Environ. Sci. Technol. 2003, 37, 1726-1730
Is a Universal Model of Organic Acidity Possible: Comparison of the Acid/Base Properties of Dissolved Organic Carbon in the Boreal and Temperate Zones J A K U B H R U Sˇ K A , * , † S T E P H A N K O ¨ HLER,‡ HJALMAR LAUDON,§ AND KEVIN BISHOP| Czech Geological Survey, Kla´rov 3, 118 21 Prague, Czech Republic, Laboratoire de Ge´ochimie CNRS (UMR 5569), Universite´ Paul Sabatier, 38 Rue des Trente-Six Ponts, 31400 Toulouse, France, Department of Forest Ecology, Swedish University of Agricultural Sciences, SE-901 83 Umeå, Sweden, and Department of Environmental Assessment, Swedish University of Agricultural Sciences, Box 7050, SE-750 07 Uppsala, Sweden
The acid/base properties of dissolved organic carbon (DOC) are an important feature of soil and surface waters. Large differences in the acid/base properties of DOC found by different studies might be interpreted as spatial and temporal differences in these properties. Different analytical techniques, however, may explain some of the differences. We used a combination of ion-exchange techniques, titration, and surface water chemistry data to evaluate DOC character from two substantially different areassthe relatively pristine boreal zone of Sweden and the heavily acidified temperate zone of the Czech Republic. We found a significantly higher site density (amount of carboxylic groups per milligram of DOC) for the Swedish sites (10.2 µequiv/mg of DOC ( 0.6) as compared to the Czech sites (8.8 µequiv/mg of DOC ( 0.5). This suggests a slightly higher buffering capacity for Swedish DOC. A triprotic model of a type commonly incorporated in biogeochemical models was used for estimating the DOC dissociation properties. For Swedish sites, the following constants were calibrated: pKa1 ) 3.04, pKa2 ) 4.51, and pKa3 ) 6.46, while the constants for Czech sites were pKa1 ) 2.5, pKa2 ) 4.42, and pKa3 ) 6.7. Despite differences in site density values, both models predict very similar dissociation and thus pH buffering by DOC in the environmentally important pH range of 3.5-5.0. This can be incorporated into models to make reliable estimates of the effect of organic acids on pH and buffering in different regions.
Introduction The acid/base properties of aquatic dissolved organic carbon (DOC) have been studied intensively during the past decades * Corresponding author e-mail:
[email protected]; telephone: +420 2 51085433; fax: +420 2 51818748. † Czech Geological Survey. ‡ Universite ´ Paul Sabatier. § Department of Forest Ecology, Swedish University of Agricultural Sciences. | Department of Environmental Assessment, Swedish University of Agricultural Sciences. 1726
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 9, 2003
with regard to the role of DOC in streamwater acidity and the balance between natural acidity and anthropogenic acidification (1-8). Studies focused on the acid/base properties of DOC have been carried out in Europe, predominantly Fenno-Scandia (5, 9-12) and Central Europe (6, 13) as well as in North America (3, 14, 15). These studies paint a conflicting picture. Some studies reported large differences in acid base properties, sometimes between quite similar and nearby localities (16, 17) or between seasons at the same site (18, 19). Other studies, however, found similar acid/base properties in waters from a variety of sites, sometimes far from each other as well as stable acid/base properties at the same site through different seasons (6, 20) or runoff events (11). Only a few researchers (3, 20-22) have successfully modeled the acid/base properties of organic acids in their studies, and models developed for one site seldom have been able to successfully predict the acid base/properties at other sites. It remains an open question as to whether the difference in measured acid/base properties and the models result from the use of different analytical techniques for organic acid separation and subsequent characterization or whether there is in fact a significant degree of similarity in the acid/base properties of DOC from site to site or from season to season. It is rare to find studies that compare acid/base properties from different regions using the same analytical method (e.g., ref 5). Since modeling the acid/base properties across a range of environments is important for a variety of research and management issues and the importance of correctly quantifying strong organic anions for charge balance calculations in natural waters has often been emphasized (e.g., ref 20), our study seeks to test the hypothesis that the acid/base properties of DOC are similar over large areas, seasons, and hydrological situations when similar characterization methods are used. We will do this by comparing the acid/base properties of streamwater and lake water DOC from the boreal zone of Sweden and the temperate zone in the Czech Republic across a range of sites, seasons, and flow conditions using identical analytical techniques.
Site Description Swedish Boreal Sites. Streamwater samples were taken from the Svartberget catchment (64°15′ N, 10°46′ E), a 50-ha headwater catchment located ca. 50 km inland from the Baltic Sea within an altitude range of 235-310 m above sea level. The mean annual temperature is 0 °C, and annual precipitation is 720 mm. The catchment is forested with Scotch pine (Pinus sylvestris) and Norway spruce (Picea abies). A total of 116 samples was taken during spring flood between April and June 1997 at three locations within the catchment, which had very different DOC concentrations and temporal variations (11). Another set of samples was taken during an experimental acidification/recovery experiment (10) of the stream in July 1997 (60 samples) and during an acidification/ alkalinization experiment (20) in September 1998 (80 samples). Forty lake water samples were taken during autumn 1997 from the boreal region of Sweden (22) between 59°10′ and 68°40′ N and 15°10′ and 21°10′ E. The general chemistry of these surface waters is summarized in Table 1. Czech Temperate Sites. Streamwater samples (145 samples) were taken monthly in the vicinity of the Lysina catchment, western Czech Republic (50°03′ N, 12°40′ E) from three high DOC streams between 1992 and 1996 (6) at elevations of 950, 930, and 813 m asl. Annual average temperature was 5° C, and the average precipitation amount was 950 mm. The area is forested with Norway spruce. An 10.1021/es0201552 CCC: $25.00
2003 American Chemical Society Published on Web 03/19/2003
TABLE 1. Summary of Measured Streamwater Chemistry in Both Regionsa Czech sites constituent
95th percentile
5th percentile
Na K Mg Ca Altotb Al Ib AlOb pHc HCO3 F SO4 Cl NO3 IS DOCd
136 28 82 168 31 16 23 4.97 bdl 14 415 53 40 1000 69
37 6 14 49 13 2 10 3.45 bdl 6 106 27 2 215 18
Swedish sites
median
95th percentile
5th percentile
median
51 14 34 78 22 5 14 3.70 bdl 9 221 41 6 496 42
252 21 94 257 20 3 13 7.34 153 8 412 93 5 1111 29
31 5 17 28 1 2 6 3.47 7 (20), we observed that organic acids dissociated and provided substantial buffering even above pH 7. Therefore, the results obtained from the inflection point titration did not measure all of the buffering capacity. By using a titration end point of pH 8.3, we titrated ca. 20% more functional groups as compared to our previous measurements of site densities in more acidic waters (6, 10, 11, 13, 22). Natural organic matter also has weaker acid groups buffering at least up to pH 11. While these functional groups certainly play a role for metal complexation, their proton buffering occurs outside the environmentally important pH range. Thus their buffering effect is negligible due to very high HCO3/CO3 concentrations. The final pH 8.3 as a titration end point was chosen arbitrarily as a practical estimate of carboxylic groups that also coincide with the titration exponent (pT) of the HCO3/CO3 buffering system. Briefly, the method involved passing an aliquot of sample through an H+-saturated cation-exchange resin. A 40-mL sample of the effluent was then titrated beyond the pH 8.3 end point with 0.005 M NaOH. During the titration, samples were constantly purged with N2. Ionic strength (IS in Table 1) was not adjusted by the addition of any salts. Ionic strengths in both regions were comparable, so it will not significantly affect titration results.
With this approach, the equivalents of base added to reach the end point (cTA) can be assumed to equal the sum of the concentrations of strong mineral acids (cSA) and carboxylic acids (cT). The total carboxylic group concentration can thus be estimated as the difference between cTA (measured) and cSA, which is the sum of the concentrations of SO42-, Cl-, and NO3- in the sample. Not all Al is removed by the cationexchange column, so some of the carboxylic groups may be masked by Al (23). Organic Acid Anions (RCOO-). The concentration of organic acid anions (RCOO-) was calculated from the discrepancy in the charge balance (e.g., refs 3, 6, and 17):
[RCOO-] ) (2[Ca2+] + 2[Mg2+] + [K+] + [Na+] + n[AlIn+] + [H+]) - (2[SO42-] + [NO3-] + [Cl-] + [F-] +
[HCO3-] + 2[CO32-]) (1)
where n is the calculated average charge of inorganic monomeric Al (for details, see ref 11). The [RCOO-] will be in µequiv/L () µmolc/L) if the other species concentrations are expressed in µmol L-1. For this calculation, the AASmeasured concentrations of Ca, Mg, Na, and K were assumed to equal ionic concentrations as it is common practice in the above-mentioned studies (3, 6, 17). Iron was not used in the charge balance calculation since Fe was assumed to occur in the form of uncharged Fe-organic complexes as has been documented in previous studies in the Czech sites (23). The average charge of AlI was determined from speciation calculations using ALCHEMI version 4.0 (24). SO4, Cl, and NO3 were determined using HPLC, while F was determined by ion-selective electrode after TISAB addition (6). Analytical error was assumed to be less than 5% for all inorganic constituents. Since the partial pressure of CO2 at concentrations commonly observed in streams can influence the pH above 4.5, both the ambient and aerated pH values of all samples were measured. Aeration of the pH samples was carried out for 20 min with standard compressed air (CO2 ) 10-3.3 atm-1) (22). For modeling purposes, only the aerated pH values were used. Bicarbonate concentrations were calculated using Henry’s law, with the aerated pH and the known pCO2. Dissolved Organic Carbon (DOC), Site Density, Charge Density, and Associated Errors. TOC was determined using platinum-catalyzed, high-temperature oxidation (TOC 5000, Shimadzu Corporation, Kyoto, Japan) on unfiltered samples. Since there is very little particulate carbon in these waters (
