Environ. Sci. Technol. 2001, 35, 1197-1201
Spectrophotometric Determination of Freshwater pH Using Bromocresol Purple and Phenol Red WENSHENG YAO AND ROBERT H. BYRNE* College of Marine Science, University of South Florida, St. Petersburg, Florida, 33701
The dissociation constants (KI ) [H+][I2-]/[HI-]) of two sulfonephthalein indicators (bromocresol purple and phenol red) were determined as function of temperature (10-30 °C) at zero ionic strength. Freshwater pH, on the free hydrogen ion concentration scale (molal units), can be precisely calculated from measurements of indicator absorbance ratios (λ2A/λ1A) using the following equations: pH ) pKI + log((R - e1)/(e2 - Re3)) and pKI ) pKI° - A∆Z2(µ1/2/(1 + µ1/2) - 0.3 µ), where R ) λ2A/λ1A, pKI ) -log KI, µ is the ionic strength, ∆Z2 ) 4, and values of A for 283 e T e 303 can be estimated from the equation: A ) 0.5092 + (T 298.15) × 8.5 × 10-4. For bromocresol purple (λ1 ) 432 nm, λ2 ) 589 nm), pKI° ) 5.226 + 378.1/T, e1 ) 0.00387, e2 ) 2.858, and e3 ) 0.0181. For phenol red (λ1 ) 433 nm, λ2 ) 558 nm), pKI° ) 5.798 + 666.7/T, e1 ) 0.00244, e2 ) 2.734, and e3 ) 0.1075. These two indicators can be used to make accurate pH measurements of freshwaters (river water, lake water, groundwater, rainwater, etc) within the range 4.5 e pH e 8.5. The precision of pH measurements using phenol red in well-buffered freshwaters is on the order of (0.001 or better.
Introduction The pH of natural waters is a master variable for describing the status of acid-base equilibria, chemical speciation, mineral solubility, and biological and kinetic processes. The demands of making appropriate environmental pH measurements vary considerably. For characterization of the carbonate system in seawater, pH precision, and accuracy on the order of (0.001 pH unit is necessary, while for some water quality purposes estimates on the order of (0.1 pH unit are sufficient. Most pH measurements in freshwaters are obtained potentiometrically using glass hydrogen-ion electrodes and reference electrodes with liquid junctions. In dilute solutions, such as rivers, lakes, or groundwaters, dilute NBS buffers (1, 2) have been widely used for electrode calibration. Nonthermodynamic assumptions (e.g., assumed constancy of residual liquid-junction potentials) are required in order to apply the NBS scale to waters having ionic strengths different from the ionic strength of the standardizing buffers: Variations in liquid-junction potentials between standardizing buffers and measured solutions are a major source of concern when precise pH measurements are required. Furthermore, even under the best of conditions, electrode potentials take several minutes to stabilize after an electrode is placed in a new solution, and long-term pH electrode drift can be as much as 0.01 pH h-1 (3). * Corresponding author e-mail:
[email protected]; telephone: (727)553-1508; fax: (727)553-1189. 10.1021/es001573e CCC: $20.00 Published on Web 02/07/2001
2001 American Chemical Society
Spectrophotometric techniques using pH indicators offer an alternative to potentiometric pH measurement methods. Indicators have the advantage of very rapid equilibration and obviate most of the problems associated with potentiometric measurements. Rapid equilibration in conjunction with the inherent precision (4) of spectrophotometric measurements allows examination of very small differences in the pH of water samples. Since spectrophotometric pH measurements are made via absorbance ratios, spectrophotometric pH is directly related to indicator (sulfonephthalein) molecular properties, and measurements do not require the use of calibrating buffers (5). Over the past 15 yr, spectrophotometric procedures developed for measurement of seawater pH (4-11) have improved the precision of at-sea measurements by more than an order of magnitude. However, few efforts have been devoted to extending spectrophotometric pH measurement protocols to freshwaters. In the present work, we examine the equilibrium and molar absorbance characteristics of bromocresol purple (BCP) and phenol red (PR) at low ionic strength. This work provides a quantitative basis for the use of BCP and PR for freshwater pH measurements.
Theory The equation used to estimate pH from spectrophotometric absorbance ratios, R, can be expressed as follows (4):
pH ) -log[H+] ) pKI + log
(
R - e1 e2 - Re3
)
(1)
where KI is defined as
KI )
[H+][I2-]
(2)
[HI-]
pKI ) -log KI; [HI-] and [I2-] are the concentrations of protonated and unprotonated indicator species, respectively; ei are molar absorbance ratios; and [H+] is the free hydrogen ion concentration (molal scale). The parameter R in eq 1 is a ratio of indicator absorbances at the absorbance maxima of I2- and HI-, and (R - e1)/(e2 - Re3) is equal to the ratio [I2-]/[HI-] (4). Absorption maxima of HI- and I2- for BCP are λ1 ) 432 nm and λ2 ) 589 nm (Figure 1); for PR, λ1 ) 433 nm and λ2 ) 558 nm. The symbols e1-e3 in eq 1 refer to indicator molar absorbance ratios at λ1 and λ2. For BCP, these molar absorbance ratios are given as
e1 ) 589HI/432HI
e2 ) 589I/432HI
e3 ) 432I/432HI
(3a)
where λI is the molar absorption coefficient of I2- at wavelength λ and λHI is the molar absorption coefficient of HI- at wavelength λ. Similarly, for PR the parameters e1-e3 are given as
e1 ) 558HI/433HI
e2 ) 558I/433HI
e3 ) 433I/433HI
(3b)
These parameters (ei) are obtained (e.g., Figure 1) by examining the absorbance properties of indicators under sufficiently alkaline conditions that the total indicator concentration (IT) is equal to [I2-] and at sufficiently acidic conditions that IT ) [HI-]. Indicator molar absorption coefficients are expressed in terms of absorbances, indicator concentrations, and path length (l) as follows: λI ) λAI/[I
2-
]l
λHI ) λAHI/[HI
-
]l
VOL. 35, NO. 6, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
(4) 9
1197
TABLE 1. pK2° Temperature Dependence for H2PO4Dissociation Calculated from pK2° ) 1979.5/T - 5.3541 + 0.019840T (14) temp (°C)
p K2 °
10 15 20 25 30
7.254 7.232 7.214 7.200 7.190
mendation for appropriate application of eq 9 is well within the e0.5 m limit of applicability suggested by Stumm and Morgan (12) and within the e0.1 m limit indicated by Millero and Schreiber (13). FIGURE 1. Absorbance vs wavelength for the basic (I2-) and acidic (HI-) forms of bromocresol purple (BCP). Indicator dissociation constants at zero ionic strength, KI°, were determined using phosphate buffer solutions (H2PO4-/HPO42-). The pH (-log[H+]) of each H2PO4-/HPO42buffer can be expressed as
pH ) pK2° - log([H2PO4-]/[HPO42-]) + log (γHPO4γH/γH2PO4) (5) where γHPO4, γH2PO4, and γH are activity coefficients of HPO42-, H2PO4-, and H+ at a given ionic strength, and [ ] denotes concentrations. The pH of the buffer solution expressed in terms of indicator characteristics is
pH ) pKI° + log((R - e1)/(e2 - Re3)) + log(γIγH/γHI) (6) where γI, γH, and γHI are activity coefficients of the indicated species. Combining eqs 5 and 6, we then obtain
pKI° ) pK2° - log([H2PO4-]/[HPO42-]) + log {(γHPO4γH/γH2PO4)/(γIγH/γHI)} - log((R - e1)/(e2 - Re3)) (7) At sufficiently low ionic strength (µ), activity coefficient terms are solely functions of ionic charge. Since the charges of HPO42- and I2- and H2PO4- and HI- are identical and µ e 0.016 M, it follows that γHPO4 = γI and γH2PO4 = γHI. Equation 7 then can be written as
pKI° ) pK2° - log([H2PO4-]/[HPO42-]) - log((R - e1)/ (e2 - Re3)) (8) BCP and PR equilibrium characteristics at 25 °C were examined using eq 8 over a range of H2PO4-/HPO42- buffer concentration and ionic strength. Using the ei and pKI° results obtained in this study, the pH of natural samples can be calculated from the following equations:
pH ) pKI° + log((R - e1)/(e2 - Re3)) - 4A(µ
1/2
/(1 +
µ1/2) - 0.3µ) (9) where R ) λ2A/λ1A, µ is the ionic strength, and
A ) 0.5092 + (T - 298.15) × 8.5 × 10-4
(10)
Equation 9 provides pH on the free hydrogen ion concentration scale with [H+] expressed in units of moles per kilogram of H2O. The final term in eq 9 accounts for the variation of I2-, HI-, and H+ activity coefficients with ionic strength using the Davies equation. We recommend use of this equation at low ionic strengths (µ e 0.05 M). This conservative recom1198
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 6, 2001
Experimental Section Absorbance measurements were obtained using an HP 8453 spectrophotometer. The temperature of the solution was controlled with a Neslab refrigerating circulator and a waterjacketed spectrophotometric cell holder. The acids, bases, and buffers used in this work (HCl, NaOH, KH2PO4, and Na2HPO4) were analytical grade reagents obtained from J. T. Baker. Stock solutions (3 × 10-3 mol dm-3) of BCP and PR were prepared by dissolving their salts (Fisher Scientific) in deionized water. Components of the phosphate buffer, KH2PO4 and Na2HPO4, were dried at 105 °C for 2 h before weighing. Concentrations of indicators in this work ranged from 3 × 10-6 to 3 × 10-5 mol dm-3. The molar absorbance ratios of e1-e3 for BCP and PR were calculated using eq 3 and measurements of λAHI and λAI (Figure 1): An indicator solution in deionized water was titrated with HCl and NaOH until the measured absorbances at the two wavelength maxima were nearly identical (λ1A ≈ 2λ2A) whereupon [HI ] ≈ [I ]. The pH of this solution was concurrently monitored with an Orion combination pH electrode (model 810200) connected to a Corning model 130 pH meter in the absolute millivolt mode. The solution was then acidified sufficiently that the indicator absorbance at λ1 reached a maximum value, whereupon [I2-] , [HI-] and [HI-] ) IT. The indicator absorbance at λ1 under this condition corresponded to λ1AHI. At constant pH (electrometric pH), the indicator concentration was increased by a factor of 10 (to increase the absorbance of HI- at wavelength λ2) and λ2AHI was measured. The parameter e1 was then calculated as
e1 ) (λ2AHI/[HI-]2)/(λ1AHI/[HI-]1)
(11)
where [HI-]2 ) 10 [HI-]1. The absorbances λ1AI and λ2AI of an indicator solution were measured at pH g 12. Under these conditions [I2-] ) IT. The parameter e2 is then calculated as
e2 ) (λ2AI/[I2-])/(λ1AHI/[HI-])
(12)
and e3 is calculated as
e3 ) (λ1AI/[I2-])/(λ1AHI/[HI-])
(13)
where λ1AHI was determined as described above. The influence of temperature on KI° was measured by monitoring R values of each indicator in a H2PO4-/HPO42buffer solution at different temperatures. At 25 °C and zero ionic strength, pK2° appropriate to the second dissociation constant of phosphoric acid is equal to 7.200 (14). Buffer solutions with concentration ratios [H2PO4-]/[HPO42-] equal to 10 and 5 were used for KI° determinations of BCP, and a buffer ratio equal to 1 was used for KI° determinations of PR. Using measured R values and the Table 1 temperature
TABLE 2. Molar Absorbance Ratios for BCP and PR indicator
e1
e2
e3
BCP PR
0.00387 0.00244
2.858 2.734
0.0181 0.1075
TABLE 3. pKI° of BCP at 25 °C and Zero Ionic Strength ionic strength
[H2PO4-]/[HPO42-]
pKI°
0.0163 0.0163 0.0160 0.0160 0.0100 0.0080 0.0040
0.0125 M/0.00125 M 0.0125 M/0.00125 M 0.010 M/0.002 M 0.010 M/0.002 M 0.00625 M/0.00125 M 0.005 M/0.001 M 0.0025 M/0.0005 M
6.493 6.493 6.499 6.499 6.492 6.491 6.492
TABLE 4. pKI° of PR at 25 °C and Zero Ionic Strength ionic strength
[H2PO4-]/[HPO42-]
pKI°
0.010 0.010 0.010 0.005
0.0025 M/0.0025 M 0.0025 M/0.0025 M 0.0025 M/0.0025 M 0.00125 M/0.00125 M
8.032 8.032 8.030 8.031
TABLE 5. Temperature Dependence of pKI° (µ ) 0) T (°C)
log KI° (BCP)
log KI° (PR)
10 15 20 25 30
6.563 6.536 6.514 6.494 6.474
8.155 8.110 8.071 8.032 8.000
FIGURE 2. Temperature dependence of bromocresol purple (a) and phenol red (b) dissociation constants.
e1 ) 0.00244
e2 ) 2.734
e3 ) 0.1075
and dependence for K2° of phosphoric acid (14), indicator KI° at different temperatures were calculated from eq 8. Finally, through iterative calculations, e1 was corrected for very small absorbance contributions from I2- at λ2. These corrections involved spectrophotometric calibration of the pH electrode and subsequent use of electrometric measurements to calculate pH when λ2AI and R were too small for accurate spectrophotometric pH measurements ([I2-] , [HI-]).
Results and Discussion
pKI°(PR) ) 5.798 + 666.7/T
(16)
For BCP (λ1 ) 432 nm, λ2 ) 589 nm), the following terms are used in eq 9:
e1 ) 0.00387
e2 ) 2.858
e3 ) 0.0181
and
pKI°(BCP) ) 5.226 + 378.1/T
The ei results for BCP and PR are given in Table 2. These values were obtained at 25 °C. Temperature was found to have an insignificant effect on the ei for both BCP and PR between 10 and 30 °C. Tables 3 and 4 present the KI° results for BCP and PR obtained at different ionic strength and 25 °C. The average pKI° value obtained for BCP at 25 °C is pKI° ) 6.494 ( 0.003. The average pKI° value for PR at 25 °C is pKI° ) 8.032 ( 0.001. Table 5 and Figure 2 show the temperature dependence of BCP and PR dissociation constants between 10 and 30 °C. Data were fitted by the following equations:
BCP
pKI° ) (378.1 ( 10.2)/T + 5.226 ( 0.035
(14)
PR
pKI° ) (666.7 ( 14.9)/T + 5.798 ( 0.051
(15)
where T ) t + 273.15. The standard deviation of the data fit represented by eq 14 is 0.0016. The standard deviation of the data fit represented by eq 15 is 0.0020. For PR (λ1 ) 433 nm, λ2 ) 558 nm), the following terms are used in eq 9:
(17)
In the application of spectrophotometric procedures for measurement of pH, it is important to note that measurements involving eqs 9, 16, and 17 are based on indicator molecular properties. If measured R values are recorded and cataloged along with temperature and ionic strength for all measurements, then any subsequent refinements in ei, pKI°, or the ionic strength correction term will allow simple refinement of calculated pH values without influencing, in any way, the precision of the measurements. As a means of illustrating the use of spectrophotometric procedures for pH measurements of freshwaters, samples were collected from the Hillsborough River within Hillsborough River State Park (Hillsborough County, FL): (i) Samples were collected directly in 10-cm path length cells. Cells were rinsed with riverwater three times and were finally filled by submerging each cell slowly through the air/ water interface. With each cell’s optical faces directed toward the surface, the cells were filled without creating bubbles. Each cell was sealed with Teflon caps while submerged. Cells were returned to the lab and analyzed within 90-180 min after sampling. Through previous experience, it was anticiVOL. 35, NO. 6, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1199
TABLE 6. Measurements of Hillsborough River Water pH Using Phenol Red over a Range of Indicator Concentrations (IT)a sample no.
1
2
3
4
5
pH (average) IT )1 µM, ∆pH IT )2 µM, ∆pH IT )3 µM, ∆pH
7.6468 -0.0005 -0.0005 +0.0010
7.6549 -0.0005 +0.0008 -0.0003
7.6466 +0.0012 +0.0009 -0.0022
7.6520 -0.0011 +0.0005 +0.0007
7.6565 -0.0010 +0.0012 -0.0001
a Each column shows the average pH and deviations (∆pH) from the sample average.
pated that PR would serve as an appropriate indicator for these samples. (ii) The ionic strength of these samples was measured using a DX-500 Dionex ion chromatograph. The major ion composition of the river-water was determined as (Na+)T ) 0.95 mM, (K+)T ) 0.11 mM, (Mg2+)T ) 0.22 mM, (Ca2+)T ) 1.31 mM, (Cl-)T ) 0.85 mM, (SO42-)T ) 0.36 mM, and (HCO3-)T ) 2.54 mM (where (HCO3-)T was determined from the cation vs anion charge balance). Carbonate ion concentrations are less than 1% of (HCO3-)T at the measured pH of the riverwater. Subsequently, the ionic strength was calculated as µ ) 6.02 × 10-3, and the pH of each sample at 25.0 °C was calculated as pH ) pKI° - 0.1430 + log ((R - e1)/(e2 - Re3)). (iii) To examine the inherent measurement reproducibility of these pH determinations, PR was added to each of five cells in three steps. A PR titrant (3 × 10-3 M) that had been adjusted to R = 1 was added to each 30 cm3 spectrophotometric cell successively to produce concentrations of 1 × 10-6, 2 × 10-6, and 3 × 10-6 M. These successive additions provide insight into the magnitude of the pH perturbations that can be created by indicator additions. The results of these measurements are given in Table 6. The standard deviation of the 15 measurements in five groups shown in Table 6 was calculated (15) as SD ) (0.0012. This estimate of the precision of spectrophotometric pH measurements for these samples is consistent with the precision of seawater pH measurements obtained using m-cresol purple (4). In measurements of many hundreds of replicate seawater samples, the calculated seawater pH measurement precision is (0.0004. The somewhat poorer precision observed in the riverwater pH measurements may be attributable to the extra sample handling in these measurements. Whereas, in the riverwater measurements, sample cells were opened and resealed three times for replicate measurements, seawater replicates are obtained from a single homogeneous source and each cell is opened, injected, and resealed only once for comparison of replicate samples. Consequently, we interpret our riverwater replicate measurements as indicating that the inherent pH measurement precision for these samples is (0.0012 or better. (iv) As an assessment of the precision of pH measurements including the natural variability of sample pH (i.e., inhomogeneity of the riverwater), an additional nine samples were processed identically by bringing the indicator concentration of each to 3 × 10-6 M. The average pH of these samples was 7.6484, and the standard deviation was (0.0027. The substantially larger standard deviation due to natural variations in riverine pH indicates that an inherent measurement precision of (0.0012 or better is sufficient for observation of small natural pH variations attributable to respiration/photosynthesis, atmospheric exchange, etc. at this location. The precision of measurements of seawater pH is expected to exhibit substantial consistency because the alkalinity of seawater is consistently on the order of 2 × 10-3 M. This is not necessarily true for freshwater samples. Consequently, for freshwater measurements in which pH may be weakly buffered, it is advisable to assess the extent of sample pH 1200
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 6, 2001
perturbations that are induced by indicator additions. If the absorbance ratio R changes substantially between the indicator stock solution and the indicator diluted within a sample, then the ratio of I2- and HI- concentrations has changed through uptake or release of H+. In cases where high precision and accuracy is desired, indicator-induced pH perturbations can be assessed through stepwise indicator additions. Indicator-induced pH perturbations are directly proportional to indicator concentrations. If significant systematic pH perturbations are observed, then direct observation of stepwise perturbations will allow linear extrapolation to a sample’s initial pH. For samples that are found to be weakly buffered it is advisable to minimize the amount of indicator added to each sample. As such, the use of 10-cm path length cells is highly recommended. The high precision and accuracy of spectrophotometric pH measurements should facilitate observations of subtle variations in the equilibrium characteristics of natural aqueous systems. Variations of interest in surface waters include, for example, diurnal CO2 system cycles driven by photosynthesis and respiration. In subsurface waters, spectrophotometric pH measurements should facilitate observations of small pH perturbations caused by mineral dissolution and allow assessment of evolving aqueous saturation states. The precision and accuracy of spectrophotometric pH measurements (e.g., ≈ (0.001) should, in some instances, be viewed as excess capability. In many, if not most instances, measurement accuracy on the order of (0.02 pH unit should represent a welcome extension of current capabilities. Viewed in the context of requirements for pH accuracy on the order of (0.02, spectrophotometric pH measurements are very robust. For the Hillsborough River waters sampled in this study, pH discrepancies as large as (0.02 substantially exceed any perturbations that could be induced by indicator additions. For water samples with lower pH (i.e., pH nearer the pK1 of carbonic acid), buffer intensities are inherently higher and, even without a detailed account of indicatorinduced perturbations, measurements accurate to well within (0.02 can be obtained for waters with alkalinities much lower than those of the Hillsborough River. Thus, while our suggested procedure involving stepwise indicator additions will allow corrections for indicator-induced pH perturbations in poorly buffered waters, perturbations as large as (0.02 pH unit should be uncommon. Possible artifacts that can give rise to spurious pH readings in addition to indicator-induced perturbations, include poor temperature control, indicator interactions with reactive trace metals, and biological processes (e.g., respiration) that occur between sampling and measurement. It has been noted previously (4) that, for samples buffered by the carbonate system, pH variation with temperature is of a similar magnitude and has the same direction as the change in indicator pKI with temperature. Consequently, even temperature perturbations as large as 2 °C will generally influence calculated pH by less than (0.01 pH unit. Indicator interactions with trace metals should generally be quite small. For most water samples, the concentrations of reactive metals will be much lower than indicator concentrations and, due to both organic and inorganic complexation of reactive metals in solution, the significance of metal-sulfonephthalein interactions will be further reduced. Furthermore, if indicators are used in samples having unusually high concentrations of reactive metals (e.g., [M]T g 1% IT), simple tests involving metal additions to buffered solutions containing sulfonephthalein indicators can demonstrate the extent to which metal-indicator interactions might perturb indicator response to pH. In addition, in such cases, the use of 1 cm cells and 10 times higher indicator concentrations can reduce the influence of metal ion complexation if it is observed. Problems with respiration-induced pH perturbation can be obviated
by prompt analysis. Field portable spectrophotometers are becoming available (16) for both prompt sample analysis and autonomous long-term in-situ monitoring. Finally, it is noted that while the two indicators calibrated in the present study allow freshwater pH measurement between about 4.5 and 8.5, PR is used to the best advantage in the range 6.8 e pH e 8.3, and BCP is most effectively used in the range 5.3 e pH e 6.8. For pH measurement above 8.3 in freshwater, it would be useful to calibrate additional indicators including cresol red, m-cresol purple, and thymol blue. At pH less than 5.3, bromocresol green would serve as a useful indicator. For precise pH measurements at the lowest effective range of PR and the highest effective range of BCP (pH ∼ 6.8), the characteristics of bromolthymol blue are ideally suited.
Acknowledgments We gratefully acknowledge the helpful comments of Dr. Andrew G. Dickson and two anomymous reviewers. This work was supported in part by contract OCE-9522878 from the National Science Foundation.
Literature Cited (1) Bates, R. G. Determination of pH, theory and practice; Wiley: New York, 1973. (2) Bates, R. G. The nature of seawater; Physical Chemistry Science Research Report 1; Dahlem Conference, 1975; p 315.
(3) Culberson, C. In Marine Electrochemistry: A practical introduction; Whitfield, M., Jagner, D., Eds.; John Wiley & Sons: New York, 1981; p 187. (4) Clayton, T. D.; Byrne, R. H. Deep-Sea Res. 1993, 40, 2115. (5) Robert-Baldo, G. M.; Morris, J.; Byrne, R. H. Anal. Chem. 1985, 57, 2564. (6) Byrne, R. H.; Robert-Baldo, G.; Thompson, S. W.; Chen, C. T. A. Deep-Sea Res. 1988, 35, 1405. (7) Byrne, R. H.; Breland, J. A. Deep-Sea Res. 1989, 36, 803. (8) King, D. W.; Kester, D. R. Mar. Chem. 1989, 26, 5. (9) Clayton, T. D.; Byrne, R. H.; Breland, J. A.; Feely, R. A.; Millero, F. J.; Campbell, D. J.; Murry, P. P.; Bobert, M. L. Deep-Sea Res. 1995, 42, 411. (10) Zhang, H.; Byrne, R. H. Mar. Chem. 1996, 52, 17. (11) Byrne, R. H. Anal. Chem. 1987, 59, 1479. (12) Stumm, W.; Morgan, J. J. Aquatic Chemistry, 2nd ed.; Wiley: New York, 1981. (13) Millero, F. J.; Schreiber, D. R. Am. J. Sci. 1982, 282, 1508. (14) Bates, R. G.; Acree, S. F. J. Res. Natl. Bureau Stand. 1943, 30, 129. (15) Kolthoff, I. M.; Sandell, E. B.; Meenan, E. J.; Bruckenstein, S. In Quantitative Chemical Analysis, 4th ed.; Macmillan Co.: London, 1969; 1197 pp. (16) Kaltenbacher, E.; Steimle, E. T.; Byrne, R. H. Underwater Technology, Proceedings of the 2000 International Symposium, 2000; p 41.
Received for review August 10, 2000. Revised manuscript received December 18, 2000. Accepted December 22, 2000. ES001573E
VOL. 35, NO. 6, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1201
