Grantee Research Project Results
Final Report: Molecular Weight Separation and HPLC/MS/MS Characterization of Previously Unidentified Drinking Water Disinfection By-Products
EPA Grant Number: R826834Title: Molecular Weight Separation and HPLC/MS/MS Characterization of Previously Unidentified Drinking Water Disinfection By-Products
Investigators: Minear, Roger A. , Barrett, Sylvia
Institution: University of Illinois Urbana-Champaign , Metropolitan Water District of Southern California
EPA Project Officer: Hahn, Intaek
Project Period: November 1, 1998 through October 31, 2001
Project Amount: $363,591
RFA: Drinking Water (1998) RFA Text | Recipients Lists
Research Category: Drinking Water , Water
Objective:
The overall goal of this project was to develop new approaches for better characterization of disinfection byproducts (DBP) molecular weight profiles by using tandem mass spectrometry (MS/MS) techniques. A prerequisite to making such procedures meaningful is the development of pre-separation procedures that will simplify the mass spectral data. The specific objectives of this research project were to: (1) obtain size exclusion chromatography (SEC)-Cl profiles of chlorinated natural organic matter (NOM) samples, which is the basis for further fractionation, concentration, and MS/MS analysis; (2) separate/fractionate chlorinated NOM and better define the chlorinated high molecular weight (MW) DBPs with ultrafiltration (UF) and SEC techniques with a combination detection of 36Cl, ultraviolet, and dissolved organic carbon (DOC); (3) roughly estimate MW distribution, and Cl/C atomic ratios of high MW chlorinated DBPs from chlorinated NOM; and (4) examine the effects of contact time, pH, and NOM on the formation of high MW DBPs. Also investigated were: the instrument's ability to perform electrospray high resolution analysis for accurate mass measurements, the ionization suppression effect in the presence of Suwannee River Fulvic Acid (SRFA), and the preliminary solid phase extraction (SPE) of mixed standards.
Summary/Accomplishments (Outputs/Outcomes):
This research has addressed various aspects of separation and characterization of unknown halogenated DBPs, especially high MW DBPs with 36Cl-UF-SEC-Electrospray (ESI)/MS/MS. A summary of the major findings of this research follows.
DBPs generated from chlorination, chloramination, ozonation, and chlorine dioxide treatment were characterized and compared. DBPs examined included 4 trihalomethanes (THMs), 9 haloacetic acids, 4 haloacetonitriles, 2 haloketones, chloropicrin, total organic halides (TOX), bromated fraction of TOX (TOBr), chlorinated fraction of TOX (TOCl), 13 aldehydes, and bromate. The primary findings are:
- Chlorination produced the largest amounts of organic halogenated DBPs (indicated by total trihalomethane (TTHM), trihalogenated acetic acids [HAAs], TOX, TOCl, TOBr, etc.). Chloramination and chlorine dioxide treatment greatly reduced the organic halogenated DBPs, compared with chlorination. Ozonation produced the lowest levels of organic halogenated DBPs but the highest levels of aldehydes and bromate. The contributions of known DBPs to TOX, TOCl, and TOBr formed by using different disinfectants were in the range of 8.2-60 percent.
- TOX level is a function of disinfectant: Cl2 > NH2Cl > ClO2 > O3.
- It appears that the reaction of humic substances with small amounts of free chlorine in equilibrium with NH2Cl constitutes an important pathway, if not the only one, for the formation of TOCl during chloramination.
- The yields of TOBr and total aldehydes produced from using each disinfectant were found to be directly related to the redox potential corresponding to each disinfectant.
Several issues related to the separation of high MW DBPs with UF and SEC were examined. These included: (1) whether low-MW DBPs and inorganic ions can be flushed out of a UF cell with a 500-dalton (Da) membrane; (2) whether low-MW DBPs can bind to large molecules such as SRFA; and (3) the elution behavior of the halogen-containing compounds with the SEC column. Major findings are:
- HAAs, chloride ions, and most phosphate ions could be effectively flushed out.
- HAAs, THMs, and chlorobenzoic acid did not bind to SRFA; trichlorophenol possibly may form a weak binding to SRFA, but such a binding could be dissociated when passing through the Frimmel SEC column.
- Halogen-containing compounds had longer retention time (RT) than their counterparts that do not contain halogen—for the same series of compounds, the more halogens, the longer the RT. Bromine-containing compounds had longer RT than their counterparts that contain the same number of chlorine atoms. Compounds that are not ionized at the pH of eluent had longer RT than their counterparts that are ionized.
UF and SEC coupled with radioisotope 36Cl were employed to characterize chlorine-containing DBPs, especially those high-MW DBPs in chlorinated NOM samples. The major findings are:
- SEC-36Cl profiles contained three major peaks and several minor peaks that represent the high-MW DBPs, Cl-, and relatively low-MW DBPs, respectively.
- UF effectively could remove almost all of the Cl- and relatively low-MW DBPs.
- Oxidation was the predominant reaction during chlorination.
- As contact time increased from 1 to 120 hours, the number of high-MW DBPs formed decreased slightly. As pH increased from 5.5 to 9.5, the formation of high-MW DBPs increased.
- Suwannee River humic acid produced DBPs with relatively higher MW compared to SRFA.
- High-MW, chlorine-containing DBPs could be adsorbed by activated carbon, but the adsorbed DBPs were not readily desorbed with 1:1 acetonitrile/water.
- The separation capabilities of three high-pressure liquid chromatography (HPLC) columns for chlorinated SRFA were compared.
- Most importantly, this work shows that 36Cl is a good indicator of whether an SEC fraction contains chlorinated DBPs, which is the basis for further separation, concentration, and MS/MS analysis.
Carrier-free, radioactive 36Cl was introduced into a SRFA sample to label the chlorine-containing DBPs. The major findings are:
- By combining the fractionation techniques of UF and SEC with the detection of 36Cl, UV, and DOC, the high-MW regions in the SEC-36Cl profiles of the chlorinated sample with and without UF were defined.
- SEC-UV and SEC-DOC profiles were found to be approximate alternatives to SEC-36Cl profiles for the high-MW region.
- The high-MW chlorinated DBPs were highly dispersed, with an average MW of around 2,000 Da based on calibration with polystyrene sulfonate standards. The Cl/C atomic ratios of the high-MW DBPs approximately were a constant (0.025), which is much lower than the ratios of the commonly known chlorinated DBPs.
A better understanding of the use of ESI for analyzing complex mixtures such as NOM was achieved. Optimized experimental conditions and calibrations were obtained for various functions, including magnet scans, voltage scans, high-resolution experiments, and MS/MS experiments using chemical standards such as dissolved combined amino acids, dichloropropane, and cardiopulmonary resuscitation. Also, appropriate MS conditions for finding specific structural information were evaluated. The parameters examined included flow rate, collision gas pressure, and collision energy. The major findings are:
- The formation of chloride ions by MS/MS requires relatively high collision energy (35-45 eV) and collision gas pressure, but obtaining structural information requires low collision energy.
- The conditions obtained were demonstrated to be effective in identifying chlorinated compounds in a simple, simulated mixture with relatively low concentrations of each component, and in a chlorinated SRFA sample.
- The mass spectra of SRFA solutions displayed an ion distribution envelope maximized at approximately 400 m/z with ion peaks at every mass. The maximum of the envelope spreads approximately from 350-500 m/z. An alternating mass pattern was observed with strong peaks at odd masses and weak peaks at even masses. The presence of half masses usually was observed between 200-400 m/z, indicating doubly charged ions. Above 1,000 m/z, the ion peaks became extremely weak in intensity and narrow in peak width, which probably indicated random ions or noise spikes. The mass spectra of two fractionated Colorado River water (CRW) fractions, the transfilic fraction (TPI) and the hydrophilic acids plus neutrals fraction (HPI), were similar to those observed for SRFA. Both fractions, however, had a few dominant ions superimposed on the ion distribution envelopes, and lowered maximums (approximately 350 m/z for the TPI fraction and approximately 250 m/z for the HPI fraction) of the ion distribution envelopes.
- The chlorinated SRFA solutions presented lower peak intensities compared to unchlorinated solutions. For the chlorinated SRFA solutions at a Cl2:total organic carbon (TOC) ratio of 1:10, the maximum of the ion distribution envelope shifted clearly toward lower masses. For the chlorinated SRFA solutions at a Cl2:TOC ratio of 1:1, the peak intensities were lowered so much that the original ion envelope was not obvious. An increase of solution pH after chlorination prior to MS analysis did not have a profound effect on the appearance of the mass spectra of chlorinated SRFA solutions. Chlorinated CRW TPI and HPI fractions showed much higher intensities than those of the chlorinated SRFA samples.
- MS/MS experiments of the chlorinated SRFA solutions and the chlorinated CRW TPI and HPI fractions showed the clear presence of both chloride ions (Cl-35 and Cl-37), together with other product ions in various precursor ions. A loss of CO2 (44 m/z) was observed from several precursor ions, which indicated the presence of carboxylic acids. Some product ions were recurring, such as 64, 80, and 81 m/z, which suggested that the corresponding precursor ions might share common structural features.
- Experiments at a high resolution of 5,000, showed strong evidence that multiple peaks might be present at one nominal mass. The phenomenon had been reported before using Fourier Transform Ion Cyclotron Resonance (FT-ICR) at a much higher resolution. This phenomenon would make structural elucidation difficult as a result of inaccurate elemental compositions provided by the system software, and product ions coming from more than one precursor ion because all the MS/MS experiments were done at a resolution of 1,000.
- Ion peaks in the UF-SEC fractionated chlorinated SRFA samples were extremely low in intensity and narrow in peak width above 800 m/z. Even though full-scan spectra were collected up to 4,000 m/z, most of the ions above 1,000 m/z probably were noise spikes. It was difficult to obtain MS/MS fragmentations of these fractions on the AutoSpec, and no chloride ions were observed.
- Average MW, weight-averaged MW (Mw), and number-averaged MW (Mn) of SRFA, CRW, TPI and HPI fractions, chlorinated SRFA samples, chlorinated TPI and HPI fractions, and UF-SEC fractionated chlorinated SRFA samples were calculated. Under the same experimental conditions, these values could be used to show trends within each set of samples, and possibly between different sample sets. Because most of the responses between 1,000 and 2,000 m/z on the AutoSpec came from background and instrument noise, however, Mw and Mn varied, even with the same set of raw data, when upper mass cutoff was set at either 1,000 m/z or 2,000 m/z.
- Two MS/MS instruments, Micromass AutoSpec tandem magnetic sector-orthogonal time of flight mass spectrometer, and Micromass Quattro triple quadrapole mass spectrometer, were used in the research. Both instruments have advantages and limitations. Whereas the AutoSpec has high-resolution capabilities and gives full range (10-1,000 Dalton) product ion mass spectra with the parent ion of interest present on the spectra, the Quattro provides better sensitivity in terms of chloride detection.
Chlorinated samples starting with high/low-dose SRFA with/without coagulation pretreatment were separated/fractionated by using UF and SEC techniques, respectively. These samples and sample fractions were characterized with the detection of UV and ESI/MS. To aid in the identification of chlorine-containing DBPs, specific ESI/MS/MS modes, including both the Parent and Daughter Ion Scans, were examined for both a standard mixture and some sample fractions. The main findings include:
- Coagulation greatly decreased the average MW of SRFA, whereas chlorination seems to increase it slightly.
- The effect of the initial concentration on the MW distribution of the final byproduct mixture was not significant if the initial carbon/chlorine ratio was the same.
- Many high-MW, chlorine-containing DBPs were found in the chlorinated sample fractions with and without coagulation.
- The results demonstrate that the ESI/MS/MS Parent Ion Scan is a powerful tool for picking and choosing the chlorine-containing compounds from a complicated mixture, and the ESI/MS/MS Daughter Ion Scan can be used to effectively confirm those compounds.
This research has enriched our knowledge of unknown halogenated DBPs, especially on the high-MW, chorine-containing DBPs (>500 Da). From the research results, several conclusions can be drawn:
- Chlorine from water chlorination is incorporated into NOM across the MW spectrum as demonstrated through chlorination of NOM surrogates, SRFA and Suwannee River humic acid, with 36Cl.
- Chlorination seems to increase the MW distribution of SRFA for a low Cl2/C ratio and decrease the MW distribution of SRFA for a high Cl2/C ratio. This needs to be confirmed, however, by preparing a series of samples with different Cl2/C ratio (what we had observed was that the distribution envelope shifted toward lower masses on the mass spectra at Cl2:TOC ratio of 1:10. The Mw and Mn calculations reflected this trend, as previously described in this report).
- High-MW, chlorine-containing DBPs can be screened out with the powerful ESI/MS/MS Parent Ion Scan. To obtain the structural information of those compounds, however, one of two methods may be used: (1) a sensitive MS instrument with higher resolution, such as FT-ICR MS; or (2) a different concentration approach, such as derivatization-extraction to eliminate salt and matrix effects.
Recommendations
Chlorine-containing DBPs with MW above 500 Da were the focus of this research. The MWs of most of the identified DBPs by gas chromatography/MS are below 250 Da. This suggests that DBPs with MWs in the range of 250-500 Da deserve more attention.
Future studies to obtain identification of individual components in complex mixtures such as chlorinated NOM may require the use of an FT-ICR MS because of its ultra-high-resolution capability combined with recent advancement of liquid chromatography compatibility and multiple mass spectrometry (MSn) capability.
Acknowledgments
Collaboration with the Center for Environmental Quality Control at Kyoto University, made possible through a grant from the National Science Foundation, International Cooperative Research Grant INT 97-26617, was of great value in the conduct of this research. Notably, we acknowledge the contributions of Drs. Yoshihisa Shimizu and Harumi Yamada, and their graduate students, at Kyoto University. Co-researchers at the Metropolitan Water District of Southern California, Cordelia Hwang and Yingbo C. Guo, were instrumental in the success of the research, as were graduate students Xiangru Zhang, Shinya Echigo, and Hongxia Lei at the University of Illinois-Urbana-Champaign.
Journal Articles on this Report : 5 Displayed | Download in RIS Format
Other project views: | All 25 publications | 10 publications in selected types | All 7 journal articles |
---|
Type | Citation | ||
---|---|---|---|
|
Zhang X, Minear RA. Decomposition of trihaloacetic acids and formation of the corresponding trihalomethanes in drinking water. Water Research 2002;36(14):3665-3673. |
R826834 (Final) R825956 (Final) |
Exit Exit |
|
Zhang X, Minear RA, Guo Y, Hwang CJ, Barrett SE, Ikeda K, Shimizu Y, Matsui S. An electrospray ionization-tandem mass spectrometry method for identifying chlorinated drinking water disinfection byproducts. Water Research 2004;38(18):3920-3930. |
R826834 (Final) |
Exit Exit |
|
Zhang X, Minear RA, Barrett SE. Characterization of high molecular weight disinfection byproducts from chlorination of humic substances with/without coagulation pretreatment using UF-SEC-ESI-MS/MS. Environmental Science & Technology 2005; 39(4):963-972. |
R826834 (Final) |
Exit |
|
Zhang X, Minear RA. Formation, adsorption and separation of high molecular weight disinfection byproducts resulting from chlorination of aquatic humic substances. Water Research 2006;40(2):221-230. |
R826834 (Final) |
Exit Exit |
|
Zhang X, Minear RA. Removal of low-molecular weight DBPs and inorganic ions for characterization of high-molecular weight DBPs in drinking water. Water Research 2006;40(5):1043-1051. |
R826834 (Final) |
Exit |
Supplemental Keywords:
tandem mass spectrometry, electrospray ionization, drinking water, chlorination, natural organic matter, Suwannee River fulvic acid, Suwannee River humic acid, 36Cl, electrospray, high-pressure liquid chromatography, HPLC, bromine, disinfection byproduct, molecular weight, halogen, Fourier transform ion cyclotron resonance mass spectrometry,, RFA, Scientific Discipline, Water, Waste, Chemical Engineering, Environmental Chemistry, Physics, Chemistry, Analytical Chemistry, chemical mixtures, Drinking Water, monitoring, alternative disinfection methods, public water systems, complex mixtures, human health effects, molecular weight separation, exposure and effects, chemical byproducts, disinfection byproducts (DPBs), exposure, community water system, HPLC, treatment, tandem mass spectrometry, toxicity, DBP risk management, water quality, drinking water contaminants, drinking water treatment, other - risk managementProgress and Final Reports:
Original AbstractThe perspectives, information and conclusions conveyed in research project abstracts, progress reports, final reports, journal abstracts and journal publications convey the viewpoints of the principal investigator and may not represent the views and policies of ORD and EPA. Conclusions drawn by the principal investigators have not been reviewed by the Agency.