DISSERTATIONES CHIMICAE UNIVERSITATIS TARTUENSIS 118 DISSERTATIONES CHIMICAE UNIVERSITATIS TARTUENSIS 118 IRJA HELM High accuracy gravimetric Winkler method for determination of dissolved oxygen Institute of Chemistry, Faculty of Science and Technology, University of Tartu Dissertation is accepted for the commencement of the Degree of Doctor philosophiae in Chemistry on June 14, 2012 by the Doctoral Committee of the Institute of Chemistry, University of Tartu. Supervisors: Research Fellow Lauri jalukse (PhD) Professor Ivo Leito (PhD) Opponent: Associate professor Jens Enevold Thaulov Andersen (D.Sc.) Technical University of Denmark Commencement: August 31, 2012 at 10:00, Ravila 14a, room 1021 This work has been partially supported by the ETF grant No 7449. This work has been partially supported by Graduate School „Functional materials and technologies” receiving funding from the European Social Fund under project 1.2.0401.09-0079 in University of Tartu, Estonia ISSN 1406–0299 ISBN 978–9949–32–069–1(trükis) ISBN 978–9949–32–070–7(PDF) Autoriõigus Irja Helm, 2012 Tartu Ülikooli Kirjastus www.tyk.ee Tellimus nr 344 TABLE OF CONTENTS LIST OF ORIGINAL PUBLICATIONS ............................................... 7 ABBREVATIONS ................................................................................ 8 1. INTRODUCTION ............................................................................ 9 2. PRINCIPLE OF THE WINKLER METHOD .................................. 12 3. EXPERIMENTAL ............................................................................ 13 3.1. General notes ...................................................................................... 13 3.2. Syringe gravimetric Winkler .............................................................. 15 3.2.1. Measurement model of the syringe gravimetric Winkler ......... 15 3.2.2. Preparing of working solution of KIO .................................... 17 3 3.2.3. Determination of the concentration of the Na S O titrant ....... 17 2 2 3 3.2.4. Sample preparation ................................................................... 18 3.2.5. Titration of the sample with the Na S O titrant ....................... 18 2 2 3 3.2.6. Determination of parasitic oxygen............................................ 19 3.2.7. Determination of iodine volatilization ...................................... 20 3.3. Flask gravimetric Winkler .................................................................. 21 3.3.1. Measurement model of flask gravimetric Winkler .................. 22 3.3.2. Preparing of standard working solutions of KIO ................... 24 3 3.3.3. Determination of the concentration of the Na S O titrant ...... 24 2 2 3 3.3.4. Sampling and sample preparation ........................................... 25 3.3.5. Titration of the sample with the Na S O titrant ...................... 26 2 2 3 3.3.6. Determination of the correction for oxygen introduced from the reagents ..................................................................... 26 3.3.7. Determination of Parasitic Oxygen ......................................... 28 3.3.8. Iodine volatilization ................................................................. 29 3.4. Saturation method for obtaining the reference DO values ................. 32 3.5. Differences between gravimetric Winkler carried out in syringes and in flasks ..................................................................... 34 4. RESULTS AND DISCUSSION ....................................................... 36 4.1. Validation of the methods .................................................................. 36 4.2. Measurement uncertainties ................................................................. 38 4.3. Comparison with the uncertainties of other Winkler methods published in the literature ................................................................... 40 4.4. Comparison of the Gravimetric Winkler method with saturation method for calibration of DO sensors ................................................ 42 CONCLUSIONS ................................................................................... 43 SUMMARY ........................................................................................... 44 SUMMARY IN ESTONIAN ................................................................ 45 5 2 REFERENCES ...................................................................................... 46 ACKNOWLEDGEMENTS ................................................................... 49 APPENDIX 1 ......................................................................................... 50 APPENDIX 2 ......................................................................................... 52 APPENDIX 3 ......................................................................................... 54 APPENDIX 4 ......................................................................................... 57 APPENDIX 5 ......................................................................................... 64 PUBLICATIONS .................................................................................. 69 6 LIST OF ORIGINAL PUBLICATIONS This thesis consists of four articles listed below and a review. The articles are referred in the text by Roman numerals I–IV. The review summarizes and supplements the articles. I. I. Helm, L. Jalukse, I. Leito, Measurement Uncertainty Estimation in Amperometric Sensors: A Tutorial Review. Sensors, 2010, 10, 4430–4455. DOI:10.3390/s100504430 II. L. Jalukse, I. Helm, O. Saks, I. Leito, On the accuracy of micro Winkler titration procedures: a case study, Accredit. Qual. Assur. 2008, 13, 575– 579. DOI: 10.1007/s00769-008-0419-1 III. I. Helm, L. Jalukse, M. Vilbaste, I. Leito, Micro-Winkler titration method for dissolved oxygen concentration measurement. Anal. Chim. Acta, 2009, 648, 167–173. DOI:10.1016/j.aca.2009.06.067 IV. I. Helm, L. Jalukse, I. Leito, A new primary method for determination of dissolved oxygen: gravimetric Winkler method. Analytica Chimica Acta, 2012, 741, 21–31. DOI: 10.1016/j.aca.2012.06.049 Author’s contribution Paper I: Main person responsible for planning and writing the manuscript. Paper II: Performed literature search and wrote large part of the text. Paper III: Main person responsible for planning and writing the manuscript. Performed all the experimental work. Paper IV: Main person responsible for planning and writing the manuscript. Performed all the experimental work. 7 ABBREVATIONS DO Dissolved oxygen FGW Gravimetric Winkler titration method, where sample preparation is performed in flasks GUM The Guide to the Expression of Uncertainty in Measurement ISO International Organization for Standardization PTFE Polytetrafluoroethene SGW Gravimetric Winkler titration method, where sample preparation is performed in syringes SI International System of Units WM Winkler titration method 8 1. INTRODUCTION Dissolved oxygen (DO) content in natural waters is an indispensable quantity whenever background data is collected for investigations of nature from hydrobiological, ecological or environmental protection viewpoint [1]. Suffi- cient concentration of DO is critical for the survival of most aquatic plants and animals [2] as well as in waste water treatment. DO concentration is a key pa- rameter characterizing natural and wastewaters and for assessing the state of environment in general. Besides dissolved CO , DO concentration is an impor- 2 tant parameter shaping our climate. It is increasingly evident that the con- centration of DO in oceans is decreasing [3–6]. Even small changes in DO content can have serious consequences for many marine organisms, because DO concentration influences the cycling of nitrogen and other redox-sensitive ele- ments [3]. Decrease of DO concentration leads to formation of hypoxic regions (or dead zones) in coastal seas, in sediments, or in the open ocean, which are uninhabitable for most marine organisms [3,7]. DO concentration is related to the changes in the ocean circulation and to the uptake of CO (including 2 anthropogenic) by the ocean [8]. All these changes in turn have relation to the climate change. Accurate measurements of DO concentration are very important for studying these processes, understanding their role and predicting climate changes. These processes are spread over the entire vast area of the world's oceans and at the same time are slow and need to be monitored over long periods of time. This invokes serious requirements for the measurement methods used to monitor DO. On one hand, the results obtained at different times need to be comparable to each other. This means that the sensors used for such measurements need to be highly stable and reproducible [9]. The performance of oxygen sensors – amperometric and (especially) optical – has dramatically improved in recent years [10]. On the other hand, measurements made in different locations of the oceans have to be comparable to each other. The latter requirement means that the sen- sors have to be rigorously calibrated so that the results produced with them are traceable to the SI. The sensors need to be calibrated with solutions of accu- rately known oxygen concentration in order to correct for sensor drift, tem- perature, salinity and pressure influences [I,11]. Oxygen is an unstable analyte thus significantly complicating sensor calibration. It has been established that if every care is taken to achieve as accurate as possible results then the accuracy of DO measurements by amperometric sen- sors is limited by calibration [11] and specifically by the accuracy of the refer- ence DO concentration(s) that can be obtained [I]. This is similar with optical sensors: their lower intrinsic uncertainty may make the relative contribution of calibration reference values even larger [10]. The issues with sensors, among them issues with calibration, have caused a negative perception about the data using sensors in the oceanography commu- 9 3 nity and because of this the recent issue of the World Ocean Atlas [12] was compiled with taking into account only DO concentrations obtained with chemical titration methods (first of all the Winkler titration method, WM) and rejecting all sensor-based data. Similar decision was taken in a recent study of DO decline rates in coastal ocean [6]. It is nevertheless clear that there is need for large amounts of data, so that the slow and clumsy titration method cannot satisfy this need. It is necessary to be able to collect data automatically and in large amounts. It is thus expected that eventually sensors will be “back in busi- ness”. In order to achieve this the accuracy of their calibration needs to be improved. There are two ways to prepare DO calibration solutions with known concen- trations: (1) saturating water with air at fixed temperature and air pressure and using the known saturation concentrations [13–15] and (2) preparing a DO solution and using some primary measurement method for measuring DO con- centration. The premier method for the second way is the WM [16] which was first described by Winkler [17] more than hundred years ago. Nowadays the use of WM as the standardizing method is even more important than measurements in the real samples [1]. Also gasometry is an old method for DO determinations, but it is a partly physical method requiring quite specific and complex experi- mental setup and is therefore not routinely used nowadays. DO measurement practitioners currently almost exclusively use the satu- ration method for calibration of DO measurement instruments. This method gives quite accurate results when all assumptions made are correct. DO values obtained with the saturation method are also used in this work for comparison with the WM values. Nevertheless, the saturation method uses ambient air – a highly changing medium – as its reference, thereby relying on the assumption that the oxygen content of the Earth’s atmosphere is constant, which is not entirely true [4]. The oxygen content of air depends on air humidity and CO 2 content, which both can change over a wide range of values. Also, this method needs careful accounting for air pressure, humidity and water temperature. It is customary to use published values of DO concentrations in air-saturated water at different temperatures. At the same time, different published values are in disagreement by up to 0.11 mg dm–3 at 20 oC and even up to 0.19 mg dm–3 at 40 oC [15]. Thus the saturation method has many factors that influence the results and it is difficult to realize it in a highly accurate way. An independent primary method, such as WM, would be free from these shortcomings. The Winkler method is known for a long time, it has been extensively studied and numerous modifications have been proposed [16,18–23]. There have, however, been very few studies using WM that report combined uncertainties taking into account both random and systematic factors influencing the measurement [II]. Usually repeatability and/or reproducibility data are presented that do not enable complete characterization of the accuracy of the methods and tend to leave too optimistic impression of the methods. Very illuminating in this respect are the results of an interlaboratory comparison 10