**5.1 Optimization of the SI-ASV system for inorganic arsenic speciation**

The optimization of the variables is a critical step in the design of new analytical methods. Optimization involves the selection of the chemical and instrumental factors which may affect the analytical signal, and the choice of the values of the variables to obtain the best response from the chemical system. For this purpose, two different strategies can be used. In the traditional univariate optimization, all values of the different factors except one are constant, and this one is the object of the examination. The alternative to this strategy is the use of chemometric techniques based mainly on the use of experimental designs (Tarley, et al. 2009).

Factorial designs are used to identify the significant variables (factors) affecting the selected response and as a tool to explore and model the responses as a function of these significant experimental factors. Two-level full factorial designs are a powerful alternative to find the adequate experimental conditions to produce the best response of the chemical system. This type of design fits the response to a linear model. For a two-factor case, the response surface is given by the linear model:

$$
\hat{y} = b\_0 + b\_1 \mathbf{x}\_1 + b\_2 \mathbf{x}\_2 + b\_{12} \mathbf{x}\_1 \mathbf{x}\_2 \tag{4}
$$

If the interaction term b12x1x2 is negligible, then the response surface is planar. The more important the interaction term, the greater is the degree of twisting that the planar response surface experiences. Chemometrical optimization commonly uses the following procedure: a) choose a statistical design to investigate the experimental region of interest, b) perform the experiments in random chronological order, c) perform analysis of variance (ANOVA) on the regression results so that the most appropriate model with no evidence of lack of fit can be used to represent the data.

The factors investigated that could affect the response for As(III) and total As determination are listed in Table 1. Two levels for each factor were selected for a complete factorial design (replicates, i.e., 23+1) allowing to identifying the critical factors. Sixteen experiments were performed in random order for each of As(III) and total As optimization. The optimal values obtained for factors B and C from the optimization of SI-ASV for As(III) were fixed for total As determination, and new factors related to the prior reduction of As(V) were included in the second design of experiments. The current of the stripping peak of arsenic (in A) using a 10.0 μg l-1 standard of arsenic was selected as the response to be optimized.

The flow rate is critical in on-line stripping methods, as it controls the dispersion between the sample and the carrier solution. This factor and the deposition time contribute to the amount of analyte deposited on the electrode surface. With respect to the reaction coil length, it has to be sufficiently long for loading and pre-treatment of sample previous to the deposition step. In addition, for total As determination the chemical parameters such as reduction time and reducer concentration must guarantee the complete reduction of the As(V) contained in the sample to As(III).

Sequential Injection Anodic Stripping Voltammetry

bold). A, B and C as in Table 1.

0.00

for each factor.

speciation in water samples.

2.00

4.00

6.00

**Mean peak height (A)**

8.00

10.00

12.00

at Tubular Gold Electrodes for Inorganic Arsenic Speciation 213

A -0.80 2.54 6.86 2.91 33.96 **403.08\***  B 2.92 34.13 **91.98\*** -0.21 0.17 2.07 C 6.15 151.04 **407.02\*** -0.40 0.63 7.45 AB -1.45 8.38 **22.58\*** -0.09 0.03 0.40 AC -2.77 30.69 **82.70\*** -0.43 0.73 **8.63\***  BC -0.60 1.42 3.82 -0.07 0.02 0.21 ABC 1.70 11.56 **31.15\*** -0.07 0.02 0.21

Effect Variance *Fcalculated* Effect Variance *Fcalculated*

Factor As(III) Total As

Residual 0.37 0.08

Table 3. ANOVA of the results of the experimental design showing the factors and/or interactions affecting significantly the peak height (Fcalculated>7.57 at 95% confidence level, in

A- A+ B- B+ C- C+ A- A+ B- B+ C- C+

**Factors and levels**

A. Flow rate (ml min-1) 0.6

C. Reaction coil length (cm) 80.0 A. Flow rate (ml min-1) 1.2

Fig. 7. Effect of control factors on the mean peak height of the stripping signal. () As(III) and ( - - -) total As variables. A, B and C meaning as in Table 1. In circles the optimal level

Analyte Variable Value

As(III) B. Deposition time (s) 40

Total As B. Reduction time (s) 120

Table 4. Optimized experimental conditions of SI-ASV system for inorganic arsenic

C. [L-cysteine] (M) 1x10-2


Table 1. Selected levels of each variable for the analysis of As(III) and total As.

The design matrix and mean values obtained for the peak height of the 10.0 g l-1 arsenic solution are provided in Table 2.


Table 2. Design matrix and experimental results (mean and RSD, in brackets; n=2). A, B and C as in Table 1.

An Analysis of Variance of the results of the experimental design revealed the mean effect of each factor on the stripping signal (Table 3). In turn, these values enabled to calculate the variance of each factor using the Yates algorithm (column 3) (Massart, et al. 1997). By comparing the variance shown by each factor with the variance of the residuals, a Fischer Ftest was then performed for each source of variation.

The F-test indicated that, at a significance level of p=0.05, the critical factors for As(III) determination were the deposition time, the length of the reactor R2 and the binary interactions flow rate-deposition time and flow rate-reaction coil length. The significance of the binary interactions is a consequence of the influence of both factors in the correct mixture of the sample solution and the electrolyte support solution. On the other hand, for total As determination, the variable with major contribution was the flow rate.

Figure 7 shows the effect of the control factors on the peak height of the stripping signal, among which the reaction coil length and the flow rate are the most important factors for As(III) and total As determination, respectively. Based on the results shown in Fig. 7, the combination of settings that generates the highest peak height was selected and is shown in Table 4.

A. Flow rate (ml min-1) 0.6 1.2

C. Reaction coil length (cm) 40.0 80.0 A. Flow rate (ml min-1) 0.6 1.2


Analyte Variable Level

As(III) B. Deposition time (s) 20 40

Total As B. Reduction time (s) 120 300 C. [L-cysteine] (M) 1x10-3 1x10-2

The design matrix and mean values obtained for the peak height of the 10.0 g l-1 arsenic

Exp A B C Mean peak height (A) and %RSD

1 - - - 2.14(5.29) 0.63(1.13) 2 - - + 13.35(1.59) 0.66(9.72) 3 - + - 5.58(7.35) 0.70(5.09) 4 + - - 7.26(2.53) 3.99(2.84) 5 - + + 12.20(1.16) 0.73(6.83) 6 + + - 4.41(3.05) 4.01(1.06) 7 + - + 9.53(2.67) 3.30(7.29) 8 + + + 8.89(1.03) 3.06(2.08) Table 2. Design matrix and experimental results (mean and RSD, in brackets; n=2). A, B and

An Analysis of Variance of the results of the experimental design revealed the mean effect of each factor on the stripping signal (Table 3). In turn, these values enabled to calculate the variance of each factor using the Yates algorithm (column 3) (Massart, et al. 1997). By comparing the variance shown by each factor with the variance of the residuals, a Fischer F-

The F-test indicated that, at a significance level of p=0.05, the critical factors for As(III) determination were the deposition time, the length of the reactor R2 and the binary interactions flow rate-deposition time and flow rate-reaction coil length. The significance of the binary interactions is a consequence of the influence of both factors in the correct mixture of the sample solution and the electrolyte support solution. On the other hand, for

Figure 7 shows the effect of the control factors on the peak height of the stripping signal, among which the reaction coil length and the flow rate are the most important factors for As(III) and total As determination, respectively. Based on the results shown in Fig. 7, the combination of settings that generates the highest peak height was selected and is shown in

total As determination, the variable with major contribution was the flow rate.

As(III) Total As

Table 1. Selected levels of each variable for the analysis of As(III) and total As.

solution are provided in Table 2.

C as in Table 1.

Table 4.

test was then performed for each source of variation.


Table 3. ANOVA of the results of the experimental design showing the factors and/or interactions affecting significantly the peak height (Fcalculated>7.57 at 95% confidence level, in bold). A, B and C as in Table 1.

Fig. 7. Effect of control factors on the mean peak height of the stripping signal. () As(III) and ( - - -) total As variables. A, B and C meaning as in Table 1. In circles the optimal level for each factor.


Table 4. Optimized experimental conditions of SI-ASV system for inorganic arsenic speciation in water samples.

Sequential Injection Anodic Stripping Voltammetry

**Working electrode** 

**geometry** 

flow Thin layer Gold

absorption spectrometry (ETAAS) for comparison.

**Technique Electrode** 

Continuous

Continuous

Continuous

Continuous

Continuous

deposition time.

at Tubular Gold Electrodes for Inorganic Arsenic Speciation 215

flow Wall-jet Gold -0.30, 300 0.15 Kopanica & Novotny,

flow Thin layer Gold film -0.2, 80 0.5 Huang & Dasgupta,

flow Wall jet Gold film -0.65, 60 100 Billing et al. 2002

SIA Tubular Gold -0.4, 40 1.0 This work Table 6. Compilation of on-line stripping techniques for As(III) determination.

flow Wall jet Gold film -0.1, 120 0.55 Muñoz & Palmero,

Listed in Table 6 there are some methods designed for on-line ASV determination of As(III). Sample introduction by continuous flow is the main strategy used for on-line ASV. This tendency is related with the effort to increase sensitivity by increasing the amount of sample used. Tubular electrodes have higher contact area; this characteristic, in combination with the use of binary sampling strategy, generates a robust and sensitive flow system which is competitive with other flow systems proposed, but using less sample volume and shorter

The developed flow system was applied to the determination of the inorganic species of arsenic in groundwater samples originating from Tierra de Pinares (Segovia, Spain) as described in the Experimental section. Five replicate determinations of both total As and As(III) were carried out on each sample by the standard additions method. Results are displayed in Table 7. Total As was also determined in samples by electrothermal atomic

Sample SI-ASV system ETAAS

1 12 24 36 30 2 37 16 53 57 3 18 19 37 39 4 41 96 137 127 5 21 17 38 38 6 44 120 164 174 7 14 4 18 24 8 116 104 220 218 9 75 85 160 164 10 42 84 126 129 Table 7. Contents (mean, n=5) of As(III), As(V) and total As, determined in groundwater

samples by the proposed SI-ASV system, and comparison between total arsenic concentrations determined by SI-ASV and ETAAS. Concentration units, µg l-1.

As(III) As(V) total As total As

**Ed (V), td**

**(s) LOD (µg l-1) Reference** 

nanoparticles -0.4, 600 0.25 Majid et al. 2006

1998

1999

2004

#### **5.2 Analytical properties of the optimized flow system**

Calibration plots for As(III) and total As (added as As(V)) were obtained in the experimental conditions described in Table 4. Three replicate measurements of each standard As solution were made and average values were used for calculations. A linear dependence of the height of the stripping signal with the injected concentration of arsenic was found in the concentration range 2-40 g l-1 for both total As and As(III) (Fig. 8.).

Fig. 8. SW-voltammograms for As(III) standards obtained by SI-ASV in 2M HCl.

The limit of detection calculated according to the IUPAC criterion as 3se/b1, where se is the square root of the residual variance of the calibration plot and b1 is the slope, resulted in values of 1 g l-1 for As(III) and 2 g l-1 for total As. These limits of detection are adequate to assess the compliance of the method with the maximum tolerable levels for As in drinking water (WHO, 2004).

The reproducibility of the procedure, expressed as relative standard deviation of six replicate determinations of a water sample containing 5.0 and 30.0 g l-1 of As(III) and total As, were 3.6 and 7.6% respectively. The regression parameters of both regression lines are tabulated in Table 5.


Table 5. Regression parameters of the calibration plots of peak current (in μA) vs. arsenic concentration (in g l-1)

Calibration plots for As(III) and total As (added as As(V)) were obtained in the experimental conditions described in Table 4. Three replicate measurements of each standard As solution were made and average values were used for calculations. A linear dependence of the height of the stripping signal with the injected concentration of arsenic was found in the

> -0.4 -0.3 -0.2 -0.1 0.0 0.1 0.2 0.3 0.4 **E (mV) vs Ag/AgCl**

The limit of detection calculated according to the IUPAC criterion as 3se/b1, where se is the square root of the residual variance of the calibration plot and b1 is the slope, resulted in values of 1 g l-1 for As(III) and 2 g l-1 for total As. These limits of detection are adequate to assess the compliance of the method with the maximum tolerable levels for As in drinking

The reproducibility of the procedure, expressed as relative standard deviation of six replicate determinations of a water sample containing 5.0 and 30.0 g l-1 of As(III) and total As, were 3.6 and 7.6% respectively. The regression parameters of both regression lines are

Parameter As(III) Total As Square root of residual variance, se 0.60 0.63 Determination coefficient, R2 0.99 0.99 Intercept confidence interval, b0t s(b0) 0.43±0.46 0.50±0.52 Slope confidence interval, b1t s(b1) 1.270.03 1.240.04 Linear range (g l-1) 3-40 6-40 Limit of detection (g l-1) 1 2 Table 5. Regression parameters of the calibration plots of peak current (in μA) vs. arsenic

Fig. 8. SW-voltammograms for As(III) standards obtained by SI-ASV in 2M HCl.

0 g l-1

5.0 g l-1

**5.2 Analytical properties of the optimized flow system** 

0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0

water (WHO, 2004).

tabulated in Table 5.

concentration (in g l-1)

**I (A)**

concentration range 2-40 g l-1 for both total As and As(III) (Fig. 8.).


Table 6. Compilation of on-line stripping techniques for As(III) determination.

Listed in Table 6 there are some methods designed for on-line ASV determination of As(III). Sample introduction by continuous flow is the main strategy used for on-line ASV. This tendency is related with the effort to increase sensitivity by increasing the amount of sample used. Tubular electrodes have higher contact area; this characteristic, in combination with the use of binary sampling strategy, generates a robust and sensitive flow system which is competitive with other flow systems proposed, but using less sample volume and shorter deposition time.

The developed flow system was applied to the determination of the inorganic species of arsenic in groundwater samples originating from Tierra de Pinares (Segovia, Spain) as described in the Experimental section. Five replicate determinations of both total As and As(III) were carried out on each sample by the standard additions method. Results are displayed in Table 7. Total As was also determined in samples by electrothermal atomic absorption spectrometry (ETAAS) for comparison.


Table 7. Contents (mean, n=5) of As(III), As(V) and total As, determined in groundwater samples by the proposed SI-ASV system, and comparison between total arsenic concentrations determined by SI-ASV and ETAAS. Concentration units, µg l-1.

Sequential Injection Anodic Stripping Voltammetry

*Chimica Acta*, Vol. 459, pp. 151-159

14, pp. 741-746

99. 77-96

*Talanta*. Vol. 50, pp. 1-24

at Tubular Gold Electrodes for Inorganic Arsenic Speciation 217

Camacho, L.M., Gutierrez, M., Alarcón-Herrera, M.T., Villalba, M.L., Deng, S. (2011).

Catarino, R.I.L., Garcia, M.B.Q., Lima, J.L.F.C., Barrado, E., Vega, M. (2002) Relocation of a

Economou, A. (2010). Recent developments in on-line electrochemical analysis-An overview

Environmental Protection Agency (EPA). (1999). *Analytical methods support document for* 

Ferreira, M.A., Barros. A.A. (2002) Determination of As(III) and arsenic(V) in natural waters

Forsberg, G., O'Laughlin, J.W., Megargle, R.G., Koirtyohann, S.R. (1975). Determination of

García-Sánchez, A., Moyano, A., Mayorga, P. (2005). High arsenic contents in groundwater

Gibbon-Walsh, K., Salaun, P., van der Berg, C.M.G. Arsenic speciation in natural waters by cathodic stripping voltammetry. Analytica Chimica Acta. Vol. 662, pp 1-8 Gómez, J.J., Lillo, J., Sahun. B. (2006). Naturally ocurring arsenic in groundwater and

Gong, Z., Lu, Z., Ma, M., Watt, C., Le, C. (2002). Arsenic speciation analysis. *Talanta*. Vol. 58,

Guerin, T., Astruc, A., Astruc, M. (1999). Speciation of arsenic and selenium compounds by

Huang, H., Dasgupta, P.K. (1999). A Field-deployable instrument for the measurement and speciation of arsenic in potable water. *Analytica Chimica Acta*. Vol. 380, pp. 27-37 Ivandini,T.A., Sato, R., Makide, Y., Fujishima, A., Einaga, Y. (2006). Electrochemical

Ivaska, A. & Kubiakb, W.W. (1997). Application of sequential injection analysis to anodic

Kopanica, M. & Novotny, L. (1998). Determination of traces of arsenic (III) by anodic

Long, J. & Nagaosa, Y. (2008). Determination of trace arsenic(III) by differential-pulse anodic

Majid, E., Hrapovic, S., Liu, Y., Male, K.B., Luong, J.H.T. (2006). Electrochemical

electrode and flow analysis. *Analytical Chemistry*. Vol. 78, pp. 762-769

*Journal of Environmental Analytical Chemistry*. Vol. 88, pp. 51-60

and southwestern USA. *Chemosphere*, Vol. 83, pp. 211-225

of the last 12 years. *Analytica Chimica Acta*. Vol. 683, pp. 38-51

voltammetry. *Analytical Chemistry.* Vol. 47, pp. 1586-1593

*Environmental Geology*, Vol. 50, pp. 1151-1170

*Analytical Chemistry.* Vol. 78, pp. 6291-6298

*Analytica Chimica Acta.* V ol. 38, pp. 211-218

stripping voltammetry. *Talanta*. Vol. 44, pp. 713-723

of central Spain. *Environmental Geology*, Vol. 47, pp. 847-854.

*arsenic in drinking water EPA -815-R-00-010*. Washington, DC. pp.57

Occurrence and treatment of arsenic in groudwater and soil in northern Mexico

Tubular Voltammetric Detector for Standard Addition in FIA. *Electroanalysis*. Vol.

by cathodic stripping voltammetry at a hanging mercury drop electrode. *Analytica* 

arsenic by anodic stripping voltammetry and differential pulse anodic stripping

identification of the geochemical surces in the Duero Cenozoic Basin, Spain.

HPLC hyphenated to specific detectors: a review of the main separation techniques.

detection of arsenic(III) using iridium-implanted boron-doped diamond electrodes.

stripping voltammetry in solutions, naturals waters and biological material.

stripping voltammetry with in-situ plated bismuth-film electrode. *International* 

determination of arsenite using a gold nanoparticle modified glassy carbon

For each groundwater sample, average total As concentrations determined by both methods were compared by means of a paired t-test. Calculated *t* value was compared with the tabulated *t* value for 9 degrees of freedom and a significance level of p=0.05 (*t*=2.26). The calculated *t* value (1.00) is lower than the tabulated one, thus the null hypothesis that the methods do not give significantly different values for the mean total As concentration is accepted.
