Open Archive TOULOUSE Archive Ouverte (OATAO) OATAO is an open access repository that collects the work of Toulouse researchers and makes it freely available over the web where possible.

This is an author-deposited version published in : http://oatao.univ-toulouse.fr/ Eprints ID : 11725

To link to this article : DOI:10.3389/fchem.2014.00019URL : http://dx.doi.org/10.3389/fchem.2014.00019

To cite this version : Pujol, Luca and Evrard, David and Groenen Serrano, Karine and Freyssinier, Mathilde and Ruffien-Cizsak, Audrey and Gros, Pierre Electrochemical sensors and devices for heavy metals assay in water: the French groups’ contribution. (2014) Frontiers in Chemistry, vol. 2 (n°19). pp. 1-24. ISSN 2296-2646

Any correspondance concerning this service should be sent to the repository

administrator: staff-oatao@listes-diff.inp-toulouse.fr

doi: 10.3389/fchem.2014.00019

Electrochemical sensors and devices for heavy metalsassay in water: the French groups’ contributionLuca Pujol1,2,3, David Evrard1,2, Karine Groenen-Serrano1,2, Mathilde Freyssinier3,Audrey Ruffien-Cizsak3 and Pierre Gros1,2*1 Université de Toulouse, UPS, INPT, Laboratoire de Génie Chimique, Toulouse, France2 CNRS, Laboratoire de Génie Chimique, Toulouse, France3 Elta, Blagnac, France

*Correspondence:Pierre Gros, Laboratoire de GénieChimique, UMR UPS/CNRS/INP5503, Université Toulouse III – PaulSabatier, 118, route de Narbonne,Bâtiment 2R1, 31062 ToulouseCedex 9, Francee-mail: gros@chimie.ups-tlse.fr

A great challenge in the area of heavy metal trace detection is the development ofelectrochemical techniques and devices which are user-friendly, robust, selective, with lowdetection limits and allowing fast analyses. This review presents the major contributionof the French scientific academic community in the field of electrochemical sensors andelectroanalytical methods within the last 20 years. From the well-known polarography tothe up-to-date generation of functionalized interfaces, the different strategies dedicatedto analytical performances improvement are exposed: stripping voltammetry, solidmercury-free electrode, ion selective sensor, carbon based materials, chemically modifiedelectrodes, nano-structured surfaces. The paper particularly emphasizes their advantagesand limits face to the last Water Frame Directive devoted to the Environmental QualityStandards for heavy metals. Recent trends on trace metal speciation as well as onautomatic “on line” monitoring devices are also evoked.

Keywords: electrochemical detection, heavy metals, carbon electrode, polarography, mercury-free electrode,chemically modified electrode, ion selective electrode, speciation

INTRODUCTIONLike many other micropollutants such as drugs or cosmetics andtheir by-products, pesticides and industrial or household chemi-cals, heavy metals represent a growing environmental (Callender,2004; Roy, 2010) and health (Musarrat et al., 2011; Prabhakaret al., 2012) problem. They may be considered as a major sourceof ecological issues due to their wide overspread in naturalmedia (Mhatre, 1995). Although naturally produced throughoutbiogeochemical processes, heavy metals occurrence in the envi-ronment mainly originates from human activities: air emissionsfrom coal-burning plants, smelters, waste incinerators, processwastes from mining, industrial and urban runoff all participateto their wide spreading (Friedman et al., 1993; Lindqvist, 1995;Nagajyoti et al., 2010). Once released to the environment, these

Abbreviations: ASV, anodic stripping voltammetry; BDD, boron doped diamond;BiFE, bismuth film electrode; CME, chemically modified electrode; CPE, carbonpaste electrode; CSV, cathodic stripping voltammetry; CV, cyclic voltammetry;DAPV, differential alternative pulses voltammetry; DLC, diamond like carbon;DME, dropping mercury electrode; DPASV, differential pulse anodic strippingvoltammetry; DPV, differential pulse voltammetry; FIA, flow injection analysis;GC, glassy carbon; HgFE, Hg film electrode; HMDE, hanging mercury drop elec-trode; ICP-MS, inductively coupled plasma mass spectroscopy; IEV, ion exchangevoltammetry; ISE, ion selective electrode; LOD, limit of detection; LOQ, limit ofquantification; LSASV, linear sweep anodic stripping voltammetry; LSSV, linearsweep stripping voltammetry; LSV, linear sweep voltammetry; MSWV, multi-ple square wave voltammetry; MWCNT, multi-walled carbon nanotube; NPV,normal pulse voltammetry; NPs, nanoparticles; SCP, stripping chronopotentiome-try; SMCPE, silica-modified carbon paste electrode; SMDE, static mercury dropelectrode; SPE, screen printed electrode; SSCP, stripping chronopotentiometryat scanned deposition potential; SWASV, square wave anodic stripping voltam-metry; SWCNT, single-walled carbon nanotube; SWCSV, square wave cathodicstripping voltammetry; SWV, square wave voltammetry; TFME, thin film mercuryelectrodes; WFD, water frame directive.

metals can remain for decades or centuries since they are notbiodegradable. Depending on the contamination pathway, theyappear at detectable levels in food resources such as vegetables,grains or fruits, and fish or shellfish throughout bioaccumula-tion all along the trophic chain, thus contaminating the finalconsumer—human being (Musarrat et al., 2011; Prabhakar et al.,2012). Another contamination way is direct intoxication fromdomestic environment, for instance lead traces in householdplumbing and old house paints.

Once penetrated inside human organism by ingestion (drink-ing or eating), inhalation or skin contact, heavy metals may beresponsible for nausea, vomiting, diarrhea or allergic reactions forshort term or low-level exposure (Martin and Griswold, 2009).They can also cause severe diseases in the case of long term orchronic high-level exposure, such as reduced growth and devel-opment, cancers, organs or nervous system damages and evendeath (Prabhakar et al., 2012). There are over 50 elements thatare classified as heavy metals, including transition metals, somemetalloids, lanthanides and actinides. Among them 17 are con-sidered to be both very toxic and relatively accessible. Lead (Pb),mercury (Hg), arsenic (As), and cadmium (Cd) are generallyconsidered as leader elements in human poisoning even at tracelevel. The general population is mainly exposed to all these met-als from air, drinking water and food, fish being a major sourceof mercury exposure. Moreover, smokers are highly exposed tocadmium (Järup, 2003). Some other heavy metals, including cop-per (Cu), zinc (Zn), nickel (Ni), cobalt (Co), selenium (Se), andbismuth (Bi) are known to play a vital role in physiological con-centrations but can also be toxic in larger doses. Depending onthe metal properties, the toxicity target may be different: kidneys

http://www.frontiersin.org/Chemistry/editorialboardmailto:gros@chimie.ups-tlse.fr

Table 1 | Environmental Quality Standards for heavy metals (also called WFD).

Substance CAS EQS-AAa EQS-AAa EQS MPCc EQS MPCc

number Inside surface watersb Other surface waters Inside surface watersb Other surface waters

nM e[µg L−1] [µg L−1] nM [µg L−1] [µg L−1]

Cadmium and itsspeciation (according towater hardness level d)

7440-43-9 ≤ 0.71 (class 1) 1.78 ≤ 4 (class 1) ≤ 4 (class 1)[0.08] [0.2] [0.45] [0.45]

0.71 (class 2) 4 (class 2) 4 (class 2)

[0.08] [0.45] [0.45]

0.8 (class 3) 5.34 (class 3) 5.34 (class 3)

[0.09] [0.6] [0.6]

1.33 (class 4) 8.09 (class 4) 8.09 (class 4)

[0.15] [0.9] [0.9]

2.22 (class 5) 13.3 (class 5) 13.3 (class 5)

[0.25] [1.5] [1.5]

Lead and its speciation 7439-92-1 34.7 34.7 Groundless Groundless

[7.2] [7.2]

Mercury and itsspeciation

7439-97-6 0.25 0.25 0.35 0.35

[0.05] [0.05] [0.05] [0.05]

Nickel and its speciation 7440-02-0 341 341 Groundless Groundless

[20] [20]

aEnvironmental Quality Standard—annual average.bInside surface waters include rivers, lakes and also water masses (artificial or seriously modified) related to them.cEnvironmental Quality Standard—maximal permissible concentration.d For cadmium and its compounds, EQS—AA values are functions of water hardness according to the five classes as follows: class 1: 1 μm), colloidal (1 nm–1 μm) and dissolved (≤1 nm) species.These latter include free metal ions, simple inorganic complexesand complexes bearing anthropogenic and natural organic lig-ands (Templeton et al., 2000). Hence speciation information onheavy metals of concern appears to be data of particular relevance(Kot and Namiesńik, 2000).

In 2000, a new European directive (“Water Frame Directive,”WFD, see Table 1) (European Directive 2000/60/EC) particu-larly pointed out four heavy metals (Hg, Cd, Pb, Ni) andhas established their maximal authorized as well as annualaverage concentration values in surface waters. As a consequence,environmental monitoring of heavy metals is of critical impor-tance for both ecological assessments and public health preser-vation. In answer there is an urgent need for in situ, real-time,and highly-sensitive sensors in order to multiply control pointsdedicated to early warning pollution alert (Suib, 2013).

Heavy metals trace detection is mainly performed using spec-troscopic techniques: atomic absorption spectroscopy (Kenawyet al., 2000; Pohl, 2009), inductively coupled plasma mass

spectroscopy (ICP-MS) (Caroli et al., 1999; Silva et al., 2009),X-ray fluorescence and neutron activation analysis are the mostcommonly used. Their main advantages are their versatility sincethey are suitable for a large panel of elements, their sensitivity andtheir limit of detection (LOD) in the femtomolar range. Howeverthey suffer from several major drawbacks: expensive materialsare required and qualified operators are needed to perform themulti-step sample preparation and complex analytical proce-dures, which are unsuitable for on-site and on time measurementsnecessary to prevent transient phenomena monitoring. Finally,only total metal concentration can be determined, and speci-ation data can be reached only by associating supplementaryextraction and separation techniques such as chromatography tothe spectroscopic detection (Feldmann et al., 2009). These addi-tional steps significantly increase the risk of contamination ofthe sample and some modifications of the speciation may occurduring sample storage or handling.

On the contrary electrochemistry represents an interestingalternative due to its numerous advantages. Electrochemicaldevices are mostly user-friendly since they require simple pro-cedures. They are also reagentless, low cost, and well-suitedfor miniaturization and automatic in situ measurements withminimal sample changes. Thus, contamination by reagents orlosses by adsorption on containers are drastically decreased.Electrochemical systems also allow quite fast analyses with

experimental data obtained mostly in real time or in a fewminutes. Hence, on-line monitoring of water samples becomespossible, providing dynamic data of relevance for biogeochem-ical survey. Nevertheless specific developments are still requiredfor such applications, particularly to improve sensitivity, LODsand automation. In this way, a large number of electrochem-ical techniques with different imposed potential or currentmodulations have been developed such as differential pulsevoltammetry (DPV), square wave voltammetry (SWV) or strip-ping chronopotentiometry (SCP). Electrochemical sensors alsoallow high temporal resolution measurements to be obtainedwhen associated to flow injection analysis (FIA) or flow elec-trochemical analysis cells, thus providing continuous in situmeasurements. Another analytical performance of high relevancewith respect to heavy metals detection concerns the selectiv-ity. In complex media, the signal of the analytical target oftenexperiences interferences due to the presence of other species(sometimes other heavy metals). To solve this problem, severalsurface functionalization strategies have been developed for manyyears to improve sensors selectivity.

Many pioneering researchers coming from several countrieshave initiated and intensified works dealing with electrochem-ical techniques for heavy metals detection and assay in nat-ural media: Jaroslav Heyrovsky (Czechoslovakia) (Heyrovsky,1960), Joseph Wang (USA) (Wang et al., 1995), Richard G.Compton (UK) (Agra-Gutierrez and Compton, 1998), GeorgeLuther III (USA) (Luther III and Ferdelman, 1993), Laura Sigg(Switzerland) (Goncalves et al., 1985), Arben Merkoçi (Spain)(Aragay and Merkoçi, 2012), Marco Mascini (Italy) (Voccia et al.,2012), and so on. French contribution to this topic is also notice-able. This is mainly due to the voluntary policy lead since 1998and the Aarhus protocol in which France contracted to limit itsrelease of Pb, Cd, and Hg at a lower level than that recorded in1990 (Commissariat général au développement durable, 2012).This goal was reached before the protocol came into effect in2003, but the situation is still worrying: over the 2007–2009period, 25 heavy metals have been detected in more than 10% ofthe analyses performed in French rivers (Commissariat généralau développement durable, 2011), whereas the contaminationof mussels and oysters, which constitute a good indicator ofcoastal water pollution, remained stable over the last 3 decades(Commissariat général au développement durable, 2012). Thisreview provides a survey of French groups’ contribution to thedevelopment of electrochemical sensors and methods aiming atheavy metals detection. The paper particularly emphasizes themultidisciplinary of French scientific investigations through thedescription of the electrochemical techniques and the evaluationof the corresponding analytical performances.

POLAROGRAPHYClassical techniquePolarography has been certainly the most studied and commonlyused electrochemical technique throughout the 20th centurysince the pioneering work of Heyrovsky in 1922 (Heyrovsky,1922). This is undoubtedly the consequence of the particularproperties of the mercury electrode: continuous renewal of theactive surface area, wide cathodic potential window due to the

high overpotential corresponding to hydrogen evolution, controlof the hydrodynamic conditions by means of mercury drop. Thesecharacteristics make polarography a very powerful electrochemi-cal technique for the study of inorganic, organic, organometallic,or biological compounds, not only from a theoretical point ofview but also for analytical applications. In this frame the assayof heavy metals has been the subject of numerous papers due tothe large inclination of mercury to form amalgams with majormetal compounds. For concentrations higher than 10−5 M lin-ear sweep voltammetry (LSV) on a dropping mercury electrode(DME) or on a static mercury drop electrode (SMDE) generatedat the end of a glass capillary is well-suited. For lower concentra-tions, the faradic current becomes smaller and the double-layercharging current is not negligible anymore. Pulse techniques, i.e.,normal pulse (NPV), differential pulse (DPV) and square wave(SWV) voltammetries have been favored to partially suppressthe background current and thus improve the LOD. In the caseof trace metals detection these potential pulse programs havebeen associated with anodic (ASV) or cathodic (CSV) strippingvoltammetries on a hanging mercury drop electrode (HMDE)inside which the analyte is pre-concentrated by constant potentialelectrolysis prior to analysis. The resulting methods, i.e., LSASV,DPASV and SWASV and their combination, allow LODs down to10−12 M to be reached (Bard and Faulkner, 2001). For instanceSuperville et al. assembled an automatic anodic stripping analysissystem with a SMDE to undertake a real-time routine analysis ofthe dynamic behavior of trace metals (Zn, Pb, Cd) in river, pondand seawater (Superville et al., 2011) (see Table 2 for quantitativefeatures). Furthermore a CSV was included to estimate simulta-neously the concentration of dissolved oxygen and reduced sulfurspecies. Magnier et al. perfected a procedure to assay lead and zincby ASV and copper by CSV in certified reference freshwater andin the French Deûle river, Cu analysis requiring the complexationwith 8-hydroxyquinoline (Magnier et al., 2011).

The need for determination of very low concentrations hasfavored the development of specific electrochemical techniqueswith new potential perturbation modes providing high resolutionand/or improved sensitivity. In this way Zlatev et al. particu-larly emphasized the advantages of differential alternative pulsesvoltammetry (DAPV) on HMDE to analyse mixtures of speciesexhibiting very close half-wave potentials (like Pb2+ and Tl+or Co2+ and Ni2+) or species couples with high concentrationratios (Zlatev et al., 2006) for which the analysis by DPV is ham-pered by complete peaks overlapping. DAPV takes advantageof the high resolution power of the second-order voltammet-ric techniques (as radio-frequency polarography) combined withthe high sensitivity and instrumental simplicity of DPV or SWV.DAPV principle is based on the superimposition on the mainelectrode potential E of a pair of single successive rectangularpulses characterized by small, equal amplitudes (

Tab

le2

|Su

mm

ary

of

anal

ytic

alp

erfo

rman

ces

and

exp

erim

enta

lco

nd

itio

ns

ob

tain

edfo

rh

eavy

met

als

det

ecti

on

inFr

ench

scie

nti

fic

acad

emic

com

mu

nit

y.

Cla

ssifi

cati

on

Ref

eren

ces

Ele

ctro

chem

ical

Det

ecti

on

Tech

niq

ue

An

alyt

e(s)

LOD

Lin

ear

An

alyt

e

pla

tfo

rmm

ediu

mco

nce

ntr

atio

nra

nge

pre

trea

tmen

tco

nd

itio

ns

PO

LAR

OG

RA

PH

Y

Onl

ine

DPA

SVTo

talZ

n2.

91nM

12.4

–23.

2nM

30s

at−1

.3V

Tota

lPb

0.03

nM1.

7–3.

2nM

60s

at−0

.7V

Mag

nier

etal

.,20

11H

MD

Em

onito

ring

river

DP

CSV

Tota

lCu

0.6

nM4.

9–7.

6nM

30s

at−1

.1V

follo

wed

byan

adso

rptio

nst

epat

−0.2

5V

durin

g15

s

Ris

oet

al.,

2006

HgF

EW

ater

sam

ples

(tre

ated

)S

CP

Fe(II

I)1.

5nM

NC

6cy

cles

of:0

.04

V(9

s)an

d−0

.4V

(1s)

Tang

uyet

al.,

2010

HgF

ES

eaw

ater

sam

ples

(tre

ated

)

SC

PS

b(III

)70

pMde

pend

ing

onsa

mpl

e30

0s

at−0

.45

V

Sla

dkov

etal

.,20

03H

gFE

0.1

MH

NO

3S

WC

SVS

e(IV

)0.

8nM

1–10

00nM

300

sat

−0.4

5V

Cu

0.7

nM

Ris

oet

al.,

1997

HgF

ES

eaw

ater

SC

PP

b14

pMN

C15

min

at−1

.1V

Cd

9pM

Cug

net

etal

.,20

09S

Pμ

EA

s0.

2M

acet

ate

(pH

=4.

5)S

WA

SVC

d(II)

11.6

nM11

.6–8

9nM

300

sat

−1V

Zn2+

4nM

Para

tet

al.,

2011

aH

MD

E0.

1M

KN

O3

AG

NE

S-S

CP

Cd2

+2.

9nM

atle

ast

25–1

00nM

1400

sto

talw

ithco

mpl

expr

oced

ure

Pb2

+4.

1nM

Para

tet

al.,

2007

Mem

bran

ean

dH

gfil

mS

PE

0.2

Mac

etat

e(p

H=

4–7)

SW

ASV

Cd(

II)2

nM5–

100

nM60

sat

−1V

Mun

tean

uet

al.,

2009

Mer

cury

mon

olay

erca

rbon

fiber

elec

trod

e

NC

SW

ASV

Pb(

II)80

fM1–

10pM

1s

at−1

.2V

Para

tet

al.,

2011

bH

gfil

mS

PE

0.2

Mac

etat

e(p

H=

4.6)

SS

CP

Cd

2.2

nMN

C60

sat

−1V

Zaou

aket

al.,

2010

aH

gfil

mS

PE

0.2

Mac

etat

e(p

H=

4.5)

SW

ASV

Cd

1.78

nM1.

78–3

56nM

60s

at−1

V

BiF

ILM

SG

uoet

al.,

2005

GC

/BiF

EM

ilkve

tch

in0.

2M

KS

CN

ASV

Zn(II

)9.

6nM

500–

3000

nM12

0s

at−1

.4V

(Con

tinue

d)

Tab

le2

|Co

nti

nu

ed

Cla

ssifi

cati

on

Ref

eren

ces

Ele

ctro

chem

ical

Det

ecti

on

Tech

niq

ue

An

alyt

e(s)

LOD

Lin

ear

An

alyt

e

pla

tfo

rmm

ediu

mco

nce

ntr

atio

nra

nge

pre

trea

tmen

tco

nd

itio

ns

Lege

aiet

al.,

2005

GC

/BiF

E0.

125

MH

NO

3

+0.

04M

H2N

SO

3H

DPA

SVC

d(II)

∼10

nM20

–100

0nM

1200

sat

−0.9

5V

Lege

aiet

al.,

2006

Cu/

Bifi

lmel

ectr

ode

0.01

Mam

mon

iabu

ffer

(pH

=9)

SW

ASV

Ni2

+N

C10

–100

0nM

900

sat

−0.7

V

Lege

aian

dVi

ttor

i,20

06C

u/N

afion

/Bi

elec

trod

e0.

01M

NaC

l+0.

001

MN

aHC

O3

DPA

SVC

d2+

6.05

nM17

.8–1

07nM

300

sat

−0.9

5VP

b2+

3nM

9.65

–86.

9nM

Urb

anov

aet

al.,

2010

Hig

hly

poro

usB

ifil

mel

ectr

odes

0.1

Mac

etat

ebu

ffer

(pH

=4.

5)

DPA

SVC

d(II)

5.34

nM17

8–11

60nM

90s

at−0

.95

VP

b(II)

6.27

nM96

.5–6

27nM

Zaou

aket

al.,

2009

Bi-C

oate

dS

Pμ

E0.

2M

acet

ate

buffe

r(p

H=

4.5)

SW

ASV

Cd(

II)11

.6nM

45–4

00nM

120

sat

−1V

Luet

al.,

2010

Bid

oped

carb

onS

PE

Air

SW

ASV

Pb(

II)as

vapo

r1

ng10

–80

ng12

0s

at−1

.2V

CA

RB

ON

ELE

CT

RO

DE

MA

TE

RIA

LS

DLC

/GC

Feie

ret

al.,

2012

Gra

phite

felt

0.1

MN

aBF 4

LSA

SVZn

(II)

50nM

1–10

0μ

M30

0s

at−1

.4V

Nas

raou

ieta

l.,20

09G

raph

itefe

lt0.

1M

LiC

lO4

LSA

SVP

b(II)

1nM

10–5

00nM

300

sat

−1V

Kha

dro

etal

.,20

09G

Cel

ectr

ode

0.1

MH

Cl

DPA

SVN

i(II)

2.56

nM8.

52–9

370

nM60

sat

−1V

0.1

Mac

etic

buffe

rH

g(II)

0.15

nM0.

5–17

40nM

Cd(

II)4.

83nM

4.83

–121

nM

Kha

dro

etal

.,20

11B

-dop

edD

LCA

ceta

te(p

H=

4.2)

SW

ASV

Pb(

II)8.

9nM

8.9–

222

nM90

sat

−1.3

V

Ni(I

I)34

.1nM

34.1

–256

nM

Hg(

II)4.

99nM

5–12

5nM

BD

DLe

etal

.,20

12B

DD

acet

ate

pH=

5.2

SW

ASV

Pb(

II)19

.3nM

96.5

–480

nM60

0s

at−1

V

Cu(

II)14

.2nM

47–3

15nM

ElT

alle

tal

.,20

07B

DD

0.01

Mac

etat

eD

PASV

Pb(

II)5.

55nM

18–2

17nM

60s

at−1

.9V

Zn(II

)25

.5nM

77–3

05nM

Cd(

II)3.

2nM

11–2

22nM

(Con

tinue

d)

Tab

le2

|Co

nti

nu

ed

Cla

ssifi

cati

on

Ref

eren

ces

Ele

ctro

chem

ical

Det

ecti

on

Tech

niq

ue

An

alyt

e(s)

LOD

Lin

ear

An

alyt

e

pla

tfo

rmm

ediu

mco

nce

ntr

atio

nra

nge

pre

trea

tmen

tco

nd

itio

ns

Cd(

II)3.

29nM

Sba

rtai

etal

.,20

12B

DD

Pota

ssiu

mci

trat

e/H

Cl

DPA

SVP

b(II)

26.5

nMN

F20

sat

−1.7

V

Ni(I

I)11

6nM

Hg(

II)11

.5nM

ISE

s

CA

LIX

AR

EN

EYa

ftia

net

al.,

2006

Cal

ix[4

]are

neC

ompl

ex(p

H=

3.5–

5)S

CP

Pb(

II)1.

4μ

M10

μM

–10

mM

No

accu

mul

atio

nan

del

ectr

olys

is

Yaft

ian

etal

.,20

07C

alix

[4]a

rene

Com

plex

(pH

=3–

7)S

CP

Pb(

II)4

nM10

nM–1

00μ

MN

oac

cum

ulat

ion

and

elec

trol

ysis

CH

ALC

OG

EN

IDE C

alie

tal

.,20

02C

u-A

s-S

KN

O3

SC

PC

u(II)

1μ

M2

μM

–10

mM

No

accu

mul

atio

nan

del

ectr

olys

is

Ess

iand

Prad

el,

2011

a,b

Cu-

Ag-

SC

ompl

ex(p

H=

3–5)

SC

PC

u(II)

1μ

MN

CN

oac

cum

ulat

ion

and

elec

trol

ysis

Mea

ret

al.,

2005

Ge 2

8S

e 60S

b 12

KN

O3

(pH

=3)

SC

PC

d(II)

1μ

M1

μM

–10

mM

No

accu

mul

atio

nan

del

ectr

olys

isC

ME

s

MIN

ER

ALS

Wal

cariu

set

al.,

1999

aS

ilica

mod

ified

CP

E0.

2M

HN

O3

SW

ASV

Cu(

II)2

nM5

nM–5

μM

600

sac

cum

ulat

ion

follo

wed

by30

sat

−0.5

V

Wal

cariu

set

al.,

2000

Sili

ca-m

odifi

edel

ectr

ode

0.1

MH

NO

3S

WA

SVH

g(II)

50nM

200

nM–1

0μ

M60

0s

accu

mul

atio

nat

open

circ

uit

follo

wed

by60

sat

−0.5

V

Wal

cariu

set

al.,

1999

bS

ever

alsi

lica/

hybr

idC

PE

with

amin

efu

nctio

naliz

atio

n

0.05

Mac

etic

acid

+0.

05M

NaN

O3

LSA

SVC

u(II)

NC

uncl

ear

Sev

eral

accu

mul

atio

ntim

ean

d24

0s

at−0

.4V

Etie

nne

etal

.,20

01O

rgan

ical

lym

odifi

edsi

lica

0.1

Mso

dium

acet

ate

SW

ASV

Cu(

II)3

nM50

–200

nM60

sat

−0.5

V

Say

enet

al.,

2003

Car

nosi

nesi

lica

hybr

idm

ater

ial

mod

ified

CP

E

0.1

MN

aNO

3+

0.01

MH

NO

3

DPA

SVC

u(II)

4nM

50–1

000

nM90

sat

−0.5

V

Wal

cariu

san

dS

ibot

tier,

2005

Am

ine-

func

tiona

lized

poro

ussi

lica

film

son

Au

0.1

MH

NO

3+

0.1

MN

aNO

3in

95%

etha

nol

DPA

SVC

u(II)

40nM

0.1–

10μ

M60

0s

accu

mul

atio

nfo

llow

edby

60s

at−0

.4V

Etie

nne

etal

.,20

07S

urfa

ctan

t-te

mpl

ated

thio

l-fun

ctio

naliz

edsi

lica

thin

film

s

0.5

MH

Cl

DPA

SVA

g(I)

6nM

0.2–

10μ

M96

0s

accu

mul

atio

nfo

llow

edby

60s

at−0

.6V

(Con

tinue

d)

Tab

le2

|Co

nti

nu

ed

Cla

ssifi

cati

on

Ref

eren

ces

Ele

ctro

chem

ical

Det

ecti

on

Tech

niq

ue

An

alyt

e(s)

LOD

Lin

ear

An

alyt

e

pla

tfo

rmm

ediu

mco

nce

ntr

atio

nra

nge

pre

trea

tmen

tco

nd

itio

ns

San

chez

and

Wal

cariu

s,20

10G

C/M

TTZ

0.1

MH

Cl

SW

ASV

Hg(

II)2

nM50

nM–1

μM

300

sat

−0.4

V

Wal

cariu

set

al.,

1998

Mes

opor

ous

pure

silic

am

odifi

edca

rbon

past

eel

ectr

ode

0.2

MH

NO

3S

WA

SV(o

rC

Vfo

rla

rger

amou

nts)

Cu(

II)30

nMN

C30

0s

accu

mul

atio

nfo

llow

edby

60s

at−0

.5V

Hg(

II)50

nMN

C

Tonl

eet

al.,

2005

Cla

ysgr

afte

dw

ithor

gani

cch

elat

ing

grou

ps(t

hiol

oram

ine)

mod

ified

CP

E

0.1

MH

NO

3D

PASV

Hg(

II)68

nM(t

hiol

)87

nM(a

min

e)

100–

700

nM18

0s

accu

mul

atio

nfo

llow

edby

60s

at−0

.4V

(or−0

.6V

depe

ndin

gon

the

med

ium

)

Tonl

eet

al.,

2011

Thio

l-fun

ctio

naliz

edcl

aym

odifi

edC

PE

0.2

MH

NO

3S

WA

SVP

b(II)

60nM

0.3–

10μ

M60

0s

accu

mul

atio

nfo

llow

edby

60s

at−0

.9V

Tchi

nda

etal

.,20

07G

C/P

CH

-SH

0.1

MH

Cl+

5%th

iour

eaD

PASV

Hg(

II)0.

4nM

4–20

nMan

d50

–80

nM12

00s

accu

mul

atio

nfo

llow

edby

180

sat

−0.7

VM

AC

RO

CYC

LIC

CO

MP

OU

ND

SR

ouis

etal

.,20

13β-k

etoi

min

eca

lix[4

]are

neon

ITO

0.05

Mam

mon

ium

acet

ate

(pH

=7)

Impe

danc

eH

g2+

NC

0.1

nM–0

.5μ

M

Gou

bert

-Ren

audi

net

al.,

2009

aC

ycla

m-

func

tiona

lized

silic

aC

PE

3M

HN

O3

SW

ASV

Cu(

II)0.

8nM

2–10

0μ

M18

00s

accu

mul

atio

nfo

llow

edby

60s

at−0

.5V

Gou

bert

-Ren

audi

net

al.,

2009

b(T

ETA

M)g

raft

edto

silic

age

land

orde

red

mes

opor

ous

silic

a

0.1

Mam

mon

ium

acet

ate

buffe

r(p

H=

7)

SW

ASV

Pb(

II)2.

7nM

10–1

00nM

900

sac

cum

ulat

ion

follo

wed

by60

sat

−0.8

V

Nas

raou

iet

al.,

2010

aTE

TRA

M-m

odifi

edgr

aphi

tefe

ltel

ectr

ode

0.1

Maq

ueou

sso

lutio

nof

LiC

lO4

LSA

SVP

b(II)

25nM

100–

250

nMar

ound

1800

sac

cum

ulat

ion

follo

wed

by30

0s

at−1

V

Nas

raou

iet

al.,

2010

bC

ycla

m-m

odifi

edgr

aphi

tefe

lt0.

5M

H2S

O4

LSA

SVP

b(II)

25nM

seve

rala

ccum

ulat

ion

time

follo

wed

by30

0s

at−1

V

Para

tet

al.,

2006

Hg

film

mod

ified

0.1

MK

NO

3LS

ASV

Cd(

II)6

nMN

C12

0s

at−1

VS

PE

Pb(

II)8

nM

(Con

tinue

d)

Tab

le2

|Co

nti

nu

ed

Cla

ssifi

cati

on

Ref

eren

ces

Ele

ctro

chem

ical

Det

ecti

on

Tech

niq

ue

An

alyt

e(s)

LOD

Lin

ear

An

alyt

e

pla

tfo

rmm

ediu

mco

nce

ntr

atio

nra

nge

pre

trea

tmen

tco

nd

itio

ns

Bet

elu

etal

.,20

07H

gfil

m+

mem

bran

e0.

01M

NaH

CO

3LS

ASV

Cd(

II)N

CN

C12

0s

at−1

V

mod

ified

SP

EP

b(II)

PO

LYM

ER

SH

eitz

man

net

al.,

2007

Poly

(pyr

role

-ED

TAlik

e)fil

m0.

1M

buffe

r(p

H=

5)S

WA

SVC

d(II)

,Pb(

II)an

dC

u(II)

NC

NC

600

sac

cum

ulat

ion

follo

wed

by40

sat

−1.2

Vfo

rP

b(II)

and

−0.9

Vfo

rC

u(II)

Hg(

II)0.

5nM

NC

Bui

caet

al.,

2009

aPo

ly(E

DTA

-like

)Film

0.1

Mac

etat

ebu

ffer

(pH

=4.

5)

DPA

SVC

u(II)

600

sac

cum

ulat

ion

follo

wed

by60

sat

−0.4

V

Bui

caet

al.,

2009

bPo

ly(p

yrro

le-E

DTA

)m

odifi

edel

ectr

ode

0.1

Mac

etat

ebu

ffer

(pH

=4.

5)

DPA

SVH

g(II)

10nM

(impr

inte

dpo

lym

ers)

10–1

000

nM(im

prin

ted

poly

mer

s)60

0s

accu

mul

atio

nfo

llow

edby

180

sat

−1.8

V

Pb(

II)0.

5nM

10–1

000

nM60

0s

accu

mul

atio

nfo

llow

edby

40s

atC

u(II)

5nM

25–2

50nM

−0.9

VH

eitz

man

net

al.,

2005

Poly

(pyr

role

-mal

onic

acid

)film

mod

ified

carb

onel

ectr

ode

0.2

Mac

etat

ebu

ffer

(pH

=4.

4)

SW

ASV

Hg(

II)50

nMN

C

Cd(

II)0.

2μ

M1–

10μ

M60

0s

accu

mul

atio

nfo

llow

edby

40s

at−1

.1V

Pb(

II)0.

5nM

10–1

000

nM

Pere

iraet

al.,

2011

Com

plex

ing

poly

mer

film

s0.

1M

acet

ate

buffe

r(p

H=

4.4)

SW

ASV

Cu(

II)5

nM25

–250

nM60

0s

accu

mul

atio

nfo

llow

edby

40s

at−0

.9V

or−1

.1V

for

Cd(

II)H

g(II)

100

nM10

0–10

00nM

Cd(

II)50

0nM

100–

1000

0nM

Riv

aset

al.,

2006

Com

plex

ing

poly

mer

film

s0.

1M

acet

ate

buffe

r(p

H=

4.8)

SW

ASV

Pb(

II)N

C0.

01–5

mM

600

sac

cum

ulat

ion

follo

wed

by40

sat

−0.6

V

Bes

sbou

sse

etal

.,20

11N

anop

orou

sβ-P

VD

Fm

embr

ane

elec

trod

e0.

1M

sodi

umac

etat

eS

WA

SVP

b(II)

0.63

nMN

C30

min

equi

libriu

mfo

llow

edby

100

sat

−0.8

V

Zejli

etal

.,20

07Po

lyth

ioph

ene

film

0.2

MK

NO

3

(pH

=5)

DPA

SVA

g(I)

0.56

μM

0.65

–9.3

μM

120

sat

−0.5

V

Yasr

iet

al.,

2011

GC

/PE

DO

T:P

SS

HC

l(pH

=2.

2)C

AP

b(II)

0.19

nM2–

100

nM30

sat

−0.6

5V

(Con

tinue

d)

Tab

le2

|Co

nti

nu

ed

Cla

ssifi

cati

on

Ref

eren

ces

Ele

ctro

chem

ical

Det

ecti

on

Tech

niq

ue

An

alyt

e(s)

LOD

Lin

ear

An

alyt

e

pla

tfo

rmm

ediu

mco

nce

ntr

atio

nra

nge

pre

trea

tmen

tco

nd

itio

ns

NP

sO

ttak

amTh

otiy

let

al.,

2012

Au/

MP

S-(P

DD

A-

AuN

Ps)

Pho

spha

tebu

ffer

(pH

=8)

DPA

SVA

s(III

)0.

48μ

MN

C

Hez

ard

etal

.,20

12a

GC

+A

uNP

s0.

01M

HC

lS

WA

SVH

g(II)

0.42

nM0.

64–4

nM30

0s

at0

V

Hez

ard

etal

.,20

12b

GC

+A

uNP

s0.

01M

HC

lS

WA

SVH

g(II)

0.4

nM0.

8–9.

9nM

300

sat

0V

BIO

SE

NS

OR

SC

hout

eau

etal

.,20

04A

lkal

ine

phos

phat

ase

10m

MTr

is-H

Cl

buffe

r(p

H=

8.5)

/1

mM

MgC

l 2

Con

duct

omet

ryC

d2+

8.9

nMN

C

Cho

utea

uet

al.,

2005

Alk

alin

eph

osph

atas

eA

cety

lcho

lines

tera

se

10m

MTr

is-H

Cl

buffe

r(p

H=

8.5)

/1m

MM

gCl 2

Con

duct

omet

ryC

d2+

89nM

NC

30m

inin

cuba

tion

Zn2+

0.15

μM

Teka

yaet

al.,

2013

Alk

alin

eph

osph

atas

e5

mM

HE

PE

Sbu

ffer

(pH

=8.

1)

Con

duct

omet

ryC

d2+

10−2

0M

NC

24h

incu

batio

n

Hg2

+

Sol

datk

inet

al.,

2012

Inve

rtas

e,m

utar

otas

e,gl

ucos

eox

idas

e

5m

Mph

osph

ate

buffe

r(p

H=

6.5)

Con

duct

omet

ryH

g2+

25nM

NC

20m

inin

cuba

tion

Ag+

100

nM

Moh

amm

adie

tal

.,20

05In

vert

ase,

mut

arot

ase,

gluc

ose

oxid

ase

0.1

Mph

osph

ate

buffe

r(p

H=

6)

Am

pero

met

ryH

g(II)

NC

10nM

–1μ

M20

min

incu

batio

nat

pH=

4

Gay

etet

al.,

1993

L-la

ctat

ede

hydr

ogen

ase

L-la

ctat

eox

idas

e

0.1

MTr

isbu

ffer

(pH

=9)

Am

pero

met

ryH

g(II)

1μ

MN

C5

min

incu

batio

n

Ag+

0.1

μM

Cd2

+10

μM

Zn2+

10μ

M

Pb2

+50

μM

Cu2

+25

0μ

M

* All

para

met

ers

give

nar

eth

eon

esfo

rth

ebe

stLO

Ds.

potential equal to the half-wave potential) with peak amplitudesproportional to the electroactive species concentration. The res-olution power of DAPV was highlighted through the analysis ofa solution containing Pb2+, Tl+, In3+, and Cd2+. The sensitivityand the LOD (54 nM) were found to be similar to those obtainedusing classical DPV but with species having half-wave potentialsdifference in the range from 28 to 50 mV and concentration ratiosfrom 1:1 up to 80:1 without any preliminary preparation of thesample.

Mercury filmDespite very good analytical performances in terms of sen-sitivity and stability of the response vs. time, the low vaporpressure and the high toxicity of mercury encouraged exten-sive researches on polarographic methods involving reducedamounts of mercury. One way consists in thin film mercuryelectrodes (TFME) electrodeposited on solid state materials likeglassy carbon (GC). The group of Riso used SCP with lowconstant current for the quantification of Fe(III) in estuarineand coastal filtered waters (Riso et al., 2006). The procedurewas proved to be highly sensitive but analysis required severalpre-treatment steps, i.e., filtration of the sample and complex-ation with solochrom violet. They also succeeded in detect-ing ultra-trace (70 pM) Sb(III) in seawater by using the sameelectrochemical method. The application of a double electrol-ysis potential during the pre-concentration step allowed theanalysis to be independent from the Cu level (Tanguy et al.,2010).

For the determination of Se(IV), insufficient reproducibilityand sensitivity of Hg film was observed by Sladkov et al. (2003).This problem was overcome by incorporating Cu(II) ions dur-ing the plating procedure on GC electrode surface. The metallicCu dissolved in the Hg film was found to play an important rolein peak current enhancement. A LOD 0.8 nM was reached bySWCSV and the relative standard deviation was 5.2% (n = 5) for1 μM Se(IV).

A potentiometric stripping method has been proposed byRiso et al. for the simultaneous measurement of Cu, Pb, andCd in ocean waters (Riso et al., 1997). The mercury coating waselectrodeposited in situ on a GC rotating electrode at the begin-ning of each analysis by applying a potential step at −1.1 V/SCEfor 10 min. Then an electrolysis-stripping cycle was carried out.Metals concentrations were compared with a reference standardsolution containing all three metals. The obtained LODs were0.7 nM, 15 pM and 9 pM for Cu, Pb, and Cd, respectively.However, it has to be noticed that the total duration of the analysiswas quite long, about 75 min.

In order to approach solid mercury-free electrodes, Munteanuet al. worked on the electrodeposition of a mercury monolayerby constant potential electrolysis with increasing electrolysis time(Munteanu et al., 2009). An exceptional sensitivity for Pb2+ assaywas obtained when the mercury monolayer-on-carbon electrodewas used with fast (v > 1 kV/s) ASV. This result was revealed to bedue to the ionization of Pb atoms in the mercury layer, which cat-alyzes the oxidation of atomic hydrogen adsorbed on the Hg layer.A remarkable LOD of 80 fM was recorded on a cylinder electrodewith a 1 s preconcentration time.

Mercury screen printed electrode (HgSPE)Potin-Gautier’s group has developed an alternative strategy basedon mercury micro arrays screen printed electrodes (SPE). Thismicro system allowed mass transport of the analytes to beenhanced compared to macro systems. This device was success-fully tested for Cd2+ detection in synthetic and river samples,providing a LOD of 11.6 nM using SWASV (Cugnet et al., 2009).More recently SCP was implemented as the second stage of theelectrochemical technique “Absence of Gradients and NernstianEquilibrium Stripping” (AGNES-SCP) for the determination offree metal concentration, namely Zn2+, Cd2+, and Pb2+, withboth an HMDE and a SPE (Parat et al., 2011a). The results showedhigher sensitivities and lower LOD and LOQ (in the nanomo-lar range) with SPE, which was linked to a higher product ofmercury volume times the gain (AGNES parameter). Finally,SPE was modified by a microwell for the assay of labile Cd2+,thus reducing the sample volume down to 200 μL. A LOD of2 nM was reached using SWASV with only a 60 s preconcen-tration step. Unfortunately, the performances of this electrodewere pH-dependent out of the pH range 4–7 (Parat et al., 2007).Nevertheless all these mercury-based techniques are promised todisappear in a few years since no mercury will be authorized from2015 (European Directive 2008/105/EC).

BISMUTHBismuth film electrode (BiFE) is often considered to be a goodalternative to Hg electrode and has been extensively used forelectroanalysis. More “environmentally friendly” and less toxicalthan Hg, Bi is considered to be a safe material, as it is a non-carcinogenic element (except for foetus and embryo) (Svancaraet al., 2010). However, in high doses, it presents similar toxic-ity to other heavy metals apart from these effects are much morereversible. The main advantage of Bi with respect to trace analy-sis is its capability to form binary or multi-component alloys withnumerous other heavy metals (Svancara et al., 2010). It has alsothe particularity to be the most diamagnetic metal, thus avoid-ing conductance problems. Another interest of Bi compared toHg is its insensitivity to dissolved oxygen, thus making unneces-sary any deaeration step. Most of the time, the Bi film is platedbefore analysis onto the electrode by potentiostatic reduction of aBi(III) solution (Wang et al., 2001), although codeposition dur-ing trace metal reduction has been also reported (Wang et al.,2000). BiFE performances compare favorably with Hg electrodes,affording high sensitivity and well-defined stripping signals. Forinstance, Guo et al. compared the anodic stripping voltammetricresponse of HgFE and BiFE obtained in a 10 mM Zn(II) solution(Guo et al., 2005). The BiFE presented a well-defined and higherstripping peak compared to HgFE. Its sensitivity was found to betwice higher than that obtained on HgFE. This electrode was thenused to detect zinc contained in milk vetch used in traditionalChinese medicine. The response was linear in the range from 0.5to 3 μM and a LOD of 10 nM was reported. Nevertheless, it hasto be noticed that the cathodic limit of the potential window ishigher than on Hg.

In 2005, Legeai et al. proposed an interesting alternative to theclassical Pt or GC substrates by electrodepositing Bi film onto Cusince the adherence of the film was found to be better in this

latter case (Legeai et al., 2005). By using DPASV as the detec-tion method, this BiFE exhibited very good performances towardCd2+ assay in acidified tap water with a linear concentrationrange between 10−8 and 10−6 M. It was also successfully testedfor the simultaneous determination of Cd2+, Zn2+ and Pb2+ ionsat 10−5 M. Good accuracy (

FIGURE 2 | Voltammograms for a-CNx (A) and BDD (B) in solutions containing different Cd2+ concentrations. Reprinted with permission from Secket al. (2012). Copyright 2012 Wiley-VCH.

radio-frequency magnetron sputtering technique. The amountof nitrogen incorporated into the film can be controlled by thecomposition of the plasma (N2/Ar ratio) used for the deposi-tion. The main property of a-CNx electrodes concerns their broadelectrochemical window, which makes it particularly suitable forelectrodetection of many species. This was exploited by Seck et al.for the simultaneous assay of Cd2+ and Cu2+ (Seck et al., 2012)(Figure 2). The presence of Cu modified the peak current relatedto Cd2+ as compared to Cd2+ detection in Cu-free solution.Anyway the a-CNx electrode allowed the detection of 2 ppb ofCd2+ with concentration of Cu2+ up to 140 ppb.

Diamond like carbonDiamond like carbon (DLC) is a carbon-based material contain-ing a mixture of sp2 (graphite) and sp3 (diamond) carbon phases.Several methods have been developed to produce DLC films:plasma enhanced chemical vapor deposition (CVD) techniques,ion beam, filtered cathodic vacuum arcs. DLC exhibits someunique properties, such as high elastic modulus, high mechani-cal hardness, very low surface roughness and chemical inertness.Khadro et al. used DLC doped with boron to improve its conduc-tivity (Khadro et al., 2011). The resulting electrode was exploitedfor the simultaneous assay of many heavy metals in water, namelyCd2+, Pb2+, Ni2+, and Hg2+, in concentration ranges up to200 nM. LODs of 8.9, 4.8, 34, and 5 nM were reached, respectively.

Boron doped diamondBoron doped diamond (BDD) is the most recent carbon-basedmaterial used for electroanalytical purpose. Diamond films canbe deposited using CVD systems involving activation of gases byeither microwave plasma or a hot filament. Traditionally electricalconductivity of diamond films is obtained through doping withboron (p-type behavior). The advantages of such material aremanifold compared to previous carbon-based ones: BDD has anextremely high chemical stability, presents a wide potential win-dow in aqueous media (−1.35 to +2.3 V/NHE at 0.1 mA cm−2 in0.5 M H2SO4) and generates a low background current. Moreoverit is extremely resistant to fouling phenomena, thus making its

surface state very reproducible with time. Unlike graphite felt,BDD exhibits however lower specific area. In order to increase thesensitivity two major ways have been envisaged. In the one hand,Le et al. associated a BDD electrode with a microelectrodialyseras a preconcentration step (Le et al., 2012). The correspondingdevice allowed the assay of Pb2+ ions with a linear concentra-tion range between 96 and 490 nM and a LOD of 19 nM. Thesame analytical device was used for the simultaneous detectionof Zn, Cd, Pb, and Cu (El Tall et al., 2007). Quantification waspossible for the first three heavy metals but the presence of Cucaused interferences. Compared to GC, BDD electrode exhibitedan enhanced sensitivity (3 or 5 times) and a longer lifetime inreal samples (El Tall et al., 2007). In the other hand, Sbartai et al.developed a new electrochemical microcell micromachined by afemtosecond laser for the simultaneous detection of Cd, Ni, Pb,and Hg (Sbartai et al., 2012) (Figure 3). Reduction of the elec-trode size resulted in mass transport amplification. LODs of 0.4,6.8, 5.5, and 2.3 nM were thus obtained, respectively. Quantitativeresults were recorded for concentrations up to 200 nM. These per-formances are comparable to those obtained on DLC by Khadroet al. for Pb2+ and more accurate for Hg2+, Cd2+, and Ni2+(Khadro et al., 2011). However, a non-linear calibration curvefor Hg was obtained in the former paper, which can be explainedby the presence of Cl− ions in the electrolyte solution, leading toHg2Cl2 formation at the electrode surface.

New carbon materials like a-CNx, DLC or BDD which exhibithigh chemical stability and offer a wide potential window holdsignificant promises for electronalytical applications. Moreover,there is a wealth of opportunities for nanoscale electrochemicaldevices based on carbon materials. Nevertheless, interceptionsbetween cations go through either the development of specificcalibrations or the chemical modification of the electrode.

ION SELECTIVE ELECTRODESIon-selective electrodes (ISEs) are potentiometric sensors thatinvolve a selective membrane which minimizes matrix interfer-ences (Bobacka et al., 2008). The response of these sensors isbased on an equilibrium state complexation reaction between the

FIGURE 3 | DPASV obtained with a BDD micromachined microcell on astandard solution of Cd (20 nM), Ni (38 nM), Pb (11 nM) and Hg(0.55 nM). Reprinted with permission from Sbartai et al. (2012). Copyright2012 American Chemical Society.

analyte and the probe with kinetic properties strongly dependingon the membrane composition. The potentiometric measure-ment as well as the nature of the molecular interactions generallyallow LODs around 1 μM to be reached and a linear concen-tration range from 10−5 to 10−2 M in a pH window from 3to 6. Innovative changes have been made in recent past yearsto improve these analytical performances. In this context severalfunctionalized materials have been used, including glass, liquidor polymer membranes. In the latter case an ionophore is gener-ally used as the sensing platform. In this way, Yaftian et al. firstsynthesized phosphorylated calix[4]arene coated on a graphiteelectrode for the assay of Pb traces (Yaftian et al., 2006). This sen-sor exhibited a particularly quite long lifetime (up to 8 weeks)and a relatively short response time (17 s). This latter was sig-nificantly shortened to 7 s by using hexahomotrioxacalix[3]areneas the ionophore, due to fast exchange kinetics complexation-decomplexation processes (Yaftian et al., 2007). Furthermore theconcentration range covered four decades (between 10−8 and10−4 M). The specificity of this ISE toward Pb2+ was success-fully tested in synthetic solution in the presence of 22 interferingspecies. Nevertheless no measurement was done in real samplewater. In both cases, the electrode was also used in potentiom-etry for the titration of Pb2+ solution using a standard EDTAsolution.

Glassy materials represent a good alternative, particularly inmicro-sensor fabrication. Beyond the advantages of all solid statedevices, vitreous materials are well-suited for the production ofhomogeneous thin layers which allow potentiometric measure-ments to be done despite their poor conductivity. Thereby, severaltypes of solid state membranes have been synthesized like thewidely used chalcogenide one, in which the conduction over themembrane is provided by halides or metallic ions. Chalcogenideglasses exhibit better chemical durability in acidic media, andin many cases, afford better selectivity and reproducibility thanarene ionophores. Several French research teams have investi-gated such ISEs toward Cu(II) determination. Cali et al. used

chalcogenide glassy-crystalline Cu-As-S alloys (Cali et al., 2002).The resulting sensor exhibited a very short response time (1–3 s)and a LOD close to 10−6 M with a linear concentration rangebetween 10−6 and 10−2 M. These results are available within a pHrange from 2 to 6. Similar analytical performances were obtainedby Essi et al., with a Cu-Ag-S thin film for the assay of Cu(II)and Ag(I) ions (Essi and Pradel, 2011a). They were satisfacto-rily compared with those obtained by ICP-MS in real samples.Furthermore the specificity of the sensor was not damaged by thesimultaneous presence of Ca2+, Mg2+, Pb2+, Cd2+, and Zn2+(Essi and Pradel, 2011b). Mear et al. investigated a thin film ofGe28Se60Sb12 chalcogenide glass including Cu in order to quantifyCu(II) ions (Mear et al., 2005). A linear range was found between10−5 and 10−3 M with a LOD of 3 μM. The concentration rangewas enlarged between 10−6 and 10−2 M for Cd(II) ions assayby coupling the membrane electrode with a pre-concentrationmodule.

CHEMICALLY-MODIFIED ELECTRODESFrom the analytical point of view, the wide and increasing suc-cess of chemically modified electrodes (CME) may be explainedby the offered possibility to purposely design the surface of con-ventional electrodes. By combining the intrinsic properties of themodifier and a given electrochemical reaction, CMEs exhibit sig-nificantly improved response compared to unmodified electrodes(Murray et al., 1987; Gilmartin and Hart, 1995; Cox et al., 1996).For heavy metals trace detection, the modification plays a criticalrole especially during the preconcentration step, by favoring selec-tive and enhanced accumulation, thus leading to higher sensitivityand lower LODs (Arrigan, 1994). During the detection step, themodification also often favors the electron transfer kinetics. Themodifier may be a mineral such as silica or clay, a polymer, aninorganic or organic compound or a metal nanoparticle basedmaterial. Depending on its nature, the modifier is associated tothe electrode by adsorption, covalent binding, coating or evendispersion into a conductive matrix.

MineralsMinerals such as silica-based materials, clays and zeolithes, are ofparticular interest for ion exchange voltammetry (IEV) (Wang,1989). Basically, they act as an ion selective film inside whichthe analytical target is preconcentrated at open-circuit potentialby an exchange process. In a second step, the analyte incorpo-rated within the ion-exchanger film is detected by using an anodicstripping technique (Walcarius, 1998).

In France, the research on silica-modified electrodes withrespect to heavy metals assay is mainly represented by the groupof Walcarius. In 1997 this group published the first report deal-ing with silica-modified carbon paste electrode (SMCPE) devotedto electroanalysis, with Cu(II) as the analytical target (Walcariusand Bessiere, 1997). By using a 10 min preconcentration step andSWASV, a LOD of 2 nM was reached. This SMCPE exhibited agood reproducibility since up to 30 detection procedures wereperformed over a period of a week without any noticeable loss ofsensitivity. However, this system suffered the classical drawbackof CPE, namely a gradual dissolution process. In this particularexample, another severe drawback was the necessary use of an

ammonia medium in order to ensure Cu(II) accumulation via itsinteraction with surface silanolate groups. This pioneering workhas been later extended to various silica-based materials and theinfluence of interfering species has been studied (Walcarius et al.,1999a). Clearly, this SMCPE failed at high ionic strength, sinceimportant cations concentration resulted in competition for theion-exchange sites. A very similar study has been reported withHg(II) as the analytical target (Walcarius et al., 2000). Using thesame procedure and materials, Hg(II) has been found to sufferno influence of ionic strength and no particular interference withother metallic species, even Ag(I). This result has been explainedtaking into account the formation of soluble Hg(II) hydroxides inthe experimental conditions used.

To overcome the drawback of adding a complexing agent inorder to ensure metal accumulation, the group of Walcarius hasdeveloped several organically modified silica CPEs. Aminopropylgroups (Walcarius et al., 1999b; Etienne et al., 2001) or a carno-sine dipeptide (Sayen et al., 2003) have been co-condensed withsilane precursors to afford the desired functionalized silica mate-rials. Cu(II) was detected with similar LODs to those reported inthe former study (Walcarius and Bessiere, 1997) without addinganything to the accumulation medium. The aminopropyl-graftedsilica CPE was successfully tested for Cu(II) detection in labora-tory tap water (Etienne et al., 2001). Cu(II) suffered importantcompetition from Co(II), Ag(I), and Hg(II) for the binding sitesof the carnosine-modified silica CPE, thus limiting the practicalusefulness of this sensor in real media (Sayen et al., 2003).

An interesting alternative to CPE has been proposed laterby using silica films coated onto Au (Walcarius and Sibottier,2005) or GC (Etienne et al., 2007; Sanchez and Walcarius,2010) electrodes. Whereas these films have been preparedmost of the time by a classical surfactant-templated synthe-sis (Etienne et al., 2007; Sanchez and Walcarius, 2010) anoriginal electrochemically-induced sol-gel deposition has been

also reported (Walcarius and Sibottier, 2005) (Figure 4). All thesefilms were functionalized either by thiol or amine groups. Cu2+(Walcarius and Sibottier, 2005), Ag+ (Etienne et al., 2007) andHg(II) (Sanchez and Walcarius, 2010) were the analytical targets,and the LODs obtained were 40, 6, and 24 nM respectively.

From a more general point of view, all these studies on silica-based modifiers proved that the key feature with respect toanalytical performances, and particularly sensitivity, is the ana-lyte diffusion inside the porous structure of silica, thus makingporosity a more predominant factor than the amount of surfacesilanol groups (Walcarius et al., 1998). Thus, mesoporous silica-based materials, which exhibit well-defined three-dimensionalstructures, appear to be much more adapted for heavy metalpreconcentration than amorphous ones (Walcarius et al., 2003),whatever they are functionalized (Ganesan and Walcarius, 2008)or not (Walcarius et al., 1998). When functionalized, the amountof organic groups is also of importance, its effect on the ana-lyte preconcentration passing by a maximum, since too muchorganic groups lead to a decrease in pore size which hampers theaccessibility of the binding sites (Etienne et al., 2007; Ganesan andWalcarius, 2008; Sanchez and Walcarius, 2010).

Clays may also be used to perform IEV. These minerals exhibitrelatively large specific surface areas and ion-exchange propertiesassociated to an ability to sorb and intercalate many compounds.In a very close approach to what has been reported for silica-basedmaterial, their surface may be functionalized by organic groups(Navratilova and Kula, 2003), affording the possibility to tunecharge selectivity of clays in IEV (Tonle et al., 2004).

In France, the group of Walcarius explored the potentiali-ties of clays functionalized by organic groups and mixed withCPE toward Hg(II) (Tonle et al., 2003, 2005) and Pb(II) (Tonleet al., 2011) determination. With respect to Hg(II), the compari-son of thiol-functionalized vs. amine-functionalized clays showedthat the former one exhibited a better LOD (68 and 87 nM,

FIGURE 4 | Typical DPASV and calibration (inset panel) curves obtained for Cu2+ using a 10% amine functionalized silica film deposited on gold.Reprinted with permission from Walcarius and Sibottier (2005). Copyright 2005 Wiley-VCH.

respectively, using DPASV with 10 min accumulation), in accor-dance with the greater affinity of Hg for sulfur groups (Tonle et al.,2005). Thiol-functionalized clays have been also coated as thinfilms onto GC (Tchinda et al., 2007, 2009). The LOD has beengreatly improved since a 6 nM value was reached using DPASVwith only 3 min accumulation, the linear range being from 50 to800 nM (Tchinda et al., 2007). It has to be noticed that increas-ing the accumulation time allowed another linear range to beobserved from 4 to 20 nM together with a remarkable LOD of0.5 nM (Tchinda et al., 2007).

Macrocyclic compoundsMacrocyclic compounds have received much attention in thelast two decades due to their particular three-dimensional shapewhich provides a suitable cavity for selective complexation ofheavy metals (Kolthoff, 1979). French groups particularly focusedtheir research on cyclams (Goubert-Renaudin et al., 2009a,b;Nasraoui et al., 2010a,b), crown-ethers (Parat et al., 2006; Beteluet al., 2007) and calixarenes (Dernane et al., 2013; Rouis et al.,2013).

With respect to these latter compounds, Dernane et al. pro-posed to detect Cd2+ thanks to a p-tert-butylcalix[8]arene mem-brane deposited onto Au electrode (Dernane et al., 2013). Inthis work the metal cation coordination was favored by the pres-ence of the four ketone groups on the calixarene, thus allowinga 0.1 μM LOD to be reached. However the results seemed a littlebit cautious since detection was performed simply by CV withoutfurther preconcentration step. Moreover, the variation of peakcurrent density as a function of the logarithm of metal concentra-tion was proposed as a calibration plot, which makes non-sensewith respect to classical equations of CV.

Rouis et al. have built an impedancemetric sensor dedicatedto Hg2+ detection by immobilizing a β-ketoimine calix[4]arenederivative in a conducting poly(p-phenylene vinylene) membranedeposited onto indium-tin oxide (ITO) electrodes (Rouis et al.,2013). The main originality of this work was the study of lightexcitation effect of the β-ketoimine calix[4]arene while opti-mizing detection parameters. Charge transfer processes at theelectrode/electrolyte interface were found to be improved underlight excitation, thus providing an enhancement of the sensitivity

toward Hg2+. The best results were obtained under blue lightemission, providing a charge transfer resistance Rct = 4.47 k�and a linear range from 0.1 nM to 0.5 μM.

Parat et al. reported the use of a thin-film mercury-coatedscreen-printed carbon electrode covered by a crown-ether basedmembrane for the determination of Cd and Pb (Parat et al.,2006). The crown-ether membrane aimed at protecting the activesurface from interferences during the analysis. The size of crown-ethers cavity has been proved to impact the performances ofthe electrode. The best results have been obtained for the small-est crown-ether, because its cavity size fitted well to metal ionsradii, although Pb was a bit favored. LODs of 6 nM and 8 nM forCd2+ and Pb2+, respectively, have been reached by using LSASVwith 2 min preconcentration. Tests were successfully conductedon river water and soil solution extracts. This system was thenapplied in a later work to the semi-continuous monitoring of Cdand Pb in tap water by FIA and afforded reproducible results for42 h (Betelu et al., 2007) (Figure 5).

Goubert-Renaudin et al. have developed a new family of func-tionalized cyclams they grafted onto silica materials and mixedto CPE (Goubert-Renaudin et al., 2009a,b). In the first workdedicated to Cu(II) detection, the functionalization aimed atstabilizing cyclams on silica in order to improve the sensor life-time. However, increasing the number of functionalizations percyclam lead to an increase in cycle rigidity, which has been cor-related to lower performances with respect to Cu(II) uptake(Goubert-Renaudin et al., 2009a). The modified CPE affordedgood stability, with a 7% relative standard deviation for 45 mea-surements and was successfully tested on tap water. In a secondwork, dealing with Pb(II) determination, the aim of the func-tionalization was to improve the analyte preconcentration step byfavoring complexation (Goubert-Renaudin et al., 2009b). Thus,the inverse trend was observed, i.e., the more functionalized thecyclam, the better the performances. The best system allowedLODs down to 2.7 nM to be reached. Among potential interferingspecies, only Hg(II), Cd(II), and Cu(II) gave rise to significantloss of signal, mainly because of competition for the bindingsites. The group of Geneste also used such kind of cyclams tofunctionalize graphite felt electrodes (Nasraoui et al., 2010a,b).Accumulation of Pb(II) at open-circuit potential by flowing the

FIGURE 5 | (A) SW Voltammogram recorded every 12 h for 42 h analysis by semicontinuous flow injection of tap water doped with Cd (B) Variation of Cd andPb peak currents over the 42 h. Reprinted with permission from Betelu et al. (2007). Copyright 2007 Wiley-VCH.

aqueous solution to analyze, followed by LSASV allowed a 25 nMLOD to be reached. However, this system exhibited relatively poorstability since a 20% decrease of the response was reported afterregeneration.

PolymersElectroactive surface modification by means of polymer deposi-tion or electrodeposition represents a broad research field, leadingto numerous papers every year. With respect to trace metals anal-ysis, polymer films allow the immobilization on the electrode ofa large number of ligands which may complex metal ions to beaccumulated (Trojanowicz, 2003; El Kaoutit, 2012; Li et al., 2012).The polymers used for surface modification may be natural orprepared purposely by chemical synthesis. However, with respectto French groups’ research activities, no work was found dealingwith natural polymers dedicated to trace metal detection. Theonly papers found in the literature were about polysaccharides(Crini, 2005) and chitosan (Vieira et al., 2011) and concernedheavy metals removal.

Concerning chemically synthesized polymers, the group ofMoutet has developed two different groups of substitutedpolypyrrole derivatives for trace metal determination. The firstone is based on “poly(pyrrole-EDTA like)” and takes advantage ofthe well-known complexing properties of EDTA to improve metalpreconcentration (Heitzmann et al., 2007; Buica et al., 2009a,b).These studies were devoted to the assay of Cu(II), Pb(II), Cd(II),and Hg(II) (Figure 6). In the first two ones, the selectivity ofthe modified electrode has been tuned by varying the accumu-lation time and the pre-structuration of the polymer. The filmthickness was also proved to influence the selectivity. However,the electrode was insensitive to Cd(II) whatever the conditionsadopted. To overcome this problem, Heitzmann et al. have chosenan imprinted polymer strategy: the polymer was electrodepositedin the presence of Cd(II) ions which were then removed fromthe metallopolymer film (Heitzmann et al., 2007). The resultingfunctionalized electrode was thus able to detect Cd(II) as well asthe other three metal cations. By introducing 4 pyrrole fragmentson the same EDTA skeleton instead of only two in the formerstudies, an enhanced stability and a better controlled dimen-sionality was conferred to the polymer, thus making the sensorresponse independent on the film thickness (Buica et al., 2009a).The global complexing capability of the polymer was also greatlyimproved by the presence of 2 amine and 4 amide coordinatinggroups per monomer unit. The second group of polymers devel-oped by the group of Moutet is based on poly(pyrrole-malonicacid) (Heitzmann et al., 2005; Pereira et al., 2011). Here, theanalyte complexation occurred with the anionic form of mal-onic acid, which is known to easily coordinate various metalions. This sensor was also tested with Cu(II), Pb(II), Cd(II), andHg(II). It allowed LODs around 10−10 M to be reached for eachmetal cation, and exhibited good stability since the same currentresponse was obtained for 2 assays at a 3 weeks interval using thesame electrode stored without any particular precaution.

Another kind of polymer based on styrene units gave riseto one report by Moutet (Rivas et al., 2006). Styrene wascopolymerized with acetamide acrylic acid or itaconic acid.These latter are hydrophilic whereas styrene is hydrophobic,

FIGURE 6 | DPV curves recorded at a poly(EDTA-like) film modifiedcarbon electrode in acetate buffer containing Hg(OAc)2 (dotted line),Cu(OAc)2, Cd(NO3)2, Pb(NO3)2,Hg(OAc)2 (full line). Reprinted withpermission from Buica et al. (2009b). Copyright 2009 Wiley-VCH.

thus providing interesting mixed properties to the resultingcopolymers. The incorporated carboxilate groups were used forthe accumulation and detection of Pb2+ but the system exhib-ited poor performances since the response was linear only in thenarrow range from 10−5 to 10−3 M.

Bessbousse et al. proposed a more sophisticated systembased on a nanoporous β-poly(vinylidene fluoride) (β-PVDF)membrane (Bessbousse et al., 2011). The nanopores were fur-ther functionalized by track-etched poly(acrylic acid) (PAA),and thin porous Au films were sputtered on each side ofthe membrane. This very sophisticated electrode has beenproved to detect Pb2+ but no very clear analytical results wereprovided.

Polythiophene also gave rise to one report (Zejli et al., 2007).The polymer was electrodeposited on a Pt electrode by cyclicvoltammety and used to detect Ag(I) by DPASV taking advan-tage of the inductive effect of the C-S dipole of thiophene units.However, the linear range was found to be very narrow, from 0.65to 9.3 μM.

A contribution from the group of Noguet has also to be noticed(Yasri et al., 2011). In this original work, a graphite electrodecoated by a 3,4-poly(ethylenedioxythiophene):poly(styrene sul-fonate) [PEDOT:poly(styrene sulfonate)] copolymer was usedto detect Pb2+ by chronoamperometry at −0.35 V. The linear-ity range was from 2 to 100 nM and the LOD was 0.19 nM for30 s accumulation at −0.65 V. The system exhibited good stabil-ity since only a slight decrease was noticed after 11 days. It wasalso successfully tested for the determination of Pb2+ in differentvegetables extracts.

In order to improve the detection limit the group of Chevaletexploited the high resolution power of a multi-pulse electro-analytical method, namely multiple square wave voltammetry(MSWV) (Fatouros et al., 1986; Krulic et al., 1990). MSWV isbased on the superimposition of several pairs of opposite pulsesof constant amplitude on each step of a staircase waveform.However MSWV differs from previously described DAPV by theelectrode response: instead of the double cathodic and anodic

current recorded in the latter case, MSWV response results inthe integration of the successive currents. The MSWV-DD (DDfor double differential) technique was used in combination witha Nafion-coated electrode for the determination of trace specieslike methylmercury (Moretto et al., 1999) and Fe (Ugo et al.,2001). The perfluorinated cation exchanger Nafion was used topreconcentrate the analyte and was simply deposited on a GCdisk by droplet evaporation. The detection capabilities of thispolymer-coated electrode combined to the sensitive MSWV-DDmethod allowed a calculated LOD down to 45 pM to be reachedfor methylmercury (Moretto et al., 1999).

Nano-scaled materialsDuring the last two decades, nano-scaled materials have arouseda great interest with respect to analytical applications dueto their specific physico-chemical properties (Murray, 2008).Improvements resulting from those nanomaterials for electro-analysis are manifold: enhanced diffusion of electroactive species,higher effective surface area of nanoparticles (NPs), electrocat-alytic and conductive properties, improved selectivity and highersignal-to-noise ratio. With respect to trace metal analysis, goldnanoparticles (AuNPs) are the most commonly used material(Lin et al., 2011; Liu et al., 2011). They can be obtained eitherby chemical or electrochemical ways.

The French contribution to this topic is very recent, andthe corresponding works all considered the structuration of thenanoparticle-based modified electrode to be a key feature withrespect to analytical performances.

Ottakam Thotiyl et al. designed a multilayer arrangement ofcitrate-capped AuNPs immobilized by a thiol group onto a Auelectrode for the detection of As(III) (Ottakam Thotiyl et al.,2012). The anionic AuNPs were deposited layer-by-layer alterna-tively with cationic polyelectrolyte to afford a layered nanocom-posite film. As(III) was detected by its electrocatalytic oxidationusing DPV without any accumulation step, and a LOD of 0.48 μMwas reported. This study evidenced a strong correlation betweenthe amperometric response and the number of layers of thenanocomposite film.

Our group has developed a Hg(II) sensor based on GC elec-trode functionalized by electrodeposited AuNPs and has studiedthe influence of the electrodeposition method on the analyticalperformances (Hezard et al., 2012a,b). Namely, cyclic voltam-metry, chronoamperometry and potentiostatic double-pulse wereused. It was shown that both the electrodeposition mode and thecharge used for the Au precursor reduction had a dramatic influ-ence on the size and density of AuNPs (Figure 7). These lattertwo parameters were strongly correlated to the analytical perfor-mances: the best results were obtained for dense deposits of smallNPs (Hezard et al., 2012b). By using a 5 min accumulation timeand SWASV, a LOD of 0.4 nM was reached.

An important development in the frame of nano-scaled mate-rials concerns single (SWCNTs) or multi-walled carbon nan-otubes (MWCNTs) since their discovery in 1991 (Ijima, 1991).Their unique structure offers very interesting properties suchas high specific surface area, high chemical stability, good elec-trical conductivity and adsorption capacity, which give rise to

FIGURE 7 | AuNPs electrodeposited onto GC from a 0.25 mM HAuCl4solution using: (A) chronoamperometry; (B) potentiostatic double pulse;(C) cyclic voltammetry. (D) SWASV response and calibration curve obtained

in the Hg(II) concentration range 0.8–9.9 nM using electrode (A). Adaptedfrom our own results published in Hezard et al. (2012b), Copyright 2012Elsevier.

wide applications in electronics, composite materials, energy stor-age, and of course sensors (Fam et al., 2011). With respect tothis latter domain the association of various CNT functional-ization procedures, including organic (polymers, proteins. . .) orinorganic (metal nanoparticles, metal oxides. . .) modifiers andseveral transducing modes allowed the elaboration and develop-ment of chemical, gas, humidity, biomedical or environmentalsensors which have been the subject of several reviews (Jacobset al., 2010; Vashist et al., 2011; Gao et al., 2012). Extensive inter-national researches have been made to assay heavy metals byusing unmodified (Yue et al., 2012) or CNTs modified eitherwith cysteine (Morton et al., 2009), thiacalixarene (Wang et al.,2012), Sb nanoparticles (Ashrafi et al., 2014) or with imprintedpolymeric nanobeads (Rajabi et al., 2013), most of the time asso-ciated with ASV. They allowed Pb, Hg, Cu, Cd, or Zn to bedetected down to 3 nM and in linear concentration ranges upto 7 μM. Surprisingly, to the best of our knowledge, no Frenchresearch team has exploited CNTs for heavy metals detectionyet. Interest has mainly focused on micro and supercapacitors(Ghimbeu et al., 2011), batteries (Carn et al., 2013) and energystorage devices (Sathiya et al., 2011) or biofuel cells (Both Engelet al., 2013; De Poulpiquet et al., 2013). Concerning sensors worksessentially dealt with gaz sensors (Bondavalli, 2010) and above allenzymatic (Singh et al., 2013), RNA (Tran et al., 2013) and DNA(Zhang et al., 2011)-based electrochemical biosensors.

BiosensorsIn general terms, a biosensor is defined as an analytical toolassociating a bioactive compound (mono- or multi-enzyme sys-tem, antibody, microorganism) which can specifically recognizespecies of interest and a transducing element (Frew and Hill,1987). Thus, it may be viewed as a particular kind of CME inwhich the modifier is a biological element. Enzymes are knownas good modifiers with respect to heavy metals since these latteroften strongly inhibit enzymatic reactions (Dennison and Turner,1995).

The group of Chovelon reported several works about conduc-tometric biosensors based on Chlorella vulgaris, a microalgae thecell walls of which bear enzymes such as alkaline phosphatasesand esterases (Chouteau et al., 2004, 2005). In both papers, themicroorganism was immobilized onto interdigitated Pt electrodesusing bovine serum albumin and glutaraldehyde as a crosslinker.In the first work, Cd2+ was detected as low as 8.9 nM but atleast one hour was needed to obtain such a result. The secondstudy (Chouteau et al., 2005) provided several improvementssince both Cd2+ and Zn2+ were detected with a 89 nM and0.15 μM LOD, respectively, but with a significantly lower expo-sure time (ca. 30 min). It has to be noticed that bioassays werefound to reach a similar LOD to the biosensor but only after 4 hexposure time, illustrating the wise enzyme immobilization strat-egy used. Finally, an originality of this latter work was that thebiosensor was also capable to detect pesticides, taking advantageof the fact that esterases were selectively inhibited by these organiccompounds whereas alkaline phosphatases were not affected.

Tekaya et al. have also reported a conductometric biosen-sor based on alkaline phosphatase from the cyanobacteriumArthrospira platensis (Tekaya et al., 2013) deposited on the ceramic

part of interdigitated Au electrodes. In this case, inhibition mea-surements were performed after 24 h incubation, allowing LODdown to 10−20 M to be reached for both analytical targets, namelyCd2+ and Hg2+. Despite this very appreciable LOD, no informa-tion was provided about sensor lifetime. This is mainly due to thefact that the authors considered their study as a proof of conceptaiming at providing a global response to the presence of heavymetals.

A last conductometric biosensor was reported 2 years ago bySoldatkin et al. (2012). This sensor consisted in a three-enzymesystem prepared by mixing invertase, mutarotase and glucose oxi-dase with bovine serum albumin. Although a bit complicated,the sensor exhibited good sensitivities toward Hg2+ and Ag+without experiencing any interference from organic compounds.After a relatively short incubation time (ca. 10–20 min), the LODwere found to be 25 and 100 nM for Hg2+ and Ag+, respec-tively. It has to be noticed that this biosensor was very selectivesince Cd2+, Zn2+, Ni2+, Pb2+, Cu2+, and Co2+ did not affect theenzymes activities. Interestingly, the authors have investigated areactivation procedure: dipping the biosensor into EDTA or cys-teine solution for 45 min allowed a significant reactivation of theenzymes together with the identification of the inhibitor metal,since EDTA reactivated Ag+ inhibition only whereas cysteine wasselective to Hg2+ inhibition.

The group of Cosnier proposed an amperometric sen-sor toward Hg(II) based on the same three enzymes whichwere entrapped in a clay matrix deposited on Pt electrode(Mohammadi et al., 2005). The electrochemical response wasprovided by the oxidation of the enzymatically produced hydro-gen peroxide. After 20 min incubation, Hg(II) was detected in therange 10 nM to 1 μM. It has to be noticed that several Hg specieshave been tested, such as methylmercury or phenylmercury, all ofthem leading to invertase inhibition. Depending of the analyti-cal target, the biosensor recovery by using cysteine was more orless effective, but never complete. The system was quite selectiveto Hg(II) since all metal cations tested induced no interferencesexcept Ag+.

Finally, the group of Burstein has reported 20 years ago abi-enzymatic system coupled to an oxygen sensor (a Clark elec-trode) which offered an original solution to the enzyme recoveryissue (Gayet et al., 1993). In this work, several enzyme com-binations were tested but the most interesting one was thatinvolving L-lactate oxidase and L-lactate dehydrogenase. Actually,the former enzyme was immobilized on the membrane of theoxygen sensor whereas the latter one remained in solution. Sinceonly L-lactate dehydrogenase was affected by the presence ofheavy metal cations, the sensor regeneration simply consisted ina renewal of the solution containing this very cheap enzyme.Unfortunately, the interest of this system was limited due to therelatively poor LOD obtained for the different analytical targets.

THE PARTICULAR ISSUE OF TRACE METAL SPECIATIONIn addition to the monitoring of total heavy metals concentra-tions in the environment, speciation analysis provides very usefulcomplementary informations. However speciation analysis oftenimplies the determination of very low concentrations of minorspecies, typically in the range of nM to pM or even lower. Hence

electrochemistry and especially anodic stripping techniques areparticularly well adapted to such kind of analysis.

For instance, Bourgeault et al. studied dissolved Cu speciationand bioavailability in freshwaters by using DPASV, diffusive gra-dient in thin films and ISE, and compared their results to thoseobtained by modeling (Bourgeault et al., 2013). It was found thatCu accumulation in aquatic mosses was better correlated to theweakly complexed Cu species measured using DPASV than to freeCu concentration measured using an ISE, thus highlighting thecontribution of electrochemical techniques to speciation studies.

In a close approach combining DPASV and conductometry,Terbouche et al. examined the complexation of Zn(II) and Cd(II)by new humic acids from Algeria, these latter class of naturalpolymers being known to have an influence on heavy metals bio-geochemical cycle and bioavailability (Terbouche et al., 2011).Based on the strong complexing capacities evidenced in this work,the authors suggested their humic acids to be used for metaluptake.

A last speciation study combining electrochemical techniqueswas provided very recently by Rotureau (2014). Here the alreadydiscussed “AGNES” technique was used for the determination offree metal in solution while SCP at scanned deposition potential(SSCP) allowed dynamic speciation information to be reached.This latter technique consists in performing classical SCP withvarying deposition potentials, and then plotting the transitiontime as a function of the deposition potential (van Leeuven andTown, 2002). Rotureau’s work dealt with Cd and Pb speciationdynamics in clay minerals, and it was shown that the interactionof Cd with clays may be described as a chemically homogeneous,labile system over a wide pH range whereas strong pH effects wereevidenced in the case of Pb.

SSCP has also been used by Parat et al.