1

A Green Analyzer for the Measurement of Total Arsenic in Drinking

Water. Electrochemical Reduction of Arsenate to Arsine and Gas Phase

Chemiluminescence with Ozone.

Mrinal K. Sengupta, Maather F. Sawalha1, Shin-Ichi Ohira2, Ademola D. Idowu3 and Purnendu K. Dasgupta*

Department of Chemistry and Biochemistry, University of Texas at Arlington

700 Planetarium Place, Arlington, Texas 76019-0065

Supporting Information

2

3

Standards and Reagents.

Stock standards of 100 mg As/L were prepared. Inorganic As(III) and As(V) were

prepared in 1 mM HCl from As2O3 and Na2HAsO4•7H2O (both from J. T. Baker),

respectively. Lower concentrations were prepared by dilutions with (18.2 MΩ

cm) Milli‐Q deionized water (DIW). Different concentrations of electrolytes used

for arsine generation were prepared from sulfuric acid (17.8 M, EMD Chemicals

Inc.). Potassium iodide (Mallinckrodt) and ascorbic acid (Mallinckrodt) were

used as reductant for reducing As(V) and sodium hypochlorite (bleach, bought

as 5.25% w/v NaOCl) was used as oxidant for oxidizing As(III). See Table S1 for

electrode material list.

Equipment

Liquid Dispensing Module: 48000‐step syringe pump SP (P/N 54022) with a 8‐

port distribution valve DV (P/N 19323) and a 10‐mL zero dead volume

UHMWPE tip glass syringe S (P/N 24139, all from www.kloehn.com) was used

for automated sample/reagent uptake, delivery and washing the ECR.

Chemiluminescence chamber (CC): CC is made from a glass test tube externally

silvered and black coated to prevent light leakage, sealed at the bottom with a

glass disc which remains uncoated and acts as a window. The tube was drilled at

three places for the entrance of arsine from top one end, ozone from the other top

end. The third end sits just above the window base; serves as the exit line from

where the reacted arsine‐O3 mixture exits. For ozone generation, a miniaturized

air pump operated at 24 V (AP, Bühler, Germany) connected with an air drying

and purification column comprising of serial beds of activated charcoal and a gel

at the inlet; supplies the purified air to a commercial silent discharge type ozone

generator (OZG; EOZ‐300Y, www.shop.enaly.com) flowing into CC at 8 sccm.

4

The CC sits atop a H5784 PMT (www.hamamatsu.com) with a built‐in high

voltage (HV) power supply serve as the detector, operating at a control voltage

of 0.85 V with a secondary stage amplification of 1000×. Details regarding the

GPCL detection can be obtained from our previous publication (ref 16 in main

text).

5

Table S1. Description and source of electrode materials tested

Lead foil, thickness 0.1 mm, 150x150 mm, www.vwrsp.com P/N AA42708-VA

Tin foil, thickness 0.25, 50x50mm, www.vwrsp.com P/N AA43233-FI

Zinc foil, thickness 0.62, 100x150mm, www.vwrsp.com P/N 100209-894 Nickel Chromium gauze, thickness 40 mesh, 0.25 mm, 75x75mm, www.vwrsp.com P/N AA40941-FL

Niobium foil, thickness 0.25, 25x25mm, www.vwrsp.com P/N AA00238-FF

Cadmium foil thickness 0.1 mm, 50x50 mm, www.vwrsp.com P/N AA11371-FI

Cobalt foil thickness 0.1 mm, 25x25 mm, www.vwrsp.com P/N AA42658-FF

Graphite foil, 0.254 mm, 150x150 mm, www.vwrsp.com P/N AA10832-VA

Molybdenum foil, thickness 0.127 mm, 100x150 mm, www.vwrsp.com P/N AA10043-GJ

Titanium foil, thickness 0.127, 25x25mm, VWR Parts No.AA13976-FF

Tungsten foil, thickness 0.1, 50x50mm, www.vwrsp.com P/N AA10416-FI

Indium foil, thickness 0.127 mm, 50x50 mm, www.vwrsp.com P/N AA12206-FI

Zirconium foil, thickness 0.127, 100x125mm, www.vwrsp.com P/N AA10594-GM

Tantalum foil, thickness 0.25, 50x50mm, www.vwrsp.com P/N AA10353-FI

Palladium foil, thickness 0.1, 25x25mm, www.vwrsp.com P/N AA11515-FF

Nickel foil, thickness 0.127 mm, 20x30 cm, www.vwrsp.com P/N AA1095-CH

Vanadium foil, thickness 0.127, 50x100mm, www.vwrsp.com P/N AA13783-FY

Neodymium foil, thickness 0.1 mm, 25x25 mm, www.vwrsp.com P/N AA13964-FF

Platinum foil, thickness 0.127 mm, 25x25 mm, www.vwrsp.com P/N AA00261-FI

Copper foil, thickness 0.1 mm, 100x100 mm www.vwrsp.com P/N AA42973-GH

Aluminum foil, Reynolds Wrap Aluminum Foil, 16 0.67 yds x 18 in).

Spectroscopic Carbon rods0.25” dia.x 12’ L, National spectroscopic carbon. Graphite rods, Fine Detail, Fine Finish EDM Rod, 0.125" dia x 12" L, www.graphitestore.com, MW001012

Stainless Steel foil, thickness 0.2 mm, 100x100 mm www.vwrsp.com P/N AA42973-GH

Nichrome Gauze, thickness 0.09inch, www.vwrsp.com P/N 66232-029

6

Table S2. Key literature describing EAG and Salient Findings Reference Documented Cathode recommended/used Observation

1 A simple procedure was developed for the speciation of inorganic arsenic by electrochemical hydride generation atomic absorption spectrometry without pre-reduction of As(V). Optimized the catholyte (H2SO4) concentration and electrolytic current to study the different in the response of As(III) and As(V). Applied this procedure to analyze inorganic As in Chinese medicine.

Glassy Carbon Primary study showed the % response from As(V) w.r.t As(III) Pb: 85 Glassy Carbon: 28 Pt: 0

2 An integrated electrochemical hydride generation cell, mainly composed of three components (a gas liquid separator, a graphite tube cathode and a reticulated Pt wire anode), was laboratory constructed and used for the detection of arsenic by coupling to atomic fluorescence spectrometry (AFS).

Graphite tube AFS signal intensity of As (V) was ~10% of As (III). After reduction by ascorbic acid - thiourea As (V) produced the same intensities of the As (III).

3 As, Bi, Ge, Sb and Se were determined using EHG followed by AFS. The effects of cathode material, shape and area of material, catholyte, sample flow rate, applied current, catholyte solution concentration and interference of transition metals on signal intensity were studied.

lead, graphite, copper, tungsten and platinum were tested. Graphite sheet as cathode in H3PO4 medium was selected

% Efficiency for As(III) Pb as cathode: 90 ± 4 Graphite: 95 ± 5, Cu: 41 ± 2 W: 79 ± 4, Pt: 33 ± 3 % Efficiency for As(V) Pb: 83 ± 3, Graphite: 72 ± 3 Cu: Not Detected (N.D) W: 45 ± 4 , Pt: N.D

4 EHG-AFS was used for the determination of As and Sb. This generator was coupled to AFS for determining the presence of As and Sb in tobacco.

W wire as cathode Pb: 87, W: 68, Cu: very low Pt: very low As(V) was reduced using ascorbic acid prior to total As measurement

7

5 Simultaneous measurements of As, Sb, Se, Sn and Ge were performed by inductively coupled plasma atomic emission spectrometry following EHG. The effects of sample flow rate, applied current and electrolytic solution concentration on response were studied and their influence on the mechanisms of hydride generation discussed.

Four materials, particulate lead, reticulated vitreous carbon (RVC), silver and amalgamated silver were tested as cathode materials. The best results were achieved with particulate lead and RVC cathodes,

% Efficiency for As(III): Pb: 97 ± 1, RVC: 81 ± 3 Ag: 93 ± 3 Ag amalgamated: 85 ± 12 % Efficiency for As(V) : Pb: 23 ± 2, RVC: 67 ± 5 Ag: 8 ± 4 Ag amalgamated: 67 ± 4

6 EHG systems with a tubular flow-through cell for sample introduction to AAS was described for simultaneous detection of As, Se and Sb and was applied for detecting Se and Sb in homeopathic medicines

Vitreous Carbon The estimated hydride generation efficiency was found for As(III) 93% and As(V) 47%.

7 A three-electrode arrangement was used to determine the hydrogen overvoltage of different cathode materials (Pt, Au, Ag, glassy carbon, Cd, Pb, and amalgamated Ag). The applicability of these cathode materials was tested for hydride formation using As(III), As(V), Sb(III), Sb(V), Se(IV), and Sn(IV).

Glassy carbon is the most suitable cathode material for hydride generation with As(III), Sb(III), Se(IV), and Sn(IV). Hg–Ag is well suited for the production of stibine and arsine. As(III), As(V), Sb(III), and Sb(V) were all converted into their hydrides with efficiencies > 90%.

Glassy C, Pb, Cd and Hg-Ag were found to be well suited for As(III), whereas the efficiency for Pb, Cd and Hg-Ag were found to be over 95%.

8

Anolyte Concentration during Operation. During recirculation, sulfate is

electrically driven into the anode compartment. Theoretically, at a current level

of 1 A assuming an anolyte initial volume of 0.75 L, the concentration of the

anolyte H2SO4 will increases 0.025 M/hour (if the current was continuously

flowing; in reality, the current is on only a small fraction of the analytical cycle).

If the catholyte contains a significantly higher acid concentration than the anolyte,

then this increase can be further augmented by water transport from the anolyte

to the catholyte. Overall, experimentally we find that with 0.5 M/0.1 M H2SO4

(0.75 L) as catholyte/anolyte, addition of 3 mL water to the anolyte every 150 min

of operation (approximately 18‐19 samples continuously run) maintains a

constant anolyte concentration. The effect of different compensatory measures

are depicted in Figure S12.

9

Figure S1. Effect of ozone flow rate into the reactor, air used for ozone generation.

50 µg/L As(III), Pt electrodes, constant current of 1.5 A.

0 5 10 15 20Flow rate, ccm

0

2

3

CL

sig

nal

, V Ozone Effect

10

0 10 20 30Voltage applied, V

1

1.1

1.2

1.3

1.4

1.5C

urr

ent,

Am

per

esCathode Area

0.5 cm2

1.0 cm2

1.5 cm2

2.0 cm2

2.5 cm2

3.0 cm2

3.5 cm2

4.6 cm2

Top ofLiquid

Figure S2a. A linear relationship between the applied voltage and the current for different cathode areas. The reciprocal of this slope is the dynamic cell resistance, plotted in the next figure.

11

0 1 2 3 4 5Cathode Area, cm2

14

16

18

20

V/

i, O

hm

s

3.2

3.6

4

4.4

4.8

5.2

5.6

As(

V)

CL

Sig

nal

, V

Right ordinateiTot 1.0 A

iTot 1.2 A

iTot 1.5 A

Figure S2b. The red circles show the change in the dynamic cell resistance with cathode area. It decreases up to an area of 2.5 cm2, past this area the resistance increases again. This behavior is caused by the dynamics of gas bubble adhesion and cannot be predicted a priori. The response for 50 g/L As(V) at three different current levels are shown for several cathode areas. Note that the absolute luminescence may be the highest at the highest applied current but precision suffers. An area of 1 cm2 and a current level of 1.0 A provides a good combination of signal intensity and precision at a modest power dissipation (18 W).

12

Figure S3. Typical response at different concentrations of As(III) and As(V); 0.1/0.5 M H2SO4 anolyte/catholyte, air based ozone generation.

0 5000 10000 15000 20000Time, s

0

2

4

6

8

10C

L s

ign

al,

VAs(III)µg/L

As(V)µg/L

Blk 25

10

20

40

50

Blk25

50

75

200

300

13

Figure S4. Calibration curve for As(III) and As(V) using a oxygen feed ozone generation system.

0 20 40 60 80 100Concentration, µg As/L

0

2

4

6

8

10C

L s

ign

al, V

As(III) O2 feed to Ozone Generator

As(V) O2 feed to Ozone generator

Signal, mV = (170.6 ± 2.8) * As(III), µg/L; r2 = 0.9983Signal, mV = (81.6 ± 1.3) * As(V), µg/L; r2 = 0.9985

14

As(III) reduction using KI – Ascorbic acid

In the present system, this is simply another reagent that is delivered by the

liquid handling module. 1 mL each of a solution containing 5% KI and 5 %

ascorbic acid was added immediately after the As(III) sample was introduced to

the ECR. The ECR exit valve was opened 3 min after this reductant introduction,

providing 2 additional minutes for the reduction of As(V) compared to the

standard protocol.

Figure S5. Calibration curve depicting the responses of As(III), As(V) and As(V) after reduction to As(III) using KI and ascorbic acid. 0.1/0.5 M H2SO4 anolyte/catholyte, air based ozone generation.

0 100 200 300Concentration, µg As/L

0

2

4

6

8

10

CL

sig

nal

, V

As(III)

As(V)

As(V) to As(III)

Signal, mV = 132±2 *As(III), g/L, r2 = 0.9983

Signal, mV = 29.5±1*As(V), g/L - 351±203, r2 = 0.9918

Signal, mV = 97.2±6 As(V)R, g/L + 327±147, r2 = 0.9830

15

Figure S6. Response of As(V) and As(III) after online addition of NaOCl. 0.1/0.5 M H2SO4 anolyte/catholyte, air based ozone generation.

0 5000 10000 15000 20000Time, s

0

2

4

6

CL

sig

na

l, V

As(V) As(III)

Blank

10 ppb

20 ppb

50 ppb

100 ppb

200 ppb

Blank

10 ppb

20 ppb

50 ppb

100 ppb

200 ppb

16

0 40 80 120 160 200Concentration, µg As/L

0

1

2

3

4

5

6

7C

L s

ign

al, V

As(V)As(III)

Signal, mV = (26.9 ± 1.3) * As(III), g/L; r2 = 0.9877 Signal, mV = (26.8 ± 1.1) * As(V), g/L;r2 = 0.9919

Figure S7. Calibration curve showing no difference in sensitivity between As(V) and As(III) after sequential addition of NaOCl to sample. 0.1/0.5 M H2SO4 anolyte/catholyte, air based ozone generation. There may be a specific dispensing pipette error in preparing the 50 and 100 g/L standards of the two of the intermediate concentration samples but since this was not the final method (oxygen-based ozone generation), the experiment was not repeated. The data as it stands shows nevertheless the equivalency of As(III) or As(V), i.e. that As(III) is completely oxidized.

17

0 40 80 120 160 200Concentration, µg As/L

0

2

4

6

8

10C

L s

ign

al, V

As(V), O2 feed to ozone generatorSignal, mV = (81.6 ± 1.3) * As(V), g/L;r2 = 0.9985 As(V), air feed to ozone generator Signal, mV = (26.8 ± 1.1) * As(V), g/L;r2 = 0.9919

Figure S8. Calibration curve for As(V) showing higher response with a oxygen feed ozone generation system compared to air feed ozone generation system.

18

0 100 200 300 400 500ICP-MS, g As/L

0

100

200

300

400

500P

rese

nt

Met

ho

d, µ

g A

s/L

10

10

Present method (g/L) = 0.9743* ICP-MS results (g/L)r2 = 0.9999

Figure S9. Comparison of the total As measured using ICPMS and the present system. Note both abscissa and ordinate are broken at 10 and restarts at 50 g/L.

19

0 100 200 300 400 500µg As/L (GPCL-Auto, GPCL-Manual)

0

100

200

300

400

500

Pre

sen

t M

eth

od

, µg

As/

L

Present method, g/L = (0.9885 ± 0.003) * GPCL-Auto, g/Lr2 = 0.9999Present method, g/L = (0.9493 ± 0.003) * GPCL-Man, g/Lr2 = 0.9999

Figure S10. Comparison between two chemical arsine generation- GPCL methods with present method for measuring total As in tap water and spiked tap water samples.

20

0

100

200

300

400

500C

on

cen

tra

tio

n, µ

g A

s(V

)/L

3

2.3

4.3

2

7.7

49.3

94.3

251.

2

486.

4

2.7

2.3

4.3

2.1

7.4

52.6

99.4

255.

7

490.

2

2.9

2.3 3.

8

2.1

6.4

49.6

103.

6

265.

5

511.

3

Present MethodGPCL-AutomaticGPCL-ManualICPMS

A B C D E F G H I

3.4

2.9 4.

4

2.6

8.2

50.2

95.4

254.

5

501.

2

Figure S11. Analytical results for 9 water samples, the present method compared

with a number of other techniques.

21

0

2

4

6C

L s

ign

al, V

8 12

45

9 12

8

7

108

13 8

Feb. 26 Mar. 01 Mar. 02

As(V) 50 µg/LNo Anolyte recirc.Anolyte recirc. w/o vol. corrAnolyte recirc. with vol. corrAnolyte recirc. with vol. & conc.corr

12 8 13 10

Feb. 25

Figure S12. The effect of different experimental variables on signal reproducibility. The number on each bar indicates the number of replicates. The height of the bar represents the average of these measurements with the indicated standard deviation. The successive bars on each block represent sequential measurements. For the data in blue, anolyte is not recirculated; fresh anolyte is used in each experiment. For the data in green, anolyte is recirculated without any further measures. These data shows continual decrease in response because the headspace in RV increases and ozone delivery is affected. For the data in grey, volume correction was made by supplying 3 mL of the temporarily stored anolyte prior to each analysis. For the data in red, aside from the volumetric correction, concentration correction is made by adding 3 mL of water after every 19 samples.

22

LITERATURE CITED (1) Li, X., Jia, J.,Wang, Z. Anal. Chim. Acta. 2006, 560,153–158. (2) Jiang, X. J., Gan, W.E., Han, S.P., Zi, H, J., He, Y. Z. Talanta, 2009, 79, 314–318. (3) Zhang, W. B., Xin-an Yang, X., Chu, X. F. Microchem. J., 2009, 93, 180–187. (4) Zhang, W. B., Gan, W., Lin, X. Q. Talanta, 2006. 68, 1316–1321. (5) Bolea, E., Laborda, F., Castillo, J.R., Sturgeon, R.E. Spectrochimica Acta Part B. 2004, 59, 505–513. (6) Laborda, F., Bolea, E., Castillo, J.R. J. Anal. At. Spectrom., 2000, 15, 103-107. (7) E. Denkhaus, E., Beck, F., Bueschler, P., Gerhard, R., Golloch, A. Fres. J Anal Chem, 2001, 370, 735–743.