Application of Online Liquid Chromatography 7 T FT-ICR Mass

May 22, 2019 - resonance mass spectrometry (FT-ICR MS), combined with quadrupolar .... concentration of 200 μg/mL in water were prepared. MS Methods...
0 downloads 0 Views 5MB Size
Download PDF
Article Cite This: Anal. Chem. 2019, 91, 7690−7697

pubs.acs.org/ac

Application of Online Liquid Chromatography 7 T FT-ICR Mass Spectrometer Equipped with Quadrupolar Detection for Analysis of Natural Organic Matter Donghwi Kim,†,‡ Sungjune Kim,† Seungwoo Son,† Maeng-Joon Jung,† and Sunghwan Kim*,†,§ †

Department of Chemistry, Kyungpook National University, Daegu 41566, Republic of Korea Analytical Research Center, Korea Institute of Toxicology, Daejeon 34114, Republic of Korea § Green-Nano Materials Research Center, Daegu 41566, Republic of Korea Downloaded via IDAHO STATE UNIV on July 19, 2019 at 01:46:49 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

‡

S Supporting Information *

ABSTRACT: In this study, Fourier transform ion cyclotron resonance mass spectrometry (FT-ICR MS), combined with quadrupolar detection (QPD), was applied for online liquid chromatography (LC) MS analysis of natural organic matter (NOM). Although FT-ICR MS has emerged as an important analytical technique to study NOM, there are few previous reports on online LC FT-ICR MS analysis of NOM due to the long acquisition time (2−8 s) required to obtain high-resolution mass spectra. The QPD technique provides a critical advantage over the conventional dipolar detection (DPD) technique for LC-MS analysis because a spectrum with the same resolving power can be obtained in approximately half the acquisition time. QPD FT-ICR MS provides resolving powers (

m ) Δm50%

of

∼300000 and 170000 at m/z 400 with acquisition times per scan of 1.2 and 0.8 s, respectively. The reduced acquisition time per scan allows increased number of acquisitions in a given LC analysis time, resulting in improved signal to noise (S/N) ratio and dynamic range in comparison to conventional methods. For example, 40% and 100% increases in the number of detected peaks were obtained with LC QPD FT-ICR MS, in comparison to conventional LC DPD FT-ICR MS and direct-injection FT-ICR MS. It is also possible to perform more quantitative comparison and molecular level investigation of NOMs with 2 μg of a NOM sample. The data presented herein demonstrate a proof of principle that QPD combined with LC FT-ICR MS is a sensitive analytical technique that can provide comprehensive information about NOM.

N

Although FT-ICR MS can provide more detailed compositional information in comparison to other analytical techniques, it has limitations such as ionization suppression like any other mass analyzer,25,26 which can be overcome by chromatographic separation before MS analysis. However, to the best of our knowledge, there have been no studies combining online LC and FT-ICR MS for NOM analysis. Fractionation by LC, followed by MS analysis, has been used to obtain molecular-level information on NOM.27−34 However, fractionations involve collection and sample treatment steps, making them time-consuming and labor-intensive. There have been a couple of online LC-MS studies on NOM, but they have been conducted by Orbitrap or even lowerresolution MS techniques, such as ion-trap and quadrupole time-of-flight mass spectrometry equipped with ion mobility.35−38 A comparison of 15 T FT-ICR and Orbitrap MS operated at different mass resolution settings for NOM

atural organic matter (NOM) is ubiquitous in aquatic and terrestrial environments and plays important roles in the environment as organic nutrients; it also finds application in carbon cycling and storage.1−7 Furthermore, the pharmacological effects of humic substances (HS) have been studied in terms of antiviral activity,8−10 anti-inflammatory properties,11 and estrogenic activity.12,13 However, some compounds in NOM can form harmful disinfection byproducts (DBPs), which can be hazardous even at low concentrations.14,15 Indepth investigations on the chemical composition of NOM are essential to understand its properties. However, the NOM constituents have not been fully characterized at the molecular level because of the complexity of NOM. Therefore, continuous development in analytical methods is needed. Ultrahigh-resolution mass spectrometry (UHRMS), utilizing a Fourier transform ion cyclotron resonance (FT-ICR) mass spectrometer, has emerged as an important analytical technique to characterize complex organic mixtures at the molecular level.16−24 The high resolving power of FT-ICR MS is essential to discriminate between various isobaric compounds in complex organic mixtures such as NOM. © 2019 American Chemical Society

Received: February 6, 2019 Accepted: May 20, 2019 Published: May 22, 2019 7690

DOI: 10.1021/acs.analchem.9b00689 Anal. Chem. 2019, 91, 7690−7697

Article

Analytical Chemistry

Figure 1. Broad-band (left column) spectra, expanded (center column) spectra, and time-domain signals (right column) obtained by direct injection (−) ESI FT-ICR MS analysis of SRFA using DPD with 2 M word data (top row), QPD with 2 M word data (middle row), and QPD with 4 M word data (bottom row).

acetonitrile (J.T. Baker, Center Valley, PA, USA), and ACSreagent-grade formic acid (Sigma-Aldrich; ≥98%) were used in this study. For direct-infusion (DI) MS measurement, SRFA was dissolved in a water/methanol (50/50, v/v) mixture at a concentration of 200 μg/mL. For LC-MS, samples with a final concentration of 200 μg/mL in water were prepared. MS Methods. A 7 T FT-ICR MS system (SolariX 2xR, Bruker Daltonik GmbH, Bremen, Germany), equipped with QPD, was used in the MS measurements. Electrospray ionization (ESI) with the standard Bruker ESI source was used as an ionization method. Details on the instrument and QPD were previously reported.48 Briefly, the instrument has a Paracell and electronics enabling QPD. With QPD, two pairs of electrodes constituting the ICR cell are used for detection. In contrast, only one pair of the electrodes is used in traditional dipolar detection (DPD). The user can switch instruments between DPD and QPD modes in the operating computer program. Shimming needs be done to minimize the unwanted harmonic signal when the detection modes are switched. Shimming is performed by applying dc bias voltages to 0, 90, 180, and 270° electrodes of the ICR cell. The presence of high harmonic peaks can affect the overall spectral shape and cause confusion in clarifying the actual analyte. For DI-MS measurement, samples were dissolved in a methanol/water (50/50, v/v) mixture to obtain final concentrations of 200 ppm and infused at a rate of 180 μL/h. MS spectra were acquired from m/z 100 to 1500 with a transient size of 2 mega (M) words, resulting in a resolving power ( m ) of ∼170000

characterization was reported, and it showed that Orbitrap MS could be used for NOM analysis. However, FT-ICR MS was superior in certain applications because about three times more peaks could be observed in comparison to that by Orbitrap MS.39 There is a dearth of research using online LC FT-ICR MS because of relatively long acquisition times ranging from 2 to 8 s required to obtain high-resolution data and difficulty in analysis of the large data sets produced by the hyphenated FTICR MS method. The long acquisition time limits the chromatographic resolution and the number of spectra obtained from LC-MS analysis. With innovations in ICR cells,40−43 higher resolution can be obtained with a given magnetic field. In addition, using the frequency-multiple detection technique can improve the resolution of FT-ICR data with the same acquisition time.44−46 In a previous study, FT-ICR MS combined with quadrupolar detection (QPD) was shown to be effective in analyzing complex organic mixtures.47 The important advantage of using QPD for LC-MS is that data with the same resolution can be obtained with half the acquisition time, which can reduce the acquisition time per scan for LC-MS applications. In this study, QPD FT-ICR MS was applied for online LC-MS analysis of NOM.

■

EXPERIMENTAL SECTION Sample Preparation. Suwannee River fluvic acid (SRFA) and Upper Mississippi River NOM (UMRNOM) were purchased from the International Humic Substances Society (IHSS). Elemental compositions of the SRFA and UMRNOM are provided in Table S1 in the Supporting Information. HPLC-grade methanol (Sigma-Aldrich GmbH, Hamburg, Germany), water (J.T. Baker, Phillipsburg, NJ, USA),

Δm50%

at m/z 400. A total of 100 time-domain transients were summed for about 15 min with DI-MS. The sweep excitation power of 18% and magnitude mode were used for all of the measurements reported in this study. The capillary voltage was 7691

DOI: 10.1021/acs.analchem.9b00689 Anal. Chem. 2019, 91, 7690−7697

Article

Analytical Chemistry

0.52 s and 170000 with QPD and 2 M words, and 1.05 s and 320000 with QPD and 4 M words, respectively. The data presented in Figure 1 show that the resolving power can be doubled for a specified acquisition time or the required acquisition time can be halved for a specified resolving power by use of QPD. Optimizing Experimental Conditions for QPD FT-ICR MS. One of the potential drawbacks of QPD is the presence of undesirable harmonic signals observed at half frequency (or double m/z) generated by off-centered ions or magnetron motion. Figure S1 shows the change in QPD mass spectra of SRFA before and after performing shimming for QPD. Without proper shimming, unwanted harmonic peaks appeared in the mass range from m/z 600 to 1200. For example, the left panel of Figure S1 shows a significant abundance of harmonic peaks at m/z 762, especially with lower number of scans. There are peaks coming from the sample at m/z 762 as well, but they cannot be observed in those spectra because of their low S/N ratios. The weak peaks originated from the sample become apparent when 50 scans were summed. However, some of the harmonic peaks remained even after accumulation of 50 scans. After shimming by applying DC correction voltages and pulsed DC bias voltages to the ICR cell, the harmonic peaks were significantly reduced in the spectra (right panel of Figure S1). The transient size and the lowest m/z limit for acquisition determine the acquisition time for FT-ICR MS analysis. Therefore, it is important to optimize them for successful LCMS setup. The FID time, time per acquisition, and resolving power at m/z 400 are presented in Table 1, depending on the

set at 3500 V, and other experimental conditions were as follows: nebulizer gas pressure 0.8 bar, drying gas temperature 200 °C, drying gas flow rate 4.0 L/min, and ion accumulation time 0.020 s. Typically, the accumulation time is set to 0.01− 0.03 s for DI FT-ICR MS data acquisition. The greater acquisition time of 0.1−0.3 s was used for the LC-MS application. This is because a smaller amount of NOM sample is spread in time on a chromatogram for the LC-MS acquisition. For LC-MS, the nebulizer gas pressure was set at 1.6 bar and ion accumulation time was set at 0.2 s, while the other parameters were the same as the previous ones. The obtained MS data of NOM were recalibrated using their corresponding homologue series having distinct molecular weights, following internal calibration using sodium trifluoroacetate (Sigma-Aldrich, 98%) in Compass DataAnalysis 5.0 software (Bruker Daltonik GmbH, Germany). Typically, the mass accuracy is within 5 ppm with external calibration and it is improved to be within 1 ppm after internal calibration. After data calibration, only peaks with S/N > 6 were extracted and assigned to chemical formulas (error within 1 ppm) using the in-house-developed software. Spectral interpretation was also performed using in-house-developed software with an automated peak-picking algorithm for more reliable and faster results.48,49 The conditions for NOM analysis were as follows: CcHhNnOoSs; c unlimited, h unlimited, 0 ≤ n ≤ 3, 0 ≤ o ≤ 30, and 0 ≤ s ≤ 3. Double bond equivalent (DBE) values were calculated using the equation DBE = c −

h n + +1 2 2

LC Methods. HPLC separation was performed on a DIONEX Ultimate 3000 system (Thermo Fisher Scientific Inc., Sunnyvale, CA,USA), equipped with an Ultimate 3000 binary pump, an Ultimate 3000 column compartment, an Ultimate 3000 autosampler, and an Ultimate 3000 diode array detector (DAD). The chromatographic separation was conducted using an Acquity UPLC BEH C18 column (1.0 × 100 mm, 1.7 μm, 130 Å, Waters) fitted with an Acquity HSS T3 VanGuard precolumn (2.1 × 5 mm, 1.8 μm, 100 Å, Waters). The mobile phase consisted of water containing 0.1% (v/v) formic acid (eluent A) and acetonitrile containing 0.1% (v/v) formic acid (eluent B) with gradient elution at a flow rate of 0.1 mL/min. The gradient program is shown in Table S2. Each 10 μL sample was injected three times to test the reproducibility, and a blank sample containing no analyte was injected between successive samples to monitor the baseline.

Table 1. Comparison of FID Time (s), Scan Rate (s), and Resolving Power (RP) at m/z 400 Obtained from LC QPD FT-ICR Spectra of SRFA with 1 or 2 M Word Data and at Mass Ranges of m/z 100, 150, and 200 mass range QPD 1 M FID scan rate RP QPD 2 M FID scan rate RP

■

100−1500

150−1500

200−1500

0.262 0.541 92,000

0.367 0.639 130,000

0.472 0.740 153,000

0.524 0.803 171,000

0.734 1.006 236,000

0.944 1.212 297,000

data size and the detected mass range. Data size and the lowest m/z value are the two important parameters determining the resolving power of data obtained with FT-ICR MS. The lowest m/z determines the sampling rate by the Nyquist limit. The spectra used to generate Table 1 are provided in Figure S2. The time per acquisition is longer than the FID time by about 0.27 s mainly because of the time required to transfer data from the mass spectrometer to data system. Under 1 M word data, the times per acquisition are 0.54, 0.64, and 0.74 s in the mass ranges of m/z 100−1500, 150−1500, and 200−1500, respectively. Total experimental times with 2 M word transients are 0.80, 1.01, and 1.21 s in the mass ranges of m/z 100−1500, 150−1500, and 200−1500, respectively. Figure S2 shows that the spectra obtained by 1 M word data did not provide enough resolution to separate the adjacent peaks around m/z 800. Two M word data obtained in the m/z

RESULTS AND DISCUSSION Comparison between Dipolar and Quadrupolar Detections. Figure 1 shows the negative-ion-mode electrospray ionization ((−) ESI) FT-ICR MS spectra of SRFA obtained by dipolar detection (DPD) and quadrupolar detection (QPD) with different acquisition times. The broad-band mass spectra of SRFA exhibit molecular weight distributions ranging from m/z 200 to 1000, centered around m/z 450. The resolving powers ( m , where m is the Δm50%

measured mass and Δm50% is the full width at half-maximum (FWHM)) were calculated and presented in the expanded spectra. In the mass spectra starting from m/z 100, the free induction decay (FID) time and resolving power (RP) at m/z 400 were 1.05 s and 160000 with DPD and 2 mega (M) words of data, 7692

DOI: 10.1021/acs.analchem.9b00689 Anal. Chem. 2019, 91, 7690−7697

Article

Analytical Chemistry

Figure 2. LC QPD FT-ICR MS data from SRFA divided into six segments marked by boxes: (a) total ion current chromatogram (red) and UV chromatogram at 274 nm (blue) and (b) broad-band and (c) expanded spectra of each segment.

Table 2. Comparison of the Number of Scans and Averaged Signal to Noise Ratios of the Top 1000 Peaks Observed in LC DPD and LC QPD FT-ICR MS DPD LC-MS time (min) scan # scan S/N QPD LC-MS time (min) scan # scan S/N

segment 1

segment 2

segment 3

segment 4

segment 5

segment 6

1.50−2.48 58−98 41 104.0

2.50−4.49 99−180 82 177.9

4.51−5.98 181−241 61 130.1

6.01−10.00 242−404 163 216.5

10.02−12.98 405−526 122 140.2

13.01−15.98 527−646 120 56.2

1.50−2.50 84−141 58 167.1

2.52−4.49 142−254 113 223.3

4.51−5.98 255−340 86 168.0

6.00−9.99 341−568 228 292.8

10.00−13.00 569−742 174 197.3

13.01−15.99 743−914 172 70.2

range of 100−1500 provided enough resolving power to resolve peaks around m/z 800 with the shortest time per scan; hence, it was determined to be the optimum condition for LC QPD MS analysis of NOM. The data presented in the rest of the paper were obtained under this condition. Comparison of Data Obtained with LC QPD MS and Conventional Methods. To evaluate the effectiveness of the proposed analytical method, the SRFA sample was analyzed by LC QPD FT-ICR MS, LC DPD FT-ICR MS, and direct injection (DI) FT-ICR MS. Figure 2 shows the LC QPD MS results for the SRFA sample. The results obtained with LC DPD and DI FT-ICR MS are presented in Figure S3a,b in the Supporting Information. As seen in Figure 2a, SRFA shows continuous UV and total ion current (TIC) chromatograms.

The averaged mass spectra of SRFA at each segment are presented in Figure 2b. The molecular weight distribution of each segment ranged from m/z 200 to 900, and the highest signal intensity was observed around m/z 400. In the expanded spectra (Figure 2c), the abundances of the peaks at low mass defect corresponding to oxygen-rich compounds decreased and those of the peaks at high mass defect increased from segments 1 to 6. Oxygen-rich compounds are expected to be more polar than oxygendeficient compounds and hence would be eluted earlier in reversed-phase chromatography. This trend in peak distribution was observed over the entire mass region. To visualize this change, the major chemical class distribution of each segment is generated and presented in Figure S4. O13 class compounds 7693

DOI: 10.1021/acs.analchem.9b00689 Anal. Chem. 2019, 91, 7690−7697

Article

Analytical Chemistry

Figure 3. van Krevelen diagrams of Ox class compounds observed from SRFA by LC QPD FT-ICR MS. Refer to Figure 2 for the spectrum obtained from each segment of LC-MS data.

are the most abundant in segment 1, whereas O11 and O9 class compounds are the most abundant in segments 4 and 6, respectively. Table 2 shows the number of scans in each segment and the averaged S/N ratios of the top 1000 peaks. The DPD and QPD data were obtained with the same resolution but different number of scans because of the difference in data acquisition times. As the number of scans in each segment were increased 1.4-fold in LC QPD MS in comparison to LC DPD MS, the averaged S/N ratio of the top 1000 peaks increased 1.2−1.6 times for LC QPD MS, consistent with the predicted theoretical value ( 1.4 ≈1.18).50 The improved S/N ratio resulted in increased number of peaks in the LC QPD MS profile. For an acquisition time of 15 min, a total of 13490 peaks assigned to CcHhNnOoSs (c unlimited, h unlimited, 0 ≤ n ≤ 3, 0 ≤ o ≤ 30, and 0 ≤ s ≤ 3) were identified for SRFA by LC QPD MS, whereas 9706 peaks and 6543 peaks were

detected in LC DPD MS and DI MS, respectively. For this calculation, compounds that were detected in two or more segments were counted only once. Therefore, the number of observed compounds increased by about 40% and 100% in LC QPD MS in comparison to that in LC DPD MS and direct injection MS. QPD was used to obtain DI-MS data, and hence the improvement of LC QPD MS over DI-MS clearly shows the advantage of coupling LC for NOM analysis. The number of identified compounds in each Ox class is presented in Table S3. The number of detected peaks in the case of LC QPD MS increased by approximately 10% and 27% in comparison with that for LC DPD MS and DI-MS, respectively. These data clearly show the expanded dynamic range of LC QPD MS. Detailed Molecular-Level Characterization of SRFA. Figure 3 shows the van Krevelen diagram of the Ox class compounds observed from each segment of the chromatogram 7694

DOI: 10.1021/acs.analchem.9b00689 Anal. Chem. 2019, 91, 7690−7697

Article

Analytical Chemistry

Figure 4. Plots showing (a) carbon number distribution and (b) double bond equivalent (DBE) of O11 class compounds observed from six segments of chromatograms in Figure 2 obtained with LC QPD FT-ICR MS.

Figure 5. (a) UV chromatograms showing the difference between SRFA and UMRNOM samples. Chemical class distribution plots for (b) Ox class compounds, (c) N1Ox class compounds, and (d) S1Ox class compounds observed in segments 1 and 4.

shown in Figure 2. Tannin-type compounds with O/C > 0.6 and H/C < 1 were dominant in the earlier segments (e.g., segments 1−3), but lignin-type compounds with O/C < 0.6 and 0.5 < H/C < 1.5 were present in segments 5 and 6. To further understand the compositional change at the molecular level, the carbon number distribution and double bond equivalent plots for O11 class compounds were generated (Figure 4). In both plots, the molecules mainly observed in segment 1 had 14−17 carbons and 9−11 DBE values, but those predominantly identified in segment 6 consisted of 26− 30 carbons and had 11−13 DBE values. The data showed that

compounds with a low carbon number and DBE were eluted earlier in the chromatogram. A similar trend was seen in the Kendrick mass defect plots for each of the six segments (Figure S5). In the Kendrick mass defect plot, compounds of the same class and DBE but different degrees of alkylation are positioned on a single horizontal line, while the compounds of different classes are displayed vertically from those of other class compounds.51−54 As shown in Figure S5, the dots on the line of the same Kendrick mass defect gradually move from left to right with decreasing polarity of the eluents. Again, the data clearly revealed separation at the molecule level according to the 7695

DOI: 10.1021/acs.analchem.9b00689 Anal. Chem. 2019, 91, 7690−7697

Article

Analytical Chemistry

QPD, as shown in Table 2. Comparison of the SRFA and UMRNOM data obtained by LC FT-ICR MS reveals that sulfur- and nitrogen-containing compounds are more abundant in the UMRNOM, consistent with UV chromatograms and elemental analysis data. LC FT-ICR MS is a sensitive technique for NOM analysis, and only 2 μg of NOM samples is used to obtain the data presented in this study. LC FT-ICR MS can be a useful tool for the analysis of NOM when sample quantities are limited. Therefore, it is concluded that LC QPD FT-ICR MS is a sensitive analytical method that allows for more quantitative comparison and molecular-level interpretation of NOM. Since this study addresses the advantage of using quadrupolar detection for LD FT-ICR MS analysis of NOM, the typical reversed-phase separation method is used. However, development of an efficient LC separation method will be beneficial in further improving the proposed analytical technique: in particular, combining with extracted ion chromatograms can be advantageous in isomeric characterization. Research is ongoing in this direction.

number of carbons in the alkyl groups. Overall, the data presented in Figures 2−4 show that the separation and detection of lignin- and tannin-type molecules existing in SRFA can be effectively done with LC QDP FT-ICR MS. Comparison of SRFA vs Upper Mississippi River NOM. Upper Mississippi River NOM (UMRNOM) from IHSS was analyzed by LC QPD FT-ICR MS, and the results obtained were compared with those for SRFA (Figure 5). The raw MS data of UMRNOM are provided in Figure S6. The scan ranges used for the segment were the same as those in Table 2. Segments 1 and 4 were selected and are shown in Figure 5 because a greater difference between SRFA and UMRNOM was observed in the segments. The UV chromatogram obtained from UMRNOM has higher intensity in segment 1 but lower intensity in segment 4 in comparison to that from SRFA (refer to Figure 5a). The summed absolute abundances of Ox, N1Ox, and S1Ox classes observed by FT-ICR MS are presented in Figures 5b−d. Ox class compounds are the most abundant in segment 4 of SRFA (refer to red bars in Figure 5b). However, N1Ox and S1Ox class compounds are the most abundant in segment 1 of UMRNOM (refer to blue bars in Figure 5c,d). N1Ox and S1Ox class compounds are polar and hence are eluted earlier in reversed-phase LC separation. The difference between the UV chromatograms in Figure 5a can be explained by the compositional differences identified by LC QPD FT-ICR MS in Figures 5b−d. The larger UV intensity of UMRNOM in segment 1 must have been caused by the larger abundances of N1Ox and S1Ox class compounds observed by FT-ICR MS. Similarly, the larger mass spectral abundances of Ox class compounds observed from SRFA must have resulted in the larger UV intensity of SRFA in segment 4. The nitrogen and sulfur contents of UMRNOM are higher than those of SRFA (refer to Table S1), and hence, this agrees with the results described above. Therefore, LC UV MS data suggest that UMRNOM is more heterogeneous and contains a greater number of polar compounds such as N1Ox and S1Ox class compounds, whereas SRFA is mainly composed of O x class compounds. Quantitation is very difficult for NOM analysis, especially when an ionization method or detection method (including UV) will have inherent biases. The data presented in Figure 5 clearly show that an orthogonal approach combining UV and mass spectral data is possible by applying LC QPD FT-ICR MS for the NOM analysis. The combined approach will be beneficial for quantitative comparison of NOM samples. The samples consumed to obtain the data presented in Figure 5 can be calculated from the sample concentration and amounts of injected samples. As mentioned in the Experimental Section, 10 μL of the samples prepared at a concentration of 200 μg/mL was used to obtain the LC QPD FT-ICR MS data, implying that only 2 μg of the NOM samples was used for the analyses.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.9b00689.

■

The effects of shimming on the mass spectra obtained by QDP FT-ICR MS, broad-band and expanded spectra of SRFA under different MS conditions, broad-band spectra of SRFA obtained by (a) LC DPD FT-ICR MS and (b) DI FT-ICR MS, the Ox class distributions obtained from each segment of LC QPD FT-ICR MS data, Kendrick mass defect plots generated from LC QDP FT-ICR MS spectra, LC QPD FT-ICR MS data obtained from UMRNOM, elemental compositions of SRFA and UMRNOM samples used in this study, gradient elution conditions for the LC-MS analysis of the NOM samples, and the number of identified Ox class compounds with LC QPD MS, LC DPD MS, and DIMS methods (PDF)

AUTHOR INFORMATION

Corresponding Author

*S.K.: e-mail, [email protected]; tel, +82-53-950-5333. ORCID

Sunghwan Kim: 0000-0002-3364-7367 Author Contributions

The manuscript was written through contributions of all authors. Notes

■

The authors declare no competing financial interest.

■

CONCLUSIONS This study demonstrates that LC QPD FT-ICR MS is an effective analytical method to study NOM at the molecular level. The improved scanning speed without compromising on the resolving power significantly increased the number of compounds observed from NOMs by FT-ICR MS. As presented in Figure 1, the detection time using QPD is 2 times shorter than that using DPD with the same resolving power. It results in increased number of scans in the given LC separation and hence increased S/N ratios of peaks detected by

ACKNOWLEDGMENTS The authors acknowledge support by a National Research Foundation of Korea (NRF) grant funded by the Korea Government (MSIP) Grant No. 2017R1A2B3003455, by the Korea Research Institute of Standsrds and Science (KRISS 2019 - GP2019-0010, and by the National Strategic ProjectFine particle of the National Research Foundation of Korea (NRF) funded by the Ministry of Science and ICT (MSIT), the Ministry of Environment (ME), and the Ministry of Health 7696

DOI: 10.1021/acs.analchem.9b00689 Anal. Chem. 2019, 91, 7690−7697

Article

Analytical Chemistry

(35) Verkh, Y.; Rozman, M.; Petrovic, M. Chemosphere 2018, 200, 397−404. (36) Patriarca, C.; Bergquist, J.; Sjöberg, P. J. R.; Tranvik, L.; Hawkes, J. A. Environ. Sci. Technol. 2018, 52, 2091−2099. (37) Sandron, S.; Davies, N. W.; Wilson, R.; Cardona, A. R.; Haddad, P. R.; Nesterenko, P. N.; Paull, B. Chromatographia 2018, 81, 203−213. (38) Lu, K.; Gardner, W. S.; Liu, Z. Environ. Sci. Technol. 2018, 52, 7182−7191. (39) Hawkes, J. A.; Dittmar, T.; Patriarca, C.; Tranvik, L.; Bergquist, J. Anal. Chem. 2016, 88, 7698−7704. (40) Nikolaev, E. N.; Boldin, I. A.; Jertz, R.; Baykut, G. J. Am. Soc. Mass Spectrom. 2011, 22, 1125−1133. (41) Park, S.-G.; Anderson, G. A.; Bruce, J. E. Int. J. Mass Spectrom. 2018, 427, 29−34. (42) Tolmachev, A. V.; Robinson, E. W.; Smith, R. D.; Leach, F. E.; Futrell, J. H.; Paša-Tolić, L. Int. J. Mass Spectrom. 2012, 325−327, 45−50. (43) Misharin, A. S.; Zubarev, R. A. Rapid Commun. Mass Spectrom. 2006, 20, 3223−3228. (44) Schweihard, L.; Lindinger, M.; Kluge, H. J. Int. J. Mass Spectrom. Ion Processes 1990, 98, 25−33. (45) Pan, Y.; Ridge, D. P.; Rockwood, A. L. Int. J. Mass Spectrom. Ion Processes 1988, 84, 293−304. (46) Pan, Y.; Ridge, D. P.; Wronka, J.; Rockwood, A. L.; Marshall, A. G. Rapid Commun. Mass Spectrom. 1987, 1, 120−121. (47) Cho, E.; Witt, M.; Hur, M.; Jung, M.-J.; Kim, S. Anal. Chem. 2017, 89, 12101−12107. (48) Hur, M.; Oh, H. B.; Kim, S. Bull. Korean Chem. Soc. 2009, 30, 2665−2668. (49) Lee, S.; Cho, Y.; Kim, S. Bull. Korean Chem. Soc. 2014, 35, 749−752. (50) Skoog, D. A.; Holler, F. J.; Crouch, S. R. Principles of Instrumental Analysis; Cengage Learning: 2017. (51) Hughey, C. A.; Hendrickson, C. L.; Rodgers, R. P.; Marshall, A. G.; Qian, K. Anal. Chem. 2001, 73, 4676−4681. (52) Wu, Z.; Rodgers, R. P.; Marshall, A. G. Anal. Chem. 2004, 76, 2511−2516. (53) Kim, S.; Kramer, R. W.; Hatcher, P. G. Anal. Chem. 2003, 75, 5336−5344. (54) Sleno, L. J. Mass Spectrom. 2012, 47, 226−236.

and Welfare (MOHW) (2017M3D8A1090658), Republic of Korea.

■

REFERENCES

(1) Brown, S.; Lugo, A. E. Biotropica 1982, 14, 161−187. (2) Kosobucki, P.; Buszewski, B. Natural Organic Matter in Ecosystems - a Review 2014, 13, 109. (3) Kononova, M. M. Soil organic matter: its nature, its role in soil formation and in soil fertility; Pergamon Press: Oxford, U.K., 1966. (4) Bronk, D. A.; Glibert, P. M. Mar. Biol. 1993, 115, 501−508. (5) Moneta, A.; Veuger, B.; van Rijswijk, P.; Meysman, F.; Soetaert, K.; Middelburg, J. J. Estuarine, Coastal Shelf Sci. 2014, 147, 78−86. (6) Letscher, R. T.; Hansell, D. A.; Kadko, D.; Bates, N. R. Mar. Chem. 2013, 148, 1−9. (7) Tulonen, T.; Salonen, K.; Arvola, L. Hydrobiologia 1992, 229, 239−252. (8) Klöcking, R.; Helbig, B.; Schötz, G.; Schacke, M.; Wutzler, P. Antiviral Chem. Chemother. 2002, 13, 241−249. (9) Helbig, B.; Klöcking, R.; Wutzler, P. Antiviral Chem. Chemother. 1997, 8, 265−273. (10) Klöcking, R.; Helbig, B.; Humic Substances in the Aquatic and Terrestrial Environment; Springer Berlin Heidelberg: Berlin, Heidelberg, 1991; pp 407−412. (11) van Rensburg, C. E. J. Phytother. Res. 2015, 29, 791−795. (12) Chen, L.; Shen, C.; Tang, X.; Chen, C.; Chen, Y. Environ. Sci. Pollut. Res. 2012, 19, 522−528. (13) Acton, Q. A. Advances in Hormone Replacement Therapy Research and Application, 2012 ed.; Scholarly Brief; Scholarly Editions: 2012. (14) Lavonen, E. E.; Gonsior, M.; Tranvik, L. J.; Schmitt-Kopplin, P.; Köhler, S. J. Environ. Sci. Technol. 2013, 47, 2264−2271. (15) Yang, X.; Guo, W.; Lee, W. Chemosphere 2013, 91, 1477−1485. (16) Reemtsma, T. J. Chromatogr. A 2009, 1216, 3687−3701. (17) D’Andrilli, J.; Dittmar, T.; Koch, B. P.; Purcell, J. M.; Marshall, A. G.; Cooper, W. T. Rapid Commun. Mass Spectrom. 2010, 24, 643− 650. (18) Bae, E.; Yeo, I. J.; Jeong, B.; Shin, Y.; Shin, K.-H.; Kim, S. Anal. Chem. 2011, 83, 4193−4199. (19) Cao, D.; Huang, H.; Hu, M.; Cui, L.; Geng, F.; Rao, Z.; Niu, H.; Cai, Y.; Kang, Y. Anal. Chim. Acta 2015, 866, 48−58. (20) Lu, Y.; Li, X.; Mesfioui, R.; Bauer, J. E.; Chambers, R. M.; Canuel, E. A.; Hatcher, P. G. PLoS One 2015, 10, No. e0145639. (21) Hertkorn, N.; Frommberger, M.; Witt, M.; Koch, B. P.; Schmitt-Kopplin, P.; Perdue, E. M. Anal. Chem. 2008, 80, 8908− 8919. (22) Cho, Y.; Ahmed, A.; Islam, A.; Kim, S. Mass Spectrom. Rev. 2015, 34, 248−263. (23) Cho, Y.; Witt, M.; Jin, J. M.; Kim, Y. H.; Nho, N.-S.; Kim, S. Energy Fuels 2014, 28, 6699−6706. (24) Cho, Y.; Birdwell, J. E.; Hur, M.; Lee, J.; Kim, B.; Kim, S. Energy Fuels 2017, 31, 7874−7883. (25) Furey, A.; Moriarty, M.; Bane, V.; Kinsella, B.; Lehane, M. Talanta 2013, 115, 104−122. (26) Ruddy, B. M.; Hendrickson, C. L.; Rodgers, R. P.; Marshall, A. G. Energy Fuels 2018, 32, 2901−2907. (27) Stenson, A. C. Environ. Sci. Technol. 2008, 42, 2060−2065. (28) Brown, T. A.; Jackson, B. A.; Bythell, B. J.; Stenson, A. C. J. Chromatogr. A 2016, 1470, 84−96. (29) Fang, Z.; He, C.; Li, Y.; Chung, K. H.; Xu, C.; Shi, Q. Talanta 2017, 162, 466−473. (30) Lv, J.; Zhang, S.; Luo, L.; Cao, D. Anal. Chim. Acta 2016, 948, 55−61. (31) Zark, M.; Dittmar, T. Nat. Commun. 2018, 9, 3178. (32) Kim, E.; Cho, E.; Ahmed, A.; Kim, Y. H.; Kim, S. Fuel 2017, 194, 503−510. (33) Sim, A.; Cho, Y.; Kim, D.; Witt, M.; Birdwell, J. E.; Kim, B. J.; Kim, S. Fuel 2015, 140, 717−723. (34) Kim, E.; Cho, E.; Moon, S.; Park, J.-I.; Kim, S. Energy Fuels 2016, 30, 2932−2940. 7697

DOI: 10.1021/acs.analchem.9b00689 Anal. Chem. 2019, 91, 7690−7697