HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (1) Citing Articles via Google Scholar Google Scholar Articles by Sikora, F. J. Search for Related Content PubMed Articles by Sikora, F. J. Agricola Articles by Sikora, F. J. Related Collections Soil pH Soil Analysis Soil Chemistry

Nutrient Management & Soil & Plant Analysis

A Buffer that Mimics the SMP Buffer for Determining Lime Requirement of Soil

Univ. of Kentucky, Division of Regulatory Services, Soil Testing Lab., 103 Regulatory Service Bldg., Lexington, KY 40546-0275

^* Corresponding author (fsikora{at}uky.edu)

	ABSTRACT

TOP ABSTRACT INTRODUCTION THEORETICAL CONSIDERATIONS MATERIALS AND METHODS RESULTS AND DISCUSSION SUMMARY Appendix A REFERENCES

 
The Shoemaker-McLean-Pratt (SMP) buffer used for determining lime requirement of soil contains chromium and p-nitrophenol, which classifies the solution as a hazardous waste. A buffer without hazardous chemicals producing the same pH as SMP buffer would eliminate hazardous waste and have no effect on agronomic interpretation. Chemicals chosen to replace chromium and p-nitrophenol were 2-(N-morpholino)ethanesulfonic acid monohydrate (MES) and imidazole. Acid titrations of SMP buffer and various mixtures of a new buffer were performed to duplicate the acid-base characteristics of the SMP buffer. The new buffer is adjusted to pH 7.70 and contains 69.6, 13.7, 31.4, 89.3, and 2000 mM of triethanolamine, imidazole, MES monohydrate, acetic acid, and KCl, respectively. Coefficients of determination (r²) for linear regressions of soil-buffer pH with the new buffer versus SMP buffer on 255 Kentucky and 87 North American Proficiency Testing (NAPT) soils were 0.974 and 0.967, respectively. The linear regression equation for new buffer versus SMP buffer lime recommendations in Mg ha^–1 was y = 0.226 + 0.994x with an r² of 0.935 for 158 Kentucky soils requiring lime and y = 1.42 + 0.876x with an r² of 0.876 for 27 NAPT soils requiring lime. Repeated soil-buffer pH measurements were made using 10 Kentucky soils and 11 NAPT soils over 150 d. Only one of the soil samples showed a significant drift in soil-buffer pH with time using the new buffer, and the drift was also observed using SMP buffer. With 88 measurements of soil-buffer pH using the new buffer on 11 NAPT soils, 2 measurements fell outside NAPT warning limits for SMP soil-buffer pH. The new buffer without hazardous chemicals mimicked the soil-buffer pH obtained with SMP buffer, provided lime recommendations similar to that with SMP buffer, and had a laboratory shelf life of at least 150 d.

Abbreviations: MAD, median absolute deviation • MES, 2-(N-morpholino) ethanesulfonic acid monohydrate • NAPT, North American Proficiency Testing • RCRA, Resource Conservation and Recovery Act • SMP, Shoemaker–McLean–Pratt.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION THEORETICAL CONSIDERATIONS MATERIALS AND METHODS RESULTS AND DISCUSSION SUMMARY Appendix A REFERENCES

 
BUFFER SOLUTIONS have been developed to react with soil acidity and aid the recommendation of lime requirement. The four most widely used buffers are Woodruff (Woodruff, 1948), SMP (Shoemaker et al., 1961), Adams-Evans (Adams and Evans, 1962), and Mehlich (Mehlich, 1976). In the 2004 reports of the NAPT program (Miller and Kotuby-Amacher, 2004), the average number of laboratories reporting values for Woodruff, SMP, Adams-Evans, and Mehlich buffers were 18, 58, 16, and 6, respectively.

The buffers used in soil testing were developed before federal laws regulating the disposal of hazardous waste. The Resource Conservation and Recovery Act (RCRA) was passed in 1976 by the U.S. Congress to improve waste management (Horinko, 2002). Hazardous waste management was further refined in 1980 with the passage of regulation controlling the disposal of chemicals considered hazardous due to ignitability, corrosivity, reactivity, or toxicity (USEPA, 1980a). Buffer constituents defined to be hazardous due to toxicity are p-nitrophenol in Adams-Evans, barium in Mehlich, and p-nitrophenol and chromium in SMP. The chromium in the SMP buffer, present as chromate (CrO₄^2–), is hexavalent and carcinogenic (USEPA, 1998). Since these buffers contain hazardous chemicals as defined by RCRA, a laboratory discarding 100 kg or more of the soil and buffer in a one-month period is considered a hazardous waste generator and has to follow certain protocols for hazardous waste disposal (USEPA, 1980b).

Studies have been conducted to develop alternative methods for making lime recommendations without the use of hazardous chemicals. Vaughan (2004) suggests replacing chromate with citric acid or succinic acid and p-nitrophenol with ethylenediamine or imidazole in the SMP buffer without specifying specific concentrations for the alternative chemicals. Wolf and Beegle (2005) have evaluated the use of a modified Mehlich buffer, with barium replaced by calcium, to replace their routine use of the SMP buffer. Huluka (2005) replaced p-nitrophenol with a potassium phosphate in the Adams-Evans buffer. The University of Georgia laboratory has discontinued use of the Adams-Evans buffer and is using direct titration with calcium hydroxide for estimating lime requirement (Liu et al., 2004, 2005). The studies on searching for alternative methods without hazardous constituents can be categorized under the emerging discipline of green chemistry (Anastas and Kirchhoff, 2002).

Replacing SMP buffer with a modified Mehlich buffer (Wolf and Beegle, 2005) is a viable option for a laboratory to eliminate the use of chromate and p-nitrophenol. However, the buffer pH values from the two buffers are not the same, which necessitates a lime calibration study or conversion to SMP buffer pH for making lime recommendations. The objective of this study was to develop a buffer to replace SMP buffer that would be free of hazardous constituents and would produce the same analytical pH value when reacted with soil. This type of replacement buffer can be readily introduced into a laboratory without modifying interpretation of the data for lime recommendations.

	THEORETICAL CONSIDERATIONS

TOP ABSTRACT INTRODUCTION THEORETICAL CONSIDERATIONS MATERIALS AND METHODS RESULTS AND DISCUSSION SUMMARY Appendix A REFERENCES

 
SMP Soil-Buffer pH as a Function of Soil Acidity
The SMP buffer contains four chemicals that buffer pH, namely triethanolamine (2.5 mL L^–1), p-nitrophenol (1.8 g L^–1), potassium chromate (3 g L^–1), and calcium acetate (2 g L^–1) (Shoemaker et al., 1961; Watson and Brown, 1998; Eckert and Sims, 1995; McLean, 1982). Calcium, supplied as calcium chloride dihydrate (53.1 g L^–1), is present to release exchangeable acidity to react with the buffer's bases. The pH of the buffer is adjusted to 7.50 with NaOH. The base forms of the four buffer chemicals react with soil acidity to reduce the pH from the buffer's initial value. The reduction in pH is directly related to soil acidity, and thus lime requirement. Shoemaker et al. (1961) reports 5 g of soil mixed with 5 mL of water to obtain a soil-water pH, followed by addition of 10 mL of SMP buffer to obtain a soil-buffer pH. Therefore, the SMP buffer is diluted with 1 part water to 2 parts buffer after being added to the soil-water suspension. Considering this dilution, the concentration of the four buffer chemicals in the liquid phase of the soil-water-buffer suspension are 12.6 mM for triethanolamine, 8.6 mM p-nitrophenol, 10.3 mM chromate, and 16.9 mM for acetate. When the SMP buffer is diluted the pH rises slightly. Fifteen observations of 20 mL of SMP buffer mixed with 10 mL of water yielded an average pH of 7.56 with a standard deviation of 0.04. Therefore, the initial pH in the liquid phase of the soil-water-buffer suspension before reaction with soil acidity is more accurately stated to be 7.56 rather than the pH of the buffer alone, which is 7.50. The ionic strength of the solution phase of the soil-water-buffer suspension is controlled by CaCl₂ in the buffer and is 0.72 M. An ionic strength of 0.5 M was used to obtain constants on chemical equilbria and activity coefficients in the theoretical development.

The following equilibrium considers the reaction that occurs with one of the SMP buffer components, namely triethanolamine, reacting with soil acidity.

[1]

The base form, on the left-hand side of the equilibria, accepts a proton to form the conjugate acid of the base, on the right-hand side of the equilibria. The proton can be directly available in soil or released from Al³⁺, Al hydrolysis products [Al(OH)²⁺ and Al(OH)₂⁺], oxides, clays, or organic matter. Smith and Martell (1975) report the equilibria constant for the reaction in Eq. [1] at an ionic strength of 0.5 as shown below.

[2]

The K is the equilibria constant defining the formation of (HOCH₂CH₂)₃NH⁺. Since Ka is the equilibria constant defining the reverse reaction which is the dissociation of (HOCH₂CH₂)₃NH⁺, log K = –log Ka = pKa. The square brackets designate concentration of chemical species in solution. Using the relationship that activity equals concentration times activity coefficient (_H+) and pH equals the negative logarithm of H⁺ activity results in

[3]

Using the Davies equation (Pankow, 1991) results in a value of 0.733 for _H+ at 0.5 M ionic strength and results in:

[4]

When pH equals 8.03, [(HOCH₂CH₂)₃NH⁺] = [(HOCH₂CH₂)₃N] and triethanolamine has its maximum potential to buffer pH on acid or alkali additions. At the initial SMP buffer pH of 7.56, the right-hand side of Eq. [4] is positive indicating the concentration of the conjugate acid is greater than the concentration of the base.

The total concentration of triethanolamine (T_TEA) in solution is equal to the concentration of the base and conjugate acid:

[5]

Solving Eq. [5] for the conjugate acid form, substituting the result into Eq. [3], and rearranging results in an equation expressing the concentration of the base as a function of total triethanolamine concentration and pH.

[6]

The difference between the concentration of base at the initial SMP pH of 7.56 and the concentration of base after the buffer reacts with soil to result in a soil-buffer pH is the concentration of base neutralized by soil acidity ([B neutr.]).

[7]

Triethanolamine is one of four bases in the SMP buffer. The concentration of base neutralized by soil acidity can be calculated for each base. The sum of neutralized bases equals the total concentration of soil acidity neutralized by the SMP buffer.

[8]

Equation [8] calculates the molar concentration of base, or soil acidity, neutralized in the volume of liquid in the soil-water-buffer suspension. Multiplying Eq. [8] by the volume of solution divided by weight of soil yields millimoles of soil acidity neutralized per gram of soil.

[9]

The theoretical development leading up to Eq. [9] is similar to theoretical calculations on pH titration of biomolecules containing several non-interactive protonation sites (Onufriev et al., 2001). The T_i values in Eq. [9] are 0.0126 M for triethanolamine, 0.0086 M for p-nitrophenol, 0.0103 M for chromate, and 0.0169 M for acetate. The pKa values for the chemical components in the SMP buffer at an ionic strength of 0.5 M are 4.50 for acetate (Martell and Smith, 1977), 5.81 for chromate (Smith and Martell, 1976), 6.85 for p-nitrophenol (Martell and Smith, 1977), and 7.90 for triethanolamine (Smith and Martell, 1975). The activity coefficient for H⁺ (_H+) is 0.733 at an ionic strength of 0.5 M according to the Davies equations (Lindsay, 1979). Using these values in Eq. [9] and plotting millimoles of soil acidity neutralized (g soil)^–1 as the independent variable and soil buffer pH as the dependent variable yields a theoretical response of soil buffer pH as a function of soil acidity neutralized (Fig. 1 ).

View larger version (24K): [in this window] [in a new window]   Fig. 1. Calculation of soil buffer pH for SMP buffer (20 mL buffer + 10 mL water + 10 g soil) as a function of soil acidity neutralized shown as closed circles. The calculation was performed according to Eq. [9] with soil acidity as the dependent variable and is the sum of the individual calculations shown for triethanolamine, p-nitrophenol, chromate, and acetate.

Glycerophosphate in the Mehlich buffer (Mehlich, 1976) was considered as a potential replacement compound. The beta form of glycerophosphoric acid has pKa values of 1.35 and 6.65 (at ionic strength = 0 M) (Martell and Smith, 1977) due to the protonation of a phosphate group bonded to glycerol. The Mehlich buffer has an initial pH of 6.60, which is less than the desired initial pH for a new buffer to mimic the SMP buffer. Glycerophosphate was not chosen for the new buffer because of potential alkaline hydrolysis of the compound in a buffer at pH 7.50 (DuBois, 1914; Castro and Rolston, 1977; Grzyska et al., 2002). A variety of sulfanilamide compounds, used as antibiotics in the pharmaceutical industry, are available with a wide range of pKa values (Boelema et al., 1982; Remko and von der Lieth, 2004). However, these compounds were not considered because of their expense.

Two chemicals to replace p-nitrophenol and chromium were selected from a list of buffers used in biochemical studies (Stoll and Blanchard, 1990) based on similarity of pKa values. Imidazole (pKa = 6.95) was chosen to replace p-nitrophenol (pKa = 6.85) and MES (pKa = 6.10) was chosen to replace chromate (pKa = 5.81). Imidazole is a common buffer used in analysis of proteins (Molina et al., 1996). MES is included in a list of buffers identified by Good and coworkers (Good et al., 1966; Good and Izawa, 1972; Ferguson et al., 1980) for use in biological studies and are commonly referred to as Good's buffers (Kandegedara and Rorabacher, 1999). The functional acidic group for imidazole and MES is a heterocyclic N atom. MES is a zwitterion since the protonated N atom is positive and a sulfonic group exists with a negative charge (Vega and Bates, 1976). MES is reported to not complex metals (Good and Izawa, 1972) because of steric hindrances (Yu et al., 1997; Kandegedara and Rorabacher, 1999). Contrary to these reports, Anwar and Azab (2001) have reported complexation of trivalent lanthanide ions by MES. Imidazole does have the ability to complex divalent cations (Kapinos et al., 1998). Values for complexation of Al by MES and imidizaole could not be found. Although there are some reports on complexation of metals by these compounds, the lack of multiple acidic functional groups should minimize the complexation ability for Al via chelation (Dynes and Huang, 1997; Tam and McColl, 1990; Hue et al., 1986). MES and imidazole are both readily available from chemical suppliers. Neither chemical is recognized as a hazardous chemical (USEPA, 1980). MES and imidazole fit the criteria to replace chromate and p-nitrophenol and were therefore selected for developing the new buffer.

Another consideration in developing a new buffer was to determine a range of SMP buffer pH values where linearity would be maintained on reaction with acidity. A buffer should maintain linearity on reacting with soil acidity so a predictive amount of acidity could be readily evaluated by a decrease in pH. The SMP buffer maintains linearity from an initial pH of 7.56 to 4.00. However, very few soils have SMP soil-buffer pH values as low as 4.00. A majority of soils analyzed from all over the world by MDS Harris Laboratories had SMP buffer pH values ranging from 6.40 to 7.20 (Vaughn, 2004). Soils from the NAPT program from the third quarter of 1999 through 2004 had a range in SMP soil-buffer pH from 5.52 to 7.74 (Miller and Kotuby-Amacher, 2004). In 15 yr of soil test data from the University of Kentucky, 99.9% of SMP soil-buffer pH values were above 5.30. Considering the common range of SMP soil-buffer pH values, the replacement buffer was developed to be linear from pH 7.56 to 5.30.

The volume of SMP buffer used is 10 mL added to 5 mL of water and 5 g of soil (Shoemaker et al., 1961; Watson and Brown, 1998; Eckert and Sims, 1995; McLean, 1982; Soil and Plant Analysis Council, 2000). The use of 20 mL of SMP buffer to 10 cm³ of soil has been reported elsewhere (Tran and van Lierop, 1993; van Lierop, 1990). In development of a new buffer, maintaining the same volume of buffer as used with the SMP buffer was not considered necessary. The important consideration was to ensure the same analytical result was obtained. To minimize volume of solution a laboratory would have to handle, the lime requirement test with the new buffer was designed with a buffer/water/soil ratio of 1:1:1 rather than 2:1:1 as used with SMP.

Potassium chloride was chosen as a salt to replace CaCl₂ in the SMP buffer. Potassium, as a univalent cation, has a lower replacement capacity than Ca²⁺ in exchanging acidity (Mengel et al., 2001). However, at a concentration of 1 M KCl in the solution phase of the proposed soil-water-buffer suspension (10 cm³ of soil, 10 mL of water, and 10 mL of buffer), the concentration of K⁺ would be the same as that used in the method to obtain exchangeable Al (Barnhisel and Bertsch, 1982) and the same concentration of the univalent NH₄⁺ ion used to determine total exchange capacity of soil (Hendershot et al., 1993). Potassium was also desired over Ca²⁺ because it was considered to have a lower probability of precipitating with anions in the buffer since the solubility products of most potassium salts are greater than the solubility product of the analogous calcium salt (Windholz et al., 1976).

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION THEORETICAL CONSIDERATIONS MATERIALS AND METHODS RESULTS AND DISCUSSION SUMMARY Appendix A REFERENCES

 
All pH measurements were made with a Beckman pH meter equipped with a Fisher accumet glass sensing electrode and an Orion single junction Ag/AgCl reference electrode. The pH meter was calibrated with two standardization buffers at pH 4.00 and 7.00 before measurements. All measurements were made at room temperature, which ranged from 22 to 27°C. All pH readings were recorded to two decimal places.

The University of Kentucky Soil Testing Program includes a laboratory in Lexington, KY and one in Princeton, KY. Tests on the new buffer were conducted by both laboratories and are identified as the Lexington or Princeton laboratory. Routine soil samples submitted to both laboratories were included in the study. Soils from the NAPT program were also included (Table 1). The NAPT program is an external quality control program that helps laboratories identify problems in their operation (Soil and Plant Analysis Council, 2000). Twenty soil samples are submitted to laboratories every year across North America. Laboratories report on results obtained from testing the samples and a report is generated containing a median and interlaboratory variance for all reported values.

The acidity was expressed on an assumed weight basis of soil so the unit could conform to the unit on the x axis in Fig. 1. With the volume of SMP buffer at 100 mL and water at 50 mL, hypothetical soil weight would be 50 g to maintain the same ratio of 2:1:1 of buffer/water/soil. Therefore, the conversion of mL of 0.5 M HCl titrant to millimoles soil acidity neutralized (g soil)^–1 was conducted with the following equation.

[10]

With the above calculation, the acidity provided by the titrant could be viewed as acidity provided by hypothetical soil and a direct comparison to the theoretical calculation in Fig. 1 could be made.

Determining Optimal Composition of the New Buffer
To determine the optimum concentration of components in the new buffer, Eq. [9] was fit to SMP buffer titration data in the pH range from 7.56 to 5.30 with triethanolamine, imidazole, MES, and acetic acid as chemicals buffering pH. TableCurve 2D (SPSS Inc., 1997) was used for the regression analysis with soil acidity neutralized (mmole g^–1) as the dependent variable and soil-buffer pH as the independent variable. The volume of water (mL), buffer (mL), and mass of soil (g) were all 10. The ionic strength of the new buffer diluted in a ratio of 1:1 with water is 1 M. The initial pKa values used for triethanolamine, imidazole, and acetic acids were those reported at 1 M ionic strength at 7.99, 7.31, and 4.60, respectively (Smith and Martell, 1975; Martell and Smith, 1977). The pKa of MES is reported to be 6.27 at 0 M ionic strength (Martell and Smith, 1982). The pKa of MES at 1 M ionic strength was estimated to be 6.50. The unknown variables in the fit of Eq. [9] to actual SMP titration data were the concentrations of triethanolamine, MES, imidazole, and acetic acid (Ti) in the solution phase of the buffer-water-soil suspension and were determined from regression analysis.

Once Ti values were estimated for a new buffer from the regression described above, 2 L of new buffer was created with double the Ti concentrations. The initial pH of the new buffer was adjusted with dropwise addition of 40% (w/w) NaOH or 50% (v/v) HCl until a mixture of 10 mL new buffer and 10 mL of water equaled a pH from 7.55 to 7.57 which is the initial pH of 20 mL of SMP buffer and 10 mL of water. The pH titration was conducted as described for the SMP buffer with the exception that 50 mL of the new buffer was diluted with 50 mL of water, rather than diluting 100 mL of buffer with 50 mL of water.

Once titration data was obtained for the first attempt on a new buffer, Eq. [9] was fit to the titration data of the new buffer to refine the values for pKa by considering the pKa values as unknowns in the regression analysis. Once pKa values were determined from the regression, Eq. [9] was fit to the SMP titration data to determine new Ti values with the refined pKa values. A second 2-L mixture for the new buffer was created with the latest Ti values and titrated. The titration data was used to further refine pKa and Ti values as described to arrive at a third and final composition for the new buffer.

The background salt in the buffer was first a combination of KCl and CaCl₂ in the first mixture to create an ionic strength similar to the SMP buffer at 1.08 M. Due to precipitation of CaCO₃ on alkali addition in the pH adjustment, KCl was used as the sole background salt in subsequent mixtures at an ionic strength of 2 M. The ionic strength was increased to 2 M to result in a 1 M concentration of KCl to react with exchangeable Al on 1:1 dilution of the buffer with water from the water pH measurement. A 1 M concentration of KCl is the concentration used to extract exchangeable Al in soil (Barnhisel and Bertsch, 1982).

The final composition of the new buffer was 69.6 mM triethanolamine, 13.7 mM imidazole, 31.4 mM MES, and 89.3 mM acetic acid, and 2.00 M KCl adjusted to pH 7.70 ± 0.01 (Appendix A). Seven titrations of the new buffer mixture were conducted. Fifty milliliters of new buffer was diluted with 50 mL of water and titrated with 0.5 M HCl. The pH of the new buffer was determined before and after dilution with water before the titration with HCl. Two titrations were conducted on one batch. Two titrations were conducted on a second batch. Three other titrations were conducted on three separate batches. As with the titrations of SMP buffer, the acidity was converted to units of millimoles soil acidity (g soil)^–1 using Eq. [10].

Comparing SMP Buffer and New Buffer in pH Measurements of Soil
Separate 10- and 2-L batches of the new buffer with constituents as described in the previous section and Appendix A were prepared. The SMP buffer was prepared according to Shoemaker et al. (1961) with the exception noted earlier in the section on pH titration of SMP buffer. Soils used to compare pH values for the SMP buffer and new buffer were routine soils submitted to the Lexington and Princeton laboratories and soils from the NAPT program. A total of 342 soils were tested including 198 from the Lexington laboratory, 57 soils from the Princeton laboratory, and 87 soils from the NAPT program. Soils from the NAPT program were those from the third quarter of 1999 through the fourth quarter of 2004 (Miller and Kotuby-Amacher, 2004). Replicate soil samples in the NAPT program during this time period were not analyzed.

Soils submitted to the Lexington and Princeton laboratories were prepared by drying at 38°C and grinding with a hammermill to pass through a 2-mm screen. Soils from the NAPT program were used as received with particle-size <1 mm for sandy loams and <0.7 mm for silt loams. Ten cubic centimeters of soil and 10 mL of water were stirred with a stir bar and let set for 10 min before determining soil-water pH. Some methods report using an equal value for volume or weight assuming soil density is 1 g cm^–3 (Tran and van Lierop, 1993; Eckert and Sims, 1995). Other methods report using a soil volume taking into account a soil density > 1 g cm^–3 (Watson and Brown, 1998; McLean, 1982; Soil and Plant Analysis Council, 2000). Soil-water pH is not presented since it was not the focus of the study. Twenty milliliters of SMP buffer was added and shaken on an end-to-end Eberbach shaker for 10 min at 260 4-cm recriprocations per minute before measurement for buffer pH. The buffer pH measurements began immediately after shaking according to the methodology outlined by Shoemaker et al. (1961) as opposed to allowing the soil-water-buffer mixture to set for 15 to 30 min after shaking as reported by others (Eckert and Sims, 1995; Watson and Brown, 1998; McLean, 1982; Soil and Plant Analysis Council, 2000; van Lierop, 1990; Tran and van Lierop, 1993). A buffer pH was obtained with the new buffer using the same protocol as used with the SMP buffer with the exception that only 10 mL of the new buffer was used.

The SMP buffer pH was compared with the new buffer pH with regression analysis and comparison of the data to a 1:1 line where SMP buffer pH exactly equaled the new buffer pH. To evaluate the extent of scatter in the data, the data was plotted with boundaries defined by interlaboratory error in SMP buffer pH from the NAPT program. The technique was similar to that used by Sikora et al. (2005) to evaluate the comparison between ICP- and colorimetric-P in Mehlich 3 soil extracts using intralaboratory error. The NAPT program defines interlaboratory error by a median absolute deviation (MAD) value (Miller and Kotuby-Amacher, 2004). The coefficient of variation was calculated as MAD/median x 100 and is plotted versus median SMP buffer pH in Fig. 2 for 87 NAPT soils from the third quarter of 1999 through the fourth quarter of 2004 (Miller and Kotuby-Amacher, 2004). The number of laboratories reporting an SMP buffer pH for the NAPT soils ranged from 37 to 71 with a median of 64. The error in SMP buffer pH increases as pH decreases. The NAPT program warns laboratories of a potential erroneous value if an SMP buffer pH is outside the range from the median –(2.5 x MAD) to median +(2.5 x MAD). The boundary of ± 2.5 x MAD was used in a plot of new buffer pH versus SMP buffer pH with the MAD value calculated according to the best-fit linear equation in Fig. 2.

[11]

[12]

[13]

View larger version (15K): [in this window] [in a new window]   Fig. 2. Interlaboratory variation in SMP buffer pH on 87 samples from the NAPT program from 1999 through 2004. Interlaboratory variation is defined as the coefficient of variation (MAD/median x 100) and is plotted versus median SMP buffer pH.

Limestone Recommendations from Soil-Buffer pH Measurements
Limestone recommendations from soil-buffer pH with SMP buffer and new buffer were compared utilizing SMP buffer limestone recommendations to achieve a target soil-water pH of 6.5 at a soil depth of 0.2 m (Soil and Plant Analysis Council, 2000). The limestone recommendations are based solely on soil-buffer pH and assume limestone has 90% calcium carbonate equivalence and a fineness of 40% < 0.150 mm (100 mesh), 50% < 0.250 mm (60 mesh), 70% < 0.833 mm (20 mesh), and 95% < 2.360 mm (8 mesh). Limestone recommendations are provided in tons acre^–1 in Soil and Plant Analysis Council (2000) but are reported here in units of Mg ha^–1 by multiplying tons acre^–1 by 2.24. If the soil-buffer pH is >6.84, limestone is not recommended. Data where soil-buffer pH was >6.84 for both the SMP buffer and new buffer would result in no lime recommendation for either buffer and thus were not used in the comparisons. Out of the 255 Kentucky soils and 87 NAPT soils, 158 and 27 soils were used in the comparisons, respectively. Limestone recommendations from soil-buffer pH using the new buffer versus the SMP buffer were compared by linear regression with an analysis of the 95% confidence intervals for slopes and intercepts. In addition, a plot of the data points was compared with a 1:1 line where recommendations are equal and error boundaries defined by ± 2.5xMAD as used in the comparison of soil-buffer pH values with the new buffer versus SMP buffer where MAD is a function of the soil-buffer pH (see Fig. 2 and Eq. [11]–[13]). Equations for the error boundaries on limestone recommendations are shown below where MAD is defined by Eq. [11].

upper boundary limestone recommendation from:

[14]

lower boundary limestone recommendation from:

[15]

The values for SMP buffer pH in the above equations ranged from 4.8 to 6.8 rounded to the tenth decimal place as shown in the limestone recommendation table in Soil and Plant Analysis Council (2000). The calculated new buffer pH was rounded to the tenth decimal place before assessing the recommended limestone value according to the table.

Measurements over Time on Soil-Buffer pH using SMP and New Buffer
The new buffer was used to evaluate long-term soil-buffer pH measurements over a period of 150 d using internal quality control soil samples. Internal quality control samples included 4 soils used in the Lexington laboratory and 10 soils used in the Princeton laboratory. The four soils used in the Lexington laboratory were soils from the NAPT program at a particle size < 1 mm for sandy loams and 0.7 mm for silt loams. The soils included a Darco loamy fine sand (2002–107), Kennan sandy loam (2002–110), Lucy loamy sand (2002–115), and Monona silt loam (2002–116) (Miller and Kotuby-Amacher, 2004) (Table 1). Internal quality control soils from Princeton were silt loams collected from various sites in Kentucky, dried and ground to pass a 2-mm screen. These samples are identified as Pri1 through Pri10. The soil-buffer pH was determined with the new buffer using the protocol described earlier and common to the routine testing of soil in the laboratories except 10 mL of the new buffer was used rather than 20 mL as used with the SMP buffer. A soil-buffer pH with the new buffer was tested on each internal quality control sample every 15 to 30 d for a total of 7 or 8 sampling periods over 150 d. During this same time period, data was collected on the soil-buffer pH using the SMP buffer on the internal quality control samples during the routine analysis of soil samples. The buffers were stored in the laboratory at room temperature in closed containers at room temperature, which ranged from 22 to 27°C.

Soil-buffer pH values using the new buffer and SMP buffer on the internal quality control samples over 150 d were statistically evaluated for stability by linear regression analysis with soil-buffer pH as the dependent variable and time in days as the independent variable. If the slope of the regression was not significantly different from zero using 95% confidence intervals, the soil-buffer pH was considered not to drift with time. Soil-buffer pH using the new buffer and SMP buffer was also compared using t tests to compare population means assuming equal variances (Freund and Wilson, 1997; Kume, 1992). The 95% confidence intervals on the difference between the mean values with the new buffer and SMP buffer according to the t tests are presented. Values for pH were recorded to the second decimal place. However, only the first decimal place is used to report a soil-buffer pH or use the pH for lime recommendations (Shoemaker et al., 1961, Soil and Plant Analysis Council, 2000). If the 95% confidence intervals on the difference between mean values included values ranging from –0.05 to +0.04, the difference between the two buffers was defined as not significant.

The new buffer was also used to obtain soil-buffer pH measurements on seven NAPT samples from 2004 with a soil pH < 6.50 (Table 1). A soil-buffer pH with the new buffer was tested on each NAPT sample every 15 to 30 d for a total of eight sampling periods over 150 d. The average and variance in soil-buffer pH with the new buffer on these seven NAPT samples and the four NAPT samples used as internal quality control samples in the Lexington laboratory were compared with the median and MAD for soil-buffer pH using SMP buffer as reported from the NAPT program. The number of individual soil-buffer pH measurements with the new buffer outside of the ranges of median ± 2.5xMAD for soil-buffer pH using the SMP buffer is reported.

	RESULTS AND DISCUSSION

TOP ABSTRACT INTRODUCTION THEORETICAL CONSIDERATIONS MATERIALS AND METHODS RESULTS AND DISCUSSION SUMMARY Appendix A REFERENCES

 
Developing the New Buffer
Developing a new buffer to replicate results of the SMP buffer began with determining if a titration of the SMP buffer could be explained from the theoretical calculation according to Eq. [9]. Acid titration of 100 mL SMP buffer diluted with 50 mL water agreed well with theortical calculation (Fig. 3 ). Since the acid-base equilibria constants of the various components of the buffer could provide a reasonable prediction of the buffer's reaction with acidity, a new buffer was designed using Eq. [9] with a set of components not including chromium or p-nitrophenol. The new buffer contained triethanolamine, imidazole, MES, and acetic acid. The concentration of these components in the first mixture are shown in Table 2 as determined from fitting Eq. [9] to the SMP titration data in Table 3 using initial estimates for pKa as shown in Table 2. The actual acid titration of the first mixture was off slightly from the titration of SMP buffer (Table 3). Therefore, refined values for pKa of the individual components were determined by fitting Eq. [9] to the experimental titration of the first mixture considering pKa values as unknowns. The refined pKa values were then used in fitting Eq. [9] to the SMP titration with Ti as unknowns to obtain component concentrations for the second mixture (Table 2).

View larger version (29K): [in this window] [in a new window]   Fig. 3. Calculation of soil buffer pH for SMP buffer (100 mL buffer + 50 mL water + 50 g soil) and new buffer (50 mL buffer + 50 mL water + 50 g soil) according to Eq. [9] compared with experimental data from seven titrations of SMP buffer (100 mL buffer + 50 mL water) and new buffer (50 mL buffer + 50 mL water). The upper x axis shows the volume of 0.5 M HCl used in the experimental titrations with equivalent values of mmol soil acidity (g soil)–1 shown in the lower x axis calculated according to Eq. [10] assuming acidity in the titration originated from 50 g of soil.

The pH values for the acid titration of the second mixture were below the pH values from the SMP titration (Table 3). Since the second mixture contained only KCl as the background salt, the pKa values were refined one more time using Eq. [9] fit to the second mixture titration data and are shown as final estimates in Table 2. Using these final pKa estimates and Eq. [9] fit to the SMP titration with Ti as unknowns, component concentrations of a third mixture was obtained (Table 2). The titration of the third and final mixture of the new buffer had pH values that were slightly below pH values from SMP titration (Table 3) but deemed close enough to proceed in evaluating the use of the new buffer in soil-buffer pH measurements.

Seven acid titrations of the new buffer were compared with theoretical calculations from Eq. [9] (Fig. 3) using final pKa estimates for the individual components (Table 2). Experimental titration of the new buffer had a tendency to be slightly below the theoretical calculation, whereas experimental titration of the SMP buffer had a tendency to be slightly above the theoretical calculation. Titrations of the new buffer and SMP buffer were compared in the linear range from pH 7.56 to 5.30 (Fig. 4 ). From seven titrations of each buffer (Fig. 4), the initial pH of the new buffer and SMP buffer before dilution was 7.70 ± 0.01 and 7.48 ± 0.02, respectively (average ± standard deviation). The pH values of the new buffer and SMP buffer after dilution with water before adding the acid titrant were 7.53 ± 0.03 and 7.57 ± 0.02, respectively. The new buffer pH values in the acid titration had slightly lower pH values than those from SMP titration (Fig. 4). However, 95% confidence intervals of the intercept and slope indicated there was no statistical difference between the two titrations.

View larger version (24K): [in this window] [in a new window]   Fig. 4. Seven titrations of SMP buffer (100 mL buffer + 50 mL water) and seven titration of new buffer (50 mL buffer + 50 mL water) with 0.5 M HCl as the titrant shown on the upper x axis. The titrant acidity was calculated as mmole soil acidity (g soil)–1 in the lower x axis using Eq. [10] assuming 50 g of soil was contributing the acidity neutralized by the buffer. The regression analysis was performed considering the lower x axis. The 95% confidence intervals (CI) are presented for the intercept (b) and slope (m).

Soil-buffer pH Measurements and Limestone Recommendations
Soil-buffer pH values with the new buffer agreed well with soil-buffer pH using SMP buffer (Fig. 5 ). The good comparison in the two soil-buffer pH values was evident by the data congregating close to the 1:1 line and high r² values of 0.974 and 0.967. In addition, the majority of the data fell within lower and upper bounds considering NAPT warning limits from interlaboratory variation. Only 1 out of 255 Kentucky soils and 3 out of 87 NAPT soils fell outside these limits.

View larger version (24K): [in this window] [in a new window]   Fig. 5. Comparison of soil buffer pH determined with the new buffer and SMP buffer on 255 Kentucky soils and 87 NAPT soils. Linear regression analysis is provided with 95% confidence intervals (CI) for intercept (b) and slope (m). Lines represent a 1:1 line where both variables are equal and an upper and lower boundary defined by NAPT interlaboratory variation (Eq. [12] and [13]).

View larger version (26K): [in this window] [in a new window]   Fig. 6. Comparison of limestone recommendations determined with new buffer and SMP buffer on Kentucky and NAPT soils with soil-buffer pH less than 6.85. Linear regression analysis is provided with 95% confidence intervals (CI) for intercept (b) and slope (m). Lines represent a 1:1 line where both variables are equal and an upper and lower boundary defined by NAPT interlaboratory variation for SMP soil-buffer pH (Eq. [14] and [15]). Most symbols represent more than one data point.

Repeated measurements of soil-buffer pH with the new buffer over 150 d were made on several NAPT samples (Table 5). There was no significant drift in pH over 150 d in any sample. The average and standard deviation of these pH values are compared with lower and upper limits from interlaboratory variation. The limits are based on warning limits reported in the NAPT program. When a value is reported outside the limit, the laboratory is issued a warning that the result may be in error. Considering each of the 8 measurements made for each of the 11 NAPT samples, only 2 out of 88 observations had a soil-buffer pH with the new buffer that fell outside the warning limit from the NAPT program (Table 5).

	SUMMARY

TOP ABSTRACT INTRODUCTION THEORETICAL CONSIDERATIONS MATERIALS AND METHODS RESULTS AND DISCUSSION SUMMARY Appendix A REFERENCES

 
A new buffer was created with triethanolaomine, imidazole, MES, and acetic acid that creates the same soil-buffer pH value as SMP buffer in the measurement of soil acidity. The new buffer does not create a hazardous waste since it does not contain chromium and p-nitrophenol that is present in the SMP buffer. Several tests were performed with the new buffer to confirm its similarity with SMP buffer. The new buffer titrated with acid produced the same results as SMP buffer titrated with acid. The new buffer produced a soil-buffer pH similar to SMP buffer on 255 Kentucky soils and 87 NAPT soils. The new buffer had a long laboratory shelf life, as indicated by no change in pH measurements over 150 d of storage. The procedure outlined here to develop a buffer to mimic SMP buffer can be utilized to develop a buffer with any desired acid-base characteristic. This procedure could prove useful if a chemical in a buffer used for making lime recommendations becomes undesirable due to safety or cost concerns.

	Appendix A

TOP ABSTRACT INTRODUCTION THEORETICAL CONSIDERATIONS MATERIALS AND METHODS RESULTS AND DISCUSSION SUMMARY Appendix A REFERENCES

 
Components of the Sikora Buffer and Methodology in Obtaining a Soil-Buffer pH for Lime Requirement Determination
The following buffer solution and methodology was designed to achieve a soil-buffer pH that mimicked soil-buffer pH obtained with the SMP buffer in a soil-buffer pH range from 7.50 to 5.30.

For every liter of solution, the following quantities of chemicals are dissolved.

Potassium chloride (KCl, mw = 74.55): 149 g

Glacial acetic acid (CH₃COOH, mw = 60.05): 5.36 g or 5.11 mL

MES (2-(N-morpholino)ethanesulfonic acid monohydrate) (C₆H₁₃NO₄S · H₂O, mw = 213.25): 6.70 g

Imidazole (C₃H₄N₂, mw = 68.08): 0.936 g

Triethanolamine [(HOCH₂CH₂)₃N, mw = 149.19]: 10.38 g or 9.23 mL

Sodium hydroxide (40% NaOH [w/w]): 5 mL

Dissolve the KCl in a volume of water that is 75% of the final intended volume. Make sure all the KCl dissolves. Weigh the other components of the buffer and add them to the solution in the order listed making sure each component dissolves before proceeding. Glacial acetic acid, triethanolamine, and sodium hydroxide are added as liquids. The other chemicals are added as solids. Adjust the volume to the final intended volume by adding water. Add drops of 40% NaOH (w/w) or 50% HCl (v/v) to achieve a pH of 7.70 ± 0.01. Allow time for the solution pH to stabilize. Place 50 mL of the buffer in a beaker and measure pH. The pH should be 7.70 ± 0.01. Add 50 mL of water to the buffer, stir, and measure pH. The pH should be 7.53 ± 0.03. Add 5 mL of 0.5 M HCl to the 1:1 dilution of buffer, stir, and measure pH. The pH should be 5.68 ± 0.06.

Soil pH is determined by stirring 10 cm³ of soil with 10 mL of water using a stir bar, letting the slurry stand for 10 min, and measuring pH in the slurry. After determining soil pH, 10 mL of the new buffer is added, then the sample is shaken for 10 min on a mechanical shaker at more than 180 oscillations per minute, followed by pH measurement of the slurry.

	ACKNOWLEDGMENTS

 
I thank Chip Zimmer, Ed Hill, and Debbie Morgan for their help with the various tests on the new buffer. Gratitude is also expressed to Danny Reid and Paula Howe for discussions on how the new buffer could be adapted to routine soil testing and to James Bartos, Greg Schwab, and Lloyd Murdock for discussions on the chemistry of the new buffer. This work is dedicated to the life and memory of my mother, Marie A. Sikora (8-21-34 to 3-11-05).

Received for publication May 25, 2005.

	REFERENCES

TOP ABSTRACT INTRODUCTION THEORETICAL CONSIDERATIONS MATERIALS AND METHODS RESULTS AND DISCUSSION SUMMARY Appendix A REFERENCES

 

• Adams, F., and C.E. Evans. 1962. A rapid method for measuring lime requirement of red-yellow podzolic soils. Soil Sci. Soc. Am. Proc. 26:355–357.
• Anastas, P.T., and M.M. Kirchhoff. 2002. Origins, current status, and future challenges of green chemistry. Acc. Chem. Res. 35:686–694.[Medline]
• Anwar, Z.M., and H.A. Azab. 2001. Ternary complexes formed by trivalent lanthamide ions, nucleotides, and biological buffers. J. Chem. Eng. Data 46:613–618.[CrossRef]
• Barnhisel, R., and P.M. Bertsch. 1982. Aluminum. p. 275–300. In A.L. Page et al. (ed.) Methods of soil analysis. Part 2. 2nd ed. Agron. Monogr. No. 9. ASA and SSSA, Madison, WI.
• Boelema, G.J., A. Bolt, H.J. Metting, B.L. Bajema, and D.A. Doornbos. 1982. Microcomputer-controlled determination of dissociation constants of sulfanilamides and stability constants of the related silver complexes. Pharm. Weekbl. Sci. 4(2):38–42.[Medline]
• Castro, C.L., and D.E. Rolston. 1977. Organic phosphate transport and hydrolysis in soil: Theoretical and experimental evaluation. Soil Sci. Soc. Am. J. 41:1085–1092.[Abstract/Free Full Text]
• DuBois, G. 1914. The chemistry and properties of glycerophosphate (glycerinophosphates). J. Ind. Eng. Chem. 6:122–128.[CrossRef]
• Dynes, J.J., and P.M. Huang. 1997. Influence of organic acids on selenite sorption by poorly ordered aluminum hydroxides. Soil Sci. Soc. Am. J. 61:772–783.[Abstract/Free Full Text]
• Eckert, D., and J.T. Sims. 1995. Recommended soil pH and lime requirement tests. p. 16–21. In J.T. Sims and A.M. Wolf (ed.) Recommended soil testing procedures for the Northeastern United States. 2nd ed. Northeast Regional Publ. No. 493. Northeast Agriculture Experiment Stations.
• Ferguson, W.J., K.I. Braunschweiger, W.R. Braunschweiger, J.R. Smith, J.J. McCormick, C.C. Wasmann, N.P. Jarvis, D.H. Bell, and N.E. Good. 1980. Hydrogen ion buffers for biological research. Anal. Biochem. 104:300–310.[CrossRef][ISI][Medline]
• Freund, R.J., and W.J. Wilson. 1997. Statistical methods. revised ed. Academic Press, San Diego, CA.
• Good, N.E., G.D. Winget, W. Winter, T.N. Connolly, S. Izawa, and R.M.M. Singh. 1966. Hydrogen ion buffers for biological research. Biochemistry 5:467–477.[CrossRef][Medline]
• Good, N.E., and S. Izawa. 1972. Hydrogen ion buffer. Methods Enzymol. 24:53–68.[Medline]
• Grzyska, P.K., P.G. Czyryca, J. Golightly, K. Small, P. Larsen, R. Hoff, and A. Hengge. 2002. Generality of salvation effects on the hydrolysis rates of phosphate monoesters and their possible relevance to enzymatic catalysis. J. Org. Chem. 67:1214–1220.[Medline]
• Hendershot, W.H., H. Lalande, and M. Duquette. 1993. Ion exchange and exchangeable cations. In M.R. Carter (ed.) Soil sampling and methods of analysis. Canadian Society of Soil Sci., Lewis Publishers, Boca Raton, FL.
• Horinko, M.L. 2002. 25 years of RCRA: Building on our past to protect our future. Office of Solid Waste and Emergency Response, U.S. Environmental Protection Agency. EPA-K-02–027.
• Hue, N.V., G.R. Craddock, and Fred Adams. 1986. Effect of organic acids on aluminum toxicity in subsoils. Soil Sci. Soc. Am. J. 50:28–34.[Abstract/Free Full Text]
• Huluka, G. 2005. A modification to the Adams-Evans soil buffer determination solution. Commun. Soil Sci. Plant Anal. 36:2005–20014.[CrossRef]
• Kandagedara, A., and D.B. Rorabacher. 1999. Noncomplexing tertiary amines as "better" buffers covering the range of pH 3–11. Temperature dependence of their acid dissociation constants. Anal. Chem. 71:3140–3144.[CrossRef]
• Kapinos, L.E., B. Song, and H. Sigel. 1998. Metal ion-coordinating properties of imidazole and derivatives in aqueous solution: Interrelation between complex stability and ligand basicity. Inorg. Chim. Acta 280:50–56.[CrossRef]
• Kume, H. 1992. Statistical methods for quality improvement. The Association for Overseas Technical Scholarship, Tokyo, Japan.
• Lindsay, W.L. 1979. chemical equilibria in soil. John Wiley & Sons, New York.
• Liu, M., D.E. Kissel, M.L. Cabrera, and P.F. Vendrell. 2005. Soil lime requirement by direct titration with a single addition of calcium hydroxide. Soil Sci. Soc. Am. J. 69:522–530.[Abstract/Free Full Text]
• Liu, M., D.E. Kissel, P.F. Vendrell, and M.L. Cabrera. 2004. Soil lime requirement by direct titration with calcium hydroxide. Soil Sci. Soc. Am. J. 68:1228–1233.[Abstract/Free Full Text]
• Madigan, M.T., J.M. Martinko, and J. Parker. 2002. Brock biology of microorganisms. 10th ed. Prentice Hall, NJ.
• Martell, A.E., and R.M. Smith. 1974. Critical stability constants. Vol. 1: Amino acids. Plenum Press, New York.
• Martell, A.E., and R.M. Smith. 1977. Critical stability constants. Vol. 3: Other organic ligands. Plenum Press, New York.
• Martell, A.E. and R.M. Smith. 1982. Critical stability constants. Vol. 5: First supplement. Plenum Press, New York.
• McLean, E.O. 1982. Soil pH and lime requirement. p. 199–224. In A.L. Page et al.(ed.) Methods of soil analysis. Part 2. 2nd ed. ASA, SSSA, Madison, WI.
• Mehlich, A. 1976. New buffer pH method for rapid estimation of exchangeable acidity and lime requirement of soils. Commun. Soil Sci. Plant Anal. 7:637–652.
• Mengel, K., E.A. Kirkby, H. Kosegarten, and T. Appel. 2001. Principles of plant nutrition, 5th ed. Kluwer Academic Publishers, Dordrecht, The Netherlands.
• Miller, R.O., and J. Kotuby-Amacher. 2004. Results of 1st, 2nd, 3rd, and 4th Quarter Exchanges of 1999, 2000, 2001, 2002, 2003, and 2004. Memos to participants of the North American Proficiency Testing (NAPT) Program. SSSA, Madison, WI.
• Molina, F., A. Rueda, J.M. Bosque-Sendra, and L. Megias. 1996. Determination of proteins in the presence of imidazole buffers. J. Pharm. Biomedical Anal. 14:273–280.[CrossRef]
• Ogawa, T., M. Usui, C. Yatome, and E. Idaka. 1989. Influence of chromium compounds on microbial growth and nucleic acid synthesis. Bull. Environ. Contam. Toxicol. 43:254–260.[Medline]
• Onufriev, A., D.A. Case, and G.M. Ullmann. 2001. A novel view of pH titration in biomolecules. Biochemistry 40:3413–3419.[CrossRef][Medline]
• Remko, M., and C. von der Lieth. 2004. Theoretical study of gas-phase acidity, pKa, lipophilicity, and solubility of some biologically active sulfonamides. Bioorg. Med. Chem. 12:5395–5403.[CrossRef][Medline]
• Shoemaker, H.E., E.O. McLean, and P.F. Pratt. 1961. Buffer methods for determining lime requirement of soils with appreciable amounts of extractable aluminum. Soil Sci. Soc. Am. Proc. 25:274–277.
• Sikora, F.J., P.S. Howe, L.E. Hill, D.C. Reid, and D.E. Harover. 2005. Comparison of colorimetric and ICP determination of phosphorus in Mehlich3 soil extracts. Comm. Soil Sci. and Plant Anal. 36: 875–887.
• Smith, R.M., and A.E. Martell. 1975. Critical stability constants. Vol. 2: Amines. Plenum Press, New York.
• Smith, R.M., and A.E. Martell. 1976. Critical stability constants. Vol. 4: Inorganic complexes. Plenum Press, New York.
• Soil and Plant Analysis Council. 2000. Soil analysis handbook of reference methods. CRC Press, Boca Raton, FL.
• SPSS, Inc. 1997. TableCurve 2D 4.0 for Windows user's manual. SPSS Inc., Chicago, IL.
• Stoll, V.S., and J.S. Blanchard. 1990. Buffers: Principles and practice. Methods Enzymol. 182:24–38.