Volume 91, Number 1, January-February 1986 Development of a Standard Reference Material for Rainwater Analysis William F. Koch, George Marinenko, and Robert C. Paule National Bureau of Standards, Gaithersburg, MD 20899 Accepted: October 9, 1985 This paper describes the development of Standard Reference Material, SRM 2694, "Simulated Rainwater," intended to aid in the analysis of acidic rainfall. Details of the formulation and preparation of the two levels of solutions (2694-I and 2694-II) are given. The 10 analytical techniques used to measure the 12 components in the solutions are described in brief. The data used in the statistical evaluation of the results are summarized and the recommended values for pH, specific conductance, acidity, fluoride, chloride, nitrate, sulfate, sodium, potassium, ammonium, calcium, and magnesium are tabulated. The instability of ammonium ion in acidic solutions is discussed. Recommendations for the use of SRM 2694, particularly with regard to the measurement of pH, are given. Key words: acid rain; acidity; ammonium; analytical chemistry; conductivity; measurement; pH; precip- 1. Introduction Wet deposition is monitored by various laboratories and agencies as part of national and international networks to record accurately the composition of rainfall. These efforts are intended to determine the extent of the problem of "acid rain," and to establish spatial and temporal trends. Discrepancies in data often occur due to differences in instruments and techniques. These discrepancies limit the conclusions which may be drawn About the Authors: William F. Koch and George Marinenko are chemists in NBS' Inorganic Analytical Research Division, part of the Bureau's National Measurement Laboratory in which Robert C. Paule, a physical scientist, serves. The work they describe was sponsored in part by the U.S. Environmental Protection Agency (National Acid Precipitation Assessment Program). from the data. To establish a common basis for chemical measurements in rainwater, a multi-year research effort has been established in the Inorganic Analytical Research Division of NBS; Center for Analytical Chemistry. This effort has resulted in the issuance of Standard Reference Material (SRM) 2694, Simulated Rainwater. The initial stages of the evolution of this SRM are described in detail in a previous article [1].' In that article, the early formulations of multicomponent solutions are given, and the problems with their stability are discussed. The stability problems inherent in these early formulations stem from both the complexity of the solutions and the use of glass ampoules as storage containers. The progression to polyethylene bottles and to simplified solutions, from which the transition and heavy metals were eliminated, resulted in solutions of greater stability and overall applicability to the measurements of pH, acidity, conductance, nitrate, and sulfate. These components are most critical to acid rain studies. 1 Figures in brackets indicate literature references. Associated difficulties with the measurement of pH and acidity are treated in other articles [2-4]. The pressing need of the scientific community for a common reference material for rainwater and the desire by us to field test such a material led to the production and distribution of a Research Material, RM 8409, Simulated Rainwater. RMs are distinguished from SRMS according to the definitions of NBS' Office of Standard Reference Materials [5]. In brief, RMs are high quality materials whose composition has been established by a single technique for each component. The composition of SRMS is certified after much more extensive testing involving at least two independent techniques for each component or analysis by a definitive method. Often an RM uncovers unforeseen problems, as was the situation in this case. This problem involved the long-term stability of the ammonium ion. The extent of the instability and its ramifications will be addressed below. The primary focus of this paper will be the preparation, analysis, and certification of SRM 2694. 2. The Preparation of SRM 2694 The decision to prepare simulated rainwater, rather than collecting natural rainfall, was based on the need to minimize contamination and unwanted components that would compromise the overall stability of the solutions. We also wanted to prepare two solutions of different concentrations of the various components, and we wanted to control the levels closely. Thus, simulated rainwater prepared by the dissolution of salts and acids in water was the best recourse. The production of this simulated rainwater involved careful coordination of several operations, including bottle cleaning, formulations, dilution, mixing, and bottling. Approximately 2000 bottles of each level were prepared. 2.1 Bottle Cleaning Based on earlier research [1], low density polyethylene (LDPE) bottles (60-mL capacity) were chosen for this project. Over 4000 bottles were cleaned using a rigorous procedure to minimize contamination. In brief, this cleaning procedure consisted of rinsing and soaking the bottles and caps for extended periods with filtered, distilled/deionized water. The bottles were then dried at 40 °C in a clean oven, and recapped until the filling operation was started. At all times, the bottles were kept away from areas with acid fumes. mulated in such a way so as to span a useful analytical range of concentrations of all components. The target values for the two levels are shown in table 1. It should be noted that the measurement of pH was the primary driving force behind the development of this SRM. Hence, the stability of the solutions with respect to pH was an overriding constraint. It has been found that unbuffered solutions at about pH 4.5 or above are extremely susceptible to fluctuations in pH and acidity due to absorption and desorption of atmospheric carbon dioxide. These processes occur even through the walls of the polyethylene bottles. For this reason, although it would have been desirable to issue a solution of pH 5.0, SRM 2694-I was targeted at pH 4.3 as a precaution. Chemical NaNO3 KNO3 CaCl2-2H2O MgSO4.7H2O NHẠC (NH4)2SO H2SO4 HNO3 HCI NaF Concentration 1 mg salt/g solution 1 mg salt/g solution 1 mg salt/g solution 1 mg salt/g solution 1 mg salt/g solution 1 mg salt/g solution 0.100 N (0.050 mol/L) 0.100 N (0.100 mol/L) 0.050 N (0.050 mol/L) 1 mg salt/g solution A 200-liter polyethylene, cylindrical tank, which had been cleaned according to the same procedure used for the bottles, was used in the final dilution of the appropriate amounts of each of the stock solutions. Graduations at 10-liter increments were made on the outside of the tank as it was being filled during the cleaning process. These marks served only as an indication of the volume, and were not intended to calibrate the volume accurately. A stirring motor with a teflon paddle was used to thoroughly mix the solutions. Figure 1 shows the apparatus in operation. Figure 1-Apparatus used in the preparation and dispensing of SRM 2694. Approximately 170 liters of 2694-I and approximately 150 liters of 2694-II were prepared in February 1985, according to the following procedure. The calculated amounts of each of stock solutions #1 through #9, (see table 3), were added to 100 liters of filtered, distilled/ deionized water in the tank. The resulting solution was mixed thoroughly. Then the calculated amount of solution #10 (sodium fluoride) was slowly added. It was necessary to delay the input of sodium fluoride to prevent the possible precipitation of calcium fluoride, which is extremely difficult to redissolve. Filtered, distilled/deionized water was then added to bring the total volume to the pre-established mark. The solution was thoroughly homogenized by intermittent vigorous stirring over a 24-hour period. Note that although care was taken to add exact amounts of each component and to dilute with the correct amount of water, there was no convenient way to accurately assess the final volume. Hence, the weights and volumes could not be used as an analytical measure of the concentrations of the various components. This was left to the analytical chemists and their myriad of techniques. 2.4 Bottling Before the bottling commenced, a sample of each level was analyzed by ion chromatography and potentiometry to verify that the target values had been met. The bottling of each level was completed in a single day. The bottles were filled manually, capped immediately, and placed sequentially in numbered cartons. Concurrent with the bottling operation,a homogeneity test was run. One out of every 120 bottles was pulled from the line and tested for specific conductance. The results, shown in tables 4 and 5, demonstrate that the two solutions are homogeneous. 3. Analysis of Simulated Rainwater After establishing the homogeneity of the solutions by the measurement of conductivity, the statistical design for the sampling and analysis of the other components was configured. Each component was to be determined in triplicate in each of three bottles (selected from the beginning, middle and end of the bottling operation) by each technique. Because of experimental exigencies of some techniques, this analysis design was not strictly adhered to in every case, but was followed whenever possible. The techniques used in the analysis of SRM 2694 were ion chromatography (IC), conductivity, potentiometry, coulometry, isotope dilution mass spectrometry (IDMS), spectrophotometry, laser enhanced ionization flame spectrometry (LEIS), flame emission spectrometry (FES), inductively coupled plasma (ICP), and flame atomic absorption spectrometry (FAAS). Details of the analytical procedures for each technique will be presented in a forthcoming publication. An overview of the methods will be given here. Ion chromatography was used to determine the concentrations of fluoride, chloride, nitrate, sulfate, sodium, potassium, and ammonium. Dual channel, dual column IC with hollow fiber chemical suppressor systems was employed. For each level and each component, three calibration points (peak height versus concentration) which bracketed closely the concentrations of the individual components were established. Chloride, nitrate, and sulfate were determined sequentially under one set of chromatographic conditions. A different set of conditions, involving an extremely weak carbonate eluent, was required to resolve the fluoride peak from the negative water-dip to allow the accurate determination of the fluoride concentrations. The cations were determined using a hydrochloric acid eluent. A concern about the stability of ammonium ion will be discussed below. -1 The specific conductance was measured using a diptype conductance cell of nominal cell constant 0.1 cm ̄ and an AC conductivity bridge operating at 1 kHz. Measurements for certifications were made at 25.0 °C by thermostating the solutions in a water-jacketed beaker, the outer chamber of which contained circulating constant temperature water to maintain the temperature at 25 °C. The exact cell constant of the cell was determined using 0.001 demal KCl, which has a specific conductance of 146.93 microsiemens per centimeter [μS/cm]. Measurements were also made at temperatures ranging from 20 °C to 28 °C to establish the temperature coefficient for the specific conductance of these solutions. The temperature coefficient for both levels was determined to be 1.5% per °C at 25.0 °C. Potentiometry was used to measure pH, fluoride, ammonium, and chloride. Measurements of pH were made with a combination glass electrode according to the procedure established at NBS. All measurements were corrected for residual liquid junction potential bias by normalizing to a dilute solution of sulfuric acid, whose pH had been determined accurately in a hydrogen cell without liquid junction. This cell is of the type used in the certification of the NBS pH buffers. Fluoride was measured with a fluoride ion-selective electrode after addition of a total ionic strength adjustment buffer. Standards which bracketed the concentrations of the rainwater samples were used to calibrate the measurement system. Ammonium ion was determined using an ammonia electrode. The method involves addition of concentrated base to the sample thereby liberating ammonia which diffuses through the semipermeable membrane of the electrode and is sensed by an internal glass |