Detection of equimolar EDTA and DTPA in spiked wastewater effluents

International Journal of Environmental Analytical Chemistry ISSN: 0306-7319 (Print) 1029-0397 (Online) Journal homepage: https://www.tandfonline.com/loi/geac20 Detection of equimolar EDTA and DTPA in spiked wastewater effluents Travis K. Sander, Astha Gautam, Sophia Sarpong-Kumankomah & Jürgen Gailer To cite this article: Travis K. Sander, Astha Gautam, Sophia Sarpong-Kumankomah & Jürgen Gailer (2019) Detection of equimolar EDTA and DTPA in spiked wastewater effluents, International Journal of Environmental Analytical Chemistry, 99:6, 541-556, DOI: 10.1080/03067319.2019.1600685 To link to this article: https://doi.org/10.1080/03067319.2019.1600685 Published online: 23 Apr 2019. Submit your article to this journal Article views: 34 View Crossmark data Full Terms & Conditions of access and use can be found at https://www.tandfonline.com/action/journalInformation?journalCode=geac20 INTERNATIONAL JOURNAL OF ENVIRONMENTAL ANALYTICAL CHEMISTRY 2019, VOL. 99, NO. 6, 541–556 https://doi.org/10.1080/03067319.2019.1600685 ARTICLE Detection of equimolar EDTA and DTPA in spiked wastewater effluents Travis K. Sander, Astha Gautam, Sophia Sarpong-Kumankomah and Jürgen Gailer Department of Chemistry and BSc Environmental Science Program, University of Calgary, Calgary/AB, Canada ABSTRACT ARTICLE HISTORY Ethylenediaminetetraacetic acid (EDTA) and diethylenetriaminepentaacetic acid (DTPA) are water ‘softening’ agents that are present in numerous household and industrial detergents. Since these particular chelating agents are not significantly degraded during conventional wastewater treatment processes, wastewater treatment plant (WTP) effluents can contain up to 19 µM of EDTA and 7 µM of DTPA. Little, however, is known about the release of EDTA and DTPA from WTPs to rivers. To gain insight, we here report on the development of a cost-effective analytical method. This method is based on the chromatography of a humic acidcadmium (HA-Cd) complex on a size-exclusion chromatography column (SEC, Sephadex G-15) while using WTP effluents from Lethbridge, Banff and Canmore which contained 10 mM Trisbuffer as the mobile phase (pH 8.2). The intact HA-Cd complex is detected by means of a flame atomic absorption spectrometer (FAAS). The addition of equimolar EDTA and DTPA up to 10 µM allowed us to observe a concentration-dependent increase of the retention time of the main Cd-peak. This behaviour was qualitatively comparable between the WTP effluents and was rationalised by the EDTA/DTPA-mediated mobilisation of Cd from the HA-Cd complex. The signal intensity that corresponded to the mobilised Cd was used to establish calibration curves with corresponding correlation coefficients in the range of 0.950–0.978. Therefore, the developed method yields robust results for realistic concentrations of equimolar EDTA/DTPA in real WTP effluents. The developed method can now be applied to analyse real WTP effluent for the presence of chelating agents, whose concentrations may be expressed as being equivalent to a particular equimolar EDTA/ DTPA concentration. Received 28 January 2019 Accepted 21 February 2019 KEYWORDS Water quality; chelating agents; EDTA; DTPA; pollution quantification; water-reuse 1. Introduction EDTA and DTPA are synthetic chelating agents (CAs) which are present in a large variety of consumer products (household/industrial detergents and personal care products) [1] and are therefore commonly used worldwide in considerable quantities [2,3]. With regard to EDTA (Figure 1), for example, ~34,000 t were consumed in Europe in 1999 [4]. While EDTA is known to be environmentally persistent [5,6] and its world production is estimated at 100,000 t per year, a critical evaluation of its chemistry in natural waters is CONTACT Jürgen Gailer jgailer@ucalgary.ca © 2019 Informa UK Limited, trading as Taylor & Francis Group 542 T. K. SANDER ET AL. Figure 1. Chemical structure of the investigated chelating agents: EDTA and DTPA. missing [7]. Surface waters, such as rivers may, therefore, contain up to 4 μM of EDTA [8] and its concentration in municipal wastewater treatment plant (WTP) effluents can reach up to 19 μM [9–11]. DTPA, a structurally related aminopolycarboxylate CA (Figure 1), is consumed at comparable rates (14,732 t in Western Europe in 1999) [4] and has been detected in WTP effluent from a paper recycling factory at concentrations of 7.3 μM [4]. Although EDTA and DTPA in themselves do not pose a problem to human health, there is increasing concern with regard to the bioavailability of the metal complexes that these CAs form in the environment [8,12,13]. To unravel the potentially unintended consequences that the release of CAs from WTPs to rivers may have on the biogeochemical cycle of toxic metals, such as lead (Pb) in the ecosystem downstream [14], it is necessary to gain insight into their temporal release from WTPs. In India and China, for example, WTP effluents are widely used for the irrigation of food crops [15] and 359,000 km2 of croplands worldwide depend on the irrigation with urban wastewater[16]. Since this practice has already been shown to adversely affect food safety [15] and will likely be used in more regions that are affected by climate change, the development of analytical methods for their quantification in WTP effluents must be a priority. While a few instrumental analytical methods for the quantification of individual CAs in surface waters have been developed [3,17–23], faster and more efficient analytical methods that can directly analyse WTP effluents without the need for time-consuming sample treatment (i.e. derivatisation [19]) would be desirable. Previously, we have reported on the analysis of a humic acid-Cd (HA-Cd) complex by size-exclusion chromatography (SEC) coupled on-line to a flame atomic absorption spectrometer (FAAS) as a Cd-specific detector [12]. The utilisation of a 10 mM Tris buffer (pH 8.0) as the mobile phase allowed us to observe the intact HA-Cd complex, while the addition of 2–20 µM of either EDTA, DTPA (Figure 1) or methylglycine diacetic acid (MGDA) to this mobile phase resulted in a chelating agent-specific increase of the retention time of the Cd-peak. This behaviour was rationalised by a partial or complete CA-mediated abstraction of Cd from HA [12]. Since this approach allowed to observe the mobilisation of Cd from HA at CA concentrations that are encountered in real WTP effluent [12], we sought to adapt this approach for the analysis of EDTA and DTPAspiked WTP effluent. We chose these particular CAs because they are most widely used and are therefore frequently present in WTP effluent [4]. After collecting WTP effluent INTERNATIONAL JOURNAL OF ENVIRONMENTAL ANALYTICAL CHEMISTRY 543 from Lethbridge, Banff and Canmore, we added Tris buffer (pH 8.0) to achieve a final concentration of 10 mM and then 1 µM, 3 µM, 5 µM, 7.5 µM and 10 µM of equimolar EDTA/DTPA. Using these spiked WTP effluents as the mobile phase, we then employed the SEC-FAAS system to observe the chromatographic behaviour of HA-Cd complex. Our main motivation was to establish if the previously observed CA-mediated abstraction of Cd2+ from HA can be observed in real WTP effluent (i.e. in the presence of a significant matrix). Based on our previous results [12] the corresponding approach should allow to detect realistic concentrations of the investigated CAs in a mixture in real WTP effluent while avoiding time-consuming extraction/derivatisation steps [3] and expensive instrumental equipment [17]. 2. Experimental 2.1. Chemicals and solutions CdCl2 (>99%), tris(hydroxymethyl)aminomethane (≥99.5%, Trizma base), ethylenediaminetetraacetic acid (EDTA) disodium salt dihydrate (>99%), diethylenetriaminepentaacetic acid (DTPA, ≥99%), sodium hydroxide (>97%) and humic acid (HA derived from natural oxidised brown coal, technical grade, Lot #BCBN1711V, pH 6.4, carbon content: 43.12%, hydrogen content: 3.95%, nitrogen content 0.99%; residue on ignition: 29.59%) were acquired from Sigma-Aldrich (St. Louis, MO, United States). All solutions were prepared with distilled water obtained from a Simplicity Water Purification System (Millipore, Billerica, MA, USA). 2.2. Collection of WTP effluent and preparation of mobile phases Effluent samples (0.85–1.0 L, 6 time points) were collected into acid-washed polypropylene bottles after the UV treatment stage from the WTP in Lethbridge (November 20) starting at 6:30 am in 2 h intervals until 3:30 pm; Banff (November 29) starting at 9:00 am in 1 h intervals until 2:00 pm and Canmore (December 13) starting at 9:00 am in 1 h intervals until 2:00 pm. The WTP effluent samples were transported to Calgary in a cooler (4º), filtered (0.45 μm; Whatman nylon filter membranes, GE Healthcare, Buckinghamshire, UK) and equal volumes from each time point were mixed to obtain a representative WTP effluent sample. Each WTP sample was then characterised by pH, total organic carbon (TOC), total nitrogen (TN) and electric conductivity (EC) (Table 1). After the addition of 2.0 mL of a 1.25 M Tris buffer (pH 8.0) to 250 mL WTP effluent aliquots, the pH of the buffered WTP effluent was in the range Table 1. Physico-chemical characterisation of the collected, representative WTP effluent samples after filtration. Wastewater treatment plant pH Lethbridge 8.06 ± 0.03 Banff 8.18 ± 0.01 Canmore 8.19 ± 0.01 a Electrical conductivity (EC) (μS/cm)a 1212 ± 2 718 ± 2 685 ± 2 Total organic carbon (TOC) Total nitrogen (TN) (ppm) (ppm) 9.1 ± 0.0 4.5 ± 0.0 5.0 ± 0.1 5.1 ± 0.0 7.4 ± 0.1 15.9 ± 0.3 For comparison the EC of tap water was 4.3 ± 0.4 μS/cm, DI water was 0.366 ± 0.001 μS/cm and 10 mM Tris buffer (pH 8.0) was 463 ± 3 μS/cm. 544 T. K. SANDER ET AL. between 8.22 and 8.25 (Symphony SB20 pH Meter, Thermo Electron Corporation, Beverly, MA, USA). Thereafter, 250 mL aliquots of each Tris-buffered WTP effluent were spiked with 1 μM, 3 μM, 5 μM, 7.5 μM and 10 μM of EDTA and DTPA by adding 167 µL, 500 µL, 833 µL, 1250 µL or 1667 µL of 1.5 mM stock solutions of equimolar EDTA and DTPA, which were prepared by dissolving 140 mg of Na2EDTA 2H2O and 148 mg DTPA in 10 mM Tris buffer pH 8.0 ± 0.02 and filling to the 250 mL mark. WTP effluents which either contained 7.5 μM EDTA or DTPA were prepared in a similar manner. 2.3. Preparation of the HA-Cd complex The HA-Cd complex was prepared as previously described [12]. In brief, 0.2 g HA were added to a 100 mL volumetric flask and 10 mM Tris buffer (pH 8.0) was added to the mark. The obtained slurry was stirred at room temperature for 16 h, filtered (0.45 μm) and the clear brown solution was kept in a refrigerator overnight. This HA solution was quantitatively transferred to a volumetric flask and 250 μL of a 5000 mg Cd L−1 solution were added in a dropwise manner while stirring in 60 s intervals. The brown solution which contained a HA-Cd complex contained 5 μg of Cd and 1.0 mg of HA per 0.5 mL (pH 7.78 ± 0.02), which based on the MW of the HA of 4 kDa corresponds to one HA molecule labelled by 5 Cd2+2ions. This HA-Cd stock solution was kept in a refrigerator (4ºC) from which aliquots (~3.0 mL) were removed for each experiment. 2.4. Analytical methods The employed HPLC system consisted of a 426 HPLC Pump (Alltech Associates, Inc., Deerfield, IL, USA), a manually packed 30 × 1.0 cm (I.D.) Sephadex G-15 size-exclusion chromatography glass column (fractionation range: 1,500–100 Da) and a Rheodyne six-port injection valve, which was equipped with a 500 μL PEEK sample-loop. The flow rate was 1.0 mL min−1, and the void volume was determined by injecting Blue Dextrane (v0 = 572 s). To detect the HA-Cd complex in the column effluent, the column exit was connected to the pneumatic nebuliser of the FAAS with PEEK tubing (20 cm). Cd-specific detection at 228.8 nm (bandpass 0.7 nm) was accomplished with a Buck Model 200A flame atomic absorption spectrometer (FAAS)(Buck Scientific, East Norwalk, CT, USA). The FAAS was operated with an air/acetylene flame (oxidant pressure: 241 kPa, fuel pressure: 83 kPa). The column was equilibrated with each WTP effluent mobile phase for 75 min at a flow rate of 1.0 mL/min before the HA-Cd complex was injected (quadruplicate analysis). After every analysis, the column was cleaned with 0.2 M NaOH (flow rate: 1.0 ml/min) for 1 h followed by rinsing with distilled water from the Simplicity Water Purification System using the same flow rate for 1 h. Raw chromatographic data (i.e. the obtained Cd-specific chromatograms) were smoothed using Sigmaplot 14.0 software and the Cd-peaks were integrated using OriginPro software (version 9.1). The obtained Cd-specific chromatograms were integrated up to 1500 s. The raw data (Salsa Software) were imported into Sigmaplot 14.0 and smoothed using the bi-square algorithm. INTERNATIONAL JOURNAL OF ENVIRONMENTAL ANALYTICAL CHEMISTRY 545 3. Results WTP effluents contain a large number of pharmaceuticals and heavy metals that can adversely affect the aquatic ecosystem downstream [24–27]. In fact, >36 polar pollutants were initially identified in European WTP effluent [28], while more recent studies have identified 160 micropollutants. These facts constitute major public health concerns all around the world [27]. CAs represent another important class of pollutants that are present in WTP effluents at comparatively high concentrations [28]. In fact, the average EDTA concentration in WTP effluents is one of the highest of synthetic chemicals in municipal WTP effluents [27]. Since CAs may, therefore, play an important role in mediating the effect that toxic metals have on the ecosystems downstream [29], it is important to gain insight into their release via the effluent stream. While the development of analytical methods to quantify individual CAs in surface waters has been reported [3,17,19], we investigated if the inherent capability of EDTA and DTPA to mobilise Cd2+ from HA [12] may be exploited to detect multiple CAs in WTP effluents. This conceptual approach offers the considerable advantage of completely avoiding time-consuming sample preparation steps and involves comparatively cost-effective equipment. To leave the physico-chemical conditions compared to our previous studies largely unchanged, we added a concentrated Tris-buffer (pH 8.0, 1.25 M) to the filtered WTP effluent to achieve a final concentration of 10 mM and then chromatographed a HA-Cd complex as a function of increasing equimolar EDTA/DTPA in the WTP effluents. In order to streamline the discussion that follows, we will shorten the term ‘WTP effluent that has been amended with 10 mM Tris buffer pH 8.0ʹ to ‘WTP effluent’. The retention time of all Cd peaks and corresponding peak areas are depicted in Table 2. To enhance the clarity of the following discussion, we will first describe the results that were obtained with WTP effluent from Lethbridge and then those obtained with WTP effluent from Banff and Canmore. Thereafter, the results of the investigated WTP effluents will be compared to each other to critically evaluate the developed SECFAAS method in terms of its advantages, disadvantages as well as its robustness to reliably detect these particular CAs in real WTP effluents from Lethbridge, Banff and Canmore. 3.1. Lethbridge effluent Owing to the considerable matrix that is present in real WTP effluents, we expected the Cd recovery̶ defined as the fraction of the total injected Cd (5 µg) that eluted from the column̶ to be adversely affected. Indeed, the total Cd-peak area that was obtained for the HA-Cd complex using WTP effluent (4.823 ± 0.117 AU s) was significantly smaller compared to that for the HA-Cd complex using 10 mM Tris buffer (pH8.0)[9.473 ± 0.251 AU s], which translates to a Cd recovery of 44%[12]. Thus, 56% of Cd were lost to the column when the HA-Cd complex was chromatographed with real WTP effluent. While this loss of Cd is undesirable, one needs to recognise that only 5 µg of Cd were injected as the HA-Cd complex. The fact that 44% of the injected HA-Cd complex remained intact during the chromatographic process, however, allowed us to systematically investigate the retention behaviour of the HA-Cd complex as a function of increasing equimolar EDTA/DTPA. One needs to be aware, however, that all obtained Cd-specific 546 T. K. SANDER ET AL. Table 2. Results obtained for the analysis of an HA-Cd complex by SEC-FAAS using mobile phases that were comprised of WTP effluents collected from Lethbridge, Banff and Canmore containing 10 mM Tris-buffer (~pH 8.2) without or spiked with EDTA and DTPA. Lethbridge Concentration of EDTA & DTPA (μM) 0 1 3 5 7.5 10 7.5 EDTA 7.5 DTPA HA-Cd control (Tris buffer) Retention time (s) pH 8.25 ± 0.01 8.23 ± 0.01 8.24 ± 0.01 8.23 ± 0.01 8.22 ± 0.01 8.24 ± 0.01 8.23 ± 0.02 8.23 ± 0.02 8.05 ± 0.03 Banff Canmore Retention time (s) Retention time (s) Peak area Peak area (AU*s) pH (AU*s) pH HA-Cd CA-Cd HA-Cd CA-Cd HA-Cd CA-Cd 588 ± 1 4.823 ± 0.117 8.38 ± 0.01 563 ± 1 3.764 ± 0.160 8.44 ± 0.02 560 ± 1 569 ± 4 7.877 ± 1.154 8.19 ± 0 564 ± 3 5.105 ± 1.321 8.41 ± 0.02 563 ± 1 574 ± 2 10.513 ± 0.991 8.41 ± 0.01 572 ± 4 11.962 ± 2.079 8.40 ± 0.01 568 ± 1 568 ± 4 724 ± 12 6.706 ± 0.391 8.38 ± 0.01 577 ± 3 714 ± 14 9.958 ± 0.477 8.36 ± 0.01 571 ± 3 693 ± 7 6.277 ± 0.434 8.38 ± 0.02 714 ± 15 9.954 ± 1.325 8.51 ± 0.02 695 ± 12 698 ± 8 10.009 ± 0.352 8.39 ± 0.01 685 ± 5 9.636 ± 0.250 8.58 ± 0.02 680 ± 2 N/D N/D 8.975 ± 1.724 8.42 ± 0.01 566 ± 5 721 ± 14 9.594 ± 0.253 8.58 ± 0.01 556 ± 6 N/D N/D 763 ± 4 10.794 ± 1.188 8.40 ± 0.01 584 ± 2 773 ± 9 11.091 ± 2.801 8.60 ± 0.01 573 ± 2 N/D 586 ± 2 9.473 ± 0.251 8.05 ± 0.03 586 ± 2 9.473 ± 0.251 8.05 ± 0.03 586 ± 2 Peak area (AU*s) 3.270 ± 1.061 6.166 ± 1.837 11.666 ± 1.311 10.517 ± 0.951 10.025 ± 0.683 11.685 ± 0.697 8.585 ± 1.145 14.447 ± 1.821 9.473 ± 0.251 INTERNATIONAL JOURNAL OF ENVIRONMENTAL ANALYTICAL CHEMISTRY 547 chromatograms are actually comprised of two contributing factors, namely the CAmediated mobilisation of Cd from the injected HA-Cd complex and the CA-mediated decreased adsorption of total Cd species to the stationary phase. The latter contributing factor is conceptually likely to result in increasing Cd recovery values as the concentration of CAs in the WTP effluent is increased. Representative Cd-specific chromatograms that were obtained with the CA-spiked WTP effluents are depicted in Figure 2. Using WTP effluent, the injection of an HA-Cd complex resulted in a brown band that migrated through the column bed and produced a single Cd-peak with a retention time (tr) of 588 ± 1 s (Figure 2(a)) and a total Cd area of 4.823 ± 0.117 AU s. The retention time of this Cd-peak was close to that obtained when the HA-Cd complex was chromatographed with 10 mM Tris buffer pH 8.0 (tr = 586 ± 2 s; total Cd area 9.473 ± 0.251 AU s). The Cd-specific chromatogram that was obtained with 1.0 µM of EDTA/DTPA in WTP effluent (Figure 2(b)) revealed a distinct Cd-peak (tr = 569 ± 4 s) followed by several poorly separated Cd-peaks of lower intensity. The distinct Cd-peak had a ~ 20 s smaller retention time than that detected with unspiked WTP effluent (Figure 2(a)) and corresponds to a fraction of the parent HA-Cd complex from which Cd2+ was not mobilised, likely due to the existence of comparatively strong Cd2+2 binding sites. The 19s smaller retention time of this Cd-peak is attributed to a conformational change that is associated with the removal of Cd2+ from the weaker HA binding sites. The low-intensity Cd-peaks that eluted thereafter correspond to Cd-CA complexes, which interacted with the stationary phase as was previously observed [12]. The total Cd-peak area of 7.877 ± 1.154 AU s indicates that the presence of 1.0 µM EDTA/DTPA in the WTP effluent notably increased the Cd recovery from 44% to 72% and is attributed to the reduced adsorption of Cd species to the stationary phase. Figure 2. Representative Cd-specific chromatograms that were obtained for the analysis of a HA-Cd complex with WTP effluent from Lethbridge that contained 10 mM Tris (pH 8.0) (a), or WTP effluent from Lethbridge that contained 10 mM Tris (pH 8.0) and was spiked with 1 µM (b), 3 µM (c), 5 µM (d), 7.5 µM (e) and 10 µM (f) of equimolar EDTA/DTPA. SEC column: Sephadex G-15 (30 x 1.0 cm); flow rate: 1.0 mL/min; injection volume: 0.5 mL (5.0 µg of Cd); detector: FAAS at 228.8 nm. 548 T. K. SANDER ET AL. Using WTP effluent which contained 3 µM EDTA/DTPA revealed the presence of the same distinct HA-Cd complex (tr = 574 ± 2 s) followed by a more intense Cd-peak cluster, which reached the baseline at ~1200 s (Figure 2(c)). The total Cd-peak area of 10.513 ± 0.991 AU s (Cd-recovery: 95%) indicates a further decrease of the matrixmediated adsorption of total Cd to the column. An increase of the WTP effluent concentration to 5 µM EDTA/DTPA resulted in a 30% decrease in the intensity of the distinct HA-Cd complex peak (tr = 568 ± 4 s) and the elution of a broad Cd-peak (tr = 724 ± 12 s), which reached the baseline at ~1000 s (Figure 2(d)) and corresponded to a Cd-peak area of 6.706 ± 0.391 AU s (Cd-recovery: 61%). Thus, less Cd eluted from the column compared to 3 µM EDTA/DTPA. While the underlying cause for this comparatively low Cd recovery is unknown, we have observed similar results in our previous experiments[12]. WTP effluent which contained 7.5 µM EDTA/DTPA resulted in the elution of essentially a single Cd-peak (tr = 693 ± 7 s; Figure 2(e)), which exhibited a very minor contribution from the parent HA-Cd complex. The Cd-peak intensity decreased to baseline at ~900 s, and the total Cd-peak area was 6.277 ± 0.434 AU s, which corresponds to a Cd-recovery of 57%. This observation is somewhat reminiscent to the results obtained with 5.0 µM EDTA/DTPA. The results obtained with WTP effluent that contained 10 µM EDTA/DTPA were qualitatively similar to those for 7.5 µM EDTA/DTPA. A single Cd-peak (tr = 698 ± 8 s) was observed, which had a total Cd-peak area of 10.009 ± 0.352 AU s (Cd-recovery: 91%). We also investigated the chromatographic retention behaviour of the HA-Cd complex with WTP effluent that contained 7.5 µM of each EDTA or DTPA and compared the results to those obtained with WTP effluent that contained 7.5 µM of both CAs (Figure 3, black line). As expected, the effect that EDTA had, when compared to DTPA, on the stability of the HA-Cd complex differed significantly because EDTA is a hexadentate ligand, which exerts a lower complex formation constant with Cd (log K = 16.4) than the octadentate ligand DTPA (log K = 19.0). Using WTP effluent that contained 7.5 µM EDTA resulted in a rather distinct HA-Cd peak, which was comparable to those observed with the WTP effluent that contained 1–5 µM of both CAs (Figure 2(bd)) and a rather broad Cd-peak cluster that did not return to baseline at 1200 s (Figure 3, black line). These results demonstrate that the on-column formed Cd-EDTA complexes have a considerable affinity for the stationary phase, which is in accord with our previous observations[12]. WTP effluent that contained 7.5 µM of DTPA produced a notably different Cd-peak cluster than that obtained with equimolar EDTA. The fact that the intensity of the Cd-peak reached baseline at ~950 s (Figure 3(b), black line) demonstrates that the on-column formed CdDTPA complexes have a lower affinity for the stationary phase. Further evidence in support of this behaviour comes for the fact that the total Cd area was 10.794 ± 1.188 AU s compared to the Cd-EDTA complexes, which was 8.975 ± 1.724 AU s. The results that were obtained with the individual CAs do not show an additive effect and cannot be used to predict the results that are obtained when both CAs were present because EDTA and DTPA compete for the finite number on Cd2+ ions bound to the HA. 3.2. Banff effluent The total Cd-peak area that was obtained for the HA-Cd complex using this WTP effluent as the mobile phase (3.764 ± 0.160 AU s) was also considerably smaller than that for the HA-Cd INTERNATIONAL JOURNAL OF ENVIRONMENTAL ANALYTICAL CHEMISTRY 549 Figure 3. Representative Cd-specific chromatograms that were obtained for the analysis of an HA-Cd complex with WTP effluents from Lethbridge, Banff and Canmore that were spiked with 7.5 µM EDTA (top), DTPA (middle) and 7.5 µM DTPA and EDTA (bottom). SEC column: Sephadex G-15 (30 x 1.0 cm); flow rate: 1.0 mL/min; injection volume: 0.5 mL (5.0 µg of Cd); detector: FAAS at 228.8 nm. complex using 10 mM Tris buffer (pH8.0) [9.473 ± 0.251 AU s] (Figure 4). These results translate to a total Cd recovery of 40%, which is slightly lower than the corresponding value with the Lethbridge WTP effluent. Using unspiked WTP effluent, the injection of a HACd complex resulted in a single Cd-peak with a retention time of 563 ± 1 s (Figure 4(a)), which was shorter than the HA-Cd complex when chromatographed with 10 mM Tris buffer pH 8.0 (tr = 586 ± 2 s; total Cd area 9.473 ± 0.251 AU s). This 23s reduced retention time correlates with the smaller EC of this WTP effluent compared to that from Lethbridge (Table 1) and can be rationalised in terms of a larger hydrodynamic radius of the HA-Cd complex. Overall, the Cd-specific chromatograms that were obtained with the CA spiked WTP effluent from Banff 550 T. K. SANDER ET AL. Figure 4. Representative Cd-specific chromatograms obtained for the analysis of a HA-Cd complex either with WTP effluent from Banff that contained 10 mM Tris (pH 8.0) (a), or WTP effluent from Banff that contained 10 mM Tris (pH 8.0) and was spiked with 1 µM (b), 2 µM (c), 5 µM (d), 7.5 µM (e) and 10 µM of equimolar EDTA/DTPA. SEC column: Sephadex G-15 (30 x 1.0 cm); flow rate: 1.0 mL/ min; injection volume: 0.5 mL (5.0 µg of Cd); detector: FAAS at 228.8 nm. (Figure 4(bf)) were rather similar compared to those from Lethbridge (Figure 2(bf)) and they will therefore not be described in detail. The total Cd areas that were obtained with the various CA concentrations are summarised in Table 2, while the results for the individual EDTA/DTPA spiked WTP effluent are depicted in Figure3 (red line). 3.3. Canmore effluent The total Cd-peak area that was obtained for the HA-Cd complex using this WTP effluent (3.270 ± 1.061 AU s) was also considerably smaller than that for the HA-Cd complex using 10 mM Tris buffer (pH8.0)[9.473 ± 0.251 AU s] (Figure 5). These results translate to a total Cd recovery of 35%, which is comparable to the value that was obtained with the Banff WTP effluent. Using unspiked WTP effluent, the injection of a HA-Cd complex resulted in a single Cd-peak with a retention time of 560 ± 1 s (Figure 5(a)), which was also shorter than that when the HA-Cd complex was chromatographed using 10 mM Tris buffer pH 8.0 as the mobile phase (tr = 586 ± 2 s). This 26s reduced retention time is in accord with the results obtained with the WTP effluent from Banff. Overall, the Cd-specific chromatograms that were obtained with the CA spiked WTP effluent from Canmore (Figure 5(bf)) were rather similar compared to those from Banff (Figure 4(bf)) and they will therefore not be described in detail. The total Cd areas that were obtained with the various CA concentrations are summarised in Table 2, while the results for the individual EDTA/DTPA spiked WTP effluent are depicted in Figure 3 (blue line). 4. Discussion The results of the studies that are presented in this manuscript represent an extension of our previous studies in which we investigated the chromatographic retention behaviour INTERNATIONAL JOURNAL OF ENVIRONMENTAL ANALYTICAL CHEMISTRY 551 Figure 5. Representative Cd-specific chromatograms obtained for the analysis of a HA-Cd complex either with WTP effluent from Canmore that contained 10 mM Tris (pH 8.0) (a), or WTP effluent from Canmore that contained 10 mM Tris (pH 8.0) and was spiked with 1 µM (b), 2 µM (c), 5 µM (d), 7.5 µM (e) and 10 µM of equimolar EDTA/DTPA. SEC column: Sephadex G-15 (30 x 1.0 cm); flow rate: 1.0 mL/min; injection volume: 0.5 mL (5.0 µg of Cd); detector: FAAS at 228.8 nm. of the same HA-Cd complex on the same SEC stationary phase using 10 mM Tris buffer (pH 8.0) which contained increasing concentration of individual CAs as the mobile phase [12]. Two notable differences of the results from this study compared to those from the previous study are recognisable. Previously, the retention time of the Cd-peak that was observed after the HA-Cd complex was chromatographed increased by 30 s or 16 s when 20 µM of EDTA or DTPA were added to the 10 mM Tris buffer (pH 8.0) mobile phase[12]. These results are in contrast to those of the present study where the retention time of the Cd-peak that was obtained when the HA-Cd complex was chromatographed with WTP effluent that was amended with 10 µM EDTA and DTPA increased by ~100 s (Table 2). A likely explanation for this behaviour is that the matrix of the WTP effluents (either the contained polar pollutants and/or the ionic matrix) increased the interaction of the oncolumn formed Cd-EDTA/DTPA complexes with the Sephadex G-15 stationary phase. Given the existence of free hydroxyl-groups on this stationary phase, it is conceivable that these functional groups play an important role in terms of mediating this increased interaction. Furthermore, in the present study the addition of 1–5 µM EDTA/DTPA to the WTP effluent mobile phase resulted in the mobilisation of Cd from weak HA binding sites, while a minor fraction of the Cd ions remained bound to strong HA binding sites. This behaviour was not observed in our previous study, [12] which was conducted with a 10 mM Tris buffer (pH 8.0) as the mobile phase. Therefore, this observation must be attributed either to the matrix and/or the increased EC of the employed WTP effluent mobile phases. Since the EC of 10 mM Tris buffer (pH 8.0) was 463 ± 3 µS/cm, whereas the EC of the WTP effluents was in the range of 685–1212 µS/cm, the higher EC may have exerted a conformational change of the HA-Cd complex when the latter was injected and immediately interacted with the mobile phase. Thereafter, EDTA/DTPA 552 T. K. SANDER ET AL. interacted with the ‘altered’ HA-Cd complex, with 4 Cd2+ [2]+ ions being mobilised from the HA, while one Cd2+2 ion remained bound to strong binding sites on the HA. The 50–65% loss of Cd to the column when three different WTP effluents were used (compared to 10 mM Tris buffer pH 8.0) is attributed to the matrix-mediated partial adsorption of Cd species to the stationary phase. Another possibility for this loss is the displacement of Cd from the HA by other metal ions that are present in the WTP effluent followed by the subsequent adsorption of the liberated Cd2+ ions to the stationary phase. Real WTP effluents contain other metal ions that may affect/participate in the complex equilibria that unfold when the HA-Cd complex is chromatographed. Since at >3 µM EDTA/ DTPA in the WTP effluent the metal concentrations will not matter as their concentration is much smaller compared to that of EDTA/DTPA, only the situation with the 1 µM EDTA/DTPA in the WTP effluents needs to be discussed. The complex stability constants (log K) of Ca2+ and Mg2+ for EDTA/DTPA are smaller than those for Cd [2, 17]. Therefore, their presence in the WTP effluents is irrelevant. WTP effluents, however, also contain Zn 2+2 and Fe3+. With regard to Zn2+ the log K for EDTA is identical to that of Cd2+ (log K = 16.4) and that for DTPA is smaller compared to that of Cd2+ (log K = 18.3 vs log K = 19.0). Thus, Cd2+ will be mobilised from the HA. With regard to Fe3+, its log K for EDTA is larger than that for Cd2+ (log K = 25.0 vs log K = 16.4) and the same holds true for DTPA (log K = 28.0 vs log K = 19.0). In the WTP effluents that were spiked with 1 µM of EDTA/DTPA, all Fe3+ will be complexed and only free/leftover EDTA/DTPA will be able to mobilise Cd2+ from the HA. Since this was observed with all WTP effluents (Figures 2, 4, 5), the Fe3+ concentration therein must have been smaller than the total concentration of EDTA and DTPA (i.e. 2 µM). In order to extract quantitative results from the obtained data, it is evident from Figures 2, 4 and 5 that the signal intensity of the second, broad Cd-peak which corresponds to CACd complexes increased with the total concentration of both CAs in the WTP effluent. We, therefore, plotted the concentration of equimolar EDTA/DTPA in the WTP effluent against the signal intensity of the second Cd-peak at a retention time of 700 s for effluent collected from Lethbridge, Banff and Canmore (Figure 6). These results revealed rather similar slopes, linear correlation coefficients in the range 0.954–0.978 indicating that the developed method is robust. We note that at 1 µM EDTA/DTPA in WTP effluent the intensity of the analytical signal, i.e. the Cd peak intensity at 700 s does not fall onto the calibration curve, but that at >2 µM EDTA/DTPA the quantification is more accurate. Collectively, these data indicate that the developed method should be able to detect realistic concentrations of CA in WTP effluent. Given that real WTP effluent may contain several CAs, the results that are obtained with the developed method would allow one to express the results that are obtained with this method in terms of ‘being equivalent to a concentration of x µM EDTA/DTPA’. Overall, the most relevant finding of our investigations is that the developed method can reliably detect ≥2 µM of equimolar EDTA/DTPA in real WTP effluents. Apart from the fact that no time-consuming sample pretreatment of the WTP effluent is necessary and that results can be obtained in < 20 min, the employed instrumental analytical equipment is considerably more cost-effective compared to previously reported analytical methods [3,17]. Arguably the most important advantage of the developed method is the fact that the underlying detection principle (i.e. the CA-mediated mobilisation of Cd from HA) allows to detect multiple CAs in WTP effluent, which is important considering at least 13 other chelating agents, which can be structurally classified as aminopolycarboxylates, INTERNATIONAL JOURNAL OF ENVIRONMENTAL ANALYTICAL CHEMISTRY 553 Figure 6. Calibration curve obtained by plotting the observed Cd-signal signal intensity at retention time 700 s (corresponding to the presence of CA-Cd complexes) with equimolar EDTA/DTPA in WTP effluents obtained from Lethbridge (blue), Banff (red) and Canmore (green). hydroxycarboxylates or organophosphonates have been detected in WTP effluent [4]. The single most important disadvantage of the developed method is that no precise information about the concentration of each CAs that is present in any given WTP effluent can be obtained. The practical value of the developed method pertains to the fact that it is ideally suited to analyse WTP effluents for CAs. This information in itself is of considerable practical use since EDTA-based soil remediation is moving from the exploratory scale to the scale of demonstrational facilities, and the focus of investigation is, therefore, shifting from remediation efficiency to possible negative environmental side-effects[30]. The rapid assessment of wastewaters for CAs is also destined to become progressively more important as water scarcity is projected to increase in various parts of the world [31]. In parts of India, China and other regions that suffer from severe drought WTP effluents are already utilised for the irrigation of food crops, [15,16] which has been demonstrated to result in an inadvertent uptake of Cd. This uptake is most likely caused by the fact that CAs form complexes with toxic Cd and Pb species which bioavailable and are uptaken by periphyton in rivers [29,32] and into food crops [12,33]. The developed method represents a useful tool to help mitigate this problem and constitutes a unique research tool to address two important environmental issues. Firstly, WTP effluent that contains comparatively high concentrations of CAs could be diverted into holding tanks for their subsequent degradation using established degradation methods to significantly increase water reuse, which is not yet widely executed in Canada [34] but may have to be embraced in the near future[35]. To this end, it has long been known that EDTA can be photochemically degraded after Fe3+ or Cu2+ 2 is added [7,36]. The developed method is ideally suited to develop faster processes to degrade CAs. 554 T. K. SANDER ET AL. Secondly, the developed method may also be employed to obtain a more detailed understanding of the mechanisms by which anthropogenic activities perturb the biogeochemical cycling of Cd downstream of WTPs [37], which is a starting point to improve the regulatory framework to curb its influx into human tissues [38,39]. 5. Conclusion Global water pollution is of increasing public health concern [26]. EDTA and DTPA are frequently present in WTP effluent at low µM concentrations, [13] which may adversely affect the aquatic ecosystems downstream of WTPs [8] and compromise food safety if it is used for the irrigation of food crops. Considering that climate change-induced water stress will force more regions to utilise WTP effluent for the irrigation of food crops, the analysis of WTP effluents for CAs will become more important. We obtained representative WTP effluents from Lethbridge, Banff and Canmore, added 1–10 μM of equimolar EDTA/DTPA to each effluent and then chromatographed an HA-Cd complex. The CA-mediated abstraction of Cd from the HA-Cd complex resulted in the elution of a new Cd-peak, the intensity of which correlated strongly with the concentration of EDTA/DTPA in the WTP effluent mobile phase (r [2]= 0.954–0.978). Although WTP effluents may contain many other CAs in addition to EDTA and DTPA [17], the developed method will detect all CAs that are present at a sufficient concentration to mobilise Cd from the HA-Cd complex. Hence, the developed method is useful to gain insight into the temporal release of CAs of WTP effluents (hourly/weekly/seasonally) into rivers. Perhaps most importantly, this method is particularly useful to help mitigate the known problems that are associated with the practice of using WTP effluent for the irrigation of food crops [15], which is inherently problematic because certain CA-Cd complexes that are present therein are known to be absorbed into the root system [14,40]. The subsequent translocation of Cd to the edible parts of food crops may ultimately compromise food safety, [15] which must be avoided at all costs since contaminated foods have become the dominant source of human exposure to environmental Cd. In fact, many durum wheats grown in North America often exceed the maximum food safety limit of 200 ppb of Cd[33]. Faced with the challenge to produce enough food to feed 9 billion people by 2050 [41], it will become progressively more important to better understand the mechanisms by which toxic metals inadvertently invade the food chain and ultimately human tissues. The role that environmentally abundant CAs [42–44] and small molecular weight thiols play in this context [45] clearly deserves more attention in the near future. To this end, it is important to better understand the phytoavailability and translocation of CA-Cd complexes at the soil–plant interface [13,46] to better hitherto unknown sources of the chronic exposure of humans to environmental chemicals [47–51]. Acknowledgments This research was funded by the Natural Sciences and Engineering Research Council (NSERC) of Canada. Farzin Malekani is gratefully acknowledged for measuring the TOC and TN concentrations in the WTP effluent samples and Duane Guzzi (Lethbridge WTP), Bryant Rathbone (Banff WTP) and Dennis Letourneau (Canmore WTP) for their help in obtaining WTP effluent samples. INTERNATIONAL JOURNAL OF ENVIRONMENTAL ANALYTICAL CHEMISTRY 555 Disclosure statement No potential conflict of interest was reported by the authors. Funding This work was supported by the National Science and Engineering Research Council of Canada [RGPIN-2014-04156]. References [1] P.C. DeLeo, M. Ciarlo, C. Pacelli, W. Greggs, E.S. Williams, W.C. Scott, Z. Wang and B. W. Brooks, ACS Sustainable Chem. Eng. 6, 2094 (2018). doi:10.1021/ acssuschemeng.7b03510. [2] I.S.S. Pinto, I.F.F. Neto and H.M.V.M. Soares, Environ. Sci. Pollut. Res. 21, 11893 (2014). doi:10.1007/s11356-014-2592-6. [3] J.J. Jimenez, Anal. Chim. Acta 770, 94 (2013). doi:10.1016/j.aca.2013.01.060. [4] T.P. Knepper, Trends Anal. Chem. 22 708–724 (2003). doi:10.1016/S0165-9936(03)01008-2. [5] B. Nowack, R. Schulin and B.H. Robinson, Environ. Sci. Technol. 40, 5225 (2006). [6] M. Bucheli-Witschel and T. Egli, FEMS Microbiol. Rev. 25, 69 (2001). doi:10.1111/j.15746976.2001.tb00572.x. [7] B. Nowack, Environ. Sci. Technol. 36, 4009 (2002). [8] C. Oviedo and J. Rodriguez, Quim. Nova. 26, 901 (2003). doi:10.1590/S010040422003000600020. [9] W.W. Bedsworth and D.L. Sedlak, J. Chromatogr. A. 905, 157 (2001). [10] M. Fuehrhacker, G. Lorbeer and R. Haberl, Chemosphere. 52, 253 (2003). doi:10.1016/S00456535(03)00037-7. [11] E. Bloem, S. Haneklaus, R. Haensch and E. Schnug, Sci. Total. Environ. 577, 166 (2017). doi:10.1016/j.scitotenv.2016.10.153. [12] A.E. North, S. Sarpong-Kumankomah, A.R. Bellavie, W.M. White and J. Gailer, J. Environ. Sci. 57, 249 (2017). doi:10.1016/j.jes.2017.02.004. [13] T. Karak, R.K. Paul, D.K. Das and R.K. Boruah, Environ. Monit. Assess. 188, 670 (2016). doi:10.1007/s10661-016-5685-5. [14] L.A. Schaider, D.R. Parker and D.L. Sedlak, Plant Soil. 286, 377 (2006). doi:10.1007/s11104006-9049-8. [15] S. Khan, Q. Cao, Y.M. Zheng, Y.Z. Huang and Y.G. Zhu, Environ. Pollut. 152, 686 (2008). doi:10.1016/j.envpol.2007.06.056. [16] A.L. Thebo, P. Drechsel, E.F. Lambin and K.L. Nelson, Environ. Res. Lett 12, 074008 (2017). doi:10.1088/1748-9326/aa75d1. [17] T.P. Knepper, A. Werner and G. Bogenschuetz, J. Chromatogr. A. 1085, 240 (2005). [18] C.Z. Xie, T. Haley, P. Robinson and K. Stewart, Int. J. Environ. Ecol. Eng. 2, 82 (2008). [19] J.B. Quintana and T. Reemtsma, J. Chromatogr. A. 1145, 110 (2007). doi:10.1016/j. chroma.2007.01.044. [20] R. Geschke, M. Zehringer and J. Fresen., Anal. Chem. 357, 773 (1997). doi:10.1007/ s002160050247. [21] R.L. Sheppard and J. Henion, Electrophoresis. 18, 287 (1997). doi:10.1002/elps.1150180218. [22] M. Sillanpaa, R. Kokkonen and M.-L. Sihvonen, Anal. Chim. Acta. 303, 187 (1995). doi:10.1016/0003-2670(94)00535-T. [23] J.J. Jimenez, Talanta. 116, 678 (2013). doi:10.1016/j.talanta.2013.07.052. [24] J. Munoz, M.J. Gomez-Ramos, A. Aguera, J.F. Garcia-Reyes, A. Molina-Diaz and A. R. Fernandez-Alba, Trends Anal. Chem. 28, 676 (2009). doi:10.1016/j.trac.2009.03.007. 556 T. K. SANDER ET AL. [25] T.J. Stewart, R. Behra and L. Sigg, Environ. Sci. Technol. 49, 5044 (2015). doi:10.1021/ es505289b. [26] R.P. Schwarzenbach, T. Egli, T.B. Hofstetter, U. von Gunten and B. Wehrli, Annu. Rev. Environ. Resour. 35, 109 (2010). doi:10.1146/annurev-environ-100809-125342. [27] J. Margot, L. Rossi, D.A. Barry and C. Holliger, WIREs Water. 2, 457 (2015). doi:10.1002/ wat2.1090. [28] T. Reemtsma, S. Weiss, J. Mueller, M. Petrovic, S. Gonzales, D. Barcelo, F. Verntura and T. P. Knepper, Environ. Sci. Technol. 40, 5451 (2006). [29] P. Bradac, R. Behra and L. Sigg, Environ. Sci. Technol. 43, 7291 (2009). [30] E. Jez and D. Lestan, Chemosphere. 151, 2020 (2016). doi:10.1016/j. chemosphere.2016.02.088. [31] D. Molden, T. Oweis, P. Stdeuto, P. Bindraban, M.A. Hanjra and J. Kijne, Agric. Water Manage. 97, 528 (2010). doi:10.1016/j.agwat.2009.03.023. [32] J.R. McCauley and J.L. Bouldin, Bull. Environ. Contam. Toxicol. 96, 757 (2016). doi:10.1007/ s00128-016-1813-8. [33] N.S. Harris and G.J. Taylor, BMC Plant Biol. 13, 103 (2013). doi:10.1186/1471-2229-13-103. [34] K. Exall, Water Qual. Res. J. Canada. 39, 13 (2004). doi:10.2166/wqrj.2004.004. [35] M.B. Masud, T. McAllister, M.R.C. Cordeiro and M. Faramarzi, Sci. Total Environ. 616–618, 208 (2018). doi:10.1016/j.scitotenv.2017.11.004. [36] M.E.T. Sillanpae, T.A. Kurniawan and W.-H. Lo, Chemosphere. 83, 1443 (2011). doi:10.1016/j. chemosphere.2011.01.007. [37] L. Malmauret, D. Parent-Massin, J.-L. Hardy and P. Verger, Food Additives and Contaminants. 19, 524 (2002). doi:10.1080/02652030210123878. [38] P.G.C. Campbell and J. Gailer, in Metal Sustainability: Global Challenges, Consequences and Prospects, edited by R. M. Izatt (John Wiley & Sions Ltd., Chichester, United Kingdom, 2016), p. 221. [39] G.A. Drasch, Sci. Total Environ. 26, 111 (1983). [40] D.L. Van Engelen, R.C. Sharpe-Pedler and K.K. Moorhead, Chemosphere. 68, 401 (2007). doi:10.1016/j.chemosphere.2007.01.015. [41] H.C.J. Godfray, J.R. Beddington, I.R. Crute, L. Haddad, D. Lawrence, J.F. Muir, J. Pretty, S. Robinson, S.M. Thomas and C. Toulmin, Science. 327, 812 (2010). doi:10.1126/ science.1185383. [42] G.R. Aiken, H. Hsu-Kim and J.N. Ryan, Environ. Sci. Technol. 45, 3196 (2011). doi:10.1021/ es103992s. [43] L. Verheyen, F. Degryse, T. Niewold and E. Smolders, Sci. Total. Environ. 426, 90 (2012). doi:10.1016/j.scitotenv.2012.03.044. [44] H. Yan, L. Yang and Q. Wang, Talanta. 84, 287 (2011). doi:10.1016/j.talanta.2011.01.019. [45] F. Liu, C. Fortin and P.C.G. Campbell, Limnol. Oceanogr. 62, 2604 (2017). doi:10.1002/ lno.10593. [46] V. Antoniadis, E. Levizou, S.M. Shaheen, Y.S. Ok, A. Sebastian, C. Baum, M.N.V. Prasad, W. W. Wenzel and J. Rinklebe, Earth Science Rev. 171, 621 (2017). doi:10.1016/j. earscirev.2017.06.005. [47] A. Kataria, L. Trasande and H. Trachtman, Nat. Rev. Nephrol 11, 610 (2015). doi:10.1038/ nrneph.2015.94. [48] A.M. Newbigging, R.E. Paliwoda and X.C. Le, J. Environ. Sci. 30, 129 (2015). doi:10.1016/j. jes.2015.03.001. [49] C.Y. Chang, H.Y. Yu, J.J. Chen, F.B. Li, H.H. Zhang and C.P. Liu, Environ. Monit. Assess. 186, 1547 (2014). doi:10.1007/s10661-013-3472-0. [50] J.M. Custos, C. Moyne, T. Treillon and T. Sterckeman, Plant Soil. 374, 497 (2014). doi:10.1007/ s11104-013-1906-7. [51] S. Sarpong-Kumankomah, M.A. Gibson and J. Gailer, Coord. Chem. Rev. 374, 376 (2018). doi:10.1016/j.ccr.2018.07.007.