	1.5700	0.2510
9	50	566	20	0.71	${\mathrm{X}}_6^2$	0.0002	1	0.0002	0.2612	0.6250
10	40	432	20	0.65	${\mathrm{X}}_4^2$	0.0094	1	0.0094	11.1000	0.0126 *
11	60	566	30	0.88	Residual	0.0059	7	0.0008
12	50	566	20	0.68	Lack of Fit	0.0036	3	0.0012	2.1300	0.2386
13	50	700	10	0.70	Pure Error	0.0023	4	0.0006
14	50	566	20	0.74	Cor Total	0.0836	16
15	50	432	10	0.78	R2	0.9291
16	50	700	30	0.75	Adj.${\mathrm{R}}^2$	0.8379
17	50	566	20	0.72

^a The results were obtained using Design Expert 13.0.1. ^b The sum of squares between the average values and the overall mean. ^c Degrees of freedom. ^d Sum of the squares divided by degrees of freedom. ^e Test for comparing term variance with residual variance. ^f Probability of the observed F-value. “*” is significant and “**” is extremely significant.

shows 17 groups of experiments designed to study the interaction of heating time, heating power and ultrasonic time on essential oil yield by Design Expert 13.0 software. Multivariate regression fitting was performed on the experimental data to obtain a second-order polynomial model between essential oil yield ($Y_{essentialoil}$) and each coding factor as follows:

$$Y_{essentialoil}=0.7080+0.0700X_7-0.0375X_6+0.0425X_4-0.0225X_7{X}_6+0.0025X_7{X}_4+0.0025X_6{X}_4-0.0178X_7^2+0.0072X_6^2+0.0473X_4^2$$

(2)

where $Y_{essentialoil}$ is the yield and $X_4$, $X_6$, $X_7$ are ultrasonic time, heating power, and heating time, respectively.

The results of the ANOVA analysis of C. citrinus essential oil yield based on response surface design are shown in Table 2. The F-value (10.19) and low p-value (0.0029) of the model showed that the model was statistically significant. The probability that the F-value of the model is caused by error is only 0.29%. In this study, the lack of fit was not significant (p > 0.05), indicating that the model fit was good and no obvious lack-of-fit phenomenon was found. The coefficient of determination ($R^2$) of the model was 0.9291, indicating that 92.91% of the essential oil yield values matched the predicted values of the model. The adjusted R² (Adj.${\mathrm{R}}^2$) was 0.8379, indicating that the regression equation was in good agreement with the actual experiments.

Table 2 shows that the linear factors ultrasonic time ($X_4$), heating power ($X_6$), heating time ($X_7$) and quadratic factors were significant (p < 0.05) for essential oil yield. Among them, the linear effect of heating time ($X_7$) was the most significant (p < 0.001), indicating that heating time was the most important factor affecting the yield of essential oil. The heating power (X₆) and ultrasonic time (X₄) also showed significant linear effects. The interaction factors were not significant (p > 0.05), indicating that the interaction between various factors had little effect on the yield of essential oil. Among the quadratic factors, only $X_4^2$ was significant, indicating a clear nonlinear relationship between ultrasound time and yield. The factors affecting the volatile oil extraction rate were sorted according to the F-value—heating time> ultrasonic time> heating power—indicating that the heating time had a greater influence on the volatile oil extraction rate and that the influence of ultrasonic time and heating power on the yield of volatile oil was not much different.

According to the regression equation, the response surface and contour plot between heating time, heating power and ultrasonic time can be plotted (Figure 2). Figure 2a,d show the interaction between heating time and heating power. The results showed that the yield of essential oil increased from 0.68% to 0.82% with the increase in heating time from 30 min to 50 min but decreased to 0.75% at 60 min after the heating time exceeded 50 min, indicating that there was an optimal heating time. At the same time, when the heating power was increased from 350 W to 450 W, the yield decreased from 0.78% to 0.74%, which indicates that too high power may have an inhibitory effect. Figure 2b,e reflect the interaction between heating time and ultrasonic time, and the results show that when the heating time is extended from 30 min to 50 min and the ultrasound time is increased from 15 min to 25 min, the yield continues to increase and reaches 0.86%. However, the growth slowed down after more than 25 min and only increased to 0.87% at 30 min, indicating that prolonging ultrasound could not further significantly improve the extraction efficiency. Figure 2c,f show the interaction between heating power and ultrasonic time. The results show that the yield increased significantly with ultrasonic time from 15–25 min, from 0.70% to 0.84%, while the heating power dropped slightly to 450 W, from 0.80% to 0.76%. Finally, ultrasound time is the main driving factor, while excessive heating power will inhibit the yield of essential oil. In conclusion, the optimal extraction conditions determined by model optimization were 57 min heating time, 563 W heating power and 23 min ultrasonic time, and the yield of essential oil reached 0.80 mL/kg DW.

2.4. GC–MS Analysis of Essential Oil Components

To clarify the chemical composition of C. citrinus essential oil (EO), this study employed gas chromatography–mass spectrometry (GC–MS) for the separation and identification of oil components. The relative area percentage (RA) in GC–MS was used to represent the relative content of the essential oil compounds. The results are shown in

Table 3. Volatile constituents of C. citrinus leaf essential oil following different inorganic ion treatments.

No.	Components	RI a	Formula	CAS	RA (%) b
					EOa	EOb	EOc	EOd	EOe
1	Isopentanol	744.0	C5H12O	000123-51-3	0.34	0.74	0.48	0.41	0.68
2	Diisopropyl ketone	794.0	C7H14O	000565-80-0	0.09	0.23	0.17	0.10	0.63
3	β-Methylbutyric acid	837.0	C5H10O2	000503-74-2	NDc	0.15	0.13	0.21	ND
4	(E)-2-Hexenal	850.0	C6H10O	006728-26-3	ND	0.18	0.16	0.14	0.15
5	Banana oil	872.0	C7H14O2	000123-92-2	0.69	0.39	0.33	0.29	0.79
6	Santen	885.3	C9H14	000529-16-8	ND	0.07	ND	0.07	ND
7	Carvomenthenal	885.3	C9H14	029548-14-9	ND	0.10	0.06	0.11	ND
8	Tricyclene	927.0	C10H16	000508-32-7	0.12	0.52	ND	ND	ND
9	1R-α-Pinene	937.0	C10H16	007785-70-8	19.45	ND	10.96	10.42	15.13
10	Camphene	945.0	C10H16	000079-92-5	0.48	0.33	0.26	0.25	0.21
11	2,4-Thujadiene	957.0	C10H14	036262-09-6	0.68	0.62	0.37	0.41	0.33
12	Sabinen	975.0	C10H16	003387-41-5	0.28	ND	0.26	ND	ND
13	β-Pinene	976.0	C10H16	000127-91-3	0.65	ND	0.51	ND	0.65
14	2-Carene	1001.0	C10H16	000554-61-0	0.23	ND	ND	0.50	ND
15	Pseudolimonen	1006.0	C10H16	000499-97-8	ND	0.37	0.43	0.50	0.45
16	Isopentyl isobutanoate	1013.4	C9H18O2	002050-01-3	0.50	ND	0.12	ND	ND
17	Camphogen	1025.0	C10H14	000099-87-6	ND	0.20	0.55	ND	ND
18	o-Cymene	1028.0	C10H14	000527-84-4	0.42	ND	ND	0.50	0.24
19	Limonene	1029.0	C10H16	000138-86-3	0.86	0.50	ND	ND	ND
20	Eucalyptol	1030.0	C10H18O	000470-82-6	36.69	33.70	32.74	31.51	34.11
21	β-Terpinen	1036.0	C10H16	000099-84-3	ND	ND	0.37	ND	0.50
22	Terpinolene	1084.0	C10H16	000586-62-9	0.07	0.10	0.10	0.07	0.05
23	α,4-Dimethylstyrene	1090.0	C10H12	001195-32-0	0.11	0.15	0.12	0.10	ND
24	Linalool	1102.0	C10H18O	000078-70-6	0.16	0.18	0.20	0.19	0.09
25	1,3,8-p-Menthatriene	1118.7	C10H14	018368-95-1	0.49	0.05	0.96	0.46	0.49
26	Fenchol	1125.6	C10H18O	001632-73-1	0.30	0.39	ND	0.37	0.27
27	(E)-Pinocarveol	1136.0	C10H16O	000547-61-5	1.66	1.44	1.73	1.85	1.22
28	L-Borneol	1160.0	C10H18O	000464-45-9	ND	ND	0.95	1.02	ND
29	Pinocarvone	1164.2	C10H14O	030460-92-5	0.86	1.18	1.15	1.20	0.89
30	Bhimsiam camphor	1168.0	C10H18O	000507-70-0	0.59	0.99	ND	ND	0.69
31	(E)-Ocimenol	1168.5	C10H18O	007643-60-9	ND	0.47	ND	ND	0.40
32	Terpinen-4-ol	1177.0	C10H18O	000562-74-3	0.31	0.46	0.47	0.49	ND
33	3,9-Epoxy-1-p-menthene	1178.0	C10H16O	070786-44-6	ND	0.09	ND	0.10	0.07
34	2-Caren-4-ol	1181.2	C10H16O	006617-35-2	ND	0.45	ND	0.43	0.36
35	L-α-Terpineol	1187.0	C10H18O	010482-56-1	ND	ND	7.14	7.04	5.56
36	α-Terpineol	1190.0	C10H18O	000098-55-5	5.10	7.12	ND	ND	ND
37	Methyl salicylate	1191.0	C8H8O3	000119-36-8	ND	0.14	0.13	0.09	ND
38	Dill ether	1191.2	C10H16O	074410-10-9	0.06	ND	0.10	ND	ND
39	Carveol	1200.0	C10H16O	000099-48-9	0.99	0.95	0.11	0.10	0.08
40	(Z)-Carveol	1217.0	C10H16O	001197-06-4	0.31	0.43	0.41	0.43	0.37
41	D-Carvone	1223.0	C10H14O	002244-16-8	ND	ND	0.15	0.15	0.13
42	Carvone	1254.0	C10H14O	000099-49-0	0.10	0.15	ND	ND	ND
43	Geraniol	1256.0	C10H18O	000106-24-1	0.08	0.10	0.15	0.13	0.11
44	Perilla alcohol	1305.0	C10H16O	000536-59-4	0.18	0.61	ND	ND	0.61
45	Adamantan-2-ol	1329.0	C10H16O	000700-57-2	ND	ND	ND	1.68	0.32
46	Biosol	1332.2	C10H14O	003228-02-2	ND	ND	0.13	0.12	0.11
47	Hydroxycineyl acetate	1345.5	C12H20O3	057709-95-2	0.21	0.43	0.42	0.44	0.39
48	(E)-Methyl cinnamate	1354.0	C10H10O2	001754-62-7	0.15	0.08	ND	0.32	0.28
49	β-Maaliene	1381.0	C15H24	000489-29-2	ND	ND	0.61	0.64	0.52
50	10s,11s-Himachala-3(12),4-diene	1399.0	C15H24	060909-28-6	0.31	ND	ND	0.04	ND
51	Caryophyllene	1417.0	C15H24	000087-44-5	ND	0.27	0.32	0.37	0.30
52	γ-Maaliene	1435.0	C15H24	020071-49-2	ND	0.10	0.10	0.12	0.10
53	Aromandendrene	1436.0	C15H24	000489-39-4	ND	2.79	2.72	3.12	2.53
54	α-Maaliene	1442.7	C15H24	000489-28-1	ND	ND	0.10	ND	0.08
55	Alloaromadendrene	1453.0	C15H24	025246-27-9	0.13	0.95	0.92	ND	0.85
56	epi-β-Caryophyllene	1465.0	C15H24	068832-35-9	ND	0.09	0.08	1.04	0.20
57	γ-Selinene	1470.0	C15H24	000515-17-3	ND	ND	0.17	0.20	0.14
58	γ-Gurjunene	1478.0	C15H24	022567-17-5	ND	0.40	0.44	0.43	0.37
59	(+)-Ledene	1489.0	C15H24	021747-46-6	0.09	0.74	0.91	0.75	0.67
60	Epizonarene	1495.0	C15H24	041702-63-0	ND	0.07	0.19	ND	0.08
61	β-Selinene	1496.0	C15H24	028624-23-9	ND	ND	ND	0.09	0.10
62	α-Selinene	1498.0	C15H24	000473-13-2	ND	0.14	ND	0.25	0.22
63	L-Calamenene	1519.0	C15H22	000483-77-2	ND	0.15	0.14	0.17	0.14
64	δ-Cadinene	1523.0	C15H24	000483-76-1	ND	0.11	0.14	0.13	0.11
65	Cadine-1,4-diene	1525.0	C15H24	016728-99-7	ND	ND	0.08	0.06	0.04
66	Flaveson	1546.0	C14H20O4	022595-45-5	ND	ND	0.30	0.31	0.23
67	Espatulenol	1574.0	C15H24O	006750-60-3	0.13	0.28	0.37	0.39	0.43
68	(-)-Palustrol	1581.0	C15H26O	005986-49-2	0.08	ND	0.29	0.34	0.36
69	Cubeban-11-ol	1600.6	C15H26O	220766-71-2	0.23	0.58	0.72	ND	ND
70	Viridiflorol	1612.0	C15H26O	000552-02-3	ND	0.64	ND	1.15	1.11
71	(+)-Rosifoliol	1613.2	C15H26O	063891-61-2	ND	0.29	0.45	0.53	ND
72	Leptospermone	1620.0	C15H22O4	000567-75-9	ND	ND	9.19	9.50	8.88
73	Epiglobulol	1629.0	C15H26O	088728-58-9	ND	0.50	0.51	ND	0.50
74	Caryophylla-4(12),8(13)-dien-5β-ol	1644.2	C15H24O	019431-80-2	ND	ND	0.07	0.07	0.06
75	Cadalene	1672.0	C15H18	000483-78-3	ND	0.13	0.14	0.13	0.12
76	Neophytadiene	1837.0	C20H38	000504-96-1	ND	ND	0.04	0.04	ND
77	Hexadecanoic acid	1964.0	C16H32O2	000057-10-3	ND	ND	0.17	0.19	0.20
78	Stigmasta-3,5-diene	2718.6	C29H48	079897-80-6	ND	ND	ND	0.14	0.10
79	Vitamin E	3138.0	C29H50O2	000059-02-9	ND	ND	ND	0.22	0.20
	Total identified	□ □	□ □	□ □	74.18	62.29	81.32	82.07	84.50
	Monoterpene				23.24	2.69	14.77	13.11	18.05
	Sesquiterpene				0.53	5.66	6.94	7.28	6.45
	Oxygenated				48.04	51.43	57.45	59.64	57.12

^a Remain index, check it out on the official website of Nist; ^b Relative peak area; ND Not detected.

. The relative peak area (RA) of each component was calculated using the area normalization method. During the detection process, four different ions (Sr²⁺, Na⁺, Ca²⁺, K⁺, corresponding to EO_a, EO_b, EO_c, and EO_d in the data table) were used in the pretreatment experiments, along with one blank group (EO_e). The GC–MS chromatogram clearly shows that the peaks of key substances in different samples are concentrated at similar retention times, indicating that they have similar compositions.

The results reveal that 79 compounds were separated and identified from C. citrinus essential oil, encompassing various types such as alcohols, ketones, esters, alkenes, and phenols. Specifically, these can be categorized into monoterpenes (16 species), sesquiterpenes (17 species), and oxygenated compounds (46 species). The primary components of the essential oil include eucalyptol, 1R-α-Pinene, Leptospermone, and L-α-Terpineol. The compounds with relatively high concentrations in the essential oil are mainly focused among monoterpenes, sesquiterpenes, and oxygenated monoterpenes. The content proportion of terpenoids is relatively stable with a small fluctuation range. In terms of relative content, the total relative content of terpenoids (including olefinic monoterpenes, sesquiterpenes, alcoholic terpenoid derivatives, etc.) in the five samples are EO_a 71.81%, EO_b 59.78%, EO_c 79.16%, EO_d 79.03%, and EO_e 81.62%. This indicates that terpenoids are the core active components of the essential oil and that the addition of salts has a slight promoting effect on the increase in the total content of terpenoids.

Monoterpenoid compounds are the primary contributors to the aroma of C. citrinus leaf essential oil, while also exhibiting physiological activities such as antibacterial and anti-inflammatory properties. Variations in their content directly affect the sensory quality and application value of the essential oil. Limonene mainly imparts a fresh citrus fragrance to the oil. The cyclic cluster heatmap of the compounds in the essential oils is shown in Figure 3. In the figure, the EO_a ring exhibits a darker color in the Limonene region, indicating a relatively high RA value. It is inferred that Sr²⁺ may promote the release of Limonene within the cells by regulating the permeability of the cell membranes of leaves, thereby reducing volatilization losses during the extraction process.

By contrast, the corresponding area in the EO_e ring appears lighter, with a lower RA value, as pure water cannot alter the cell membrane structure, which results in a relatively lower content of Limonene compared to the strontium chloride treatment group. 1R-α-Pinene imparts a fresh, pine-needle-like woody aroma to essential oils and exhibits strong antioxidant activity. In the EO_c zone, the 1R-α-Pinene region shows a darker color and higher RA values, which may be attributed to Ca²⁺ acting as a key ion in plant cell signal transduction. Ca²⁺ can activate enzymes related to 1R-α-Pinene biosynthesis, such as terpene synthase, within the leaves and also enhance cell rupture efficiency, which leads to a more complete release of 1R-α-Pinene. By contrast, in the EO_b zone, the RA values of 1R-α-Pinene are lower because Na⁺ has a weaker regulatory effect on cell membrane permeability and cannot effectively activate terpene synthase, which results in no significant promotion of 1R-α-Pinene synthesis or release [18] (Schilcher, 1993).

Based on the specific chemical composition of each sample, the biological activities of different essential oils also vary. Regarding antimicrobial activity, EO_a may exhibit the strongest antibacterial properties due to its higher content of monoterpene alcohol compounds, which are generally known for their excellent antibacterial efficacy. In terms of antioxidant activity, EO_c and EO_d may hold greater advantages, as they are rich in ketone compounds, which generally demonstrate potent antioxidant capabilities. Regarding anti-inflammatory activity, EO_c may demonstrate more pronounced effects, closely linked to its abundant sesquiterpene content. Sesquiterpene compounds have been extensively documented to exhibit significant anti-inflammatory effects [19] (Cho et al., 2000).

In summary, different ionic treatments may significantly influence the RA of compounds such as monoterpenes, sesquiterpenes, and alcohols in essential oils by regulating the permeability of C. citrinus leaf cell membranes, activating relevant synthases, or inhibiting oxidative degradation processes. Changes in the content of these components not only directly determine the aroma characteristics of essential oils (such as citrus, woody, and sweet floral scents) but also relate to their biological activity (such as antibacterial and anti-inflammatory properties). Further comparative analysis of different metal salts revealed that, compared to the control group, strontium chloride (EO_a) and sodium chloride (EO_b) favored the accumulation of monoterpenes, while potassium chloride (EO_d) and calcium chloride (EO_c) promoted the release of sesquiterpenes and oxygen-containing compounds. This indicates that ion treatment methods can be selected based on the requirements of target bioactive compounds to selectively enhance functional properties such as antioxidant and antibacterial activity in essential oils. Furthermore, for easily oxidized components like terpenes, their relative content and color changes under different ion treatments suggest potential stability differences. Because terpenes readily oxidize, leading to essential oil degradation and off-flavor formation, these results also indicate that ion treatment methods significantly impact the storage stability of essential oils.

2.5. Principal Component Analysis

To systematically evaluate the overall impact of different metal salt pretreatments on the chemical composition of C. citrinus leaf essential oil, principal component analysis (PCA) was performed on the GC–MS relative content data of five essential oil samples (EO_a–e). The analysis results are presented in Figure 4, which is a PCA biplot showing both sample distribution and compound contributions. The PCA results indicated that the cumulative variance contribution rate of the first two principal components reached 97.8%. This demonstrates that these two principal components can fully capture and explain most of the variation information in the original GC–MS data, confirming high model reliability.

In Figure 4, both EO_a and EO_b are located on the positive semi-axis of PC1 but separated along the PC2 axis, which indicates that the effects of the Sr²⁺ and Na⁺ pretreatments on the chemical composition of the essential oil have both similarities and specificity. The three samples (EO_c, EO_d, and EO_e) cluster on the negative semi-axis of PC1, which suggests that their chemical compositions are relatively similar but significantly different from those of EO_a and EO_b, especially EO_e, which has the largest projection in the negative direction of PC2. Key compounds in the right half of the figure, such as eucalyptol and α-pinene, have loading vectors pointing to the positive direction of PC1, which is fully consistent with the GC–MS data showing that EO_a and EO_b have a higher eucalyptol content and that EO_a is rich in α-pinene. Compounds in the upper and left halves of the figure make a significant negative contribution to PC1 and PC2, which indicates that the Ca²⁺ and K⁺ pretreatments, as well as the control treatment, are more likely to promote the dissolution of oxygenated compounds. Therefore, Sr²⁺ and Na⁺ pretreatments tend to enhance the release of monoterpene olefins, while Ca²⁺ and K⁺ pretreatments are closely related to the content of specific oxygenated terpenoids and ketone compounds, with the chemical profile of the blank control group falling between the two.

2.6. Microscopic Change Analysis

2.6.1. Electron Scanning Microscopy (SEM) Analysis

Figure 5 shows the microscopic morphology of raw material leaf powder and extracted filter residue visualized with scanning electron microscopy. Figure 5a,c,e show micrographs extracted without ultrasound at different scales, and Figure 5b,d,f correspond to SEM images of C. citrinus leaf powder that has undergone ultrasonic extraction at different scales. Under SEM, the surface of the sample after ultrasonic extraction showed significant wrinkles, cracks, and holes, compared to the relatively intact and smooth surface of the sample without sonication. This was due to the powerful ultrasonic cavitation effect that produces instantaneous high pressure and shock waves [20] (Shen et al., 2023), which physically break the solid cell wall and effectively destroy the microstructure of the C. citrinus leaves, which increases the mass transfer area and channels [21] (Shi et al., 2025), making the essential oil more easily soluble and extracted by solvents.

2.6.2. Fourier Transform Infrared Spectroscopy Analysis

Fourier transform infrared spectroscopy (FTIR) analysis confirmed the successful extraction of essential oils by ultrasound-assisted steam distillation. As shown in Figure 6, the structural analysis of raw leaf powder (a) and extracted filter residue (b) of C. citrinus by Fourier transform infrared spectroscopy shows that the new absorption peaks observed at 2918.8${\mathrm{c}\mathrm{m}}^{-1}$ and 2847.0${\mathrm{c}\mathrm{m}}^{-1}$ are attributed to C-H telescopic vibrations, indicating the presence of fat chain structures unique to terpenoids. The strong and wide absorption bands at 3305.8${\mathrm{c}\mathrm{m}}^{-1}$ and 3329.4 ${\mathrm{c}\mathrm{m}}^{-1}$ are attributed to the O-H telescopic vibration, indicating the presence of phenolic or alcoholic components in the essential oil. The characteristic peak at 1686.9${\mathrm{c}\mathrm{m}}^{-1}$ corresponds to the C=O telescopic vibration, indicating the presence of carbonyl-containing compounds such as aldehydes, ketones, or esters. The displacement of the C-O telescopic vibration peak to 1031.4 ${\mathrm{c}\mathrm{m}}^{-1}$ further demonstrates the successful extraction of oxygen-containing terpenoids. These spectral results indicate that essential oils contain characteristic functional groups of terpenes and their oxygen-containing derivatives, confirming the effectiveness of the extraction process.

3. Materials and Methods

3.1. Materials and Reagents

Fresh plant materials were collected in May 2025 from areas surrounding the campus of Jiaying University, Guangdong, China. The following reagents were used in this study: choline chloride (C₅H₁₄ClNO, purity ≥ 98%; Yuanye Bio-Technology Co., Ltd. (Shanghai, China), catalog No. S30169-500 g), sodium hydroxide (NaOH, purity ≥ 97%; Shanghai Zhanyun Chemical Co., Ltd. (Shanghai, China)), and hydrochloric acid (HCl, purity 36–38%; Xilong Scientific Co., Ltd. (Guangzhou, China)). These reagents were employed for the preparation of deep eutectic solvents (DESs).

The inorganic salts used for pretreatment included sodium chloride (NaCl, purity ≥ 99.5%; Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China)), strontium chloride (SrCl₂, purity ≥ 99.5%; Tianjin Zhonglian Chemical Reagent Co., Ltd. (Tianjin, China)), calcium chloride (CaCl₂, purity ≥ 96%; Tianjin Beichen Fangzheng Reagent Co., Ltd. (Tianjin, China)), and potassium chloride (KCl, purity ≥ 99.5%; Tianjin Baishi Chemical Co., Ltd. (Tianjin, China)). The deep eutectic solvents (DESs) used in this study were prepared by combining choline chloride, which acted as the hydrogen bond acceptor, with dilute hydrochloric acid or sodium hydroxide as hydrogen bond donors to construct eutectic systems with different pH values. The components were mixed at predetermined ratios and heated at 80 °C under continuous magnetic stirring until a transparent and homogeneous liquid was obtained, which indicated the formation of a stable eutectic phase.

After cooling to room temperature, inorganic salts (NaCl, KCl, SrCl₂, or CaCl₂) were directly dissolved into the prepared DES to obtain ion-containing pretreatment solutions with controlled pH and ionic composition.

3.2. Instruments and Equipment

The main instruments and equipment used in this research include: electronic analytical balance (JM-B5003, Zhuji Chaoze Weighing Instrument Equipment Co., Ltd., Shaoxing, China), grinding crusher (2500C, Yongkang Hongtaiyang Electromechanical Co., Ltd., Yongkang, China), temperature regulating heating mantle (DZTW 3000 mL, Shanghai Lizhan Bangxi Instrument Technology Co., Ltd., Shanghai, China), numerical control ultrasonic cleaning instrument (Shenzhen Peninsula Medical Co., Ltd., Shenzhen, China), distillation apparatus (1000 mL, Sichuan Shubo Co., Ltd., Sichuan, China), scanning electron microscope (Zeiss, Carl Zeiss AG, Shanghai, China), gas chromatography–mass spectrometry (Agilent 6890GC, Agilent Technologies, Inc., Beijing, China) and Fourier transform infrared spectrometer (Nicolet Magna-IR 560, Thermo Fisher Scientific, Inc., Shanghai, China).

3.3. Optimization of the Experimental Process

In the process of optimizing the extraction process, the Plackett–Burman design (PBD) was used to screen the key influencing factors based on the results of the univariate test. Seven factors—pH (X₁), ion type (X₂), ion mass (X₃), ultrasonic time (X₄), ultrasonic power (X₅), heating power (X₆) and heating time (X₇)—were selected as the investigation variables. All the factors were set into two levels: high (+1) or low (−1). The factor codes and levels are shown in Table S1. Subsequently, the steepest climbing experiment was carried out, starting from the center point of PBD and moving along a path that gradually increased in heating time, heating power and ultrasonic time. The experimental design and results are shown in Table S2. The optimal parameters for the extraction process were a heating time of 50 min, heating power of 566 W, and ultrasonic time of 20 min, while the extraction efficiency approaches the maximum response range. On this basis, the three-factor three-level Box–Behnken design (BBD) was further adopted, and the heating time, heating power and ultrasonic time were selected as independent variables, and the volatile oil yield was used as the response value, and the response surface was analyzed and optimized using Design-Expert 13.0 software. The factor coding and level are shown in Table S3.

3.4. The Characterization of the Samples

3.4.1. GC–MS Characterization of Essential Oil Components

The essential oil components were characterized by gas chromatography–mass spectrometry (GC–MS) under different ionic treatments using a DB-5 capillary column (30 m × 0.25 mm, 0.25 μm). The carrier gas (He) was maintained at a flow rate of 1.0 mL/min, and the column temperature for injecting 1 μL of sample in split mode was set to an initial temperature of 50 °C, warmed up to 180 °C at a rate of 3 °C/min for 10 min, and then warmed to 280 °C at a rate of 10 °C/min. The injection temperature was set at 280 °C. Compound identification was achieved by matching mass spectrometry to NIST libraries.

3.4.2. Micromorphology of Extracts

In order to explore the failure mechanism of ultrasound-assisted pretreatment of different inorganic salts on plant cell structure, a morphology analysis of dried leaf powder before extraction and filter residue after steam distillation was performed. After drying and sieving, the sample was fixed on the conductive glue and sprayed with gold, and the structure of the sample was observed by scanning electron microscopy (SEM) with an acceleration voltage of 3 kV and a working distance of 7.0 mm or 7.2 mm.

3.4.3. Fourier Transform Infrared Spectroscopy Analyzes the Chemical Structure of Extracted Samples

Liquid nitrogen was used to homogenize the freeze-dried samples (leaf powder raw material and post-extraction residue) into fine powders, and approximately 1 mg of each powder sample was thoroughly mixed with 100 mg of spectral-grade potassium bromide (KBr) and pressed into transparent pellets under hydraulic pressure. Fourier transform infrared spectroscopy (FTIR) measurements were performed using a Nicolet Magna-IR 560 Fourier transform infrared spectrometer. The test conditions were as follows: the wavenumber range was set to 4000–400 cm⁻¹, the resolution was set to 4 cm⁻¹, and each sample was scanned 64 times to ensure that the signal-to-noise ratio met the analysis requirements. The characteristic functional groups were identified by comparing the obtained absorption peaks with the standard spectra reported in the existing literature. All assays were repeated three times independently to ensure the reproducibility and reliability of the experimental data.

4. Conclusions

It should be noted that the novelty of the present work lies primarily in the process–composition relationship rather than in the extraction technique itself. While the general extraction framework has been previously described, this study emphasizes how ion-assisted pretreatment, coupled with multivariate optimization, influences the distribution of key volatile components in C. citrinus essential oil. Such an approach extends conventional yield-focused optimization toward a more composition-driven perspective. In this study, an ultrasound-assisted metal salt pretreatment strategy was optimized to enhance the extraction efficiency of C. citrinus essential oil and to elucidate the regulatory effects of different metal ions on its chemical composition. Response surface methodology indicated that heating time was the dominant factor influencing essential oil yield. Furthermore, GC–MS combined with PCA analysis demonstrated that metal ions with different valence states selectively modulated the chemical profile of the extracted essential oils. Specifically, Sr²⁺ and Ca²⁺ favored the enrichment of monoterpenoids, whereas Na⁺ and K⁺ promoted the release of certain sesquiterpenes and oxygenated compounds. Overall, this study establishes a green and efficient approach for both the extraction and compositional regulation of C. citrinus essential oil and demonstrates that its chemical profile can be tailored through appropriate salt pretreatment. These findings provide an experimental basis for the targeted production of essential oils with specific aroma characteristics or potential functional properties.