Determination of Urinary Hydroxyl PAHs Using Graphene Oxide@Diatomite Based Solid-Phase Extraction and High-Performance Liquid Chromatography

A diatomite supported graphene oxide composite (GO@Dt–NH2) was fabricated and explored as a solid-phase extraction adsorbent coupled with high performance liquid chromatography to determine the trace hydroxyl polycyclic aromatic hydrocarbons (2-hydroxy-naphthalene, 2-hydroxy-fluorene, 1-hydroxy-phenanthrene, and 1-hydroxy-pyrene) in urine samples. The fabricated composites were characterized by X-ray powder diffractometry and scanning electron microscopy. GO@Dt–NH2 offered enhanced adsorption affinity towards the analytes compared with the bare diatomite. The amount of graphene oxide and the factors affecting solid-phase extraction were investigated in detail. Under the optimized conditions, the method gave good linearity (0.30–200 ng/mL) and a low detection limit (0.10–0.15 ng/mL) for the hydroxyl polycyclic aromatic hydrocarbons. The average recovery for spiked urine samples with three levels ranged from 90.6% to 100%. The intra-day and inter-day relative standard deviations were in the range of 1.8–6.4% and 2.7–11.8%, respectively. Besides, the GO@Dt–NH2 provided enrichment factors of 18–20 and superior purification ability. The developed method was successfully applied to the determination of hydroxyl polycyclic aromatic hydrocarbons in urine samples from smoking volunteers.


Introduction
Polycyclic aromatic hydrocarbons (PAHs) are a class of ubiquitous environmental contaminants originating from the incomplete combustion and pyrolysis of organic materials such as wood, fossil fuels, and tobacco [1]. Human exposure to these compounds occurs in multiple sources, including the environment (polluted air and drinking water), occupation (coke oven or aluminum and steel manufacture), personal habits (smoking), and diet (barbecue, roasting, broiling, or cooked foods) [2]. Chronic exposure to PAHs has been associated with lung cancer [3][4][5], peripheral arterial disease [6], female breast cancer [7], and colorectal cancer [8]. Given their wide distribution and their carcinogenic, mutagenic, and teratogenic toxicity, PAHs can pose potential risks to human health. However, the direct assessment of PAHs remains a great challenge because of their various entry routes. Hydroxylated PAHs (OH-PAHs), as metabolites of PAHs in urine, are commonly used as stable internal biomarkers to comprehensively estimate the total external PAH exposure level and health risk [9,10].
The measurement of OH-PAHs in urine is usually performed by gas chromatography-mass spectrometry (GC-MS) or high-performance liquid chromatography (HPLC) with MS or a fluorescence

Fabrication and Characterization of the GO@Dt-NH 2 Composite
In this study, GO nanosheets were developed to load on the Dt-NH 2 via covalent coupling of the carboxy groups of GO and the amino groups of Dt-NH 2 using N, N -dicyclohexylcarbodiimide (DCC) as a coupling agent. 2-Hydroxy-naphthalene (2-OHNap), 2-hydroxy-fluorene (2-OHFlu), 1-hydroxy-phenanthrene (1-OHPhe), and 1-hydroxy-pyrene (1-OHPyr) were selected as testing analytes to investigate the effect of the mass percentage of GO in the reaction mixture (0.0%, 0.4%, 0.8%, 1.2% and 1.6%) on the extraction efficiency (Figure 1a). To guarantee the complete adsorption equilibrium, 100 mg of each synthesized composite was added to 2.0 mL of a 20 ng/mL spiked aqueous sample solution. The adsorption efficiency was evaluated by the difference in the peak areas of OH-PAH standard solutions and that detected in the effluents of GO@Dt-NH 2 . As shown in Figure 1a, the adsorption efficiencies of Dt-NH 2 for OH-PAHs were in the range of 0-59.5%. The increase in the mass percentage of GO in the reaction mixture from 0.4% to 1.6% dramatically enhanced the extraction performance of all the analytes and reached the maximum (100% for each analyte) when the percentage of GO exceeded 1.2%. It also definitely revealed the presence of GO in the as-prepared composites. Furthermore, 1.2% of the GO in the reaction mixture provided adequate adsorption for each analyte, significantly confirming the superiority of the hybrid GO and Dt-NH 2 adsorbent. Thus, 1.2% of the mass percentage of GO in the reaction mixture was selected to prepare the GO@Dt-NH 2 composites.
Molecules 2019, 24, x FOR PEER REVIEW 3 of 13 solvent, and volume were optimized. Moreover, the GO@Dt-NH2 composite was applied for the SPE of OH-PAHs in urine samples.

Fabrication and Characterization of the GO@Dt-NH2 Composite
In this study, GO nanosheets were developed to load on the Dt-NH2 via covalent coupling of the carboxy groups of GO and the amino groups of Dt-NH2 using N, N′-dicyclohexylcarbodiimide (DCC) as a coupling agent. 2-Hydroxy-naphthalene (2-OHNap), 2-hydroxy-fluorene (2-OHFlu), 1-hydroxy-phenanthrene (1-OHPhe), and 1-hydroxy-pyrene (1-OHPyr) were selected as testing analytes to investigate the effect of the mass percentage of GO in the reaction mixture (0.0%, 0.4%, 0.8%, 1.2% and 1.6%) on the extraction efficiency (Figure 1a). To guarantee the complete adsorption equilibrium, 100 mg of each synthesized composite was added to 2.0 mL of a 20 ng/mL spiked aqueous sample solution. The adsorption efficiency was evaluated by the difference in the peak areas of OH-PAH standard solutions and that detected in the effluents of GO@Dt-NH2. As shown in Figure 1a, the adsorption efficiencies of Dt-NH2 for OH-PAHs were in the range of 0-59.5%. The increase in the mass percentage of GO in the reaction mixture from 0.4% to 1.6% dramatically enhanced the extraction performance of all the analytes and reached the maximum (100% for each analyte) when the percentage of GO exceeded 1.2%. It also definitely revealed the presence of GO in the as-prepared composites. Furthermore, 1.2% of the GO in the reaction mixture provided adequate adsorption for each analyte, significantly confirming the superiority of the hybrid GO and Dt-NH2 adsorbent. Thus, 1.2% of the mass percentage of GO in the reaction mixture was selected to prepare the GO@Dt-NH2 composites. The prepared Dt-NH2, GO, and GO@Dt-NH2 were investigated by XRD measurements. As shown in Figure 1b, for GO@Dt-NH2, the existence of the typical peak of GO at 10.2° (001) and the The prepared Dt-NH 2 , GO, and GO@Dt-NH 2 were investigated by XRD measurements. As shown in Figure 1b, for GO@Dt-NH 2 , the existence of the typical peak of GO at 10.2 • (001) and the appearance of the characteristic peaks of Dt-NH 2 at 21.8 • (101), 28.3 • (111), 31.3 • (102), and 35.9 • (112) further confirmed the successful combination of GO and Dt-NH 2 . The morphology of the synthesized Dt-NH 2 and GO@Dt-NH 2 composites (7.5 mg GO) was characterized by SEM. SEM images (Figure 1c) indicated Dt-NH 2 with a characteristic disk-like shape. Regularly spaced rows of pores (~300-500 nm in diameter) located along the frustule shell wall were clearly observed, indicating the porosity and large specific surface area of Dt. For GO@Dt-NH 2 (Figure 1d), the Dt-NH 2 was encapsulated completely by GO sheets. The SEM images demonstrate that GO sheets were immobilized onto the Dt-NH 2 surface, as indicated by the arrows.
The adsorption capacities of Dt-NH 2 and GO@Dt-NH 2 for the OH-PAHs are listed in Table 1. Dt-NH 2 gave adsorption capacities of 0.0-17.8 mg/g and the GO@Dt-NH 2 provided greatly enhanced adsorption capacities ranging from 181.8 to 409.6 mg/g. In addition, the adsorption ability of GO@Dt-NH 2 for the selected compounds followed an increasing order of their octanol-water partition coefficient (logP) values (logP 2-OHNap > logP 2-OHFlu > logP 1-OHPhe > logP 1-OHPyr ) and their numbers of benzene rings. These results likely resulted from the hydrophobic effect and π-π interactions between the GO and OH-PAHs. appearance of the characteristic peaks of Dt-NH2 at 21.8° (101), 28.3° (111), 31.3° (102), and 35.9° (112) further confirmed the successful combination of GO and Dt-NH2. The morphology of the synthesized Dt-NH2 and GO@Dt-NH2 composites (7.5 mg GO) was characterized by SEM. SEM images ( Figure 1c) indicated Dt-NH2 with a characteristic disk-like shape. Regularly spaced rows of pores (~300-500 nm in diameter) located along the frustule shell wall were clearly observed, indicating the porosity and large specific surface area of Dt. For GO@Dt-NH2 (Figure 1d), the Dt-NH2 was encapsulated completely by GO sheets. The SEM images demonstrate that GO sheets were immobilized onto the Dt-NH2 surface, as indicated by the arrows. The adsorption capacities of Dt-NH2 and GO@Dt-NH2 for the OH-PAHs are listed in Table 1. Dt-NH2 gave adsorption capacities of 0.0-17.8 mg/g and the GO@Dt-NH2 provided greatly enhanced adsorption capacities ranging from 181.8 to 409.6 mg/g. In addition, the adsorption ability of GO@Dt-NH2 for the selected compounds followed an increasing order of their octanol-water partition coefficient (logP) values (logP2-OHNap > logP2-OHFlu > logP1-OHPhe > logP1-OHPyr) and their numbers of benzene rings. These results likely resulted from the hydrophobic effect and π-π interactions between the GO and OH-PAHs.

Optimization of the SPE Conditions
In order to optimize the SPE conditions for OH-PAHs, several parameters, including the loading volume, eluting solvent, and its volume were investigated in detail. All of the optimization experiments were performed in triplicate.

The Loading Volume
In this work, 100 mg of GO@Dt-NH2 adsorbent was selected to load the sample solution into the cartridge. To facilitate the pretreatment process and obtain high sensitivity, the loading volume was a crucial factor that influenced the extraction capacity and enrichment ability.
Different loading volumes from 4 to 10 mL were investigated, with the constant spiked OH-PAH amount in loading samples. The adsorbed analytes were eluted using 3 mL of ACN. As shown in Figure 2a, recoveries ranging from 92.6% to 104% were obtained for loading volumes up to 8 mL, and further increases in the loading volume decreased the analyte recoveries to 47.8%-82.6%. (112) further confirmed the successful combination of GO and Dt-NH2. The morphology of the synthesized Dt-NH2 and GO@Dt-NH2 composites (7.5 mg GO) was characterized by SEM. SEM images ( Figure 1c) indicated Dt-NH2 with a characteristic disk-like shape. Regularly spaced rows of pores (~300-500 nm in diameter) located along the frustule shell wall were clearly observed, indicating the porosity and large specific surface area of Dt. For GO@Dt-NH2 (Figure 1d), the Dt-NH2 was encapsulated completely by GO sheets. The SEM images demonstrate that GO sheets were immobilized onto the Dt-NH2 surface, as indicated by the arrows. The adsorption capacities of Dt-NH2 and GO@Dt-NH2 for the OH-PAHs are listed in Table 1. Dt-NH2 gave adsorption capacities of 0.0-17.8 mg/g and the GO@Dt-NH2 provided greatly enhanced adsorption capacities ranging from 181.8 to 409.6 mg/g. In addition, the adsorption ability of GO@Dt-NH2 for the selected compounds followed an increasing order of their octanol-water partition coefficient (logP) values (logP2-OHNap > logP2-OHFlu > logP1-OHPhe > logP1-OHPyr) and their numbers of benzene rings. These results likely resulted from the hydrophobic effect and π-π interactions between the GO and OH-PAHs.

Optimization of the SPE Conditions
In order to optimize the SPE conditions for OH-PAHs, several parameters, including the loading volume, eluting solvent, and its volume were investigated in detail. All of the optimization experiments were performed in triplicate.

The Loading Volume
In this work, 100 mg of GO@Dt-NH2 adsorbent was selected to load the sample solution into the cartridge. To facilitate the pretreatment process and obtain high sensitivity, the loading volume was a crucial factor that influenced the extraction capacity and enrichment ability.
Different loading volumes from 4 to 10 mL were investigated, with the constant spiked OH-PAH amount in loading samples. The adsorbed analytes were eluted using 3 mL of ACN. As shown in Figure 2a, recoveries ranging from 92.6% to 104% were obtained for loading volumes up to 8 mL, and further increases in the loading volume decreased the analyte recoveries to 47.8%-82.6%. (112) further confirmed the successful combination of GO and Dt-NH2. The morphology of the synthesized Dt-NH2 and GO@Dt-NH2 composites (7.5 mg GO) was characterized by SEM. SEM images ( Figure 1c) indicated Dt-NH2 with a characteristic disk-like shape. Regularly spaced rows of pores (~300-500 nm in diameter) located along the frustule shell wall were clearly observed, indicating the porosity and large specific surface area of Dt. For GO@Dt-NH2 (Figure 1d), the Dt-NH2 was encapsulated completely by GO sheets. The SEM images demonstrate that GO sheets were immobilized onto the Dt-NH2 surface, as indicated by the arrows. The adsorption capacities of Dt-NH2 and GO@Dt-NH2 for the OH-PAHs are listed in Table 1. Dt-NH2 gave adsorption capacities of 0.0-17.8 mg/g and the GO@Dt-NH2 provided greatly enhanced adsorption capacities ranging from 181.8 to 409.6 mg/g. In addition, the adsorption ability of GO@Dt-NH2 for the selected compounds followed an increasing order of their octanol-water partition coefficient (logP) values (logP2-OHNap > logP2-OHFlu > logP1-OHPhe > logP1-OHPyr) and their numbers of benzene rings. These results likely resulted from the hydrophobic effect and π-π interactions between the GO and OH-PAHs.

Optimization of the SPE Conditions
In order to optimize the SPE conditions for OH-PAHs, several parameters, including the loading volume, eluting solvent, and its volume were investigated in detail. All of the optimization experiments were performed in triplicate.

The Loading Volume
In this work, 100 mg of GO@Dt-NH2 adsorbent was selected to load the sample solution into the cartridge. To facilitate the pretreatment process and obtain high sensitivity, the loading volume was a crucial factor that influenced the extraction capacity and enrichment ability.
Different loading volumes from 4 to 10 mL were investigated, with the constant spiked OH-PAH amount in loading samples. The adsorbed analytes were eluted using 3 mL of ACN. As shown in Figure 2a, recoveries ranging from 92.6% to 104% were obtained for loading volumes up to 8 mL, and further increases in the loading volume decreased the analyte recoveries to 47.8%-82.6%. (112) further confirmed the successful combination of GO and Dt-NH2. The morphology of the synthesized Dt-NH2 and GO@Dt-NH2 composites (7.5 mg GO) was characterized by SEM. SEM images (Figure 1c) indicated Dt-NH2 with a characteristic disk-like shape. Regularly spaced rows of pores (~300-500 nm in diameter) located along the frustule shell wall were clearly observed, indicating the porosity and large specific surface area of Dt. For GO@Dt-NH2 (Figure 1d), the Dt-NH2 was encapsulated completely by GO sheets. The SEM images demonstrate that GO sheets were immobilized onto the Dt-NH2 surface, as indicated by the arrows. The adsorption capacities of Dt-NH2 and GO@Dt-NH2 for the OH-PAHs are listed in Table 1. Dt-NH2 gave adsorption capacities of 0.0-17.8 mg/g and the GO@Dt-NH2 provided greatly enhanced adsorption capacities ranging from 181.8 to 409.6 mg/g. In addition, the adsorption ability of GO@Dt-NH2 for the selected compounds followed an increasing order of their octanol-water partition coefficient (logP) values (logP2-OHNap > logP2-OHFlu > logP1-OHPhe > logP1-OHPyr) and their numbers of benzene rings. These results likely resulted from the hydrophobic effect and π-π interactions between the GO and OH-PAHs.

Optimization of the SPE Conditions
In order to optimize the SPE conditions for OH-PAHs, several parameters, including the loading volume, eluting solvent, and its volume were investigated in detail. All of the optimization experiments were performed in triplicate.

The Loading Volume
In this work, 100 mg of GO@Dt-NH2 adsorbent was selected to load the sample solution into the cartridge. To facilitate the pretreatment process and obtain high sensitivity, the loading volume was a crucial factor that influenced the extraction capacity and enrichment ability.
Different loading volumes from 4 to 10 mL were investigated, with the constant spiked OH-PAH amount in loading samples. The adsorbed analytes were eluted using 3 mL of ACN. As shown in Figure 2a, recoveries ranging from 92.6% to 104% were obtained for loading volumes up to 8 mL, and further increases in the loading volume decreased the analyte recoveries to 47.8%-82.6%.

Optimization of the SPE Conditions
In order to optimize the SPE conditions for OH-PAHs, several parameters, including the loading volume, eluting solvent, and its volume were investigated in detail. All of the optimization experiments were performed in triplicate.

The Loading Volume
In this work, 100 mg of GO@Dt-NH 2 adsorbent was selected to load the sample solution into the cartridge. To facilitate the pretreatment process and obtain high sensitivity, the loading volume was a crucial factor that influenced the extraction capacity and enrichment ability.
Different loading volumes from 4 to 10 mL were investigated, with the constant spiked OH-PAH amount in loading samples. The adsorbed analytes were eluted using 3 mL of ACN. As shown in Figure 2a, recoveries ranging from 92.6% to 104% were obtained for loading volumes up to 8 mL, and further increases in the loading volume decreased the analyte recoveries to 47.8-82.6%. Considering the recovery and the sensitivity of the proposed method, a sample volume of 8 mL was selected for further work.
Considering the recovery and the sensitivity of the proposed method, a sample volume of 8 mL was selected for further work. The elution process must make sure all analytes are desorbed from the sorbent with minimal carryover. Thus, the eluting solvent and its volume have a significant effect on the recovery of the method. Various elution solvents, including acetone, ethyl acetate, methanol (MeOH), and acetonitrile (ACN), were studied. Figure 2b indicates that ACN exhibited the best elution ability (recovery: 92.6-98.2%). A proper elution volume guarantees adequate desorption of the analytes from the adsorbent with minimum consumption of the solvent. The volume of elution solvent from 1

Eluting Solvent and Eluting Volume
The elution process must make sure all analytes are desorbed from the sorbent with minimal carryover. Thus, the eluting solvent and its volume have a significant effect on the recovery of the method. Various elution solvents, including acetone, ethyl acetate, methanol (MeOH), and acetonitrile (ACN), were studied. Figure 2b indicates that ACN exhibited the best elution ability (recovery: 92.6-98.2%). A proper elution volume guarantees adequate desorption of the analytes from the adsorbent with minimum consumption of the solvent. The volume of elution solvent from 1 to 4 mL was also studied. Figure 2c shows that the recovery of OH-PAHs increased as the volume increased from 1 to 3 mL. An elution volume of 4 mL caused a slight decrease in the recovery of OH-PAHs. A large eluting volume is unfavorable as it extends the time for nitrogen (N 2 ) evaporation which would also cause a loss of OH-PAHs. Thus, 3 mL of ACN was employed for the elution of OH-PAHs.

Validation of the GO@Dt-NH 2 Based SPE HPLC-FLD Method
The analytical evaluation parameters including the linear range, correlation coefficients (r), limits of detection (LODs), limits of quantification (LOQs), accuracy, and precision were studied under the optimized experimental conditions. The results are listed in Table 2. The calibration curves were constructed using OH-PAHs at seven concentration levels from 0.30 to 200 ng/mL. For each concentration level, three replicate extractions and determinations were performed. As shown in Table 2, good linearity (r = 0.999) was obtained in the range of 0.50-200 ng/mL for 2-OHNap, 0.30-150 ng/mL for 2-OHFlu and 1-OHPhe, and 0.40-200 ng/mL for 1-OHPyr. The LOD (S/N = 3) was found to be 0.15 ng/mL for 2-OHNap, 0.10 ng/mL for 2-OHFlu and 1-OHPhe, and 0.12 ng/mL for 1-OHPyr, respectively. The LOQ (S/N = 10) in urine was found to be 0.50 ng/mL for 2-OHNap, 0.30 ng/mL for 2-OHFlu and 1-OHPhe, and 0.40 ng/mL for 1-OHPyr, which shows that the present method is sufficiently sensitive to monitor four OH-PAHs. Moreover, the enrichment factor (EF) was used to evaluate the enrichment ability of the method, which is defined as the ratio of the calibration curve slope of the analyte performing via SPE to that without SPE. The EFs of the four OH-PAHs were in the range of 18-20. The results were further confirmed by the recovery experiments. The accuracy was evaluated using the recovery of the proposed method which was assayed on spiked urine samples at three analyte levels (0.5, 1, and 2 ng/mL). The precision was described based on the intra-day and inter-day relative standard deviations (RSDs). As shown in Table 2, the average recoveries of four OH-PAHs were in the range of 90.6-100%. The intra-day and inter-day RSDs were in the range of 1.8-6.4% and 2.7-11.8%, respectively. The reproducibility of the material was evaluated by calculating the RSDs of peak areas of OH-PAHs. The column-to-column RSDs and batch-to-batch RSDs for three parallel varied from 3.9-8.7% and 7.0-14.5%, respectively. The GO@Dt-NH 2 adsorbents not only offer excellent extraction ability for OH-PAHs but also provide good precision and reproducibility. In conclusion, these results highlight the potential feasibility of GO@Dt-NH 2 for the SPE of OH-PAHs in urine.

Urine Sample Analysis and Comparison with Commercial C 18 Adsorbent
The proposed method was further applied for the analysis of OH-PAHs in urine samples collected from ten smokers. As indicated in Table 3, the concentration of 1-OHPhe was found to be 1.37 ± 0.08 ng/mL in only one sample, 2-OHFlu was detected in five cases ranging from 1.04 ± 0.07 to 3.18 ± 0.04 ng/mL. The concentration of 2-OHNap was found to be 0.91 ± 0.26 to 4.02 ± 0.08 ng/mL in nine samples and 1-OHPyr was detected in nine cases with a concentration of 1.23 ± 0.15 to 4.80 ± 0.01 ng/mL. The recoveries of the OH-PAHs obtained by spiking the 2 ng/mL standards in the urine samples ranged from 86.4% to 108%.  Besides, the extraction performance of the GO@Dt-NH 2 adsorbent was compared with that of the commercial C 18 adsorbent. The evaluation employed 100 mg of each adsorbent with 8 mL of the urine sample solution spiked with 2 ng/mL of 2-OHNap and 1-OHPyr and 1 ng/mL of 2-OHFlu and 1-OHPhe. As shown in Figure 3b,c, both sorbents enable the purification and enrichment of the analytes from the urine sample. The interfering chromatographic peaks appear in the range of 1.5-3.0 min. Their intensities in Figure 3c are obviously lower than those shown in Figure 3b, indicating the superior removal ability of the GO@Dt-NH 2 adsorbent for the interfering molecules. Moreover, the recoveries of four OH-PAHs obtained by commercial C 18 adsorbent were 59.3-92.3% while the corresponding results were 89.0-98.4% using the GO@Dt-NH 2 adsorbent. In addition, the starting material of natural Dt for the preparation of the GO@Dt-NH 2 sorbent is readily available and rather inexpensive. These comparisons sufficiently confirmed that GO@Dt-NH 2 has great advantages as an effective and economical adsorbent.
while the corresponding results were 89.0%-98.4% using the GO@Dt-NH2 adsorbent. In addition, the starting material of natural Dt for the preparation of the GO@Dt-NH2 sorbent is readily available and rather inexpensive. These comparisons sufficiently confirmed that GO@Dt-NH2 has great advantages as an effective and economical adsorbent.
Stock solutions of standards (1.0 mg/mL for each) were prepared by dissolving four OH-PAHs with MeOH and kept at 4 • C in the dark. Working solutions were prepared daily by diluting the standard solution with MeOH.

Instruments
Chromatographic analysis was performed on an Agilent 1200 HPLC (Santa Clara, CA, USA), consisting of a G1322A degasser, a G1311A pump system, a G1329A autosampler, a G1316A temperature control center, and a G1321A FLD. The XRD patterns were acquired on a D/Max 2500 PC single crystal X-ray diffractometer (Rigaku, Akishima, Tokyo, Japan). SEM images of the prepared composites were obtained by a FEI JEM-2800F dual beam focused ion beam/field emission scanning electron microscope (Hillsboro, TX, USA).

Preparation of GO@Dt-NH 2 Composites
Dt-NH 2 was synthesized in accordance with Jiang et al. [42] with small modifications. Dt particles were initially treated with 2 mol/L HCl solution for 1 h and then washed thoroughly with ultrapure water and dried. The collected Dt (1 g) was mixed into 100 mL ethanol solution containing 1 mL APTES by sonication for 30 min to form a homogeneous suspension. After stirring at 60 • C for 8 h, the solids were collected by centrifugation and washed with ultrapure water and ethanol several times. The obtained Dt-NH 2 product was dried under a vacuum at 60 • C for 12 h.
GO@Dt-NH 2 composites were prepared via the covalent coupling of the amino groups of Dt-NH 2 and the carboxy groups of GO using DCC as a coupling agent. Typically, 7.5 mg of GO was dispersed in 100 mL of DMF under 1 h ultrasonication. Then, 500 mg of Dt-NH 2 and 105 mg of DCC were added. The mixture was stirred at 50 • C for 24 h. The obtained gray slurry was washed with ultrapure water and MeOH repeatedly, and finally dried in a vacuum oven at 80 • C for 12 h. Figure 4 indicates a schematic illustration of the fabrication of GO@Dt-NH 2 composites. Research Institute (Tianjin, China). Acetone, hydrochloric acid (HCl), and ethanol were provided by Kermer Chemical Company (Tianjin, China). β-Glucuronidase/arylsulfatase from HELIX POMATIA (Type HP-2, aqueous solution, ≥100000 unites/mL) was purchased from Sigma-Aldrich (Milan, Italy). HPLC-grade MeOH and ACN were obtained from Sigma-Aldrich (Milan, Italy). Ultrapure water was provided by a Milli-Q water purification system (Millipore, Burlington, MA, USA

Instruments
Chromatographic analysis was performed on an Agilent 1200 HPLC (Santa Clara, CA, USA), consisting of a G1322A degasser, a G1311A pump system, a G1329A autosampler, a G1316A temperature control center, and a G1321A FLD. The XRD patterns were acquired on a D/Max 2500 PC single crystal X-ray diffractometer (Rigaku, Akishima, Tokyo, Japan). SEM images of the prepared composites were obtained by a FEI JEM-2800F dual beam focused ion beam/field emission scanning electron microscope (Hillsboro, TX, USA).

Preparation of GO@Dt-NH2 Composites
Dt-NH2 was synthesized in accordance with Jiang et al. [42] with small modifications. Dt particles were initially treated with 2 mol/L HCl solution for 1 h and then washed thoroughly with ultrapure water and dried. The collected Dt (1 g) was mixed into 100 mL ethanol solution containing 1 mL APTES by sonication for 30 min to form a homogeneous suspension. After stirring at 60 °C for 8 h, the solids were collected by centrifugation and washed with ultrapure water and ethanol several times. The obtained Dt-NH2 product was dried under a vacuum at 60 °C for 12 h.
GO@Dt-NH2 composites were prepared via the covalent coupling of the amino groups of Dt-NH2 and the carboxy groups of GO using DCC as a coupling agent. Typically, 7.5 mg of GO was dispersed in 100 mL of DMF under 1 h ultrasonication. Then, 500 mg of Dt-NH2 and 105 mg of DCC were added. The mixture was stirred at 50 °C for 24 h. The obtained gray slurry was washed with ultrapure water and MeOH repeatedly, and finally dried in a vacuum oven at 80 °C for 12 h. Figure 4 indicates a schematic illustration of the fabrication of GO@Dt-NH2 composites.

Urine Sample Collection and Preparation
Human urine samples (15 mL) were collected in the morning from 10 smoking volunteers. All volunteers signed informed consent forms for this study. This study was approved by the Ethics Committee of North China University of Science and Technology (Tangshan, China). All samples were frozen and stored at −80 °C until further analysis.
Prior to enzymatic hydrolysis, the urine sample was thawed at room temperature and briefly agitated. Five milliliters of urine was transferred into a 15 mL polypropylene centrifuge tube, and 5

Urine Sample Collection and Preparation
Human urine samples (15 mL) were collected in the morning from 10 smoking volunteers. All volunteers signed informed consent forms for this study. This study was approved by the Ethics Committee of North China University of Science and Technology (Tangshan, China). All samples were frozen and stored at −80 • C until further analysis.
Prior to enzymatic hydrolysis, the urine sample was thawed at room temperature and briefly agitated. Five milliliters of urine was transferred into a 15 mL polypropylene centrifuge tube, and 5 mL of sodium acetate buffer (0.5 mol/L, pH = 5.0) and 10 µL of β-glucuronidase/arylsulfatase enzyme were added. The mixture was incubated for 12 h at 37 • C to release OH-PAHs from the conjugated forms. Finally, the resulting sample was centrifuged at 1500 rpm for 10 min. The supernatant was collected and stored at 4 • C.

SPE Procedure
The SPE cartridges were obtained by packing 100 mg of GO@Dt-NH 2 adsorbents in 6 mL empty cartridges using polypropylene upper and lower frits to avoid adsorbent loss. For SPE of OH-PAHs from urine, the GO@Dt-NH 2 cartridge was preconditioned with 2 mL of ACN and 2 mL of water. Eight milliliters of urine hydrolysate was loaded on the cartridge at a flow rate of 0.2 mL/min. The cartridge was kept in a vacuum for 5 min to remove the residual solvent. Subsequently, the cartridge was eluted with 3 mL ACN at a flow rate of 0.2 mL/min. The eluent was collected and evaporated to dryness under N 2 steam. Finally, the residues were re-dissolved in 200 µL MeOH for subsequent HPLC analysis. Each extraction was carried out for three replicates for parallel testing.

Conclusions
A novel GO@Dt-NH 2 adsorbent was fabricated and applied for the SPE of OH-PAHs in urine samples with wide linear ranges, low detection limits, and satisfactory recoveries. Compared with the bare Dt-NH 2 particles, the adsorption affinity of the developed GO@Dt-NH 2 sorbent obtained for OH-PAHs was significantly enhanced. GO@Dt-NH 2 sorbent obtained with a minimax ratio of 1.2% of GO in the polymerization provided an analyte enrichment factor of 18-20. Besides, the sorbent offered a superior purification ability compared with commercial C 18 adsorbent. GO@Dt-NH 2 composites are promising as effective and economical adsorbents for the SPE of trace analytes from biological samples.

Conflicts of Interest:
The authors declare no conflict of interest.