Elsevier

Analytica Chimica Acta

Volume 896, 8 October 2015, Pages 120-127
Analytica Chimica Acta

The nature of the salt error in the Sn(II)-reduced molybdenum blue reaction for determination of dissolved reactive phosphorus in saline waters

https://doi.org/10.1016/j.aca.2015.09.021Get rights and content

Highlights

  • SnCl2 molybdenum blue reaction for PO43 determination in saline waters is studied.

  • The reaction is inhibited by complexation of Sn(IV) with Cl causing a salt error.

  • Dilution of the sample matrix prior to SnCl2 addition is recommended at low salinity.

  • The ascorbic acid/Sb(III) method is recommended at high salinity.

Abstract

Sn(II) is a well-known reductant used in the formation of phosphomolybdenum blue for the determination of dissolved reactive phosphorus (DRP) in waters because it provides rapid and quantitative reduction. However, in saline waters, this method suffers from a salt error which causes a significant decrease in sensitivity. This phenomenon has never been adequately explained in the literature. The Murphy and Riley method, which uses Sb(III) and ascorbic acid for the reduction step, is preferred for DRP determination in saline waters because it is unaffected by salinity, but it exhibits a sensitivity approximately 30% lower than that when Sn(II) is used as the reductant without Cl interference. This study investigates the processes causing the salt error and possible ways of minimizing it, so that the benefits of Sn(II) reduction on the molybdenum blue reaction rate and sensitivity may be exploited in the determination of low levels of DRP in marine and estuarine waters. It has been established that the salt error is caused by the formation of Sn(IV) chloro-complexes which compete with the formation of Sn(IV)-substituted phosphomolybdenum blue, forcing the reaction to proceed via the much slower, less favourable process of direct reduction that occurs in methods using organic reductants such as ascorbic acid.

Introduction

The determination of orthophosphate is of key importance in environmental monitoring, as it is the most bioavailable form of phosphorus and a limiting macronutrient for primary production in freshwaters [1]. High concentrations of orthophosphate promote eutrophication, and monitoring of this nutrient is therefore necessary for assessment of water quality, the efficiency of wastewater treatment, the size of anthropogenic phosphorus inputs, and for the study of nutrient cycling [1], [2], [3]. Dissolved reactive phosphorus (DRP) is the fraction of phosphorus species which reacts with acidic molybdate to form the reducible 12-molybdophosphoric acid species in the well-known molybdenum blue reaction, and is often used as a surrogate measure of orthophosphate (and hence bioavailable P) despite the potential for overestimation due to hydrolysis of other labile P fractions [1].

Of the published molybdenum blue procedures for DRP determination, the two most common reduction methods use either a combination of ascorbic acid and Sb(III) [4] or Sn(II) (as SnCl2) in tandem with a sacrificial reductant such as hydrazine sulfate [5]. The so-called Murphy and Riley method [4] is the most highly cited paper in Analytica Chimica Acta with over 4600 citations as of June 2015. However, the reduction process in the ascorbic acid/Sb(III) method is slower than that involving SnCl2 and forms a product with a molar absorptivity ca. 20–30% lower than that of SnCl2 methods or ascorbic acid methods that use heating [4], [6], [7]. This is a substantial difference in sensitivity, especially considering the need to determine DRP at low μg L−1 P concentrations in fresh and marine waters. However, ascorbic acid/Sb(III) methods are generally assumed to be superior for sample matrices with high or variable salinities such as marine or estuarine waters, because SnCl2 reduction methods suffer from significant subtractive interference (the salt error) in the presence of even millimolar chloride concentrations [4], [8], [9], [10], [11]. This is particularly problematic in the analysis of samples with varying salinities, e.g. estuarine waters, as the magnitude of the salt error increases with the chloride concentration [8].

The ‘phosphomolybdenum blue’ (PMB) product reported in numerous published methods is not a single species. The use of organic reductants with heating (e.g. ascorbic acid [4], [14], [15], hydrazine sulfate [5], [16], [17], 4-(methylamino)phenol sulfate [10], 1-amino-2-naphthol-4-sulfonic acid [18], [19]), ascorbic acid with Sb(III) [4], [6], [7], and SnCl2 [8], [18], [20], [21] produce different products with distinctive UV–visible spectra (Fig. 1), which are discussed at length in the review by Nagul et al. [22]. The colour of all PMB species originates from intense inter-valence charge transfer (IVCT) transitions between Mo(V) and Mo(VI) within their Keggin structure [23]. Perturbation of this process through changes in the solvation shell, substitution of a metal into the structure, or the formation of charge-transfer complexes will alter the spectral profile of the product. Despite the utility of SnCl2 as a reductant in the MB reaction, the chemistry of the corresponding reduction products has been given little attention in the analytical literature, and an in-depth understanding of this chemistry is required for meaningful discussion of the salt error and formulation of possible ways of minimising it.

Orthophosphate reacts with acidified molybdate to produce 12-molybdophosphoric acid (12-MPA, H3PMo12O40) in a mixture of its α- and β-isomers, of which only the α-isomer is thermodynamically stable [24]. In most MB methods a four-electron reduction of 12-MPA takes place yielding the PMB species [HnPMo12O40](7−n)−, abbreviated as PMB(4e), which exhibits an absorption maximum at 820 nm (Fig. 1) [24], [25], [26], [27]. Immediately after reduction, PMB(4e) exists as the reduced α-isomer but spontaneously changes to the β-isomer which is more stable in the reduced state. However, when SnCl2 is used as the reductant, it has been shown via analogy with silicomolybdate species that after the reduction step, two Sn4+ cations immediately substitute for two Mo(VI) cations in α-PMB(4e) before it can isomerise to the β-form, producing the unstable species α-[PMo10Sn2O37]5−, abbreviated as Sn2PMB(4e), which may then lose one Sn4+ cation to form the poorly characterised and more intensely coloured α-SnPMB(4e) species [9], [28], [29]. This species is tentatively suggested to be α-[PMo11SnO39]7− because Mo(VI) is 6-coordinate in the Keggin structure and Sn(IV) is expected to only be 4-coordinate. Importantly, the substitution of Mo(VI) with Sn(IV) in 12-MPA does not take place before the reduction has occurred [29], and it is critical to note that whilst SnCl2 is capable of reducing both isomers of 12-MPA, Sn4+ only substitutes into α-PMB(4e), and never into the more stable β-isomer [28], [29]. Sn2PMB(4e) possesses an absorption maximum at ca. 710 nm with a distinct shoulder at 620 nm [29]. SnPMB(4e) also has its absorption maximum at around 710 nm, although with a much greater molar absorptivity, and the shoulder at 620 nm is almost entirely obscured by the main peak [28], [29]. Both species are evident in the spectra of SnCl2 PMB products reported in the literature [4], [13], [20], [25], [30], as is a weak shoulder at 820 nm due to β-PMB(4e). SnCl2 is often used in the presence of a sacrificial reductant such as hydrazine sulfate to protect it from passive oxidation without altering the reduction product [5], [8], [31].

The elucidation of the salt error is the objective of the present work. Hesse and Geller demonstrated in 1968 that the absorbance of the PMB formed by SnCl2 at 730 nm progressively decreased by increasing the concentration of Cl [8]. Furthermore, they also showed that the salt error was specific to Cl and was not an ionic strength effect since various salts of SO42, NO3 and ClO4 did not interfere at concentrations below about 0.1 M. However, significant Cl interference was observed at concentrations an order of magnitude lower than this [8]. Early work on the salt error by Cooper [32] reported not only a decrease in the absorbance but also a change in the shape of the absorption spectrum. Kobayashi and Nakamura have since demonstrated through the use of solvent extraction that the ratio of Sn: P in the extracted PMB decreased linearly with increasing Cl concentration [9], and they concluded that Cl caused the formation of tin chloro-complexes which prevented Sn from associating with PMB. No such interaction was observed in the case of the ascorbic acid/Sb(III) method. This is in good agreement with thermodynamic studies on Sn(II) and Sn(IV) equilibria in Cl solutions which demonstrate the formation of strong complexes with Sn in both oxidation states [33], [34]. However, the Sn(IV) complexes are more stable than those of Sn(II), which leads to an increase in the reducing power of Sn(II) in the presence of Cl [35]. Complexation with chloride thus decreases the standard potential of the Sn(IV)/Sn(II) redox couple. It is implausible that this change in redox potential alone could be the source of the salt error, since it would be expected to actually enhance the reduction of 12-MPA.

The present work aims to confirm the above findings through a spectrophotometric investigation of the 12-MPA reduction products formed by SnCl2 in the presence and absence of Cl, in order to provide a comprehensive explanation of the salt error with a view to developing possible mitigation strategies.

Section snippets

Reagents

All reagents employed were used as received. Deionised water (18.2 MΩ cm, Millipore, Synergy 185, France) was used in the preparation of all solutions. A 1.0% (w/v) stock solution of ammonium heptamolybdate tetrahydrate was prepared by dissolving (NH4)6Mo7O24.4H2O (AR, Fisons, Australia) in deionised water. A stock reductant solution containing 0.60% (w/v) N2H6SO4 (AR, Ajax, Australia) and 0.06% (w/v) SnCl2 (Technical grade anhydrous SnCl2, 99.9%, Ajax, Australia) was similarly prepared. A

UV–vis spectral changes

The spectra shown in Fig. 2 indicate that as the reduction of 12-MPA occurs in progressively more concentrated Cl solutions, the peak at 710 nm (ε = 26,000 L mol−1 cm−1) corresponding to SnPMB(4e) markedly decreases in intensity. Coincidentally, the shoulder at 820 nm due to PMB(4e) increases in intensity as the 710 nm band weakens. The observation that this shoulder also contains an isosbestic point which moves with the Cl concentration demonstrates that the total concentration of these

Conclusions

By examining the UV–visible spectra of 12-MPA reduced by SnCl2/N2H6SO4 under a variety of conditions, it was determined that the salt error experienced when SnCl2 was used as the reductant in the presence of Cl was caused by the formation of Sn(IV) chloro-complexes, which inhibited the normally favourable process of forming SnPMB(4e) and forced the reaction to proceed via the much slower, less favourable path of direct reduction to PMB(4e). The amount of product lost, and the concomitant

Acknowledgements

E. Nagul is grateful for the award of an Australian Postgraduate Award. The authors wish to thank Dr. Stephen Best and Dr. Chris Ritchie for their helpful discussions.

References (42)

  • J.P. Launay et al.

    Gradual reduction of molybdosilicates and related compounds

    J. Less-Common Met.

    (1974)
  • T. Rupasinghe et al.

    Pervaporation-flow injection determination of arsenic based on hydride generation and the molybdenum blue reaction

    Anal. Chim. Acta

    (2001)
  • B. Müller et al.

    Spectrophotometric determination of the stability of tin(II) chloride complexes in aqueous solution up to 300 °C

    Geochim. Cosmochim. Acta

    (2001)
  • I.D. McKelvie et al.

    Elimination of the Schlieren effect in the determination of reactive phosphorus in estuarine waters by flow-injection analysis

    Anal. Chim. Acta

    (1997)
  • J.Z. Zhang et al.

    Optimization of performance and minimization of silicate interference in continuous flow phosphate analysis

    Talanta

    (1999)
  • I.D. McKelvie et al.

    Techniques for the quantification and speciation of phosphorus in natural waters

    Anal. Proc. Incl. Anal. Commun.

    (1995)
  • G. Hesse et al.

    Phosphorus (nucleotide phosphorus) determination with a stannous chloride hydrazine sulfate reagent by the Hurst method

    Mikrochim. Acta

    (1968)
  • H. Kobayashi et al.

    Mechanism of salt error on the color development of phosphomolybdenum blue

    Bunseki Kagaku

    (2004)
  • J.D. Burton et al.

    Determination of soluble phosphate, and total phosphorus in sea water and of total phosphorus in marine muds

    Mikrochim. Acta

    (1956)
  • J. Murphy et al.

    A single-solution method for the determination of soluble phosphate in sea water

    J. Mar. Biol. Assoc. U. K.

    (1958)
  • APHA

    Standard Methods for the Examination of Water & Wastewater

    (1999)
  • Cited by (8)

    • A compact hydrodynamic sequential injection system for consecutive on-line determination of phosphate and ammonium

      2019, Microchemical Journal
      Citation Excerpt :

      This reaction involves the formation of colored phosphomolybdenum blue species, from orthophosphate and molybdate in acidic medium and followed by the reduction of the complex with a reductant. Different reductants that have been reported for this reaction include ascorbic acid [25], antimony (Sb) [27], tin (II) chloride [28], hydrazine sulfate [29] and sodium sulfide [30]. Some have drawbacks that make the development of automatic analysis systems more complicated.

    • Determination of phosphorus in natural waters: A historical review

      2016, Analytica Chimica Acta
      Citation Excerpt :

      At this wavelength, absorbance can be measured using simple solid state detectors that utilise red light emitting diodes as the light source [72,73], and this approach has been widely adopted for flow analysis systems intended for field application [74]. However, chloride inhibits the reduction process by complexing with Sn(IV), and a decrease in sensitivity of ca. 15% occurs in a seawater matrix compared with that of deionised water [75]. Despite this limitation, reduction with tin(II) chloride has been favoured for automatic flow analysis of non-saline waters because of the faster reaction kinetics and higher sensitivity.

    • The molybdenum blue reaction for the determination of orthophosphate revisited: Opening the black box

      2015, Analytica Chimica Acta
      Citation Excerpt :

      Whilst a small (approx. 1–4%) decrease in analytical signal is generally expected at high ionic strengths due to the decrease in activity coefficients [50,176], the decrease in signal is usually reported to be between 15 and 20% in marine samples [50,64,85,176] and even higher in some cases [68,97], implying a dependence on method conditions as well. The degree of analytical signal suppression varies with the Cl− concentration in the matrix [68,97] as well as the orthophosphate concentration [97], and method linear range is attenuated with increasing Cl− concentration. As such, the salt error is troublesome not only because of the loss of method sensitivity, but because this loss varies in magnitude between sample matrices of different salinities and P concentrations.

    View all citing articles on Scopus
    View full text