Iminodiacetic acid functionalised organopolymer monoliths: application to the separation of metal cations by capillary high-performance chelation ion chromatography

ORIGINAL PAPER

Iminodiacetic acid functionalised organopolymer monoliths:application to the separation of metal cations by capillaryhigh-performance chelation ion chromatography

Áine Moyna & Damian Connolly & Ekaterina Nesterenko &

Pavel N. Nesterenko & Brett Paull

Received: 8 June 2012 /Revised: 10 August 2012 /Accepted: 14 August 2012# Springer-Verlag 2012

Abstract Lauryl methacrylate-co-ethylene dimethacrylatemonoliths were polymerised within fused silica capillariesand subsequently photo-grafted with varying amounts ofglycidyl methacrylate (GMA). The grafted monoliths werethen further modified with iminodiacetic acid (IDA), result-ing in a range of chelating ion-exchange monoliths of in-creasing capacity. The IDA functional groups were attachedvia ring opening of the epoxy group on the poly(GMA)structure. Increasing the amount of attached poly(GMA),via photo-grafting with increasing concentrations of GMA,from 15 to 35 %, resulted in a proportional and controlledincrease in the complexation capacity of the chelatingmonoliths. Scanning capacitively coupled contactless con-ductivity detection (sC4D) was used to characterise andverify homogenous distribution of the chelating ligandalong the length of the capillaries non-invasively.Chelation ion chromatographic separations of selected

transition and heavy metals were carried out, with retentionfactor data proportional to the concentration of grafted poly(GMA). Average peak efficiencies of close to 5,000 N/mwere achieved, with the isocratic separation of Na, Mg(II),Mn(II), Co(II), Cd(II) and Zn(II) possible on a 250-mm-longmonolith. Multiple monolithic columns produced to thesame recipes gave RSD data for retention factors of <15 %(averaged for several metal ions). The monolithic chelatingion-exchanger was applied to the separation of alkaline earthand transition metal ions spiked in natural and potablewaters.

Keywords Porous polymer monolith . Iminodiacetic acid .

Photo-grafting . Capillary chelation ion chromatography

AbbreviationsAIDA Acetyliminodiacetic acidGMA Glycidyl methacrylateGMA-co-EDMA Glycidyl methacrylate-co-

ethylene dimethacrylateIDA Iminodiacetic acidLMA-co-EDMA Lauryl methacrylate-co-

ethylene dimethacrylateVAL Vinyl azlactone

Introduction

The advantages and disadvantages associated with the useof polymer monoliths in separation science have been wellreported over the past decade [1–5]. On the positive side, itis mostly agreed that their ease of preparation, particularly incapillaries and micro-channels, together with relative ease ofsubsequent surface modification, count greatly in their fa-vour (e.g., they can be surface-functionalised via thermal or

Published in the topical collection Monolithic Columns in LiquidPhase Separations with guest editor Luis A. Colon.

Á. Moyna :D. ConnollyMarine and Environmental Sensing Technology Hub(MESTECH), National Centre for Sensor Research,Dublin City University,Glasnevin,Dublin 9, Ireland

Á. Moyna :D. Connolly : E. Nesterenko : B. Paull (*)Irish Separation Science Cluster (ISSC),National Centre for Sensor Research, Dublin City University,Glasnevin,Dublin 9, Irelande-mail: [email protected]

P. N. Nesterenko :B. PaullAustralian Centre for Research on Separation Science (ACROSS),School of Chemistry, University of Tasmania,Sandy Bay,Hobart 7001, Australia

Anal Bioanal ChemDOI 10.1007/s00216-012-6361-4

photo-initiated post-polymerisation reactions, with the latterapproach amenable to photo-masking and spatially isolatedsurface modification). Other notable advantages are theirlower operating backpressures, pH stability and format flex-ibility (e.g. they can be polymerised in a wide variety ofmoulds, without requiring retaining frits). On the negativeside, reproducibility of production and limited surface area,are still aspects in need of further attention, particularly inthe area of small molecule separations.

To date, surface modification of polymer monoliths hasoften taken the form of attachment of ionogenic function-alities for application to ion-exchange based separations,both for large biomolecules and, to a lesser degree, for polarand charged small molecules. Hence, there have been sev-eral recent review articles and book chapters on such phases,which have been published in recent years [1–5]. For theseparation of small molecules using polymer monoliths, thetrend over recent years is to move away from co-polymerisation procedures to post-polymerisation surfacemodification. This essentially means the initial productionof the monolithic base substrate, followed by either thermalor photo-initiated attachment or grafting of functional orreactive groups upon the monolith surface. This approachusually results in a greater surface density and availability offunctional groups, particularly in the case of surface graftedpolymeric chains.

Modification of the surface of the polymer monolith post-polymerisation ensures that the analyst can control the con-centration of the reactive groups at the monolith surface(thus controlling column capacity), without affecting thepore structure or morphology of the polymer monolith itself.However, as mentioned above, one of the main drawbacksin the use of polymer monoliths (particularly for chromato-graphic analysis of small ions) is the relatively small surfacearea exhibited by the polymer monolithic structure [6].Unfortunately, the direct surface modification of the mono-lith through covalent attachment of individual functionalgroups does little to remedy this underlying problem, andso attention has turned to photo-grafting techniques to in-crease the density of surface functional groups, via theirincorporation within surface grafted polymer chains [7–9].Photo-grafting allows the formation of branch like structuresemanating from one reactive surface site on the polymermonolith, thus increasing the surface density of the func-tional groups, in an ordered manner, through controlledgrowth of the polymeric branches.

The advantages to this approach were clearly demonstrat-ed in recent work by Potter et al. [10], who modifiedglycidyl methacrylate-co-ethylene dimethacrylate (GMA-co-EDMA) monoliths with phenylboronic acid groups forthe separation of glycoproteins. The surface of GMA-co-EDMA polymer monoliths was first directly modified bynucleophilic attack of the epoxide groups on the GMA

structure with p-hydroxyphenolboronic acid. However, analternative approach to the preparation of boronate function-alised phase was investigated, which involved the photo-grafting of GMA onto the monolith surface in the presenceof benzophenone, a free radical initiator. This involvedflushing a solution containing both GMA and benzophe-none, through the monolith, where upon photo-grafting,took place using UV irradiation. Boronate groups weresubsequently attached using the method described previous-ly. The authors applied both types of monolith to the sepa-ration of ribonucleotides. As expected, the polymermonolith which was prepared by photo-grafting GMA onthe monolith surface yielded longer retention times in com-parison to the monolith prepared by direct attachment ofboronate groups on the GMA surface.

The direct attachment of iminodiacetic acid (IDA) viathe ring opening of the epoxy group in GMA-co-EDMApolymer monoliths (Fig. 1a) has recently been reportedin the literature [11, 12], and applied to the separation ofproteins using immobilised metal affinity chromatogra-phy (IMAC). The concentration of IDA upon the surfaceof the GMA-co-EDMA polymer monoliths was restrictedto the number of GMA sites presenting themselves at thesurface. To address this limitation, Moyna et al. [13]recently reported the photo-initiated grafting of vinylazlactone (VAL) onto lauryl methacrylate-co-ethylenedimethacrylate (LMA-co-EDMA) polymer monoliths fol-lowed by the covalent attachment of acetyliminodiaceticacid (AIDA) (Fig. 1b). Using this approach, a range ofchelating ion-exchange monoliths could be produced,with capacity, as demonstrated by retention and chelationion chromatographic separation of metal ions, being pro-portional to concentration of VAL monomer graftingsolution.

Gillespie et al. [14] had earlier reported the immobilisa-tion of a similar molecule to IDA, namely N-(2-acetoami-do)iminodiacetic acid (ADA), to a poly(VAL)-modifiedpolymer monolith, through the reaction of the amido groupof ADA with VAL, resulting in the formation of a di-imidetype linkage (shown in Fig. 1c). The presence of this linkergroup reduces the coordination strength of the nitrogen,connected to the two carboxylic acid groups within theligand, and results in the ability of the ligand to form com-plexes with metal ions via mainly O,O type coordination.This results in weaker complexation and potentially alteredselectivity for metal ions, due to the lack of the usual O,N,Otype coordination. This effect should be even more pro-nounced in the case of immobilised AIDA, in which acarbonyl group is located directly next to the coordinatingnitrogen (Fig. 1b). Indeed, retention data supporting suchweaker complexation characteristics of immobilised AIDAwas clearly shown by the authors in their previous study[13].

Á. Moyna et al.

In the following paper, an alternative approach to IDAimmobilisation upon surface polymer grafted monoliths ispresented (Fig. 1a), using poly(GMA) for the attachment ofIDA onto the surface of a LMA-co-EDMA polymer mono-lith. Using this approach, attached IDA groups maintaintheir ability to complex cations via the usual O,N,O coordi-nation structure, and the avoidance of amide formationeliminates the question of potential bond stability in thefinal monoliths. Selectivity of the resultant IDA functional-ised monoliths of varying capacity is compared to that seenpreviously with the above-mentioned AIDA phases, and thevarious capacity phases applied directly to the chromato-graphic separations of alkaline earth and transition metalcations.

Materials and methods

Materials

Lauryl methacrylate (LMA), ethylene dimethacrylate(EDMA), glycidyl methacrylate (GMA), 1-propanol, 1,4-butanediol, 2,2-dimethoxy-2-phenylacetophenone (DAP),sodium hydroxide, iminodiacetic acid (IDA), benzophe-none, methanol, 3-(trimethoxysilyl)-propyl methacrylate,acetone, copper nitrate, cobalt chloride, nitric acid andhydrochloric acid were all purchased from Sigma Aldrich(Gillingham, UK). Cadmium nitrate, nickel chloride, man-ganese chloride, magnesium sulphate-6-hydrate, calciumchloride dehydrate, barium chloride dehydrate and stron-tium nitrate were purchased from VWR (Dublin, Ireland)and zinc chloride from Merck (Darmstadt, Germany). Stock

standards were prepared weekly to concentrations of1,000 mg/L in 1 mM nitric acid. Working chromatographicstandards were prepared daily using dilution within theappropriate eluent. Deionised water was obtained from aMillipore water purification system (Bedford, MA, USA)and filtered using a 0.22-μm nylon filter. Teflon-coated100 μm I.D. fused silica capillary was purchased fromComposite Metal Services Ltd. (Shipley, UK).

The pump used for the vinylisation of the fused silicacapillary was a KDS-100-CE, KD Scientific syringe pump(Cole Parmer, IL, USA). The monolithic columns werepolymerised and photo-grafted using a Spectrolinker XL-1000 UV Crosslinker at 254 nm (Spectronics Corp., NY,USA). The pump used during formation of the IDA mono-liths was a Knauer Smartline 100 high-pressure analyticalpump (Knauer, Bedfordshire, UK). Modification of GMA-grafted monoliths with IDA was carried out at 70 °Cusing a Mistral column heater (Spark Holland, Emmen,Netherlands). Chromatographic separations were carriedout using an Ultimate 3000 capillary chromatography sys-tem (Thermo Scientific/Dionex, Sunnyvale, CA, USA) witha 100 nL injection loop. Eluents of either 0.1 or 0.2 mMnitric acid were delivered at 1 μL/min. The separationcolumn was connected to the injector valve using 1/16″O.D.×20 cm×25 μm I.D. PEEKsil tubing (internal volumeof 98 nL) using a zero dead-volume adapter (MicroTight,Upchurch Scientific, Oak Harbour, WA, USA), designed tocreate a zero dead-volume connection between 1/16″ O.D.tubing and fused silica capillary tubing (360 μm I.D.). On-column capacitively coupled contactless conductivity detection(C4D) was employed for the study of metal retention using aTraceDec C4D detector (Innovative Sensor Technologies

Fig. 1 a Attachment of IDA topoly(GMA) photo-graftedpolymer monoliths; b IDA topoly(VAL) grafted monoliths(forming acetyl-IDA) and cADA to poly(VAL) graftedmonoliths

Iminodiacetic acid functionalised organopolymer monoliths

GmbH, Innsbruck, Austria), as well as for the characterisationof the stationary phase itself using scanning C4D.

Methods

Formation of polymer monoliths

UV-transparent fused silica capillaries were vinylised using3-(trimethoxysilyl)-propyl methacrylate, following a meth-od precisely described elsewhere [15]. The fabrication of theLMA-co-EDMA polymer monolith was carried out as de-scribed by Collins et al. [16]. Briefly, a monomer mixtureconsisting of 24 wt% LMA, 16 wt% EDMA, 45.5 wt% 1-propanol, 14.5 wt% 1,4-butanediol and 4 mg of DAP (1 %weight with respect to the monomers) was prepared, de-oxygenated for 10 min with nitrogen and filled into thepreviously vinylised UV-transparent fused silica capillary(30 cm×100 μm) by capillary action. The filled capillarieswere sealed with rubber septa followed by irradiation with2 J/cm2 of UVenergy at 254 nm. The resultant monolith waswashed with methanol at 2 μL/min for 1 h.

Photo-grafting of poly(GMA) on the monolith surface

The monolith was flushed with a 5 %w/v solution of ben-zophenone in methanol, previously de-oxygenated with ni-trogen for 10 min, at a flow-rate of 0.2 μL/min for 1 h.Following irradiation with 1 J/cm2 of UVenergy at 254 nm,the monolith was washed with methanol at 1 μL/min for 1 h.Following this, de-oxygenated GMA solutions (15–35 % v/v) in methanol were flushed through the polymer monolithsat 1 μL/min for 1 h and 1 J/cm2 of UVenergy at 254 nm wasapplied in order to graft GMA to the benzophenone-activated surface.

Covalent attachment of IDA to the poly(GMA)photo-grafted monolith

A 100 mM IDA solution (pH adjusted to 9.0 with 1 MKOH) was flushed through the polymer monolith at 70 °Cfor 16 h. Figure 1a shows the attachment of IDA throughring opening of the GMA structure. Five polymer monolithswere grafted with various concentrations of GMA (15, 20,25, 30 and 35 % GMA) and referred to hereafter as GMA15,GMA20, GMA25, GMA30 and GMA35.

Characterisation of IDA-modified monoliths by scanningC4D

At selected stages during the monolith modification, scan-ning C4D was used for evaluation of the distribution of IDAfunctional groups along the column length as describedpreviously [17]. Specifically, while continuously flushing

the monolith with a 1 mM ethanolamine buffer (pH9.8),the detector head was moved at 2-mm increments along thecolumn and the detector response recorded at each detectorlocation.

Sample preparation for applications

Tap water samples were collected and spiked with 10 ppmCo(II), Cd(II) and Zn(II), and injected without further dilu-tion. A bottled water sample was obtained and diluted 100-fold using the appropriate eluent and filtered prior to use.Saline samples (seawater samples) consisting of approxi-mately 0.5 M (33,000 ppm) NaCl, 400 ppm Ca(II) and1100 ppm Mg(II) were collected and diluted tenfold usingthe appropriate eluent. The samples were spiked with10 ppm Co(II), 10 ppm Cd(II) and 10 ppm Zn(II) prior toanalysis.

Results and discussion

Formation of IDA-modified monoliths

Previously, Moyna et al. [13] prepared LMA-co-EDMAchelating polymer monoliths by surface modification ofthe LMA structure with photo-grafted poly(VAL). IDAwas covalently attached to the poly(VAL) grafts, formingpendant AIDA groups, which were then applied to theseparation of selected metal cations. In this present work,GMA-surface-grafted LMA-co-EDMA polymer monolithswere modified with IDA through the ring opening of theepoxy groups on the poly(GMA) structure, providing adense surface coverage of poly(GMA–IDA). This couplingapproach is shown in Fig. 1a, in comparison to the previ-ously reported poly(VAL) grafted monoliths (Fig. 1b and c).Figure 1 illustrates the potential advantages of the use ofpoly(GMA)-based modification over the previous proce-dures, including greater flexibility in the polymer linkergroups, reducing steric hindrance, which could potentiallyaffect metal ion complexation.

Characterisation of IDA-modified polymer monolithsusing scanning C4D

To confirm surface modification, and to monitor the effectof increasing GMA concentration during the photo-graftingprocedure, the functionalised polymer monoliths were char-acterised by scanning C4D (sC4D) techniques, according toprocedures outlined by Connolly et al. [17], following theprocedure detailed by Moyna et al. [13]. For each polymermonolith prepared, a C4D scan was performed both beforeand after photo-grafting and modification with IDA. Duringthe scanning process, differences in the backpressure

Á. Moyna et al.

exhibited before and after the modification of the basemonolith with poly(GMA–IDA) were noted. MonolithGMA30 for example, showed a backpressure of 23 barduring the C4D scan of the unmodified polymer monolithwith a 1 mM ethanolamine buffer. Following modificationwith poly(GMA–IDA), the backpressure exhibited duringthe second C4D scan was ~90 bar, demonstrating a 400 %increase in backpressure. This increase in backpressure isconsistent with previous reports, where grafted polymerchains containing ionogenic functional groups have dis-played higher backpressures after modification due to “elec-trolyte responsive flow permeability” [7]. This phenomenonresults in a reversible swelling of the charged polymer grafts(proportional to eluent ionic strength), thus reducing theeffective pore volume.

Figure 2a shows an overlay of the sC4D profiles obtainedfor IDA functionalised monoliths produced using 15–35 %GMA, namely GMA15, GMA20, GMA25, GMA30 andGMA35. The data shown in this plot was obtained afternormalisation of each data set, via subtraction of the profileobtained from the scan of the unmodified base polymermonolith. As clearly shown in the normalised plots, each

column, with the possible exception of GMA15, demonstratedexcellent longitudinal homogeneity following modification,showing no signs of uneven distribution of functional groups.Following each scan, % RSD values (calculated using theC4D response at each pre-defined location along the lengthof the polymer monolith) were calculated, with resultsobtained as follows: GMA20 0.2 %, GMA25 0.6 %,GMA30 0.8 % and GMA35 0.3 %. The % RSD valueobtained for monolith GMA15 was 8 %. As can be seen fromFig. 2a, an increase in the C4D response was noted between 13and 15 cm during the scan of the IDA-modified GMA15polymer monolith. As these are normalised plots, the distur-bance was obviously not observed in the scan of the polymermonolith prior to modification, therefore suggesting that theanomaly occurred during the modification process.

Figure 2b shows a typical sC4D profile (obtained formonolith GMA30), where the bottom trace was obtainedby scanning the unmodified base polymer monolith and thetop trace obtained following photo-grafting with GMA andmodification with IDA. As expected, the scan of the un-modified phase yielded a low response due to the absence ofcharged groups on the monolith surface.

One of the main advantages of using sC4D techniques isthe ability to measure axial homogeneity of the immobilisedIDA ligand along the entire length of the column, non-invasively and non-destructively. The excellent longitudinalhomogeneity achieved in this current work is a substantialimprovement upon comparable data obtained for the variouspoly(VAL) grafted AIDA polymer monoliths presented pre-viously [13]. This improvement suggests a more quantita-tive conversion (to bonded IDA groups) of availableattachment sites with grafted poly(GMA), than achievedwith the previous poly(VAL) grafted monoliths.

Figure 3 shows a comparison of the averaged normalisedC4D responses obtained during the scans of all poly(GMA–IDA)-modified polymer monoliths and the poly(VAL–

0

500

1000

1500

2000

2500

4 6 8 10 12 14 16 18

Detector position on column (cm)

Det

ecto

r re

spon

se (

mV

)

0

200

400

600

800

1000

1200

1400

1600

1800

4 9 14 19

Det

ecto

r res

pons

e (m

V)

Detector position on column (cm)

(a)

(b)

Fig. 2 a Overlay of scanning C4D profiles obtained for GMA15 ( ),GMA20 ( ), GMA25 ( ), GMA30 ( ) and GMA35 ( ). b ScanningC4D profile obtained for GMA30 where ( ) is the profile of theunmodified polymer monolith and ( ) is the profile obtained aftermodification with IDA. All scans were carried out using a 1 mMethanolamine buffer (pH9.8) at a flow-rate of 1 μL/min

0

500

1000

1500

2000

2500

10 15 20 25 30 35 40

Monomer concentration (v/v %)

Det

ecto

r re

spon

se (

mV

)

poly(GMA)

poly(VAL)

Fig. 3 Differences in normalised and averaged C4D responseobtained during the C4D scanning of the poly(GMA–IDA)photo-grafted monoliths and previously reported poly(VAL–AIDA)photo-grafted monoliths [13]

Iminodiacetic acid functionalised organopolymer monoliths

AIDA)-modified polymer monoliths produced previously[13] (for the poly(VAL–AIDA) data shown the monomerconcentration (v/v%), as with the current GMA-graftedmonoliths, refers to % VAL solution used during photo-grafting). The same detector conditions and ethanolaminebuffer were used during the C4D scans of the sets of chelat-ing polymer monoliths, allowing for a direct comparison ofboth modification methods. As can be seen from Fig. 3,VAL15 and VAL20 monoliths exhibited a higher C4D re-sponse in comparison to GMA15 and GMA20, whereasGMA25, GMA30 and GMA35 showed large increases inresponse compared to VAL25, VAL30 and VAL35.However, it was the more proportional increase in signalseen with the poly(GMA–IDA)-modified monoliths, whichwhen compared to the poly(VAL–AIDA) data, showed howthe new modification approach provided greater control andpredictability in the fabrication of the final monolith.

Selectivity for divalent metal ions on poly(GMA–IDA)photo-grafted monoliths

Using the poly(GMA–IDA)-grafted polymer monoliths, re-tention data for a range of selected transition and heavymetals were obtained using dilute HNO3 eluents, with de-tection achieved using on-capillary indirect C4D. This re-tention factor (k) data for each of the five modifiedmonoliths is presented in Table 1. The eluent used forGMA15, GMA20 and GMA25 was 0.1 mM nitric acid.Using this eluent with the lowest capacity monolith,GMA15, Mn(II) was unretained, however no peak for Cu(II) was observed after 45 min, while Co(II) and Cd(II)showed similar k values. The selectivity shown, namelyMn(II)<Co(II)0Cd(II)<Zn(II)<<Cu(II), is constant forGMA15, GMA20 and GMA25, and is typical of that seen

with other IDA-modified stationary phases used in chelationion chromatography previously [18, 19]. Retention of Mn(II) was achieved using GMA20 and each subsequent highercapacity monolith, although as with GMA15, Co(II) and Cd(II) exhibited similar k values using GMA20 and could notbe resolved. GMA25 also showed similar k values for Cd(II)and Co(II), while Cu(II) was retained>120 min, with the0.1 mM HNO3 eluent.

Retention data for transition and heavy metals withGMA30 and GMA35 was carried out using a 0.2 mM nitricacid eluent, due to the increased complexation capacity.Interestingly, the increased capacity and use of the moreacidic eluent resulted in the resolution of Co(II) and Cd(II)(in that order), which were also well resolved from both Zn(II) and Cu(II). Once more Cu(II) was retained strongly, inexcess of 60 min using both GMA30 and GMA35monoliths.

Table 1 also shows a comparison of k values obtained forsome of the poly(VAL)-grafted AIDA monoliths and theIDA-modified poly(GMA) grafted polymer monoliths. Anumber of clear selectivity changes between the two typesof chelating monoliths can be seen, although it is the clearreduced selectivity for both Zn(II) and Cu(II) on the poly(VAL–AIDA) relative to the poly(GMA–IDA) monolithswhich is most obvious. This selectivity change results inthe elution order Mn(II)<Co(II)<Zn(II)<Cd(II)<<Cu(II) onthe VAL25 monolith, and Mn(II)<<Cd(II)0Co(II)<<Zn(II)<<<Cu(II) on the GMA25 monolith. Additionally, selectiv-ity for Mn(II) on the poly(VAL–AIDA) monoliths, unlike allother cations, is relatively high compared to the poly(GMA–IDA) phases. The above selectivity differences meant thatusing the poly(VAL–AIDA)-modified polymer monoliths, itwas possible to elute Cu(II) in the same run as Mn(II), Co(II), Zn(II) and Cd(II), although not all could be baseline

Table 1 α and k values for transition and heavy metals using poly(GMA–IDA) and poly(VAL–IDA) (data from Ref. [13]) photo-grafted monolithsof increasing complexation capacity

Metal ion GMA15 GMA20 VAL20 GMA25 VAL25 GMA30 VAL30 GMA35

k α k α k α k α k α k α k α k α0.1 mM nitric acid 0.2 mM nitric acid

Mn(II) –a 0.82 2.8 1.53 3.44 1.61 2.46 2.21

– 2.67 1.04 4.42 1.46 3.63 – 4.53

Co(II) 0.72 2.19 2.9 6.77 5.01 5.85 –b 10.02

1.00 1.01 1.94 0.98 1.98 1.40 – 1.31

Cd(II) 0.72 2.21 5.63 6.61 9.93 8.18 4.94 13.18

1.83 2.01 0.60 2.011 0.53 1.35 – 1.51

Zn(II) 1.32 4.45 3.36 13.29 5.23 11.04 –b 19.88

– – 3.50 – 4.49 – – –

Cu(II) >20 >20 11.76 >50 23.5 >25 9.94 >30

a Not retainedb Data not available

Á. Moyna et al.

separated. It was not possible to elute Cu(II) isocratically ina reasonable retention time on any of the poly(GMA–IDA)monoliths. This is as expected for IDA functionalised sta-tionary phases, as it is known that Cu(II) forms the moststable surface complexes with IDA of all the metals studiedin this work, according to known stability constant data[20]. Figure 4 shows two representative chelation ion chro-matograms obtained using the two types of modified che-lating monoliths, namely (a) poly(VAL–AIDA) (15 %) and(b) poly(GMA–IDA) (25 %). Under the conditions shown,Zn(II) would elute between Mn(II) and Cd(II) on the poly(VAL–AIDA), whereas it is much more strongly retained onthe poly(GMA–IDA) phase. Similarly, Cu(II), seen elutingat 11 min on the poly(VAL–AIDA) monolith, is retained forover 60 min on the poly(GMA–IDA) phase under identicaleluent conditions.

The retention of alkaline earth metals using GMA25,GMA30 and GMA35 was also studied. The effective col-umn length for the separations of transition and heavymetals shown previously was 140 mm to allow direct com-parison of each monolith. An advantage of detection usingon-column C4D is the ability to place the detector cell alongthe column, thus for the weaker retained alkaline earthmetals, the detector cell was placed as close to the end ofthe column outlet as possible, giving an effective columnlength of ~250 mm. Table 2 shows the chromatographic

performance data obtained for Mg(II), Ca(II), Sr(II) andBa(II) upon GMA25, GMA30 and GMA35 monoliths. Itcan be seen from the table that all four cations were virtuallyunretained on GMA25, with increasing retention seen onGMA30 and significant retention on monolith GMA35.Baseline resolution of all four cations was not possible onany of the three monoliths, although some minor improve-ments in selectivity were observed with increasing columncapacity.

Using the highest capacity monolith, with a 0.2 mMHNO3

eluent, the baseline separation of Mg(II) and Ca(II) was pos-sible (resolution of 1.63). An elution order of Mg(II)<Ca(II)<Sr(II)<Ba(II) was obtained, typical of retention dominated bysimple ion-exchange, which is as expected for cations whichform only very weak complexes with IDA under acidicconditions.

Comparison with alternative IDA phases

The selectivity for metal cations on IDA functionalisedadsorbents depends on three basic parameters: (1) chemistryof IDA immobilisation; (2) matrix effects, including pres-ence of reactive groups, e.g. silanols in silica based phasesor various polar groups in organo-polymers; and (3) con-centration of IDA groups and their distribution within thebonded layer, which can cause coordination of separatedmetal ions via more than one chelating group.

As shown previously, an elution order of Mn(II)<Co(II)<Zn(II)<Cd(II)<<Cu(II) was achieved using the 25 % poly(VAL–AIDA) monolith (using a 0.1 mM nitric acid eluent)[13]. This corresponds exactly to the elution order reported byBarron et al. [18] using a ProPac IMAC-10 analytical column

Fig. 4 a Separation of 5 ppm Mn(II), 10 ppm Cd(II) and 10 ppm Cu(II) on a poly(VAL–AIDA) functionalised monolith (VAL15). Effec-tive column length: 250 mm. Eluent: 0.2 mM nitric acid; flow-rate:1 μL/min: detection: on-capillary C4D. b Separation of 5 ppm Mn(II),10 ppm Cd(II) and 10 ppm Zn(II) on a poly(GMA–IDA) functionalisedmonolith (GMA25). Effective column length 140 mm. Eluent: 0.1 mMnitric acid. Other conditions: as in (a)

Table 2 Chromatographic performance data for alkaline earth metalsusing various GMA–IDA monoliths

Monolith (eluent HNO3) Metal ion k α N/m

GMA25 (0.1 mM) Mg(II) 0.43 1.23 1,040

Ca(II) 0.53 1.04 1,170

Sr(II) 0.55 1.09 1,110

Ba(II) 0.60 1,190

GMA30 (0.2 mM) Mg(II) 1.48 1.21 3,300

Ca(II) 1.79 – 2,400

Sr(II) – – –

Ba(II) 1.80 – 2,350

GMA35 (0.2 mM) Mg(II) 2.29 1.34 1,290

Ca(II) 3.06 1.06 1,890

Sr(II) 3.24 1.16 3,620

Ba(II) 3.76 4,460

Chromatographic conditions: Eluent: as specified in Table, metalsstandard concentration: 5 ppm. All other conditions as in Fig. 5

– not determined

Iminodiacetic acid functionalised organopolymer monoliths

(10 μmpolymeric beads coated with a poly(acrylate) layer withcovalently attached poly-IDA groups), when using a similardilute HNO3 eluent (although under these conditions theauthors noted that Pb(II) and Cu(II) were retained for over 3 h).

As shown in Table 1, using the GMA30 polymer mono-lith, the elution order Mn(II)<<Co(II)<Cd(II)<<Zn(II)<<<Cu(II) was obtained. This elution order matches closely thatobtained recently by Sugrue et al. [19] and Jones et al. [20]using IDA-modified silica phases. Sugrue et al. [19] used asilica monolith, surface-modified with IDA functionalitiesand obtained an elution order of Mn(II)<Co(II)<Cd(II)<Zn(II)<<<Cu(II), albeit using an eluent comprising of 0.065 MKNO3 and 0.035 M KCl. Jones et al. [20] reported anelution order of Mn(II)<Co(II)<Cd(II)<Zn(II)<<Cu(II)with an 8 mM HNO3 eluent, a column packed with 5 μmsilica particles with bonded IDA (150 mm×4 mm I.D.). Inaddition, earlier studies by Jones and Nesterenko on threedifferent IDA functionalised substrates, namely, DiasorbIDA silica, polymethacrylate based Tosoh TSK GelChelate 5 PW, and chelating (IDA) dye-coated poly(sty-rene-divinylbenzene) particles [21, 22], all showed similarseparation selectivity to that obtained with GMA35.

Table 3 shows a comparison of the column efficiencyobtained for GMA30, VAL30 and the above-mentionedIDA functionalised phases reported in the literature.Clearly, the efficiency obtained on the monolithic polymerIDA phases falls below that of both silica and polymericparticle type, and silica monolithic type IDA. However, thedata shows reasonable efficiencies are possible with poly-meric monoliths for small cations using the surface graftingapproach, even for chelation ion chromatography, wherepeak efficiencies are known to be relatively low in all casesdue to slower exchange kinetics associated with complexformation and dissociation (in comparison to retention-based upon non-covalent interactions, such as surface ad-sorption or simple electrostatic ion-exchange). Indeed, inmost practical applications of chelation ion chromatographyon the phases included within Table 1, step gradient elutionprogrammes are often used to elute cations with large differ-ences in selectivity for the IDA phases, in which case peakefficiencies are not such a significant issue. Furthermore, therelatively low operating backpressure across the polymermonolithic phases, together with their relative ease of pro-duction, does allow considerably longer capillary columnsto be produced in the future, which would provide highercapacity, and thus allow use of lower pH eluents (whichgenerally improves peak efficiency through improved ex-change kinetics), and obviously a higher total number oftheoretical plates (N). However, what is worthy of note fromthe data shown currently is the average ~40 % improvementin efficiency seen with the poly(GMA–IDA) monoliths,compared to the previously reported poly(VAL–AIDA)phases, which is a significant step forward. T

able

3Com

parisonof

efficiency

forpo

ly(G

MA–IDA)-mod

ifiedmon

olith

sandpo

ly(VAL–A

IDA)mon

olith

swith

otherID

Afunctio

nalised

substrates

Statio

nary

phase

GMA30

VAL15

ProPac

IMAC-10

Silica-IDA

mon

olith

Silica-IDA

5μm

Dim

ension

s(140

mm×10

0μm

I.D)a

(250

mm×10

0μm

I.D)a[13]

(50mm×2mm

I.D)b[18]

(100

mm×4.6mm

I.D)c[19]

(100

mm×4mm

I.D)d[18]

Detectio

nOn-columnC4D

On-columnC4D

Post-columnph

otom

etric

Post-columnph

otom

etric

Post-columnph

otom

etric

Mn(II)

2,31

04,60

03,47

016

,300

48,480

Co(II)

7,48

0–

4,50

027

,530

22,510

Cd(II)

5,90

04,20

010

,060

9,40

043

,280

Zn(II)

4,58

03,10

010

,420

16,450

37,720

Cu(II)

–2,77

0–

––

Av.

5,07

03,67

07,110

17,420

37,998

–data

notavailable

aEluent,0.2mM

HNO3

bEluent,0.25

mM

HNO3–10mM

KCl

cEluent,0.2M

KCl

dEluent,0.01

MHNO3

Á. Moyna et al.

Separation of metal cations using IDA-modified poly(GMA) polymer monoliths

Separations of metal standard samples were carried out on allfive poly(GMA–IDA)-modified polymer monoliths.Figure 5a and b shows a series of separations of Mn(II), Co(II), Cd(II) and Zn(II) using monoliths GMA15 to GMA35.All separations were carried out using a 0.1 or 0.2 mM HNO3

eluent with on-column indirect C4D detection. Peak tailingwas clearly evident upon the higher capacity monoliths, typ-ical of retention dominated by surface complexation. This wasless obvious on the lower capacity phases, and points to apossible change in the nature of the complex formation uponthe surface as the ligand density increases. The likelihood ofmetal ion coordination with more than one IDA group willincrease with higher ligand density and this would likely causea further decrease in exchange kinetics, and resultant bandbroadening. The highest capacity column, GMA35, was theonly monolith capable of providing baseline resolution of Mn(II), Co(II), Cd(II) and Zn(II), with Mn(II) removed from thelarge injection void peak. Retention factor precision data (n05injections) for separations of Mn(II), Co(II), Cd(II) and Zn(II)using GMA30 and GMA35 monoliths were calculated as<2.6 % for all cations shown.

Figure 6a shows a separation of Na, Mg(II), Mn(II), Co(II), Cd(II) and Zn(II) achieved using monolith GMA30(with an effective column length of 250 mm). Baselineresolution of Mn(II) with Mg(II) or Ca(II) was not possible,

and overall separation time rather long, however the sepa-ration shows the potential of the monolith for determinationof Co(II), Cd(II) and Zn(II) within samples containing highlevels of alkali and alkaline earth metal ions.

Column-to-column reproducibility

Having demonstrated the ability to produce poly(GMA–IDA) monoliths of controlled capacity, it was important todemonstrate column-to-column reproducibility. Therefore,four GMA30 monoliths and four GMA35 monoliths wereindividually prepared. Separations of selected metal ionswere obtained upon each monolith and column-to-columnvariation calculated, based upon peak retention for Mn(II),Co(II), Cd(II) and Zn(II). The chromatograms obtained us-ing the four GMA30 monoliths can be seen in Fig. 6b. ForGMA30, average column-to-column reproducibility (basedupon peak retention times) was 13 %; for GMA35, thisvalue was slightly higher at 16 %. Peak shape reproducibil-ity was similarly consistent between monoliths, with peakasymmetry data displaying an RSD of 14 % for the fourcolumns (calculated from the Co(II) and Cd(II) peaks).

Column-to-column reproducibility is greatly affected by thecomplexity of the monolith fabrication and modification pro-cedure. In this work, the process required four individual steps(monolith polymerisation, benzophenone immobilisation,GMA photografting, IDA attachment) and as such is relatively

Fig. 5 Separation of selected transition metal ions using a GMA15,GMA20, GMA25 and b GMA30, GMA35 monoliths. Chromatograph-ic conditions: Eluent: a 0.1 mM HNO3, b 0.2 mM HNO3: flow-rate:1 μL/min: injection volume: 100 nL: detection: on-column C4D:effective column length: 140 mm. Peaks (1) 5 ppm Mn(II), (2)10 ppm Co(II), (3) 10 ppm Cd(II) and (4) 10 ppm Zn(II)

Fig. 6 a Separation of Mg(II), Mn(II), Co(II), Cd(II) and Zn(II).Chromatographic conditions: Eluent: 0.2 mM HNO3: Effectivecolumn length: 250 mm (GMA30). All other conditions as inFig. 5. b Overlay of separations of Mn(II), Co(II), Cd(II) and Zn(II) using 4 individually prepared GMA30 monoliths. Chromato-graphic conditions: as in Fig. 4(b)

Iminodiacetic acid functionalised organopolymer monoliths

complex. The work described here can be compared withpreviously published work by Connolly et al. [9] who prepareda methacrylate-based monolith which was grafted with [2(methacryloyloxy)ethyl] trimethylammonium chloride and ap-plied to the ion-chromatographic separation of selected inor-ganic anions. The authors used only a two-step process(monolith polymerisation and one-step photografting) andachieved column-to-column retention time reproducibility of9 %. In this context, therefore, an average 15 % reproducibilityfor a process which has twice as many steps may be consideredto be quite reasonable.

Comparison of surface area measurements for the fourprepared columns was not possible due to the very low massof material present in capillary formats. Although acceptedpractise is to prepare a larger mass of monolith (in a largermould) for BET analysis, this option was not feasible in thiswork due to photo-initiated polymerisation being the meth-od of choice. A larger mould with a correspondingly longerlight-path would result in polymerisation occurring at adifferent rate due to the different flux of photons throughthe monomer mixture relative to that in a fused silica capil-lary, resulting in incorrect surface area estimates.

Application of poly(GMA–IDA) polymer monoliths

Determination of Mg(II) and Ca(II) in a bottled watersample

As the GMA35 had the capacity to separateMg(II) and Ca(II),the GMA35 monolith was first applied to the determination ofMg(II) and Ca(II) in a bottled water sample. Standards wereprepared in the range of 0.1–20 ppm for Mg(II) and 0.5–20 ppm for Ca(II), and calibration curves generated. R2 valuesof >0.999 were obtained for both cations. Figure 7a shows (1)a blank injection, overlaid with (2) a Mg(II) and Ca(II) stan-dard (0.8 ppm Ca(II) and 0.25 ppm Mg(II)), and (3) thechromatogram obtained for the bottled water sample (100 folddilution). A 100-fold dilution of the bottle water sample wasrequired to obtain baseline resolution between Mg(II) and Ca(II). Using peak areas, the concentration of Ca(II) and Mg(II)in the diluted water sample was calculated to be 77 ppm Ca(II)and 28 ppmMg(II). According to the manufacturer’s label, theconcentration of Ca(II) was 80 ppm and the concentration ofMg(II) was 26 ppm.

Separation of transition metals in seawater samples

The determination of transition metal ions in saline samplesusing conventional ion exchangers has the disadvantage thatthe high concentration of salt ions present in the sample matrixcan swamp the ion-exchange sites available on the stationaryphase causing ‘self-elution’, and reduced retention and resolu-tion [23]. The use of chelating ion exchangers can partially

overcome this problem, due to the relatively low affinity of thechelating ligand for alkali and alkaline earth metals.

Therefore, to demonstrate this aspect of chelation ion chro-matography on the IDA-modified monoliths, the retention oftransition and heavy metals was investigated when spiked intoreal seawater samples (tenfold diluted). A seawater sample wasobtained (Quays beach in Co. Clare) and following dilution,spiked with 10 ppmCo(II), Cd(II) and Zn(II). Figure 7b showsthree overlaid chromatograms, obtained using the GMA30monolith with an effective column length of just 100 mm.The figure shows (1) the un-spiked ×10 diluted seawater, (2)a 10 ppm Co(II), Cd(II) and Zn(II) standard and (3) a spiked×10 diluted seawater sample. The chromatograms demonstratethe ability of the chelating polymer monolith to separate Co(II), Cd(II) and Zn(II) in the presence of ~3,300 ppm NaCl aswell as in the presence of 40 ppm Ca(II) and 110 ppm Mg(II),which are unretained at this particular effective column length.For the spiked seawater sample, differences in retention timesof 2.2 % for Co(II), 1.4 % for Cd(II) and 4.5 % for Zn(II) wereobserved when compared to the retention times obtained forthe metal ion in the standard mixture.

Conclusions

LMA-co-EDMA polymer monoliths were fabricated andmodified with IDA following photo-grafting of poly(GMA)to the monolithic surface. Attachment of the IDAwas possible

Fig. 7 a Overlay of chromatograms obtained for (1) blank sample(0.2 mM nitric acid), (2) 0.8 ppm Ca(II) and 0.25 ppmMg(II) standardand (3) 100-fold dilution of a commercial bottled water sample usingmonolith GMA35. Chromatographic conditions: eluent: 0.2 mMHNO3: flow-rate: 1 μL/min: injection volume: 100 nL: detection: on-column C4D: effective column length: 250 mm. b Seawater sample(Quays beach, Co. Clare, dil. ×10) (1) Un-spiked seawater, (2) 10 ppmCo(II), Cd(II) and Zn(II) standard solution, and (3) a spiked (10 ppm)seawater sample. Chromatographic conditions: column: GMA30: ef-fective column length: 100 mm. All other conditions as in (a)

Á. Moyna et al.

through the ring opening of the epoxy group on the GMAstructure, and sC4D techniques were used to verify the con-trolled and uniform modification of a series of increasingcapacity chelating ion-exchange monoliths. Comparison ofselectivity and chromatographic performance to previouslyreported chelating ion-exchange monoliths, prepared usingsurface grafted poly(VAL), demonstrated the advantages ofthe current approach and provided new insight into the effectupon selectivity the choice of coupling chemistry can play insuch phases.

The poly(GMA–IDA)-modified polymer monoliths pro-duced in this work have been applied to the separation ofdivalent metal cations for the first time in capillary chelationion chromatography format, with on-column C4D detectionproving be a particularly simple and versatile detection optionfor this type of chromatography using dilute acid eluents.

Acknowledgements The authors would like to acknowledge the Beau-fort Marine Research Awards which is carried out under the Sea ChangeStrategy and the Strategy for Science Technology and Innovation (2006-2013), with the support of the Marine Institute, funded under the MarineResearch Sub-Programme of the National Development Plan 2007–2013and Science Foundation Ireland (Grant Number 08/SRC/B1412) forresearch funding under the Strategic Research Cluster programme.

References

1. Nischang I, Teasdale I, Brüggemann O (2011) Anal Bioanal Chem400:2289–2304

2. Svec F (2010) J Chromatogr A 1217:902–9243. Paull B, Nesterenko PN (2005) Trends Anal Chem 24:295–3034. Nordborg A, Hilder EF, Haddad PR (2011) Ann Rev Anal Chem

4:197–2265. Wang PG (2010) Monolithic chromatography and its modern

applications. ILM, UK6. Nischang I, Brüggemann O, Svec F (2010) Anal Bioanal Chem

397:953–9607. Rohr T, Hilder EF, Donovan JJ, Svec F, Frechet JMJ (2003)

Macromolecules 36:1677–16848. Eeltink S, Hilder EF, Geiser L, Svec F, Frechet JMJ, Rozing GP,

Schoenmakers PJ, Kok WT (2007) J Sep Sci 30:407–4139. Connolly D, Paull B (2009) J Sep Sci 32:2653–2658

10. Potter O, Breadmore MC, Hilder EF (2006) Analyst 131:1094–109611. Luo Q, Zou H, Xiao X, Guo Z, Kong L, Mao X (2001) J Chro-

matogr A 926:255–26412. Sproß J, Sinz A (2011) J Sep Sci 34:1958–197313. Moyna Á, Connolly D, Nesterenko E, Nesterenko PN, Paull B

(2012) J Chromatogr A 1249:155–16314. Gillespie E, Connolly D, Nesterenko PN, Paull B (2009) J Sep Sci

32:2659–266715. Connolly D, O’Shea V, Clarke P, O’Connor B, Paull B (2007) J

Sep Sci 30:3060–306816. Collins D, Nesterenko E, Connolly D, Vasquez M, Macka M,

Brabazon D, Paull B (2011) Anal Chem 83:4307–431317. Connolly D, Floris P, Nesterenko PN, Paull B (2010) Trends Anal

Chem 29:870–88318. Barron L, O’Toole M, Diamond D, Nesterenko PN, Paull B (2008)

J Chromatogr A 1213:31–3619. Sugrue E, Nesterenko PN, Paull B (2004) J Sep Sci 27:921–93020. Jones P, Nesterenko PN (2008) J Chromatogr A 1213:45–4921. Nesterenko P, Jones P (1996) J Liq Chromatogr Rel Tech 19:1033–

104522. Jones P, Challenger OJ, Hill SJ (1992) Analyst 117:1447–145023. Rahmi D, Takasaki Y, Zhu Y, Kobayashi H, Konagaya S, Haraguchi

H, Umemure (2010) Talanta 81:1438–1445

Iminodiacetic acid functionalised organopolymer monoliths