Silicon oxide surface functionalization by self-assembled nanolayers for microcantilever transducers

Page 1

A

cfscaraotc

s©

K

1

sTpodc

dala

0d

Available online at www.sciencedirect.com

Colloids and Surfaces A: Physicochem. Eng. Aspects 321 (2008) 87–93

Silicon oxide surface functionalization by self-assemblednanolayers for microcantilever transducers

F. Gambinossi a, L. Lorenzelli b, P. Baglioni a, G. Caminati a,∗a Department of Chemistry and CSGI, University of Florence, Via della Lastruccia 3, 50019 Sesto Fiorentino (Florence), Italy

b FBK-irst, Biomems Group, Via Sommarive 18, 38050 Povo, Trento, Italy

Received 10 July 2007; received in revised form 17 October 2007; accepted 17 January 2008Available online 2 February 2008

bstract

We developed a functionalization procedure for silicon oxide surfaces used in microcantilever-based sensors dedicated to the detection of foodontaminants in fluid matrices. In particular we focused on the determination of heavy metal ions and of agricultural pesticides. The surfaceunctionalization was obtained by direct self-assembly of long chain molecules bearing at one end a complexing moiety for metal ions. Theelected chelating molecule, the nitrilotriacetic acid (NTA), was immobilized onto silicon oxide surfaces using a three-step process involving theonsecutive addition of an organosilane, glutaraldehyde and a NTA derivative solutions. The formation of the self-assembled nanostructure (SAN)t the surface was traced by means of quartz crystal microbalance with dissipation monitoring (QCM) measurements as a function of time. Theesults indicated that the functionalized molecule forms a rigid self-assembled film on silicon dioxide. Data analysis provided the layer thicknessnd the molecular orientation of the chemisorbed layers at the interface. The optimized procedure was tentatively applied to functionalize the siliconxide outer surface of an array of microwells each containing four microcantilevers. Quantitative determination of the metal ions complexation at

he surface was achieved adding the desired solution in the QCM measuring chamber and recording the adsorbed mass change as a function ofoncentration.

The above self-assembled system was further exploited for the detection of pesticides in fluid matrices monitoring the variation in the QCMignal upon addition of the analyte in the measuring chamber.

2008 Elsevier B.V. All rights reserved.

pative

iaomstcatf

eywords: Self-assembled nanostructure; Microcantilever-based sensors; Dissi

. Introduction

Nanotechnology research on chemical and biochemical sen-ors burst out in the attempt to reach ‘lab-on-chip’ devices [1,2].he microelectromechanical systems (MEMS) [3,4] are used asarts of nanofluidic devices and the development of new typesf MEMS may result in advances in this kind of applications,ue to their ultra-sensitivity, ease of mass production, and lowost [5].

The aim of this work is to develop a modification proce-ure for silicon oxide surfaces of microcantilevers to be used

s sensors for food contaminants in fluid matrices. In particu-ar, we focused on the determination of heavy metal ions andgricultural pesticides in water.

∗ Corresponding author. Tel.: +39 055 4573040/3024; fax: +39 055 4573036.E-mail address: [email protected] (G. Caminati).

tcap

cs

927-7757/$ – see front matter © 2008 Elsevier B.V. All rights reserved.oi:10.1016/j.colsurfa.2008.01.030

quartz crystal microbalance; NTA; Heavy metal ions

The surface functionalization of SiO2 was obtained by chem-cal self-assembly of long chain molecules bearing at one endcomplexing moiety for metal ions since it has long been rec-gnized that self-assembled monolayers (SAM) provide an easyethod to prepare surfaces of the desired chemical composition,

tructure and thickness [6,7]. The selected chelating molecule ishe nitrilotriacetic acid (NTA), known to form stable octahedralomplexes with many metal ions [8]. Three carboxylate groupsnd the tertiary amine group of NTA are used for the coordina-ion with the metal ions, leaving two binding sites accessible tourther binding. Surface modification and metal ions complexa-ion at the interface was followed by means of dissipative quartzrystal microbalance (QCM) measurements. The surface char-cterization of the self-assembled nanostructures (SANs) was

erformed by Ellipsometric Imaging.

The experimental findings were modelled in order to elu-idate the complexation mechanism and the kinetics of theelf-assembly formation in the bi-dimensional confined systems.

Page 2

8 A: Ph

Te

2

2

NAGNp

awh

rmoaw

2

qwH

2

ia(so(tTT2

teio4mtof

H

3

3s

stttagtla

wnrddn

Adttt

8 F. Gambinossi et al. / Colloids and Surfaces

he data obtained from the QCM measurements were used toxtract the stability constants of the surface complex.

. Experimental

.1. Materials

Nα,Nα-Bis(carboxymethyl)-l-lysine hydrate, aminobutyl-TA (AB-NTA), and (3-aminopropyl)trimethoxysilane,PTMS, were obtained from Fluka. Glutaraldehyde solutionrade I, 50% in water, was purchased from Sigma andiSO4·6H2O was supplied by Aldrich and used without furtherurification.

A buffer containing 200 mM phosphate and 150 mM NaCldjusted to pH 7.4 (PBS) was used for the preparation of theorking solutions. The concentrations for APTMS, glutaralde-yde and AB-NTA were 10 mM in PBS buffer.

The chemical structures of all the precursor molecules areeported in Fig. 1. The ultrapure water used for all the experi-ents and for all the cleaning steps was obtained by reversed

smosis (Milli-RO; Millipore GmbH) followed by ion exchangend filtration steps (Milli-Q, Millipore GmbH); the resistivityas better than 18 M� cm and pH 5.6.

.2. Cleaning procedures

The substrates used for the deposition were SiO2-coateduartz crystals (KSV, Finland). The silicon slides were cleanedith chromic acid and immersed for 30 s in a solution of 2.5%F, then rinsed thoroughly with ultrapure water and finally dried.

.3. Methods

Surface modification and metal ion complexation at thenterface was followed by means of a quartz crystal microbal-nce instrument with dissipation monitoring, model QCM-Z500KSV, Finland) employing a 5 MHz crystal with a frequencytability and resolution of ±0.05 Hz in liquid. Changes in the res-nant frequency of the oscillator (�f) and changes in dissipation�D = Edissipated/2�Estored) were recorded for the fundamen-

al as well as for the other overtones as a function of time.he temperature of the measuring cell was kept constant at= 20 ± 0.02 ◦C with a Peltier unit, room temperature was0 ± 0.1 ◦C with an average humidity = 56%.

Fig. 1. Chemical structures of APTMS, glutaraldehyde and AB-NTA.

dus

[spefwlacm

f

ysicochem. Eng. Aspects 321 (2008) 87–93

SANs coverage was checked by ellipsometric mapping withhe EP3 modular system (NFT, Nanofilm Technologie GmbH),mploying a 532 nm Nd-YAG laser (50 mW) at an angle ofncidence of 60◦. The imaging system is composed of a 10×bjective and a CCD camera and has a field of view of00 �m × 590 �m with a lateral resolution of 2 �m. Measure-ents were taken in air directly on the QCM sensor covered with

he SAN, unless otherwise specified. The ellipsometric thicknessf the multilayers was calculated by using a three-layers modelor the fitting of the experimental data [9].

Semi-empirical calculations were run using the softwareyperChem 5.1 (HyperCube, U.S.A.) [10].

. Results and discussion

.1. Formation of the self-assembled nanostructure onilicon dioxide surfaces

An optimized modification of a previously reported three-tep process [11] was used to immobilize the chelating agent tohe SiO2 supports. The procedure consisted in the immersion ofhe substrate in an organosilane solution (APTMS), followed byhe activation with glutaraldehyde and finally by the addition ofsolution of the AB-NTA. An analogous method that employslutaraldehyde was previously used by other authors [12,13]o immobilize proteins to silicon substrates by covalent cross-inking between the primary amine on a surface and a primarymine on the protein.

The formation of such self-assembled nanostructures on SiO2as followed by means of dissipative QCM. Fig. 2 shows theormalized frequency shift and the changes in dissipation valueselative to the third harmonic. �f values for other harmonicso not superimpose when normalized by the overtone numberuring the process of surface modification, whereas a uniqueormalized behaviour is observed after buffer rinsing.

The sequential addition of APTMS, glutaraldehyde andB-NTA in the measurement chamber causes an initial rapidecrease in frequency with a corresponding increase in dissipa-ion. Such behaviour is due to the formation of multilayers athe interface although the different bulk viscosity of the solu-ion may contribute to the observed changes in frequency andissipation. Upon rinsing with PBS buffer the dissipation val-es get close to baseline level indicating that only the covalentlyelf-assembled molecules remain attached to the surface.

Under the rigid film approximation, the Sauerbrey relation14] can be applied to obtain additional information about theurface coverage of the support. In Table 1 the most relevantarameters, i.e. the mass density, the area per molecule, thexperimental and the theoretical thicknesses, are summarizedor each step of the self-assembly formation. In the same tablee also reported the bulk density value, ρf, at 20 ◦C for each

ayer. For the second and the third step of the process wedopted a ρf value of 1.03 and 1.24 g cm−3, respectively, as

alculated by using a quantitative structure-activity relationshipodel (QSAR) [15].Upon chemisorption of APTMS we found an area occupied

or a single molecule on the surface of A = 19 A2; this value

Page 3

F. Gambinossi et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 321 (2008) 87–93 89

tor ch

cmclwmc[

atmgtabca2w(

A(nost

p0oSrtwbd

ibblsbps

3i

l

TS

AAA

Fig. 2. Normalized frequency shift (left axis) and dissipation fac

an be compared to the theoretical area estimated from energyinimization assuming a compact monolayer coverage, in this

ase we obtained A = 25 A2. Since the experimental value isower than what expected for a monolayer at maximum packinge may hypothesize the build-up of incomplete multilayers or,ost probably, the formation of siloxane bonds between adja-

ent organosilane molecules as also reported by other authors16,6].

Application of a rigid film model to all the recorded overtonesfter buffer rinsing allowed to determine a thickness of 12 A forhe APTMS adlayer, such value is in accordance with the ellipso-

etric thickness determined on the final one-layer assembly thatave a thickness for the first layer = 13 A. These results excludehe presence of multilayer patches since the theoretical value foruniform self-assembled monolayer of APTMS is reported toe 10 A [17–19]. The addition of glutaraldehyde increases thehain length of the molecule resulting in an increase of the avail-ble area per molecule at the surface. The experimental area of4 A2 and the thickness value of 16 A are in a perfect accordanceith a vertical disposition of the molecule at maximum packing

see Table 1).After the third step, i.e. AB-NTA addition, a final area

= 41 A2 is obtained, this value is close to the theoretical oneA = 38 A2). However, the low value of the calculated thick-

ess (see Table 1) suggests an incomplete chemical reactionf the aminobutyl-NTA with the aldehydic moiety due to theteric hindrance of the bulky NTA. If we consider separatelyhe decrease of frequency due to the addition of NTA to the

ifia

able 1urface parameters

�m/A(ng/cm2)

Experimental area(A2/molecule)

Theoretica(A2/molec

PTMS 155 19.2 25PTMS + Glutaraldehyde 190 24.3 26PTMS + Glutaraldehyde + AB-NTA 220 40.6 38

a Calculated by using a quantitative structure-activity relationship model (QSAR).

anges (right axis) for the three-step formation of SAN on SiO2.

reformed two-layers SAN, we obtain a surface density of.0069 molecules/A2 that results in a surface coverage of 17%f the SAM, this implies that NTA molecules are linked to theAM in a 1:6 ratio. Applying the rigid film model to the data cor-esponding to the third layer formation we obtain a final QCMhickness value of 18 A, interestingly this value is lower thanhat obtained for the ellipsometric thickness, this result maye explained by a partial re-arrangement of the surface whenried.

The above functionalization procedure is currently under test-ng on silicon microcantilever-based sensor arrays manufacturedy some of us [20,21]. As shown in Fig. 3 they are composedy 4 wells with 4 cantilevers each. The cantilevers, 300 �mong, were placed in a 4-beam array with integrated piezore-istors. Stress-induced displacement experiments [22,23] wille performed on microcantilever functionalized with the aboverocedure and the response upon contact with heavy metal ionsolutions will be determined [24].

.2. Surface complexation through metal–ligandnteraction

Metal complexes formation at the surface of SANs was fol-owed in water by dissipative QCM. Typical results are reported

n Fig. 4a for the variation of the normalized �f and �D as aunction of time for the sequential addition of increasing nickelon concentration to the SiO2/SAN system. The time to reach

steady-state value for the adsorption was found to be close

l areaule)

QCM thickness(nm)

Ellipsometricthickness (nm)

Theoreticalthickness (nm)

Bulk density(g cm−3)

1.2 1.3 1.0 0.961.6 1.8 1.5 1.03a

1.8 2.2 2.5 1.24a

Page 4

90 F. Gambinossi et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 321 (2008) 87–93

ntilev

ttdrps

i

mtdd

F1

Fig. 3. Microca

o 1 h. For low Ni (II) concentration we found that dissocia-ion of the complex was readily obtained upon PBS rinsing; aetailed study of the reversibility of surface complexation will beeported in a separate paper that fully describes the complexation

rocess studied by combined QCM and electronic absorptionpectroscopy [25].

Determination of the equilibrium frequency shifts afternjection of the corresponding metal ion concentration in the

tcNe

ig. 4. (a) Normalized frequency shift (left axis) and dissipation factor changes (× 10−5 M; (C) 1 × 10−4 M; (D) 1 × 10−3 M. (b) Mass density values as a function

er array layout.

easuring chamber resulted in a decrease of �f up to concen-rations of 5 × 10−3 M, after this value we observed an abruptecrease in frequency together with a decrease in dissipationue to the change of bulk density and viscosity. Fig. 4b shows

he mass density coverage values as a function of nickel ionsoncentration up to 5 × 10−3 M. The stability constant of thei–NTA complexes at the surface was determined by fitting the

xperimental data with a Langmuir type model [26,27] using the

right axis) upon addition of nickel ions on SiO2/SAN: (A) 1 × 10−6 M; (B)of nickel ions concentration (triangles); Langmuir fit (solid line).

Page 5

A: Ph

f

wihuboatb

ndl

cStsbsfia

ahtou

F(

F. Gambinossi et al. / Colloids and Surfaces

ollowing equation:

�m

A= (�m/A)sat[M2+]

[M2+] + kd(1)

here (�m/A)sat is the mass density shift upon saturation and kds the dissociation constant of the complex. This model assumes aomogeneous distribution of the NTA molecules, energeticallyndistinguishable adsorption sites and absence of interactionsetween the bound metal ions. As a result of the fitting webtained a dissociation constant kd = (1.0 ± 0.3) × 10−5 M andmass density saturation value equal to 43 ± 3 ng/cm2. Other

heoretical models were tentatively applied to fit the QCM data,ut an unsatisfactory fit was always obtained.

Complexation behaviour at the surface was found to be sig-ificantly different from what observed in bulk solution. Theissociation constant for the Ni–NTA complex at the surface wasower than in solution [28–30] probably because of the different

rcN

ig. 5. (a) Normalized frequency changes upon addition of carbamate solutions on tD) 2 × 10−5 M. (b) Mass density values as a function of the carbamate concentration

ysicochem. Eng. Aspects 321 (2008) 87–93 91

omplex geometry induced by the confinement at the surface.eparate experiments on a variety of heavy metal ions showed

hat surface complexation occurs but with different binding con-tants [25]. Electrochemical-QCM experiments are currentlyeing implemented on these systems in order to address the sen-or selectivity issue: simultaneous determination of the redoxngerprint of the complex will be coupled to the nanogravimetricpproach.

When Nickel (II) concentrations above adsorption saturationre used, the Ni–NTA complex was found to be stable for manyours after buffer rinsing. Therefore, the resulting nanostruc-ures can be used to detect other analyses such as contaminantsf wine and fruit; in particular we focused on pesticides heavilysed in grape production.

Fig. 5a shows the decrease of the normalized frequencyecorded upon the addition of a typical pesticide used in agri-ulture (carbamate) to the QCM measuring chamber using thei–NTA-coated sensor. The variation of dissipation values as a

he Ni–NTA-coated sensor: (A) 2 × 10−6 M; (B) 5 × 10−6 M; (C) 1 × 10−5 M;(circles); Langmuir fit (solid line).

Page 6

9 A: Ph

fttwdoAsiawccr

ufdp

4

mirswrrlnt

aifshrf

A

C

R

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[[

2 F. Gambinossi et al. / Colloids and Surfaces

unction of time was omitted in this figure because it was closeo zero. Addition of the pesticide solution to the SAN exposinghe Ni–NTA functional groups induced the replacement of theater molecule linked to the Ni–NTA complex by the carbamateue to its higher stability constant. Fig. 5b reports the variationf mass density as a function of the pesticide concentration.s for the complexation of nickel ions with the NTA-coated

ensor we successfully applied Eq. (1) to the QCM data obtain-ng a dissociation constant value of kd = (6 ± 2) × 10−6 M and

mass density saturation value of 55 ± 6 ng/cm2. Moreover,e obtained a value of 2 × 10−6 M as the minimal detectable

oncentration of pesticides in water. Interestingly, such con-entration is one order of magnitude lower than the maximumesidue limit for such kind of pesticides in wine and fruits.

Thus, the functionalization procedure we proposed could besed for the detection of low level of pesticides in agriculturaloods by means of QCM with dissipation monitoring. Moreover,etection of carbamate pesticides employing stress-induced dis-lacement method is currently in progress [24].

. Conclusions

Dissipative QCM investigations permit to monitor the for-ation of complex self-assembled nanostructures, to extract

nformation on the mass-density properties and to followeactions occurring at the surface. QCM and Ellipsometric mea-urements indicated that the silicon dioxide surface was coveredith the desired nanostructure. Such architecture exposes the

eacting groups towards the solution phase so that complexationeactions at the surface are readily accomplished. In particu-ar, QCM data allowed monitoring surface complexation withickel ions and to determine surface dissociation constant forhe Ni–NTA complex and the reaction kinetics.

The functionalized self-assembled nanostructures representstarting point for the determination of organic toxic residues

n aqueous solution upon further complexation at the SAN sur-ace. In fact, preliminary results indicate that Ni–NTA modifiedurfaces may reversibly recognize pesticide or antibiotics withigh sensitivity. The results are a significative step towards theealization of microcantilevers for the detection of residues ofood contaminants in liquid matrices.

cknowledgements

The work was supported by FIRB (RBNE01ZB7A) andSGI/CNR-FUSINT 2006.

eferences

[1] D.R. Reyes, D. Iossifidis, P.A. Auroux, A. Manz, Micro total analysissystems. 1. Introduction, theory, and technology, Anal. Chem. 74 (2002)2623–2636.

[2] P.A. Auroux, D. Iossifidis, D.R. Reyes, A. Manz, Micro total analysis sys-

tems. 2. Analytical standard operations and applications, Anal. Chem. 74(2002) 2637–2652.

[3] A.C.R. Grayson, R.S. Shawgo, A.M. Johnson, N.T. Flynn, Y. Li, M.J. Cima,R. Langer, A BioMEMS review: MEMS technology for physiologicallyintegrated devices, Proc. IEEE 92 (2004) 6–21.

[

ysicochem. Eng. Aspects 321 (2008) 87–93

[4] D.L. Polla, E. Erdman, W.P. Robbins, D.T. Markus, J.D. Diaz, R. Rinz, Y.Nam, H.T. Brickner, Microdevices in medicine, Ann. Rev. Biomed. Eng. 2(2000) 551–576.

[5] T. Thundat, P.I. Oden, R.J. Warmack, Microcantilevers sensors, MicroscaleThermophys. Eng. 1 (1997) 185–199.

[6] D.K. Aswal, S. Lenfant, D. Guerin, J.V. Yakhmi, D. Vuillaume, Self assem-bled monolayers on silicon for molecular electronics, Anal. Chim. Acta 568(2006) 84–108.

[7] J.C. Love, L.A. Estroff, J.K. Kriebel, R.G. Nuzzo, G.M. Whitesides, Self-assembled monolayers of thiolates on metals as a form of nanotechnology,Chem. Rev. 105 (2005) 1103–1170.

[8] D.L. Rubinstein, R.J. Kula, Ligand-exchange kinetics solution equilibriaof cadmium, zinc, and lead nitrilotriacetate complexes, J. Am. Chem. Soc.91 (1969) 2492–2503.

[9] Principles of nulling and imaging ellipsometry, nanofilm surface analysis,Technical Release Note, 04/2002.

10] J.J.P. Stewart, Optimization of parameters for semiempirical methods. I.Method, J. Comput. Chem. 10 (1989) 209–220.

11] J. Maly, E. Illiano, M. Sabato, M. De Francesco, V. Pinto, A. Masci, D.Masci, J. Masojidek, M. Sugiura, R. Franconi, R. Pilloton, Immobilisationof engineered molecules on electrodes and optical surfaces, Mater. Sci.Eng. C 22 (2002) 257–261.

12] P.F. Luckham, K. Smith, Direct measurement of recognition forces betweenproteins and membrane receptors, Faraday Discuss. 111 (1998) 307–320.

13] R.A. Williams, H.W. Blanch, Covalent immobilization of protein mono-layers for biosensor applications, Biosens. Bioelectron. 9 (1994) 159–167.

14] F. Hook, M. Rodahl, P. Brzezinski, B. Kasemo, Energy dissipation kineticsfor protein and antibody-antigen adsorption under shear oscillation on aquartz crystal microbalance, Langmuir 14 (1998) 729–734.

15] ACD/ChemSketch Freeware, version 10.00, Advanced Chemistry Devel-opment Inc., Toronto, ON, Canada, http://www.acdlabs.com, 2006.

16] F. Zhang, M.P. Srinivasan, Self-assembled molecular films of aminosi-lanes and their immobilization capacities, Langmuir 20 (2004) 2309–2314.

17] J. Orbulescu, C.A. Constantine, V.K. Rastogi, S.S. Shah, J.J. DeFrank,R.M. Leblanc, Detection of organophosphorus compounds by covalentlyimmobilized organophosphorus hydrolase, Anal. Chem. 78 (2006) 7016–7021.

18] G.C. Allen, F. Sorbello, C. Altavilla, A. Castorina, E. Ciliberto, Macro-, micro- and nano-investigations on 3-aminopropyltrimethoxysilane self-assembly-monolayers, Thin Solid Films 483 (2005) 306–311.

19] P.A. Heiney, K. Gruneberg, J. Fang, C. Dulcey, R. Shashidhar, Structure,Growth of chromophore-functionalized (3-aminopropyl)triethoxysilaneself-assembled on silicon, Langmuir 16 (2000) 2651–2657.

20] A. Adami, M. Decarli, V. Guarnieri, L. Lorenzelli, M. Zen, Microcantilever-based sensor array for amine detection in agro-food applications, in:Proceeding of the International Conference on MEMS and SemiconductorNanotechnology, Kharagpur, India, 20–22th December, 2005.

21] M. Decarli, A. Adami, L. Lorenzelli, N. Laidani, MEMS-based SiliconCantilevers Arrays for combinatorial analysis of thin-film materials, in:Proceedings of the conference: Eurosensors XX, Gothenburg, Sweden,17–20th September, 2006.

22] T. Thundat, R.J. Warmack, G.Y. Chen, D.P. Allison, Thermal and ambient-induced deflections of scanning force microscope cantilevers, Appl. Phys.Lett. 64 (1994) 2894–2903.

23] J.R. Barnes, R.J. Stephenson, M.E. Welland, C. Gerber, J.K. Gimzewski,Photothermal spectroscopy with femtojoule sensitivity using a microme-chanical device, Nature 372 (1994) 79–81.

24] L. Lorenzelli, F. Gambinossi, P. Baglioni, G. Caminati, in preparation.25] F. Gambinossi, T. Viitala, G. Caminati, Complex formation at the

surface of nano-assemblies monitored by dissipative QCM and UV–

VIS–NIR absorption spectroscopy, Sens. Actuators B, submitted forpublication.

26] T. Hianik, V. Ostatna, Z. Zajacova, The study of the binding of globular pro-teins to DNA using mass detection and electrochemical indicator methods,J. Electroanal. Chem. 564 (2004) 19–24.

Page 7

A: Ph

[

[

[29] G. Anderegg, Critical survey of stability constants of NTA complexes, Pure

F. Gambinossi et al. / Colloids and Surfaces

27] K. Matsuura, A. Tsuchida, Y. Okahata, T. Akaike, K. Kobayashi, A

quartz-crystal microbalance study of adsorption behaviours of artificialglycoconjugate polymers onto chemically modified gold surfaces and theirinteractions with lectins, Bull. Chem. Soc. Jpn. 71 (1998) 2973–2977.

28] E.R. Souaya, W.G. Hanna, E.H. Ismail, N.E. Milad, Studies on some aciddivalent-metal nitrilotriacetate complexes, Molecules 5 (2000) 1121–1129.

[

ysicochem. Eng. Aspects 321 (2008) 87–93 93

Appl. Chem. 54 (1982) 2693–2758.30] K.L. Cheng, K. Ueno, T. lmamura, CRC Handbook of Organic Analytical

Reagents, 2nd ed., CRC Press, Boca Raton, Florida, 1982.