Standard Dilution Analysis: Willis B. Jones, George L. Donati, Clifton P. Calloway, JR., and Bradley T. Jones
Standard Dilution Analysis: Willis B. Jones, George L. Donati, Clifton P. Calloway, JR., and Bradley T. Jones
Standard Dilution Analysis: Willis B. Jones, George L. Donati, Clifton P. Calloway, JR., and Bradley T. Jones
pubs.acs.org/ac
standard reference solution were added to identical amounts of mACA and SI = mICI), the ratio of the analyte to internal
urine sample, and a constant concentration of an internal standard signals is described by
standard was also added to each solution. While the results
were improved by the combination of the methods, the SA mC m (C std + CAsam) m C sam m C std
= A A = A A = A A + A A
solution preparation was tedious. SDA uses a similar approach, SI mIC I mIC I mIC I mIC I
with the advantage of simplicity. Because of the gradual dilution (1)
obtained by mixing only two solutions, many standard
Notice that this is the equation for a line if we plot y as (SA/
calibration points are obtained in a single run.
SI) versus x as (1/CI). The slope and intercept of this line are
In practice, SDA is performed by continually monitoring two
given by
signals for a solution containing a constant amount of sample
and varying amounts of a standard solution containing both mA CAsam mA CAstd
analyte and internal standard. This may be accomplished in slope = intercept =
mI mIC I (2)
molecular absorption spectrometry by partially filling a cuvette
with a solution containing 50% sample and 50% standard The SDA method is performed by combining a sample
mixture. As data are collected for this solution, a second containing an unknown amount of analyte, with a standard
solution containing 50% sample and 50% solvent is slowly solution containing a fixed ratio of analyte to internal standard
added. Analyte and internal standard signals are collected as the (CAstd/CI = constant). If the relative amount of sample in the
solutions mix in the cuvette. In this case, the relative amount of solution remains constant during the mixing process, then the
sample remains constant (50%) while the relative amount of matrix effects are constant and (mA/mI) is constant. Therefore,
standards added decreases (the standards are diluted). the intercept is also constant.
Similarly, elements may be determined by ICP OES if a The concentration of analyte in the sample is contained in
solution containing 50% sample and 50% standard mixture is the value for the slope of the line and is given simply by
added to the typical sample reservoir. As the peristaltic pump
delivers this sample to the ICP, a solution containing 50% slope C std
CAsam = × A
sample and 50% solvent is added to the same reservoir. Data intercept CI (3)
are collected as the two solutions are mixed. The relative
amount of sample remains constant, while the standards are The final term in eq 3 is the concentration ratio of analyte to
diluted. internal standard in the prepared standard mixture, so the
Gradient dilution techniques have been applied in flow concentration of analyte in the sample is easily obtained from
systems in the past. A calibration procedure for flow injection the calibration plot. Eq 1 demonstrates the types of interference
flame atomic absorption spectrometry was reported in 1998.7 that SDA may correct. Note that the concentration of both the
Later, a gradient ratio standard addition method was reported analyte and the internal standard must be within the linear
for the same technique.8 Some of these ideas were discussed dynamic range of the traditional calibration curve method. Also,
among a set of univariate calibration techniques for flow the signals measured for the analyte and internal standard (SA
injection analysis.9 The same research group has described a and SI) must not include contributions from any other species.
manifold for calibration in flow injection analysis,10 and also a Given these conditions, the SDA method will correct any
technique for flow injection gradient titration.11 More recently, matrix effects that affect the sensitivity of the analyte, any
a gradient dilution method that overcomes matrix interferences chemical interferences that might affect the sensitivity of the
in ICP OES has been described.12 This approach uses an analyte, and any fluctuations caused by variations in the flows of
HPLC pump to dilute the sample provided to the ICP while liquids or gases, light source power, or sample position. The
monitoring two emission lines continually. This allows an SDA method will not correct for spectral interference unless
automated approach to finding the highest sample concen- the magnitude of that interference could be measured
beforehand and subtracted from the appropriate signals.
■
tration where the matrix effects are minimized.
SDA, on the other hand, requires no flow since the dilution is
performed in a single container. To demonstrate the viability of
EXPERIMENTAL SECTION
the SDA method, six commercial mouthwash samples have Visible Absorption Spectrometry. FD&C Blue No. 1
been analyzed for the presence of FD&C dye Blue No. 1, using (B1) and FD&C Yellow No. 6 (Y6) crystalline dyes were
FD&C dye Yellow No. 6 as the internal standard. Subsequently, obtained from Rainbow Colors, LLC (Windsor, CT). Caution
a suite of eight metals (Al, Cd, Co, Cr, Cu, Fe, Ni, and Pb) has is advised when handling the solid dyes to prevent accidental
been determined in mouthwash, wine, cola, nitric acid, and inhalation. Blue no. 1 also may be harmful to the aquatic
water by ICP OES using Y as the internal standard. environment. An aqueous solution was prepared containing a
■
mixture of 6.03 μM B1 (analyte) and 37.6 μM Y6 (internal
standard). Concentrations were recorded to three significant
MATHEMATICAL APPROACH
figures. A modular UV−vis apparatus consisting of a sample
The mathematical approach is straightforward. The analyte holder with integrated light source (tungsten coil lamp), a glass
species (A) in the sample is combined with a standard mixture cuvette, fiber optic, and an USB4000 spectrometer (Ocean
containing both the analyte (A) and an internal standard Optics, Dunedin, FL) was used to analyze the samples. Solution
species (I). Analyte signal (SA) arising from both the sample 1, containing 50% of the standard mixture and 50% sample, was
(sam) and the standard (std) mixture is monitored at one placed in the glass cuvette (1.0 mL aliquot), and the signals at
wavelength, while the internal standard signal (SI, from the 635 nm (analyte) and 480 nm (internal standard) were
standard mixture only) is monitored at a different wavelength. recorded using a 1-s integration time. Then 2.0 mL of solution
Given that the signal in each case is equal to the calibration 2, containing 50% sample and 50% distilled water, were slowly
sensitivity (m) times the concentration of each species (SA = added to the same cuvette using a micropipette with no manual
B DOI: 10.1021/ac504152x
Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry Article
Figure 2. SDA plot of the four replicates for mouthwash C (Figure 1) plotted as a single run.
in SDA begin with a relatively large analyte signal from the method may randomly result in artificially high or low values,
standard. This might help minimize small signal fluctuations the average amount found is similar to that for SDA. However,
from the background, which would be more prominent in the the average percent error is significantly worse for SA (Table
absence of an analytical signal. Furthermore, if the measure- 3). In addition, SDA provides improvements in precision,
ment is shot-noise-limited, as might be the case for an emission calculated as percent relative standard deviation, when
measurement using an internal standard, a larger initial signal compared to all three calibration methods.
should result in a lower LOD. On the other hand, the SDA For this proof of concept work, we have chosen two different
method is time-limited (a single point per second is collected in techniques that are widely used and that can suffer from severe
ICP OES, for example). The traditional methods might employ matrix effects. ICP-based methods are especially sensitive to
longer integration times to improve detection limits. such effects,18,19 and we have intentionally chosen complex
Table 3 displays the results of all four calibration methods matrices, with diverse compositions, to demonstrate the
separated by element. Individual entries in the table are an capabilities of SDA in comparison to the traditional calibration
methods.
Table 3. Concentration of Eight Metals, Accuracy, and
Precision Found Using Different Calibration Methods
EC IS SA SDA n
■ CONCLUSION
Standard dilution analysis is simple and effective, providing
Average Concentration Found, mg/L (10 mg/L added)
superior accuracy, precision, and sample throughput when
Al 12.1 12.6 9.5 9.5 15
compared to other traditional calibration methods. It offers the
Cd 11.7 12.2 11.0 10.5 15 best of both worlds: it simultaneously corrects for matrix effects
Co 11.8 12.3 10.2 10.0 15 due to the sample concomitants (as with standard additions)
Cr 11.0 11.6 10.3 10.2 15 and for fluctuations due to changes in sample size, orientation,
Cu 11.6 12.3 9.5 9.7 15 or instrumental parameters (as with internal standard). SDA
Fe 11.2 11.8 10.5 10.3 15 requires no instrument modifications or special strategies for
Ni 11.1 11.7 10.4 10.2 15 widely used techniques such as molecular absorption
Pb 11.0 11.6 10.1 9.9 15 spectrometry and ICP OES. It can be applied to most
average 11.4 12.0 10.2 10.1 120 instrumental analysis techniques that will accept liquid samples
Accuracy (average percent error) and are capable of simultaneous multianalyte determinations.
Al 24.8 26.1 4.8 5.0 15 The novel approach proposed by SDA is rapid, requiring less
Cd 20.8 22.2 16.0 7.3 15 than 3 min to analyze each sample by mixing two solutions in a
Co 19.6 23.3 11.5 4.2 15 single container. That container may be the cuvette in a
Cr 15.5 16.4 10.4 4.3 15 molecular absorption experiment or the sample tube in an ICP
Cu 16.4 23.4 7.8 3.8 15 OES experiment. Since the same amount of sample is always
Fe 20.1 17.6 10.8 4.5 15 present during the measurement, the relative effect of the
Ni 18.7 17.5 12.4 4.9 15 matrix on analytes and internal standard is not a concern. This
Pb 18.9 15.8 12.0 4.0 15 makes the selection of a suitable internal standard trivial. As
average 19.3 20.3 10.7 4.7 120 shown in this proof of concept work, the SDA method may be
Precision (percent relative standard deviation) successfully used in molecular and atomic spectrometry
techniques.
■
Al 30.5 17.4 2.9 3.5 15
Cd 24.4 10.9 18.8 7.7 15
Co 21.0 8.0 13.9 5.1 15 AUTHOR INFORMATION
Cr 18.7 5.9 12.7 4.5 15 Corresponding Author
Cu 15.6 9.7 7.6 3.3 15 *E-mail: donatigl@wfu.edu. Tel: +1 336 758-4815. Fax: +1 336
Fe 25.6 12.7 13.5 5.0 15 758-4656.
Ni 22.6 9.4 14.9 5.8 15
Notes
Pb 22.4 9.3 13.8 4.9 15
The authors declare no competing financial interest.
■
average 19.8 9.3 13.3 5.8 120
REFERENCES
average of 15 SDA runs: five sample matrixes, n = 3. As (1) Bader, M. J. Chem. Educ. 1980, 57, 703−706.
observed in Table 2, SDA and standard additions present (2) Harvey, D. J. Chem. Educ. 2002, 79, 613−615.
similar results. Purely from a recovery standpoint, SDA (3) Veening, H. J. Chem. Educ. 1966, 43, 319−320.
provides results of a slightly better accuracy than those of (4) Rice, G. W. J. Chem. Educ. 1987, 64, 1055−1056.
standard additions and significantly better than those obtained (5) Magee, J. A.; Herd, A. C. J. Chem. Educ. 1999, 76, 252.
with the other methods. However, when average percent errors (6) Davis, A. C.; Alligood, B. W.; Calloway, C. P., Jr.; Jones, B. T.
are compared, SDA accuracies are superior to the standard Appl. Spectrosc. 2005, 59, 1300−1303.
addition (SA) values by a factor of 2 and to the other methods (7) Koscielniak, P. Anal. Chim. Acta 1998, 367, 101−110.
by a factor of 4. This may be explained by the number of data (8) Koscielniak, P.; Kozak, J. Anal. Chim. Acta 2002, 460, 235−245.
(9) Koscielniak, P. Anal. Chim. Acta 2001, 438, 323−333.
points used to construct the plots for each method. The SDA (10) Koscielniak, P.; Wierczorek, M.; Kozak, J.; Herman, M. Anal.
plots are prepared using 30 data points, one collected each Chim. Acta 2007, 600, 6−13.
second during the SDA region of the dilution. The SA method (11) Wojtowicz, M.; Kozak, J.; Koscielniak, P. Anal. Chim. Acta 2007,
employs a calibration curve prepared from only five solutions. 600, 78−83.
Therefore, the best fit is superior with the SDA method and the (12) Cheung, Y.; Schwartz, A. J.; Hieftje, G. M. Spectrochim. Acta,
accuracy is better. Since the quality of the best fit in the SA Part B 2014, 100, 38−43.
F DOI: 10.1021/ac504152x
Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry Article
G DOI: 10.1021/ac504152x
Anal. Chem. XXXX, XXX, XXX−XXX