assembled monolayer on gold electrode

1 downloads 0 Views 504KB Size Report
Oct 20, 2010 - This study is supported by a European project (HYDRONET, http://www.hydronet-project.eu). H. Tobias is acknowledged for the Cd analysis.
See discussions, stats, and author profiles for this publication at: http://www.researchgate.net/publication/49692720

Determination of low levels of cadmium ions by the under potential deposition on a selfassembled monolayer on gold electrode ARTICLE in ANALYTICA CHIMICA ACTA · JANUARY 2011 Impact Factor: 4.52 · DOI: 10.1016/j.aca.2010.10.021 · Source: PubMed

CITATIONS

9

2 AUTHORS: Tomer Noyhouzer

Daniel Mandler

McGill University

Hebrew University of Jerusalem

5 PUBLICATIONS 19 CITATIONS

198 PUBLICATIONS 3,829 CITATIONS

SEE PROFILE

SEE PROFILE

Available from: Tomer Noyhouzer Retrieved on: 20 August 2015

Analytica Chimica Acta 684 (2011) 1–7

Contents lists available at ScienceDirect

Analytica Chimica Acta journal homepage: www.elsevier.com/locate/aca

Determination of low levels of cadmium ions by the under potential deposition on a self-assembled monolayer on gold electrode Tomer Noyhouzer, Daniel Mandler ∗ Institute of Chemistry, The Hebrew University of Jerusalem, Jerusalem 91904, Israel

a r t i c l e

i n f o

Article history: Received 24 August 2010 Received in revised form 9 October 2010 Accepted 13 October 2010 Available online 20 October 2010 Keywords: Cadmium Under potential deposition Self-assembled monolayer Subtractive anodic square wave voltammetry

a b s t r a c t The electrochemical determination of low levels of Cd using a self-assembled monolayer (SAM) modified Au electrode is reported. Determination was based on the stripping of Cd, which was deposited by under potential deposition (UPD). A series of short alkanethiol SAMs bearing different end groups, i.e., sulfonate, carboxylate and ammonium, were examined. Lowest level of detection (ca. 50 ng L−1 ) was achieved with a 3-mercaptopropionic acid (MPA) monolayer using subtractive anodic square wave voltammetry (SASV). Additional surface methods, namely, reductive desorption and X-ray photoelectron spectroscopy, were applied to determine the interfacial structure of the electrodeposited Cd on the modified electrodes. We conclude that the deposited Cd forms a monoatomic layer, which bridges between the gold surface and the alkanethiol monolayer associating with both the gold and the sulfur atoms. © 2010 Elsevier B.V. All rights reserved.

1. Introduction Cadmium poses a potential environmental hazard [1]. The metal and its compounds are extremely toxic even at low concentrations because of their ability to bioaccumulate in living organisms as well as in the environment, e.g., in sediments [2]. Yet, cadmium is still widely used in different industrial processes such as painting and plating as well as in batteries and microelectronics. At least one major cadmium pollution event has been documented during the recent years; the pollution of the Jinzu river in Toyama prefecture, Japan [3]. The maximum allowed levels of cadmium in drinking water are 5 ␮g L−1 [4], which is derived from the standard method of detection, i.e., graphite furnace atomic absorption (GF-AA) and ICP-MS, vide infra. The levels of Cd in several rivers and lakes in Europe are as high as 40 ␮g L−1 whereas the concentration of total cadmium in unpolluted rivers or in the ocean can be as low as 5 ng L−1 determined mainly by atomic absorption spectroscopy (AAS) [5,6]. Determination of Cd in different waters is routinely measured nowadays by GF-AA and ICP-MS. The level of detection (LoD) of cadmium as determined by the European Environment Agency (EEA) is 10 and 100 ␮g L−1 by ICP-MS and GF-AAS, respectively [7,8]. Nevertheless the LoD can be decreased to a few ␮g L−1 by changing the operational conditions. Evidently, these levels are not sufficiently low for monitoring Cd in sites that are either mildly or unpolluted.

∗ Corresponding author. Tel.: +972 2 6585831; fax: +972 2 6585319. E-mail address: [email protected] (D. Mandler). 0003-2670/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.aca.2010.10.021

In fact, the method of detection and its limit serve as determining factors in legislation. The development of sensitive methods for heavy metals in general and cadmium in particular is therefore driven by public concern as well as by the scientific challenge of measuring heavy metals in the environment. Cadmium like most heavy metals is electroactive and therefore can be detected electrochemically. Electrochemical methods offer very high sensitivity and at the same time instrumentation is inexpensive and portable, which makes electrochemistry an attractive approach for monitoring Cd in the environmental. Mercury was the most used electrode for Cd determination due to its high reproducibility, sluggish kinetics towards water reduction and the formation of amalgam that allowed Cd accumulation upon reduction. Dropping mercury electrode had allowed LoD of 1 ␮g L−1 , mercury film electrodes enabled detection as low as 8 ng L−1 [9,10], while the lowest LoD (1–0.01 ng L−1 ) using Hg electrode was reported by a hanging drop mercury electrode [11,12]. However the general concern about mercury toxicity and its practical inconvenience for on-site measurements have led to extensive search for new and improved electrode materials. Bismuth modified electrodes (BMEs) proved to be very successful and enabled detection levels as low as 1–100 ng L−1 [13]. The behavior of BMEs was shown to compare favorably to that of Hg with attractive properties such as high sensitivity and well-defined stripping signals and high reproducibility. Yet, the major disadvantage of BMEs is the adsorption of macromolecules in particular organic matter which fouls the electrode surface [14,15]. In addition, at pH > 4 Bi(OH)3 is formed on the electrode surface, leading to irreproducibility [16].

2

T. Noyhouzer, D. Mandler / Analytica Chimica Acta 684 (2011) 1–7

Chemically modified electrodes have also been used for Cd determination [17]. For example, Nafion-graphene nanocomposite film reached a detection limit of 10 ng L−1 , whereas a Bi electrode modified with Nafion was capable of detecting 1 ng L−1 of Cd [9,10]. Nafion coating minimized the interferences from surface active materials and prevented the formation of Bi(OH)3 . Higher selectivity and lower levels of detection can be achieved using under potential deposition (UPD). UPD involves deposition of metals on foreign substrates at potential more positive than their thermodynamic reversible potential. UPD is usually limited to a monolayer [18], yet making it highly sensitive due to the accumulation effect and the Faraday constant. Arrigan reviewed the determination of trace metals by UPD stripping voltammetry [19]. UPD was primarily used for determination of Cu and Pb [19]. The UPD of Cd on Au was studied by Gewirth [20–22] and others [23,24] who investigated the UPD mechanism of several metals. They concluded that the Cd UPD on Au(1 1 1) is a two step process during which an alloy formation between Cd and Au occurs. We are aware of only a few studies, which applied the UPD of Cd on Au and Ag as a means of electrochemical determination of Cd [25–31]. Pemberton and Chow [26,31] used either an immobilized single strand DNA or a monolayer of glutathione as a means of increasing the selectivity towards Cd2+ by voltammetry, while Bonfil and Kirowa-Eisner applied either silver or gold bare electrodes [27–30]; however, introduced a rotating disk electrode and subtractive square wave anodic stripping voltammetry (SASV) for decreasing the level of detection. Here we present a detailed study whereby a gold electrode modified by different self-assembled monolayers (SAMs) based on short ␻-functionalized alkanethiols, X–(CH2 )n –SH (n = 2 or 3) and X = NH2 , SO3 H and CO2 H) has been used for determination of low levels of Cd by UPD. The LoD of Cd using either 2mercaptoethanesulfonic acid (MES) or 3-mercpatopropanoic acid (MPA) was as low as 10 ng L−1 using SASV and the precision was better than 3%. The application of the short SAMs improved the signal and prevented fouling of the electrode surface, which allowed using the same electrode for more than one month. The different parameters that affected the electrochemical signal, such as the pH and nature of electrolyte, deposition time and potential, were studied and optimized. The electrode showed a wide linear range and highly reproducibility over a period of more than a week.

2. Experimental 2.1. Instrumentation Cyclic voltammetry (CV) and subtractive anodic stripping square wave voltammetry (SASV) experiments were preformed with an Autolab PGSTAT10 potentiostat using GPES software, version 4.9 (EcoChemie, Utrecht, The Netherlands). Measurements were conducted with home-made Au disk electrodes, which were assembled by sealing a gold rod (3 mm diameter, Holland-Moran, Israel) in Teflon tube under pressure. Pt wire and Ag/AgCl (KCl sat’) were used as a counter and reference electrodes, respectively. pH measurements were carried out with Cyberscan 510 pH meter and a pH electrode (Eutech instruments, Singapore). Surface analysis of the modified electrodes was performed by X-ray photoelectron spectroscopy (XPS) using a Kratos Axis Ultra spectrometer (Kratos Analytical, Manchester UK). The XPS spectra were acquired with monochromatic Al K␣ (1486.7 eV) under a chamber pressure of ca. 2 × 10−9 Torr. After modification samples were immediately introduced into the vacuum chamber to avoid as much as possible oxidation and impurities adsorption. The highresolution XPS spectra were acquired for Cd 3d, S 2p, C 1s and Au 4f levels with pass energy of 20 eV and step size of 0.1 eV. Data

analyses and processing were carried out using CasaXPS software (Casa Software Ltd). ICP-MS measurements were preformed using a SCIEX Elan-6000 spectrometer (PerkinElmer, USA). 2.2. Chemicals Cadmium nitrate tetrahydrate (≥99%) 3-mercaptopropanoic acid (MPA>99%), 2-mercaptoacetic acid (MAA > 97%) and 11-mercaptoundecanoic acid (MUA 95%) were purchased from Sigma–Aldrich. 2-Mercaptoethaneamine (Cys ≥ 98%), 2mercaptoethanesulfonic acid, sodium salt (MES ≥ 98%) and disodium hydrogen phosphate were purchased from Fluka (Switzerland). Sulfuric acid (GR grade) and phosphoric acid (HPLC grade) were obtained from Mallinckrodt (MO, USA) hydrochloric acid (30%, suprapure) from Merck and ethanol was from J.T. Baker (Deventer, Holland). Sodium phosphate monohydrate was purchased from Alfa Aesar (MA, USA). Solutions were prepared from deionized water (Barnstead Easypure UV system). 2.3. Procedures Gold electrodes were modified with SAMs following a conventional procedure. Namely, the electrodes were polished with alumina (1 and 0.05 ␮m, Buehler, IL, USA) washed with deionized water and electrochemically scanned in 0.1 M H2 SO4 (10 cycles with a scan rate of 0.1 V s−1 from −0.65 to 1.85 V vs. Ag/AgCl). Then, they were immersed in an ethanolic solution containing 1 mM of the thiol for ca. 10 h. The electrodes were carefully washed with ethanol and water prior to application. SASV analysis was performed using the project option in the GPES software as described by Bonfil [28,29]. All solutions were bubbled with nitrogen for at least 10 min before electrochemical experiments. SWV commenced by applying −0.4 V for a given time (15–360 s) followed by scanning between −0.4 V and 0.5 V with an amplitude of 0.01 V, step potential of 0.025 V and a frequency of 25 Hz. SASV is based on subtracting a second scan, where the deposition time was eliminated from the original scan. A constant potential of 0.0 V was applied for 10 s in between both scans. Real samples were analyzed without nitrogen bubbling. The solutions were diluted 1:1, v/v with deionized water and the pH was adjusted with HCl to 3. The deposition potential was −0.5 V, all other conditions were as describe above. Reductive desorption experiments were conducted using two sets of bare and SAMs modified Au electrodes. CV was performed with the first set of electrodes between −0.4 and 1.4 V (0.1 V s−1 ) in NaOH 0.1 M solution. The second set of electrodes was first dip in a solution of 1 mM of Cd2+ and 1 mM HCl and a deposition potential of −0.4 V was applied for 120 s to form a UPD layer of Cd. After deposition the electrodes were rinsed with TDW and were scanned in NaOH solution similarly to the first set. The XPS measurements were carried out with two Au evaporated and flamed annealed [32] electrodes. Both electrodes were inspected after deposition of Cd by UPD (−0.4 V for 120 s in 1 mM of Cd2+ and 1 mM HCl). One electrode was covered by a MPA SAM while the other was bare. 3. Results and discussion Fig. 1 shows the CV of 1 mM of Cd2+ in 0.1 M H2 SO4 using a polycrystalline gold electrode. The potential window shown in Fig. 1A encompasses the UPD as well as bulk deposition of Cd, while Fig. 1B focuses only on the UPD region. Electrochemical bulk deposition of Cd on Au does not show a distinct wave but rather a shoulder (1) at ca. −0.67 V. Yet, the stripping of bulk deposited Cd reveals a clear peak (1 ) at ca. −0.51 V. This stripping peak has been used for the determination of Cd [33]. Recently, we have shown that the

T. Noyhouzer, D. Mandler / Analytica Chimica Acta 684 (2011) 1–7

20

6

1'

A

4

22'

10

B

2'

2

0 -10

0

2

I/µA

I/µA

3

-20

-4

1

-30

-2

-40

-6

-50

-8

-60

-10 -0.8

-0.6

-0.4

-0.2

0.0

0.2

E/Volt Vs Ag/AgCl

0.4

2 -0.4

-0.2

0.0

0.2

0.4

0.6

0.8

E/Volt Vs Ag/AgCl

Fig. 1. Cyclic Voltammetry of 1 mM of cadmium nitrate in 0.1 M H2 SO4 using a polycrystalline Au electrode. Scan rate was 0.1 V s−1 . (A) Potential window of UPD and bulk deposition. (B) UPD region only.

stripping wave becomes much sharper and distinct upon adsorbing a self-assembled monolayer of relatively short n-alkanethiols or ␣,␻-alkanedithiol [34]. Nevertheless, we have been unsuccessful in using such modified gold electrodes for the detection of low levels, i.e., below 10 ␮g L−1 , of Cd2+ . The UPD of Cd shows well-defined and almost reversible reduction-oxidation waves centered at ca. −0.05 V (Fig. 1A and B (peaks 2 and 2 )). The shape of the peaks is clearly indicative of a surface-confined process. A second set of UPD peaks at ca. 0.3–0.4 V is hardly seen with polycrystalline Au surface but is evident on Au(1 1 1). These UPD peaks were studied by Gewirth and co-workers [20,22] and Inzelt and Horanyi [24]. They suggested that Cd can form on Au multiple structures which can interconvert as a function of the electrode potential. The UPD structure at 0.3–0.4 V, which is open-packed, can be transformed to a closepacked structure at more negative potentials. The anodic stripping peak can readily be used for the quantification of deposited Cd and therefore of Cd2+ in the solution. This requires adjusting and optimizing all other parameters, such as the time and potential of deposition as well as the electrolyte and its pH. This will be described in the following section. 3.1. Deposition time Fig. 2 shows the effect of deposition time on the UPD of 1 ␮g L−1 of Cd2+ in phosphate buffer (pH 2.0) using SASV. This method, which was developed by Kemula [35] and more recently by Kirowa-

8 6 4 2 0 12

-4

10

-6

8

I/µA

I/µA

-2

-8

6 4

-10

2

-12

0

60

120

180

240

300

360

Depositon Time/sec

-14 -0.4

-0.2

0.0

0.2

0.4

E/Volt vs Ag/AgCl Fig. 2. SASV of 1 ␮g L−1 of Cd in phosphate buffer pH 2.0 using a bare Au electrode at different deposition times: 15, 30, 60, 120, 300 and 360 s. Edep = −0.4 V. Inlet: peak currents as a function of tdep . Frequency was 25 Hz and step potential 0.025 V.

Eisner and co-workers [29,36] is based on recording the stripping voltammetry after a certain time of deposition, followed immediately by a second stripping cycle with no deposition time. The output is the subtraction of the voltammetry with no deposition time from the first stripping voltammetry. The main advantage of this procedure lies in the elimination of most of the background current, which increases significantly the signal to noise ratio as shown by Kirowa-Eisner et al. A constant potential of −0.4 V was applied with different deposition times (tdep ) ranging from 15 to 360 s. The height of the peak obtained should linearly be depended on the amount of Cd2+ in the solution providing that a complete monoatomic layer of Cd is not formed. It can be seen (Fig. 2 and the inset) that in spite of the fact that the peak at ca. 0 V increases as a function of tdep , the SASV current does not yield satisfactory linearity. It is inconceivable that saturation of the surface by an adlayer of Cd was reached as is indicated by our calculations, vide infra. Furthermore, another negative peak at ca. −0.2 V is formed, which suggests that the surface changes upon deposition of Cd. Interestingly, the UPD at 0.3–0.4 V can be clearly seen with this method. To improve the sensitivity and avoid fouling of the surface we decided to examine the effect of assembling short SAMs on the Au surface. This approach has previously been reported by us [37,38] and others [19,39,40] and comprises the formation of a disorganized SAM made of short ␻-functionalized alkanethiols. The short thiols tend to form disorganized layers due to relatively low van der Waals interactions and/or steric effects. The latter are a result of bulky functional groups, e.g. sulfonate. This, on one hand, does not prevent electron transfer and, on the other hand, still prevents electrode fouling. Hence, we examined the UPD of Cd on Au electrodes modified with short alkanethiols bearing different functional groups: MAA, MPA, MES and Cys. While MAA and MPA are not charged at pH 3, MES is negatively charged and Cys is positively charged. Fig. 3 shows the SASV of Cd recorded with Au surfaces modified with MPA (the SASV of MAA and MES modified electrodes is shown in the supplementary information: Fig. S1). The UPD solution and all other conditions were identical to those detailed in Fig. 2. The two UPD processes discussed in Fig. 2 can be clearly seen in Fig. 3. Moreover, the UPD peaks are significantly sharper and narrower than the signals of the bare Au shown in Fig. 2. The sharp peaks also simplify the quantification of the Cd concentration in the solution. In addition, the modified electrodes show an excellent linear correlation between the peak current and tdep . Finally, the UPD peak at ca. 0.3 V is also better seen and affected by the tdep in particular in MPA and MES modified electrodes. Moreover, the negative peak that grew with time of deposition using bare Au (Fig. 2) is hardly seen as a result of modifying the electrode surface. On the other

4

T. Noyhouzer, D. Mandler / Analytica Chimica Acta 684 (2011) 1–7

2.5

2.5

2.0

I/µA

1.5

2.0

1.0

R2=0.982

0.5 0.0

1.5

0

60

120 180 240 300 360

I/µA

Depositon Time/sec

1.0 0.5 0.0

3.3. Electrolyte

-0.5 -0.4

-0.2

0.0

0.2

0.4

0.6

E/Volt vs Ag/AgCl Fig. 3. SASV of 1 ␮g L−1 of Cd in phosphate buffer pH 2.0 using Au electrode modified with MPA SAM. All other conditions are as described in Fig. 2.

hand, the Cys modified Au (not shown) did not show any improvement as compared with the bare surface. This suggests that the positive charge of the SAM repels the Cd2+ and interferes with the UPD. It should be added that applying a much longer thiol, i.e., 11mercaptoundecanoic acid, blocked electron transfer and UPD was not detected [41]. These results encouraged us to continue working with the Au electrodes modified with MAA, MPA and MES. 3.2. pH The second parameter that was examined and optimized was the pH of the solution. It is well known that the acidity has a profound effect on the reduction process of Cd. At pH > 9 Cd2+ hydrolyses [42]. Fig. 4 shows the effect of pH on the UPD stripping peak current of Cd for a bare and modified Au surfaces. The solution contained 100 ␮g L−1 of the metal and 0.1 M phosphate buffer and the potential and time of deposition were −0.4 V and 360 s, respectively. SASV was employed as described above. It is evident that the UPD stripping peak current increases with the acidity of the solution for all electrodes. Cd2+ has a tendency to form hydroxides at increasing pH. Cd2+ has three stable hydroxides: CdOH+ , Cd(OH)2 and Cd2 (OH)3 + ; however, their formation constant suggests that they are formed at pH > 9 [20]. In order to understand the trend shown in Fig. 4, we ran pH simulation using HYDRAQL [43]. Our simulations indicated that cadmium sulfate and 2.5

pH is not the only significant parameter, which influences the shape and size of the peak of Cd stripping and therefore the analysis quality. We found that the electrolyte in which the UPD and stripping processes are carried out also affects the measurement. Bonfil et al. already investigated the effect of different anions on the stripping of Cd [30]. They reported that UPD is strongly affected by the concentration and type of the anion. The UPD peaks of cadmium and lead on gold can be separated with an addition of at least 4 mM of chloride. Therefore, we added 10 mM NaCl to all the solutions. Two sets of SASV experiments, using different acids and electrolytes were conducted. One set was conducted in HCl pH 3.0 and 10 mM NaCl into which increasing concentrations of Na2 SO4 (0, 1, 10, 100 mM) were added. The second set was performed in H2 SO4 pH 3.0 with increasing concentrations of NaCl (0, 1, 10, 100 mM). N2 was bubbled for 10 min in all solutions to maintain a deaerated atmosphere. Each SASV measurement was repeated 10 times consecutively with deposition time of 120 s and a continuous stirring. Usually, the first five scans were not highly reproducible and therefore we report here the average of the last five scans (for the complete results, see Table S1). It is quite evident that the reproducibility of the modified electrodes was better than the bare Au. In average the MES modified surface yielded the best results. The necessity of the addition of an electrolyte is evidence by the low reproducibility of the experiments where only low concentrations of acid were added. Yet, no clear correlation was observed between the amount of electrolyte added and the precision obtained for the modified electrodes. Based on Table S1 we decided to pursue our stripping analysis using MES and MAA modified Au electrodes in 1 mM HCl (pH 3.0) and 1–10 mM Na2 SO4 . It is worth mentioning that Gewirth [22] reported that the presence of sulfate anions can induce different Cd structures on an Au surface. The sulfate anions coadsorb with the Cd adatoms and influence the attained structures. This affects the UPD coverage of Cd. 3.4. Calibration curve

2.0

1.5

I /µA

cadmium carbonate species are formed at pH ≥ 7. Yet, in carbonate and sulfate-free environment the concentrations of Cd hydroxide species are not negligible. For example, at pH 8 more than 60% of the cadmium in the solution is in the form of cadmium hydroxide. Hence, we attribute the change of the UPD stripping current to both the decrease of the available Cd2+ species that is prone to electrochemical deposition by UPD as well as to the adsorption of hydroxide unknown species on the gold surface. The latter is supported by the observation that a bare Au electrode is similarly affected by the change of pH.

1.0

0.5

0.0 2

3

4

5

6

7

8

pH Fig. 4. Effect of pH on the UPD oxidation peak current obtained by SASV using: bare Au (black) and Au modified with MAA (red), MPA (blue) and MES (green). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)

After optimizing the system a calibration curve was obtained. The three types of modified electrodes (MPA, MAA, MES) were tested while the bare gold electrode was used as a reference. All the experiments were conducted in a solution containing 1 mM HCl (pH 3), 1 mM Na2 SO4 and 10 mM NaCl (as described before) and the current of the UPD stripping peak obtained from the SASV experiments was used for the quantitative analysis. The MAA electrode showed an excellent linear correlation between 10 and 150 ␮g L−1 (R2 = 0.999) and a poor correlation between 150 ng L−1 to 10 ␮g L−1 (R2 = 0.915). The MES showed a good linear correlation between 1 and 50 ␮g L−1 (R2 = 0.984) and a poor correlation between 50 ng L−1 and 1 ␮g L−1 (R2 = 0.903). Best results were achieved using the MPA electrode which gave satisfactory linear correlation (R2 = 0.954) of a wide range of concentrations and a low LoD of 50 ng L−1 to 3 ␮g L−1 (Fig. 5). The bare gold electrode demonstrated an inferior linear correlation (R2 < 0.84) for concentrations below 10 ␮g L−1 .

T. Noyhouzer, D. Mandler / Analytica Chimica Acta 684 (2011) 1–7

5

1.1 1.0 0.9 0.8

I / µA

0.7 0.6 0.5

Fig. 7. Possible scenarios for the Cd UPD on a thiol covered Au: (A) the cadmium cleaves the Au–S bond and forms an Au–Cd–S configuration, (B) Cd deposits on the exposed Au sites yet interacting with the sulfur and (C) as (B) but Cd does not interact with the thiol.

0.4 0.3 0.2 0.0

0.5

1.0

1.5

2.0

2.5

3.0

-1

Cd / µg L

Fig. 5. Calibration curve for Cd2+ obtained by a MPA modified electrode. The current was extracted from the anodic stripping peak of the UPD peak of cadmium using SASV in solution containing 1 mM HCl, 10 mM NaCl and 1 mM Na2 SO4 . The deposition potential and time were −0.4 V and 120 s, respectively.

3.5. Measuring a real sample In order to test our probe we measured the concentration of dissolved cadmium in tap water. Clearly, there are tens of different substances in tap water, which could have interfered with the measurement of Cd. Yet, SASV should, in principle, eliminate the different matrix complications and at the same time the SAM was expected to prevent fouling of the electrode. Nevertheless we repeated the experiments described in Table S1 in order to optimize the concentration and nature of the electrolyte to be used for the tap water analysis. Best results were achieved in 1 mM of HCl and Na2 SO4 each (for more information see Table S2). These conditions are somewhat different than those obtained for artificial solutions (1 mM Na2 SO4 and 10 mM NaCl). Another difference in the protocol was the potential of deposition, which was changed from −0.4 V to −0.5 V. These conditions gave a good linear correlation of R2 > 0.93 over a wide a range of Cd concentrations (10 ng L−1 to 10 ␮g L−1 ). Determination of Cd2+ in tap water using the standard addition method with a MPA modified Au electrode is shown in Fig. 6. The concentration of Cd2+ in tap water was ca. 0.37 ± 0.06 ␮g L−1 (R2 = 0.979). The same sample was also analyzed with ICP-MS and gave a concentration of 0.35 ± 0.01 ␮g L−1 . Evidently, the electrochemical measurement matches quite well with that of ICP-MS. 7

3.6. Studying the solid–electrolyte interface Clearly, the UPD of the Cd takes place on the gold surface. Yet, we were intrigued by the way the thiol affects the UPD. Alkanethiols spontaneously adsorb on Au(1 1 1) and form an organized √ √ 3 × 3R30◦ adlayer. Such layer covers only one third of the entire surface [44]. We are aware of a relevant study by Yoneyama at el. where the UPD of Cd on Au covered by alkanethiols was examined [45]; however they used alkanethiols bearing a hydroxyl end group and did not focus on Cd. There is, however, evidence of the effect of a short disorganized layer of thiol on the UPD of Cd on Ag [25]. Fig. 7 shows three possible scenarios for the cadmium-thiolAu electrode interactions. It is plausible that cadmium penetrates across the thiol and deposits on the Au surface interacting with the sulfur (Fig. 7A–C). We have recently shown that Cd deposits beneath the thiol leading to the formation of an Au-Cd alloy in the electrochemical bulk deposition of Cd onto alkanethiol and alkanedithiol SAMs [34]. The fact that the presence of a MPA monolayer does not reduce the charge associated with the UPD of Cd (based on the stripping peak) strongly suggests that the same applies here. This does not rule out any of the scenarios depicted in Fig. 7A–C. At the same time, it is conceivable that Cd2+ interacts with both the thiol as well as the carboxylic acid as we have shown in the past [37] ruling out the configuration shown in Fig. 7C. 3.7. Reductive desorption The interaction between the thiol and the gold surface can be studied by the reductive desorption in strong alkaline solutions [46,47]. This involves the cleavage of S–Au bond and the dissolution of the thiolate:

6

RS–Au + e− → RS− + Au

5

Previous studies used this approach to determine the thermodynamics and coverage of SAMs on Au and Hg surfaces [11,46] Since the reductive desorption depends on the reducible Au–S bond, we anticipated that it will be affected by an adlayer of Cd providing that it interacts with the sulfur. Fig. 8 shows the reductive desorption of bare and MAA Au electrodes before and after Cd was electrodeposited via UPD. It should be mentioned that the other SAMs, e.g., MPA, showed similar behavior and therefore are not shown. Two major differences can be observed between the CVs of the bare and modified electrodes. The reductive desorption waves at ca. −0.8 and −1.0 V vs. Ag/AgCl are clearly seen with the MAA modified gold (Fig. 8B). These waves are associated with the stripping of the thiol (Eq. (1)) and match well with previous reports [48]. Once Cd was electrochemically deposited via its UPD (at −0.4 V in acidic solutions) using a MAA modified electrode these peaks completely disappear (Fig. 8B). In

4

I /µA

Moreover, this analysis demonstrates the advantage of the SASV methods, which enables determination of real samples containing a variety of different substances in a range of concentrations.

3 2 1 0 -0.4

0.0

0.4

0.8

1.2

1.6

2.0

2.4

2.8

3.2

3.6

-1

Cd / µg L

Fig. 6. Standard addition method for Cd2+ in tap water obtained by a MPA modified electrode. The solution contained 1 mM HCl and 1 mM Na2 SO4 . The deposition potential and time were −0.5 V and 360 s, respectively.

(1)

6

T. Noyhouzer, D. Mandler / Analytica Chimica Acta 684 (2011) 1–7

0

0

A

-5

-20

I/µA

-10

I/µA

B

-10

-15

-30 -40

-20 -50 -25 -60 -1.4

-1.2

-1.0

-0.8

-0.6

-0.4

-1.4

-1.2

E/Volts vs Ag/AgCl

-1.0

-0.8

-0.6

-0.4

E/Volt vs Ag/AgCl

Fig. 8. CV of Au bare (A) and MAA SAM modified (B) electrode in 0.1 M NaOH and scan rate of 0.1 V s−1 before (red) and after (black) Cd was electrochemically deposited via UPD (−0.4 V for 120 s in 1 mM HCl containing 100 ␮g L−1 of Cd2+ ). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)

x102

x102

Cd 3d 5/2

Cd 3d 5/2

50

40

40

CPS

CPS

Cd 3d 3/2

30

Cd 3d 3/2

30

20 20 10 10

415

410

405

Binding Energy (eV)

415

410

405

Binding Energy (eV)

Fig. 9. Cadmium 3d XPS signal of a bare (A) and MPA modified (B) gold surfaces after UPD of cadmium carried out in solution of Cd2+ (1 mg L−1 ) in HCl (1 mM, pH 3).

addition, a reduction wave at ca. −0.6 V is clearly detected at a bare Au electrode on which Cd was previously deposited through its UPD (Fig. 8A). This irreversible surface confined reductive wave can presumably be attributed to the electrochemical reduction of Cd(OH)n species on the electrode surface. We are not aware of a similar study on the reduction of cadmium under basic conditions using an Au electrode. We can point out to previous studies where the electrochemical reduction of a Cd electrode was examined under basic conditions and showed similar behavior [49,50]. These experiments clearly indicate that a monoatomic layer of Cd atoms (on Au) strongly affects the reductive desorption of thiol SAMs. This must be due to an interaction between the Cd layer and the sulfur atom. Such interaction is expected to be stronger than Au–S association and therefore is likely to shift the reductive desorption of the Cd–S to more negative potentials. Furthermore, the disappearance of the Cd(OH)n reduction peak (Fig. 8A) in the presence of a thiol SAM suggests that the thiols substitute the hydroxide to form a more stable Cd–SR compound that is less reducible. Hence, we can conclude that the most probable configuration of the UPD of Cd on an Au surface modified by a thiol SAM, is that shown in Fig. 7A. 3.8. XPS XPS was further used to disclose the interfacial structure. In principle, the binding energies of two elements are relevant in this study, namely, the S and Cd. Yet, we found that the binding energies of sulfur of short alkanethiols associated with either Au

or Cd, are very similar and therefore cannot provide reliable evidence for changes in the binding of the thiol moiety [51]. Hence, we present the XPS spectra of the 3d Cd binding energies on Au before and after adsorbing a MPA SAM (Fig. 9). Two distinct peaks can be observed for the Cd 3d3/2 and 3d5/2 at 411.7 and 405.0 eV, respectively (Fig. 9A). The binding energies of these two peaks are in good agreement with pervious reports of Cd deposited on gold [51]. As a result of the formation of a MPA SAM these two peaks are slightly split (Fig. 9B). Two additional peaks at 412.3 and 405.6 eV can be detected upon deconvolution. We attribute this blue shift of ca. 0.6 eV to the formation of a Cd–S bond [51]. We have reproduced these measurements several times. Since full coverage of the surface by Cd is impossible using UPD, we are still expecting to find sulfur atoms that do not bind to Cd. These results are in good agreement with the reductive desorption experiments indicating that the Cd deposits below the sulfur on the gold surface and bridges between the surface and the alkanethiol. 4. Conclusion The electrochemical determination of low levels of Cd(II) has been accomplished through its under potential deposition using a gold electrode modified with short alkanethiols. The latter enhanced the signal to noise ratio and made it possible to determine Cd(II) down to 50 ng L−1 in artificial solutions. Subtractive anodic square wave voltammetry exhibited superior signals using the SAM modified electrodes compared with the bare Au. The determination of Cd in tap water gave comparable result to a standard method.

T. Noyhouzer, D. Mandler / Analytica Chimica Acta 684 (2011) 1–7

Reductive desorption and XPS analysis of the bare and modified Au surfaces shed light on the intimate structure of the interface. We conclude that the electrodeposited Cd penetrates the organic monolayer and bridges between the gold surface and the sulfur group. This system is currently being examined as a sensor for monitoring Cd in aquatic systems. Acknowledgments This study is supported by a European project (HYDRONET, http://www.hydronet-project.eu). H. Tobias is acknowledged for the Cd analysis. We are indebted to the Harvey M. Krueger Family Center for Nanoscience and Nanotechnology of the Hebrew University. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.aca.2010.10.021. References [1] K. Arisawa, A. Nakano, H. Saito, X.J. Liu, M. Yokoo, M. Soda, T. Koba, T. Takabashi, K. Kinoshita, International Archives of Occupational and Environmental Health 74 (2001) 255. [2] M. Nendza, T. Herbst, C. Kussatz, A. Gies, Chemosphere 35 (1997) 1875. [3] K. Aoshima, Tohoku Journal of Experimental Medicine 152 (1987) 151. [4] U.S. Environmental Protection Agancy (EPA), 2009, http://www.epa.gov/ safewater/contaminants/index.html. [5] World Health Organisation (WHO), Environmental Health Criteria 134, 1992. [6] C.-G. Elinder, Cadmium and Health: A Toxicological and Epidemiological Appraisal, CRC Press, Inc., Boca Raton, Florida, 1985. [7] SCA (Ed.), Oxidized Nitrogen in Waters, HMSO, UK, 1981. [8] European Environmental Agency (EEA), 2009, http://www.eea.europa.eu/ publications/92-9167-023-5/page023.html. [9] H. Xu, L.P. Zeng, S.J. Xing, Y.Z. Xian, G.Y. Shi, Electroanalysis 20 (2008) 2655. [10] J. Li, S.J. Guo, Y.M. Zhai, E.K. Wang, Electrochemistry Communications 11 (2009) 1085. [11] L.M. De Carvalho, P.C. Do Nasciniento, A. Koschinsky, M. Bau, R.F. Stefanello, C. Spengler, D. Bohrer, C. Jost, Electroanalysis 19 (2007) 1719. [12] A.M. Beltagi, M.M. Ghoneim, Journal of Applied Electrochemistry 39 (2009) 627. [13] L.Y. Cao, J.B. Jia, Z.H. Wang, Electrochimica Acta 53 (2008) 2177. [14] J. Wang, Electroanalysis 17 (2005) 1341. [15] A. Economou, P.R. Fielden, Analyst 118 (1993) 1399. [16] S. Legeai, K. Soropogui, M. Cretinon, O. Vittori, A.H. De Oliveira, F. Barbier, M.F. Grenier-Loustalot, Analytical and Bioanalytical Chemistry 383 (2005) 839.

7

[17] H.M. Dong, H. Zheng, L. Lin, B.X. Ye, Sensors and Actuators B-Chemical 115 (2006) 303. [18] J.O.M.a.K. Bockris, U.M. Shahed (Eds.), Surface Electrochemistry, Plenum Press, New York, N.Y., 1993. [19] G. Herzog, D.W.M. Arrigan, Trac-Trends in Analytical Chemistry 24 (2005) 208. [20] J.C. Bondos, A.A. Gewirth, R.G. Nuzzo, Journal of Physical Chemistry 100 (1996) 8617. [21] M.H. Ge, A.A. Gewirth, Surface Science 324 (1995) 140. [22] B.K. Niece, A.A. Gewirth, Langmuir 13 (1997) 6302. [23] P. Abiman, G.C. Wildgoose, A. Crossley, R.G. Compton, Electroanalysis 21 (2009) 897. [24] G. Inzelt, G. Horanyi, Journal of Electroanalytical Chemistry 491 (2000) 111. [25] J.W.F. Robertson, D.J. Tiani, J.E. Pemberton, Langmuir 23 (2007) 4651. [26] E.L.S. Wong, E. Chow, J.J. Gooding, Electrochemistry Communications 9 (2007) 845. [27] Y. Bonfil, M. Brand, E. Kirowa-Eisner, Analytica Chimica Acta 464 (2002) 99. [28] Y. Bonfil, M. Brand, E. Kirowa-Eisner, Reviews in Analytical Chemistry 19 (2000) 201. [29] Y. Bonfil, M. Brand, E. Kirowa-Eisner, Analytica Chimica Acta 387 (1999) 85. [30] Y. Bonfil, E. Kirowa-Eisner, Analytica Chimica Acta 457 (2002) 285. [31] E. Chow, D.B. Hibbert, J.J. Gooding, Analyst 130 (2005) 831. [32] H.J. Himmel, C. Woll, R. Gerlach, G. Polanski, H.G. Rubahn, Langmuir 13 (1997) 602. [33] I. Sinko, J. Dolezal, Journal of Electroanalytical Chemistry 25 (1970) 299. [34] E. Malel, J.K. Sinha, I. Zawisza, G. Wittstock, D. Mandler, Electrochimica Acta 53 (2008) 6753. [35] W. Kemula, Pure Appl. Chem. 15 (1967) 283. [36] E. Kirowa-Eisner, M. Brand, D. Tzur, Analytica Chimica Acta 385 (1999) 325. [37] I. Turyan, D. Mandler, Analytical Chemistry 66 (1994) 58. [38] I. Turyan, D. Mandler, Israel Journal of Chemistry 37 (1997) 225. [39] G. Herzog, D.W.M. Arrigan, Electroanalysis 17 (2005) 1816. [40] H. Shen, J.E. Mark, C.J. Seliskar, H.B. Mark, W.R. Heineman, Journal of Solid State Electrochemistry 1 (1997) 241. [41] H. Shen, J.E. Mark, C.J. Seliskar, H.B. Mark, W.R. Heineman, Journal of Solid State Electrochemistry 1 (1997) 148. [42] W.J. Tomlinson, N. Wardle, Corrosion Science 15 (1975) 663. [43] C. Papelis, K.F. Hayes, J.O. Leckie, Dep. of Civil Eng., Stanford Univ., Stanford, CA, 1988. [44] C. Vericat, M.E. Vela, R.C. Salvarezza, Physical Chemistry Chemical Physics 7 (2005) 3258. [45] D. Oyamatsu, S. Kuwabata, H. Yoneyama, Journal of Electroanalytical Chemistry 473 (1999) 59. [46] M.M. Walczak, D.D. Popenoe, R.S. Deinhammer, B.D. Lamp, C.K. Chung, M.D. Porter, Langmuir 7 (1991) 2687. [47] C.A. Widrig, C. Chung, M.D. Porter, Journal of Electroanalytical Chemistry 310 (1991) 335. [48] S. Imabayashi, M. Iida, D. Hobara, Z.Q. Feng, K. Niki, T. Kakiuchi, Journal of Electroanalytical Chemistry 428 (1997) 33. [49] Y. Duhirel, B. Beden, J.M. Leger, C. Lamy, Electrochimica Acta 37 (1992) 665. [50] N. Simic, E. Ahlberg, Journal of Electroanalytical Chemistry 462 (1999) 34. [51] NIST X-ray Photoelectron Spectroscopy Database, Version 3.5 (National Institute of Standards and Technology, Gaithersburg, 2003); http://srdata.nist.gov/xps. And references there in.