View PDF Version
 
DOI: 10.1039/D5AY00221D (Paper) Anal. Methods, 2025, Advance Article

Sustainable headspace-in-needle microextraction analysis of phthalates using biomass-derived carbon-coated needles†

Jooyoung Kim and Sunyoung Bae*
Department of Chemistry, Seoul Women's University, Seoul 01797, Korea. E-mail: sbae@swu.ac.kr

Received 9th February 2025 , Accepted 13th July 2025

First published on 16th July 2025

Abstract

Phthalates are used in various products as plasticizers and pose environmental and health risks owing to their endocrine disruption potential. The detection of phthalates requires appropriate sample preparation, such as adsorption-based extraction. This study explores the conversion of withered flowers into activated hydrochar through hydrothermal carbonization and subsequent activation to utilize it as an adsorbent in in-needle microextraction for phthalates. The process involved the hydrothermal carbonization of rose petals and stems; subsequently, activation was optimized by applying the Box–Behnken design and response surface methodology. The resulting activated hydrochar was applied as a coating on the interior of a needle via sol–gel polymerization to form a polydimethylsiloxane/activated hydrochar composite. The optimal conditions for hydrothermal carbonization and activation were identified as a reaction temperature of 210 °C for 18 h using 7.5 g of withered rose, followed by activation at 600 °C with a melamine-to-biomass ratio of 1[thin space (1/6-em)]:[thin space (1/6-em)]1 and a potassium hydroxide-to-biomass ratio of 5[thin space (1/6-em)]:[thin space (1/6-em)]1, with subsequent analysis via GC-MS. The synthesized adsorbent was characterized using various analytical techniques, including BET surface area analysis, FE-SEM, FT-IR spectroscopy, and TGA. This innovative in-needle microextraction approach optimized for headspace extraction demonstrated efficient phthalate extraction. The method's efficacy was validated through parameters such as the LOD, LOQ, linearity, and recovery, and the method can be considered a sustainable and effective sample preparation technique. This study shows that this method is easy to fabricate, convenient for storage, cost-effective, and can detect phthalates for advanced sample preparation.

1. Introduction

Phthalates are used as plasticizers in industrial processes to make plastics flexible.1 They are frequently detected in food packaging materials, car parts, clothes, children's toys, and many other products. Since the chemical structure of phthalates is very similar to that of human hormones, continuous exposure could cause endocrine disruption.2 The direct injection of phthalates dissolved in various media into analytical instruments might be difficult because the sample matrix is not simple but rather complex. To avoid compounds that interfere with detection, the sample requires proper pretreatment to increase the concentration of analytes prior to analysis.3 Extraction of phthalates from the media through adsorption on adsorbents is considered a significant technique.

Conventional methods for phthalate extraction include liquid–liquid extraction (LLE), solid-phase extraction (SPE), solid-phase microextraction (SPME), and Soxhlet extraction.3–6 Although these methods have been widely used, they are often labor-intensive and time-consuming and require large volumes of organic solvents. Furthermore, these extraction techniques might not effectively isolate volatile and semi-volatile phthalates from matrices such as air; therefore, alternative approaches are required for accurate quantification. An innovative approach within this framework involves the use of adsorbent materials for headspace extraction. These adsorbents can be packed into a needle or coated on a wire to offer a versatile and efficient means of phthalate extraction.

In-needle microextraction (INME) is a method of coating the interior of a needle with an adsorbent and analyzing target compounds using a syringe. The adsorbent coating is performed via the sol–gel, electrochemical, and dip-coating methods.7–15 The adsorbent is in the semi-fluid kneaded sol state initially but forms a gel state with reduced fluidity after thermal curing9–12,15 or electrochemical deposition on a wire.7,8,13,14 In addition, a microbore tunnel was generated in the middle of the needle, further enhancing the adsorption efficiency. By coating the interior of the needle with the adsorbent, the probability of the adsorbent being damaged by exposure was reduced. The thermal desorption of the adsorbed phthalates allows for their direct introduction into detection systems, such as GC-MS, enhancing both the sensitivity and precision of the analysis. This method also minimizes the risk of sample loss or contamination, which can be problematic in liquid-phase extraction methods.

Recent studies have demonstrated the effectiveness of polymer-based adsorbents, activated carbon, and molecularly imprinted polymers (MIPs) in extraction.14–16 In particular, activated hydrochar (AHC), a porous carbonaceous material derived from biomass through hydrothermal carbonization (HTC) and subsequent activation, has been selected for extraction.16–22 Its surface chemistry, characterized by abundant oxygen-containing functional groups, provides enhanced interaction with semi-volatile organic compounds. The mesoporous structure of activated hydrochar facilitates the adsorption of larger organic molecules to make it particularly effective for phthalates with relatively high molecular weights.17 They exhibit high binding affinity and selectivity for phthalates, particularly suitable for headspace extraction applications.5,9 Additionally, the low detection limits achieved with adsorbent-based techniques enable the analysis of trace levels of phthalates in complex environmental and biological matrices.7–9,23

In this study, an INME needle was fabricated by coating a mixture of polydimethyl siloxane (PDMS) and AHC via sol–gel polymerization to synthesize PDMS/AHC. Adsorbent synthesis conditions, including the HTC reaction and activation reaction, and analysis conditions of the headspace (HS)-INME-PDMS/AHC method were optimized by Box–Behnken design to efficiently extract phthalates and validated through limit of detection (LOD), limit of quantification (LOQ), linearity, and recovery.

2. Experimental section

2.1. Chemicals and reagents

Sulfuric acid (95.0%), ethyl alcohol (94.5%), potassium hydroxide (95.0%), melamine (99.0%), and hydrochloric acid (35.0–37.0%) were used to generate AHC. Methylene blue, sodium thiosulfate pentahydrate, sodium carbonate, iodine, potassium iodide, potassium iodate, and soluble starch were used to measure the methylene blue and iodine numbers. A solution of 0.05 M sodium thiosulfate pentahydrate was prepared by mixing sodium thiosulfate pentahydrate and sodium carbonate, while a triiodide ion solution was prepared by combining iodine and potassium iodide. These solutions were standardized before use.

PDMS (Sylgard 184A) and PDMS curing agent (Sylgard 184B) were used to fabricate the INME adsorbent. A Hamilton 9022 needle, 290 μm O.D. nichrome wire, and 1 mL disposable syringe were used to coat the adsorbent in the needle.

Phthalate standard solutions, including dimethyl phthalate (DMP), diethyl phthalate (DEP), dibutyl phthalate (DBP), and di(2-ethylhexyl)phthalate (DEHP), were prepared with methanol at a concentration of 1000 mg L−1 and diluted to different concentrations.

2.2. Optimization of experimental conditions using design of experiments

A two-step preparation strategy was adopted to enhance both the structural control and surface functionality of the adsorbent. The initial HTC step allows for the formation of a stable, oxygen-rich carbon framework at a relatively low temperature, which is well-suited for further activation. Subsequent chemical activation using KOH and melamine introduces micropores into the mesoporous structure and incorporates nitrogen functionalities. This dual modification significantly improves the surface area and adsorption performance. Furthermore, such hierarchical porous carbons are widely explored as adsorbents and as electrode materials for supercapacitors owing to their high conductivity and electrochemical stability.24,32

The design of the experiment (DOE) method was used to optimize the experimental conditions efficiently. Optimization processes were conducted using the Box–Behnken design (BBD) in Minitab 19 (Minitab Inc., State College, Pennsylvania, USA). It is the number of experimental runs while still providing sufficient information for estimating the main effects and interactions.20–22 The optimization process was performed for each condition of the HTC reactions, activation reaction, and INME-GC/MS analysis. For each optimization process, three factors (X1, X2, and X3) were set in the range of −1, 0, and 1 levels for each factor. Three levels of the three factors generate a total of 17 conditions to be investigated, as shown in Table S1.† The specific values for each experimental condition generated by the DOE are illustrated in Table 1.

Table 1 Experimental design factors and levels of the chosen variables for the optimization of hydrothermal carbonization conditions
Factors Level
Low (−1) Middle (0) High (+1)
(A) Hydrothermal carbonization conditions
Reaction temperature (X1, °C) 200 210 220
Reaction time (X2, h) 6 12 18
Amount of sample (X3, g) 2.5 5.0 7.5
[thin space (1/6-em)]
(B) Activation conditions
Reaction temperature (X1, °C) 600 700 800
Melamine ratio (X2, unitless) 1[thin space (1/6-em)]:[thin space (1/6-em)]1 1[thin space (1/6-em)]:[thin space (1/6-em)]2 1[thin space (1/6-em)]:[thin space (1/6-em)]4
KOH ratio (X3, unitless) 1[thin space (1/6-em)]:[thin space (1/6-em)]1 1[thin space (1/6-em)]:[thin space (1/6-em)]3 1[thin space (1/6-em)]:[thin space (1/6-em)]5
[thin space (1/6-em)]
(C) HS-INME-GC/MS analysis conditions
Saturation temperature (X1, °C) 30 50 70
Adsorption time (X2, min) 10 30 50
Desorption time (X3, min) 1 3 5

2.2.1. Experimental condition 1: reaction conditions of HTC. The factors for the HTC reaction conditions include the reaction temperature, reaction time, and mass of the raw material. The HTC reaction temperature was set between 200 °C and 220 °C, the reaction time between 6 h and 18 h, and the mass of the withered flower between 2.5 g and 7.5 g. For adsorbent synthesis, the reaction temperature (X1), reaction time (X2), and amount of raw material (X3) were investigated (Table 1, (A)). Withered rose parts (flowers, leaves, and stems) were obtained from a local flower shop near the university campus. Each part was cut into 1 cm pieces and dried at 80 °C for 24 h. These dried pieces were sealed in a container within a desiccator. All the collected rose materials were processed as a single batch and used consistently throughout the study to eliminate batch-to-batch variations and ensure the reproducibility of the adsorbent properties. In a 200 mL stainless steel reactor, dried roses (X3) at each level (−1, 0, and 1 for 2.5 g, 5.0 g, and 7.5 g, respectively) were combined with a 5% H2SO4 solution to achieve an 80% water content. After the reaction, the resulting hydrochar (HC) was washed with distilled water and dried at 80 °C for 12 h.

To optimize the HTC reaction conditions, the methylene blue number of HC was measured for comparison. After the methylene blue standard solution (1–6 ppm) was prepared, the absorbance was measured at 665 nm using a UV-vis spectrometer (UV-2600, Shimadzu, Kyoto, Japan) to establish a calibration curve.22 HC (0.01 g) generated from each HTC condition was placed in 10 mL of the methylene blue solution (10–1000 ppm) and vortexed. Then, it was incubated in a shaking incubator (25 °C, 180 rpm). After 24 h of shaking, the reaction solution was centrifuged at 4000 rpm for 1 min and filtered with a PTFE syringe filter (0.2 μm).

2.2.2. Experimental condition 2: reaction conditions of activation. For activation conditions, three factors including activation temperature (X1), melamine rations (X2), and potassium hydroxide ratios (X3) were investigated. Melamine addition was expected to have more meso-sized pores on the adsorbent.23,24 For the activation process, HC, melamine, and potassium hydroxide were mixed in various ratios and heated at the reaction temperature for 1 h. Then, the AHC was filtered, washed with a 1 M HCl solution, and dried at 105 °C for 12 h.

The iodine number of AHC was determined to optimize the thermochemical activation conditions described in the previous study.22 To calculate the iodine number, 0.3 g of the AHC and 0.05 M triiodide solution (30 mL) were mixed in a centrifuge tube (50 mL) at 180 rpm for 30 min in a 25 °C shaking incubator. Following the reaction, centrifugation was performed for 2 min at 3500 rpm. After filtration, 10 mL of the filtrate was placed in a conical flask, and 0.05 M sodium thiosulfate pentahydrate solution was added dropwise until the solution changed from deep brown to light yellow. The addition of starch solution (3 mL) turned the solution from pale yellow to dark blue. The titration was continued until the dark blue solution became colourless. The titre value was used to calculate the iodine number.

2.2.3. Experimental condition 3: HS-INME-GC/MS analysis condition. The optimization of HS-INME-GC/MS using an INME needle fabricated under optimal conditions was conducted by adjusting the saturation temperature (X1), adsorption time (X2), and desorption time (X3). The INME needle coated inside with PDMS/AHC (100[thin space (1/6-em)]:[thin space (1/6-em)]1 ratio) was prepared similarly to previous reports.13,14 The INME needle with PDMS/AHC was formulated by attaching a disposable syringe to a Hamilton 9002 needle and placing it in the PDMS and AHC mixture solution to extract it by 0.01 mL. Then, the nichrome wire was inserted into the needle where the mixture solution was filled, placed on the septum and cured at 100 °C for 15 min. After gelation was almost complete, the wire was removed so that PDMS/AHC was coated inside the needle. Then, it was kept in a desiccator until needed. The INME needle was conditioned to eliminate impurities by heating to 230 °C for 30 min prior to extraction (Fig. 1).
image file: d5ay00221d-f1.tif
Fig. 1 The fabrication of the INME needle coated on the interior with PDMS/AHC.

2.3. Headspace in-needle microextraction (HS-INME) and GC/MS phthalate analysis

HS-INME method using adsorbent-coated needles was similarly performed as in previous studies.11,12 The headspace of the vial was saturated with the vapor of aqueous phthalates by heating an aliquot of the phthalate mixture in a 70 mL vial. It was then placed in a sand bath and maintained at 50 °C for 60 min. The pre-purified INME needle was firmly attached to an air-tight syringe and reciprocating pump at a speed of 6 cycles min−1. Then, the INME needle was inserted into the headspace for 50 min to extract phthalate vapors. After extraction, the INME needle was immediately transferred to the GC/MS injector for the thermal desorption of the phthalates.

The HP-5MS column (30 m × 0.25 mm × 0.25 μm, (5%-phenyl)-methylpolysiloxane) column was used for separation. The oven temperature was set by heating from 60 °C (2 min) to 210 °C at 15 °C min−1 and 210 °C to 290 °C (8 min) at 5 °C min−1, and the splitless injection mode was used. This analysis condition was determined by referring to the conditions in previous studies.25

To ensure that no background contamination affected the phthalate analysis, all experimental materials, such as latex gloves, micropipette tips, vials, vial caps, and septa, were subjected to blank testing. These materials were exposed to 50 °C and 100 °C under the same HS-GC/MS conditions, and the four target phthalates (DMP, DEP, DBP, and DEHP) were not detected. This confirms that no significant cross-contamination occurred in the laboratory materials used in this study.

2.4. Characterization of HC, AHC, and PDMS/AHC

FT-IR spectroscopy (Spectrum 100, PerkinElmer, Waltham, USA) was used to measure the functional groups of the raw material (RM, withered rose), HC, and AHC. Surface area analysis was conducted using BET (3flex, Micromeritics, USA) N2 adsorption/desorption isotherms. The Barrett–Joyner–Halenda (BJH) method was applied to calculate the pore volume and diameter. FE-SEM (JSM-6700F, JEOL Ltd, Tokyo, Japan) was used to observe the morphologies of RM, HC, and AHC with platinum coating for analysis. TGA (DTG-60H, Shimadzu, Japan) was utilized to assess thermal stability. The procedure involved transferring approximately 10 mg of the sample to an aluminum oxide crucible and heating it from 20 °C to 800 °C at a rate of 10 °C min−1 under a nitrogen atmosphere.14

2.5. Validation of the method

The optimized HS-INME method was applied to an actual sample for the recovery test of the four phthalates. Phthalate solutions at various concentrations were spiked on the stickers as children's toys since phthalate detection was reported.26 Intra-assay and inter-assay were performed to confirm reproducibility. LOD, LOQ, working range, sensitivity, and reproducibility were calculated for the validation process.27

3. Results and discussion

3.1. Physicochemical characterization of HC, AHC, and PDMS/AHC

When HC was produced under optimized conditions, the yield was 43.63 (±7.19)%, while the yield of bioliquid, a liquid product generated by the HTC process, was 63.6 (±9.8)%, and the average volume of the gas was 668.9 (±157.8) mL. Physicochemical characterization was performed using FT-IR, FE-SEM, BET, and TGA to confirm the characteristics of RM, HC, AHC, and PDMS/AHC and the reaction completion.

The differences in the functional groups of RM, HC, and AHC were confirmed by FT-IR (Fig. S1†). As the HTC process at a temperature of 200 °C or higher forms an aromatic structure,28 the –CH3 (2924 cm−1) stretching in HC was reduced compared to that in RM. In addition, the –OH stretching peak (3401 cm−1) of HC was reduced compared to that of RM because water loss occurred owing to the high temperature of the HTC process.29 The peaks in the range of 1400–900 cm−1 of RM and HC are attributed to the stretching of the C[double bond, length as m-dash]C bond that appeared owing to lignin degradation at 200–700 °C.30 Since activation was carried out at a high temperature (600–800 °C), most of the functional groups of HC disappeared. It was confirmed that the n-doping process during the activation was successfully performed with the N–O stretching peak at 1559 cm−1.

The FE-SEM image observed the surface of RM, which changed during the HTC and activation processes (Fig. 2). RM initially exhibited a relatively smooth surface without any observable pore structures (Fig. 2(A)). Conversely, HC displayed a few similarly sized pores, resulting in a rough surface texture (Fig. 2(B)). For AHC, featuring interconnected pores of varying sizes, it led to its notably extensive surface area (Fig. 2(C)). The pores exhibited an irregular and connected configuration, prompting the calculation of pore volume for each diameter using the desorption branch.31 Pores were categorized into three types (macro, meso, and micro) based on their diameter, with those exceeding 50 nm classified as macropores, pores ranging from 2 nm to 50 nm categorized as mesopores, and those under 2 nm designated as micropores.32


image file: d5ay00221d-f2.tif
Fig. 2 SEM images (×3500) of (A) raw material (RM), (B) hydrochar (HC), and (C) activated hydrochar (AHC).

The BET measurements determined the specific surface area (m2 g−1), pore volume (cm3 g−1), and pore diameter (Å) for RM, HC, and AHC. Analyzing N2 adsorption–desorption isotherms shown in Fig. S2-(A),† the pore volume of AHC significantly differs from those of RM and HC. The pore shape corresponds to H4 in all three materials, featuring a narrow slit-like pore. RM exhibited a pore distribution of 77.24% macro, 20.41% meso, and 2.35% micropores, while HC exhibited 80.72% macro and 19.28% mesopores. The pore size distribution depicted in Fig. S2-(B)† reveals that micropores were not present in HC likely owing to structural changes during the HTC process. AHC showed a unique pattern with 22.08% macro, 57.56% meso, and 20.36% micropores.

Specific surface areas were 1.112 m2 g−1 for RM, 6.696 m2 g−1 for HC, and 1082 m2 g−1 for AHC, representing a six-fold increase for HC and a remarkable 1000-fold increase compared to RM. As illustrated in Fig. S2-(B),† AHC's micropore volume significantly surpassed those of RM and HC.

The addition of AHC to PDMS enhances thermal stability. The measurements conducted by TGA showed that PDMS retained 99% of its mass up to 242 °C, with decomposition beginning at 430 °C, similar to that of a previous study.33 In contrast, PDMS/AHC maintained about 99% of its mass up to 295 °C with decay starting at 449 °C (Fig. S3†). This indicates that PDMS/AHC possesses superior thermal stability compared to PDMS without additives. Notably, the decomposition temperature of PDMS/AHC is higher than the GC inlet temperature of 270 °C. Its enhanced thermal stability makes it a suitable adsorbent for phthalate analysis using GC/MS. In addition, the extraction efficiency of the four phthalates using HS-INME-PDMS/AHC was more than ten times higher than that of HS-INME-PDMS (data not shown).

3.2. Optimization of experimental conditions using the DOE method

3.2.1. Hydrothermal carbonization (HTC) and activation condition. To determine the optimized HTC reaction condition, the methylene blue number was chosen because its molecular size is similar to mesopores, ensuring effective adsorption.24 Methylene blue number and iodine number were measured to assess mesoporosity and microporosity, respectively, to guide the selection of optimized HTC and activation conditions. These values reflect the general structural characteristics of the adsorbent and were not intended to predict phthalate extraction performance directly. Phthalate-specific extraction results were obtained separately using the HS-INME-GC/MS system. The methylene blue number was calculated using the following equation:
 
Methylene blue number (mg g−1) = (C0 − Ce) × V/M, (1)
where C0 is the initial concentration of the methylene blue solution (ppm), Ce is the equilibrium concentration (ppm), V is the solution volume (L), and M is the HC mass (g) used in the experiment.

Fig. S4-(A)† displays the optimization results of HTC reaction conditions according to DOE, showing correlation among three factors: reaction temperature (X1), reaction time (X2), and amount of sample (X3). As a result, the sample amount was significant with a p-value of 0.055, compared with the reaction temperature (p-value of 0.943) and time (p-value of 0.788) among the three factors.34 Consequently, additional experiments were conducted to fine-tune the reaction temperature and time. It was concluded that a reaction temperature of 210 °C and a reaction time of 18 h with 7.5 g of the sample were ideal. These conditions were selected for the optimal HTC reaction, and prediction model suitability was confirmed, as shown in Table 2, where the relative standard deviation was under 10% between the calculated and actual methylene blue numbers.

Table 2 Methylene blue number and iodine number prediction formula, calculated/actual value, and relative error obtained through the DOE in the optimization of hydrothermal carbonization and activation conditions
  Prediction formula
Calculated value Actual value Relative error (%)
a MBN: methylene blue number.b IN: iodine number.
MBNa 4.522 − 0.025X1 − 0.092X2 + 0.759X3 − 0.223X12 + 0.808X22 + 0.419X32 + 0.132X1X2 − 0.367X1X3 + 0.363X2X3
7.00 7.86 10.9
INb 2077.14 + 4.53X1 − 4.30X2 + 7.30X3 − 1.6X12 − 7.6X22 − 7.2X32 + 1.7X1X2 − 16.4X1X3 − 16.7X2X3
2102.61 2098.90 0.17

The DOE method was also used to optimize the activation conditions. Comparison factors were set as reaction temperature (°C), melamine ratio, and KOH ratio (Table 1, (B)). For optimization, a 3-level BBD model was used, and 17 experiments were performed. IN was selected as the optimization criterion for the activation condition, in which the size of the iodine molecule was small enough to enter the micropore.24 The IN was calculated using the following equation:24

 
Iodine number (mg g−1) = (10B × f × 2.69)/S, (2)
where B is the volume of the 0.05 M Na2S2O3·5H2O solution used for titration (mL), f is the normality of the 0.05 M Na2S2O3·5H2O solution, and S is the amount of AHC (g) used in the experiment.

The optimization results are depicted with the correlation among the three factors shown in the response surface analysis diagram, as depicted in Fig. S4-(B).† However, it was observed that the p-values for the reaction temperature, melamine ratio, and potassium hydroxide ratio were all higher than 0.05, indicating that they were not significant factors.34 Overall, experiments conducted with a reaction temperature of 600 °C, a melamine ratio of 1[thin space (1/6-em)]:[thin space (1/6-em)]1, and a potassium hydroxide ratio of 1[thin space (1/6-em)]:[thin space (1/6-em)]5 optimized by applying the DOE method yielded both high IN values and a high yield. Table 2 shows the model formula for predicting methylene blue numbers and iodine numbers under conditions optimized by applying the DOE method. It was confirmed that the optimization model for activation conditions was appropriate, as the relative standard deviation between the calculated value of the optimal IN condition and the actual experimental IN value was less than 10%. A yield of AHC, 33.50 (±4.66)%, was achieved under the optimal activation conditions.

3.2.2. Optimization of the HS-INME condition. A total of 17 experiments were also conducted, and the peak area of four phthalates was analyzed using HS-INME with PDMS/AHC adsorbent, followed by GC/MS. The response surface design analysis results in Minitab 19 were used to optimize the analysis conditions for maximizing the peak area. The correlation of the three factors including saturation temperature (X1, °C), adsorption time (X2, min), and desorption time (X3, min) was visualized in the diagram of response surface analysis (Fig. 3). It was observed that the peak areas of DMP, DEP, and DBP were primarily influenced by adsorption time (min) during the DOE optimization, while the peak area of DEHP was more affected by saturation temperature than other factors.
image file: d5ay00221d-f3.tif
Fig. 3 Response surface diagram of four phthalates and correlation among the saturation, adsorption, and desorption times.

Table 3 shows the peak area prediction formula, calculated/actual value, and relative error of each phthalate obtained through DOE. When the analysis was performed under optimal conditions, the relative standard deviations of the actual peak area and the calculated peak area were all less than 10% for the four target phthalate compounds, confirming the accuracy of the model.

Table 3 Peak area prediction formula, calculated/actual value, and relative error of each phthalate obtained through the DOE in the optimization of HS-INME-PDMS/AHC analysis conditions
Compound Prediction formula
Predicted peak area Actual peak area Relative error (%)
Dimethyl phthalate (DMP) 2700323 + 133534X1 + 540499X2 + 78483X3 − 1706525X12 − 757914X22 − 659348X32 − 47498X1X2 − 214889X1X3 + 465282X2X3
2367905 2588510 6.29
Diethyl phthalate (DEP) 2769877 + 121666X1 + 1512946X2 + 710648X3 − 2466573X12 + 187287X22 + 201258X32 − 233028X1X2 − 232912X1X3 + 1746353X2X3
7140380 6484835 6.80
Dibutyl phthalate (DBP) 5058625 + 377673X1 + 256243X2 + 1304353X3 − 4625734X12 + 489831X22 + 1390808X32 − 523865X1X2 − 1108476X1X3 + 3380598X2X3
11946041 16607191 5.63
Di(2-ethylhexyl) phthalate (DEHP) 964419 + 449676X1 − 102960X2 + 84767X3 − 707542X12 + 219163X22 + 533121X32 − 273365X1X2 − 257573X1X3 − 859463X2X3
2560182 2779776 5.82

To provide a clearer interpretation of the influence of each variable on extraction efficiency, we performed an analysis of variance (ANOVA) using the results obtained from the DOE. ESI Table S4† shows the detailed statistical outcomes. Among the three variables, adsorption time (X2) had a statistically significant effect on DMP (p = 0.034), DEP (p = 0.037), and DBP (p = 0.084). These results indicate that a sufficient duration for interaction between the analyte vapor and the adsorbent surface is essential for efficient extraction. In the case of DEHP, saturation temperature (X1) showed the strongest influence (p = 0.016). This finding reflects the lower volatility and larger molecular size of DEHP, which requires an adequate temperature to ensure sufficient vapor pressure for transfer into the needle. The results also revealed that several quadratic terms, such as X12 and X22, had statistically significant effects on extraction performance. This observation suggests the existence of non-linear relationships between these variables and peak area. When temperature or adsorption time exceeds a certain optimal point, extraction efficiency tends to decline. In addition, the interaction between saturation temperature and desorption time (X1X3) was particularly important for DEHP (p = 0.004). This effect likely resulted from the temperature-enhanced desorption behavior that influenced analyte release during GC injection.

The R2 values ranged from 0.7033 to 0.9008 across the four phthalates, indicating that the regression models effectively described the experimental outcomes. The highest R2 was observed for DEHP (0.8822). These results confirm that the response surface models are statistically valid and suitable for identifying optimal extraction conditions. Based on these findings, the optimal conditions were determined as 50 °C saturation temperature, 20 minutes of adsorption time, and 5 minutes of desorption time for the HS-INME-GC/MS method.

3.3. Validation of analytical method

The values of LOD and LOQ were calculated according to ISO standards27 using a calibration curve. Validation results for HS-INME-PDMS/AHC are summarized in Table 4. A calibration curve with 6 points (n = 3) provided the regression equation, linearity (0.99 or higher for all target compounds), LOD, LOQ, and dynamic range. Among the compounds, DBP showed the highest sensitivity, followed by DEP, DMP, and DEHP. Calculated LOD and LOQ ranged from 19.8 ng to 1930 ng and 60.1 ng to 5480 ng, respectively, with a dynamic range of 19.8–2500 ng. Recovery was determined using eqn (3):
 
Recovery (%) = 100(Ac + AI)/As, (3)
where Ac represents the amount of each phthalate adsorbed on stickers (g), AI is the amount adsorbed on INME-PDMS/AHC (g), and As is the amount spiked onto the stickers (g). The recovery percentages for each phthalate fell within the 90–110% range (Table S2†). The method's reproducibility was evaluated for the four phthalates in repeated measurements (n = 5), with relative standard deviation values for both intra-assay and inter-assay being less than 10% (Table S2†).
Table 4 Validation data of HS-INME, followed by GC/MS, regression equation, linearity (r2), the limit of detection (LOD), the limit of quantification (LOQ), and dynamic range
Compound Regression equation r2a LODb (ng) LOQc (ng) Dynamic range (ng)
a r2: coefficient of determination.b LOD: limit of detection.c LOQ: limit of quantification.
Dimethyl phthalate (DMP) y = 2.55 × 106x + 1.17 × 106 0.992 1.98 × 101 6.01 × 101 1.98 × 101 to 2.50 × 103
Diethyl phthalate (DEP) y = 4.64 × 106x + 1.17 × 106 0.992 9.42 × 101 2.85 × 102 9.42 × 101 to 2.50 × 103
Dibutyl phthalate (DBP) y = 1.16 × 107x + 4.78 × 106 0.993 1.60 × 103 4.85 × 103 3.21 × 101 to 2.50 × 103
Di(2-ethylhexyl) phthalate (DEHP) y = 9.71 × 105x + 7.12 × 105 0.999 1.93 × 103 5.84 × 103 2.92 × 102 to 2.50 × 103

In addition, the reusability of the PDMS/AHC-coated needle was evaluated by consecutively applying the same needle in five repeated HS-INME-GC/MS runs. The relative standard deviations (RSDs) of the peak areas for DMP, DEP, DBP, and DEHP were 3.86%, 4.12%, 1.21%, and 9.01%, respectively, confirming the consistent performance and reusability of the coated adsorbent.

3.4. Comparison of extraction efficiency

The static and dynamic INME methods were compared to compute the enrichment factor.17 The dynamic method differs from the static method because it uses a pump to move the air saturated with phthalate standards in the needle. The EF was determined based on the peak area of the four phthalates using the following equation:
 
Enrichment factor = A1/A0, (4)
where A1 is the peak area obtained by applying the dynamic method (unitless) and A0 is the peak area obtained by applying the static method (unitless). The calculated EF values are summarized in Table S3.†

EF values ranged from 1.47 to 1.96, and as the molecular weight increased, the EF values tended to decrease. It is expected that the vapor pressure of the four phthalates decreases as the molecular weight increases.25 The use of overly polar or non-polar adsorbents can hinder the simultaneous analysis of phthalates with different polarities. Therefore, phthalate analysis benefits from a moderately polar adsorbent.35 SPME, a common method, requires changing expensive SPME fibers for each phthalate type, while PDMS/AHC offers cost-effective flexibility, allowing for easy adjustment according to phthalate types. Comparing INME-PDMS/AHC to SPME-PDMS/DVB under optimized INME conditions, PDMS/AHC displayed lower extraction efficiency (data not shown) for DMP, DEP, and DBP, but increased DEHP extraction efficiency by approximately 28%. The PDMS/AHC adsorbent might be effective for semi-volatile organic compounds, such as DEHP, owing to its large surface area and porosity, hydrophobic interaction, high chemical stability, and affinity on the surface via sorption.36,37 The advantage of PDMS/AHC lies in its ability to extract DEHP even at lower temperatures, even at 50 °C. This difference in extraction performance can be explained by the physical configuration of the extraction devices. In the SPME method, the sorbent is located externally and directly faces the headspace environment, which results in more efficient contact with volatile analytes. In contrast, the INME system contains the sorbent on the inner surface of a narrow needle, which may limit the diffusion of smaller or highly volatile compounds. However, this enclosed configuration minimizes external contamination, improves structural stability, and allows for repeated use without significant loss of efficiency.7–15

To further evaluate the analytical performance of the developed HS-INME-GC/MS method, key validation parameters were compared with those reported in previous studies that used HS-SPME-GC/MS for phthalate analysis.38,39 As shown in Table 5, the LOD, recovery, and precision of our method were comparable to or better than those of established SPME-based techniques. Notably, the INME method offered competitive sensitivity and high reproducibility while enabling solvent-free extraction and potential for repeated use.

Table 5 Comparison of validation parameters for the developed HS-INME method and previously reported HS-SPME methods for phthalate analysis
Parameter This study (HS-INME) Cao et al., 2008 (HS-SPME)38 Holadová et al., 2007 (HS-SPME)39
LOD (μg L−1) 0.003–0.085 0.003–0.085 10–100
Recovery (%) 92.83–98.46 84–103 79–99
Intra-assay RSD (%) 1.21–9.01 5–10 5–15
Sorbent type PDMS/AHC (INME needle) PDMS/DVB (SPME fiber) PDMS (SPME fiber)

4. Conclusions

In this study, an adsorbent was made using biomass, withered roses, through hydrothermal carbonization to analyze four phthalates, including DMP, DBP, DEP, and DEHP, using the PDMS/AHC coated inside the needle. The withered roses were converted into activated porous carbon through the HTC reaction and activation process and mixed with PDMS to be coated on the inner wall of the INME needle using the sol–gel method. Gaseous phthalate saturated in the headspace of the vial was circulated inside the INME needle using a homemade pump, extracted from the adsorbent, and then desorbed directly to the inlet of GC/MS for analysis without additional solvent applied. Manual and DOE methods were used to determine the optimal conditions for HTC, activation, and HS-INME-GC/MS analysis. The optimized conditions were determined as a saturation temperature of 50 °C, saturation time of 60 min, adsorption time of 50 min, and desorption time of 5 min. The proposed method was validated by generating the calibration curve, linearity (r2), LOD, LOQ, and dynamic range through HS-INME-PDMS/AHC, followed by GC/MS. This confirmed that INME-PDMS/AHC is easy to fabricate, convenient for storage, and economically advantageous, and can analyze phthalates at a relatively low temperature.

Data availability

All data generated or analyzed during this study are included in this published article and its ESI file.†

Author contributions

Jooyoung Kim: methodology, investigation, data curation, formal analysis, visualization, validation, writing-original draft; Sunyoung Bae: conceptualization, supervision, validation, visualization, writing-review and editing, project administration, funding acquisition.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This research work has been supported by a grant from Seoul Women's University (2025-0203).

References

  1. A. Zhao, L. Wang, X. Pang and F. Liu, Environ. Res., 2022, 204, 112041 CrossRef CAS PubMed .
  2. Y. Fan, H. Chen, H. Liu, F. Wang, S. Ma, A. Latipa, S. Wang and C. Wang, Anal. Methods, 2017, 9, 370–380 RSC .
  3. F. Freitas, M. J. Cabrita and M. G. da Silva, Molecules, 2023, 28, 7628 CrossRef CAS PubMed .
  4. S. Bai, X. Wang, F. Hu, T. Wang, P. Cheng and Z. Zhou, Anal. Methods, 2014, 6, 7361–7366 RSC .
  5. R. Perestrelo, C. L. Silva, M. Algarra and J. S. Câmara, J. Agric. Food Chem., 2020, 68, 8431–8437 CrossRef CAS PubMed .
  6. G. Cinelli, P. Avino, I. Notardonato, A. Centola and M. V. Russo, Food Chem., 2014, 146, 181–187 CrossRef CAS .
  7. Y. Hwang, Y. Lee, S. Ahn and S. Bae, Anal. Sci. Technol., 2020, 33, 76–85 CAS .
  8. S. Lee, J.-H. Yoon, S. Bae and D.-S. Lee, Food Anal. Methods, 2018, 11, 2767–2777 CrossRef .
  9. H. H. Son, S. Bae and D. S. Lee, Anal. Chim. Acta, 2012, 751, 86–93 CrossRef CAS PubMed .
  10. S. Ahn and S. Bae, Molecules, 2023, 28, 3517 CrossRef CAS PubMed .
  11. H. L. Jeon, H. H. Son, S. Bae and D. S. Lee, Bull. Korean Chem. Soc., 2015, 36, 2730–2739 CrossRef CAS .
  12. Y. Bang, Y. Hwang, S. Lee, S. Park and S. Bae, J. Sep. Sci., 2017, 40, 3839–3847 CrossRef CAS PubMed .
  13. S. Kim and S. Bae, Molecules, 2022, 27, 4795 CrossRef CAS PubMed .
  14. S. Kim, S. Bae and D. S. Lee, Talanta, 2022, 245, 123463 CrossRef CAS PubMed .
  15. Y. Lee, H. Jung, S. Park and S. Bae, Anal. Sci. Technol., 2024, 37, 295–305 Search PubMed .
  16. Y. Bouhamidi, F. Kaouah, L. Nouri, S. Boumaza and M. Trari, Int. J. Environ. Sci. Technol., 2017, 14, 271–284 CrossRef CAS .
  17. K. S. Ro, J. R. Flora, S. Bae, J. A. Libra, N. D. Berge, Á. Álvarez-Murillo and L. Li, ACS Sustain. Chem. Eng., 2017, 5, 7317–7324 CrossRef CAS .
  18. A. B. Fuentes, R. L. S. Canevesi, P. Gadonneix, S. Mathieu, A. Celzard and V. Fierro, Ind. Crops Prod., 2020, 155, 112740 CrossRef .
  19. S. Masoumi and A. K. Dalai, J. Clean. Prod., 2020, 263, 121427 CrossRef CAS .
  20. S. Román, J. Libra, N. Berge, E. Sabio, K. Ro, L. Li, B. Ledesma, A. Álvarez and S. Bae, Energies, 2018, 11, 216 CrossRef .
  21. D. Jaruwat, P. Udomsap, N. Chollacoop, M. Fuji and A. Eiad-ua, AIP Conf. Proc., 2018, 2010, 020016 CrossRef .
  22. T. P. de Araújo, H. B. Quesada, R. Bergamasco, D. T. Vareschini and M. A. S. D. de Barros, Bioresour. Technol., 2020, 310, 123399 CrossRef PubMed .
  23. M. Choi, S.-E. Choi, S. Han and S. Bae, Anal. Sci. Technol., 2015, 28, 40–46 CrossRef .
  24. C. A. Nunes and M. C. Guerreiro, Quim. Nova, 2011, 34, 472–476 CrossRef CAS .
  25. D. Armada, M. Celeiro, T. Dagnac and M. Llompart, J. Chromatogr. A, 2022, 1668, 462911 CrossRef CAS PubMed .
  26. M. Al-Natsheh, M. Alawi, M. Fayyad and I. Tarawneh, J. Chromatogr. B:Anal. Technol. Biomed. Life Sci., 2015, 985, 103–109 CrossRef CAS .
  27. T. Wenzl, J. Haedrich, A. Schaechtele, R. Piotr, J. Stroka, G. Eppe and G. Scholl, Guidance Document on the Estimation of LOD and LOQ for Measurements in the Field of Contamination in Feed and Food, Publications Office of the European Union, Luxembourg, 2016 Search PubMed .
  28. N. A. de Figueredo, L. M. da Costa, L. C. A. Melo, E. A. Siebeneichlerd and J. Tronto, Rev. Cienc. Agron., 2017, 48, 395–403 Search PubMed .
  29. L. Suárez-Hernández and R. Barrera-Zapata, Rev. Fac. Ing., 2017, 26, 123–130 CrossRef .
  30. M. Jouiad, N. Al-Nofeli, N. Khalifa, F. Benyettou and L. F. Yousef, J. Anal. Appl. Pyrolysis, 2015, 111, 183–190 CrossRef CAS .
  31. G. Leofanti, M. Padovan, G. Tozzola and B. Venturelli, Catal. Today, 1998, 41, 207–219 CrossRef CAS .
  32. P. Jia, H. Tan, K. Liu and W. Gao, Appl. Sci., 2018, 8, 1903 CrossRef .
  33. T. S. Radhakrishnan, J. Appl. Polym. Sci., 2006, 99, 2679–2686 CrossRef CAS .
  34. H. E. Park and K. H. Row, Appl. Chem. Eng., 2013, 24, 299–304 CAS .
  35. D. M. Lunderberg, K. Kristensen, Y. Liu, P. K. Misztal, Y. Tian, C. Arata, R. Wernis, N. Kreisberg, W. W. Nazaroff and A. H. Goldstein, Environ. Sci. Technol., 2019, 53, 7337–7346 CrossRef CAS PubMed .
  36. D. Widel, K. Jedynak and Z. Witkiewicz, Desalin. Water Treat., 2021, 232, 91–104 CrossRef CAS .
  37. C. M. Ma, T. J. Yu, C. H. Yang and G. B. Hong, Ecotoxicol. Environ. Saf., 2021, 228, 112972 CrossRef CAS PubMed .
  38. X. L. Cao, J. Chromatogr. A, 2008, 1178, 231–238 CrossRef CAS PubMed .
  39. K. Holadová, et al., Anal. Chim. Acta, 2007, 582, 24–33 CrossRef PubMed .

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d5ay00221d
This journal is © The Royal Society of Chemistry 2025
Click here to see how this site uses Cookies. View our privacy policy here.