A critical review of analytical methods for the analysis of polar compounds in water

(1)

Literature Thesis

Master Chemistry, track Analytical Sciences

A critical review of analytical methods for the analysis of polar

compounds in water

by

Sanne Brekelmans

10505172

15 April 2021

12 EC January - April

Daily supervisor: dr. E.N. Pieke (eelco.pieke@hetwaterlaboratorium.nl) Examinator: prof. dr. M. Lamoree (marja.lamoree@vu.nl)

(2)

(3)

Acknowledgements

I would like to express my gratitude to Ruud Steen and Eelco Pieke from Het Waterlaboratorium for providing me the opportunity to do this project on the topic of polar compounds. Thank you, Eelco, for guiding me and proofreading my work. Having been given a topic that I genuinely liked has shown me how good my academic writing can become. I would also like to thank my roommates for providing me work at home quietly as these difficult pandemic times left me working from home and in my student room for many hours. Thank you for constantly supporting me mentally when I needed advice or needed a drink after a long day of writing. Thank you, my friend Bob, who has read my work and provided me with feedback and spellchecks. Finally, I would like to thank Marja Lamoree and Rob Haselberg for being there when I coped with difficulties or had a question. Besides, thank you for being my examiners. I hope you enjoy reading my thesis.

(4)

Abstract

The presence of polar compounds in environmental waters might threaten the water quality as they are persistent, mobile and potentially toxic. They can cross natural riverbanks or anthropogenic barriers, spread along the water cycle, and reach surface or even drinking water, posing a potential risk to human health. Purification methods such as sorption processes may not cover polar compounds as their sorption capacities are not high enough, resulting in incomplete removal in the water treatment plant. Therefore, together with protecting the water sources, it is necessary to monitor polar compounds in water. However, limited research and a lack of analytical methods challenge the desire for the commercial availability and routine analysis of polar compounds. In this literature review, multiple analytical methods for analysing polar compounds in various water matrices are given to obtain an overview of commercial purposes' most suitable methods. Different sample pre-treatment, separation and detection methods are explained and discussed with the possibility of implementing methods in commercial laboratories, outside research facilities. Evaporative techniques and mixed-bed solid-phase sorbents were found to be most successful in capturing polar compounds. For separation, HILIC-MS showed the most promising results and highest potential for commercial laboratories as it is complementary to RPLC, and instruments are available. SFC proved to be complementary to 2D chromatography and showed remarkable results in polar analyte coverage although it requires special equipment and is not favourable for commercial laboratories. MMLC provides unique selectivity that cannot be reproduced by single-mode chromatography, but many stationary phases are not commercially available. CE or IC might be amenable for ionic compounds, although specific applications and further research are required. Although many analytical methods are being suggested and discussed, which can cover many (very) polar compounds, not a single method fits all requirements and implementing methods commercially in the short-term remains a laborious task.

(5)

List of abbreviations

EDTA Ethylenediaminetetraacetic acid PMT Persistent, mobile, toxic

PMOC Persistent and mobile organic chemicals PFAS Per- and polyfluoroalkyl substances EU European Union

WFD Water framework directive POP Persistent organic pollutants

REACH Registration, Evaluation, Authorisation and Restriction of Chemicals ECHA European Chemicals Agency

SVHC Substances of very high concern GC Gas chromatography

LC Liquid chromatography

RPLC Reversed phase liquid chromatography

HILIC Hydrophilic interaction liquid chromatography MS Mass spectrometry

MMLC Mixed-mode liquid chromatography HRMS High-resolution mass spectrometry SFC Supercritical fluid chromatography IC Ion chromatography

CE Capillary electrophoresis LLE Liquid-liquid extraction SPE Solid-phase extraction EC Evaporation concentration

QuEchERS Quick, easy, cheap, effective, rugged, and safe DLLME Dispersive liquid-liquid microextraction EME Electro membrane concentration

CNT Carbon nanotubes

HLB Hydrophilic-lipophilic balance PAH Polyaromatic hydrocarbons mlSPE multi-layered SPE

PCB Polychlorinated biphenyl WAX Weak anion exchanger SAX Strong anion exchanger WCX Weak cation exchanger SCX Strong cation exchanger MAX Mixed-mode anion exchanger MCX Mixed-mode cation exchanger TFA Trifluoracetic acid

TP Transformation product HMSA Halomethane sulfonic acid ACN Acetonitrile

VEC Vacuum-assisted evaporation ESI Electrospray ionisation

(8)

TRIS Tromethamine HF Hollow fibre

LPME liquid-phase microextraction SLM Supported liquid membrane UV Ultra-violet

LOQ Limit of quantification MWCNT Multiwalled carbon nanotubes TSM Thifensulfuron-methyl GCB Graphitized carbon black PGC Porous graphitic carbon IL Ionic liquids

HESI Heated electrospray ionisation QQQ Triple quadrupole

PFAA Perfluoroalkyl acid TOF Time of flight

NMR Nuclear magnetic resonance

NPLC Normal phase liquid chromatography MRM Multiple reaction monitoring

ICM Iodinated contrast media DoE Design of experiments CO Carbon dioxide NP Normal phase

HPLC High-performance liquid chromatography PCA Principal component analysis

MM Mixed mode

IEC Ion exchange chromatography CX Cation exchange

AX Anion exchange LOD Limit of detection

AMPSA 2-acrylamino-2-methylpropane sulfonate OPE Organophosphate ester

TMP Trimethyl phosphate TEP Triethyl phosphate EOF Electroosmotic flow HAA Halogenated acetic acid PFA Perfluoro alkoxy BGE Background electrolyte

AMPA α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid SRM Selected reaction monitoring

MPPA 3-hydroxy(methyl)phosphinoyl propionic acid 2D Two-dimensional

LR Low-resolution HR High-resolution

(9)

1. Introduction

1.1. Relevance of polar compounds

Increasing usage of pharmaceuticals, surfactants, cytostatics, consumer products, pesticides, and other chemicals by industry and consumers has led to increased pollution of rivers and raw waters as chemicals can enter the aquatic environment by various sources1_{. Compounds can be} discharged directly through effluent outfalls from factories or treatment plants, or indirectly by contaminants entering the water from soils or groundwater systems, for example residues of human agricultural practices. A schematic overview of possible emission sources leading to water pollution is depicted in Figure 1.

Figure 1. Possible emission sources contributing to the pollution of rivers and surface waters (adapted from ref [2]).

The consequence of chemicals entering the aquatic environment is the increased pressure on (waste) water treatment plants. Especially the more polar compounds, for example glyphosates, sulfonates, or complexing agents such as ethylenediaminetetraacetic acid (EDTA), are of serious concern as they travel longer through the water cycle and are challenging to remove in treatment plants as the only process leading to decreased concentration is dilution3_{. Nonpolar} compounds can effectively be removed through treatment processes such as coagulation, membrane filtration, oxidation, ozonation, adsorption with activated carbon or a combination of them4,5_{. However, sorption processes cannot remove polar compounds as their sorbent} potential to soils and sediment is not high enough, causing mobility in water and ineffective removal6_{. This group of polar organic compounds is persistent in the environment and mobile} in the aquatic environment and are also known as PM or PMOC compounds7_{. If they also} exhibit toxic effects, they are denoted as PMT compounds8_{. Examples are per-}

(10)

and polyfluoroalkyl substances (PFAS), which are industrial chemicals and contaminate the surface- and drinking water9_{. PMT and PMOC substances are since 2015 of high interest and} can be described as chemicals having a high propensity to be transported easily in the aqueous environment6_{. They pass treatment plants' natural barriers such as river banks, can reach surface} water and possess a potential risk to drinking water6,3,10,11_{. The low degradation potential of} these polar compounds, which causes persistency and long half-live times in water, together with the low sorption potential to soil and sediment, contribute to the challenging removal of polar compounds in treatment plants, causing exposure to humans or other organisms, leading to a potential risk for human health3_.

1.2. Challenges in studying polar compounds 1.2.1. Definitions

Investigating (very) polar compounds in aquatic environments is challenging as there is a lack of proper definitions, regulation, monitoring, and analytical methods to detect these substances12,13_{. To establish if a chemical is polar, it is important to consider the various} physical and chemical properties of a compound. Each chemical has various characteristics based on its molecular structure, which might affect its polarity. Examples of these properties are mobility, persistency, dipole moments, bioaccumulation, degradability, toxicity and octanol/water distribution coefficients14_{. The octanol/water distribution coefficient is based on} the distribution of a compound between octanol and water, meaning the lower its value, the higher the mobility of a compound3_{. However, the definition of polar compounds remains} slightly abstract. To create a more robust definition, this review classifies a compound polar if the chemical has an octanol/water distribution coefficient or log Dow value equal to or below two, which is explained in section 2.2. Defining the scope of polar compounds

1.2.2. Regulation and monitoring

Regulation and monitoring are other challenging topics in investigating polar compounds. Various organisations in the world, including the European Union, regulate the protection of water for the production of healthy drinking water. The EU Water Framework Directive (WFD) registers hazardous substances that have to be banned or require reduced emission based on experimental data to protect (fresh) water sources15_{. Another essential document is the} Stockholm Convention, an alliance with 152 countries which protect human health and the environment from persistent organic pollutants (POPs)16_{. REACH, a regulation of the European}

(11)

Union which falls under the European Chemicals Agency (ECHA), stands for Registration, Evaluation, Authorisation and Restriction of Chemicals, and provides also information on chemicals to protect citizens' health and the environment17_{. According to REACH, a list of} substances of very high concern (SVHCs) is available for chemicals that are identified as hazardous18_{. The PMT, PMOC, POP and SVHC terms all describe the problematic intrinsic} fate and hazard properties of persistent chemicals in combination with mobility8_{. However, all} these lists are based on experimental data, which indirectly depend on used analytical techniques and methods. As the analytical methods on polar compounds are not widely developed yet, and the focus is more on nonpolar compounds, it is not surprising that most of the candidates on these lists are nonpolar or slightly polar substances3_{. An overview of these} lists related to log D values is depicted in Figure 2.

Figure 2. Box and whisker plots of calculated log D values at pH 7.4 for contaminants regulated by the Stockholm Convention, candidates of substances of very high concern (SVHCs) according to REACH, the

list of priority substances according to the water framework directive (WFD) and the watch list of the WFD. In red, the polarity range of polar compounds (or gap compounds) is shown (adapted from ref [3]).

From Figure 2, it is shown that regulation of high-risk candidates is based on nonpolar compounds with log D values higher than one. However, as the demand for tightened regulation is rising, monitoring data on polar compounds is desired13_{. Experimental data originates from} routine-based laboratories that extensively monitor emerging substances which is important for the water quality. However, new analytical methods require excessive sample pre-treatment as the preconcentration of highly polar analytes is challenging13_{. Also, new methods need to be} fully developed, require high expertise, and can be expensive, making it unattractive for

(12)

commercial labs to work with as they do not fit in their routine-based methods. Thus, monitoring polar compounds is an important aspect for establishing risk assessments and monitoring strategies which then relates back to the WFD or REACH regulators for establishing new contaminant candidate lists. Therefore, implementation of methods available for commercial labs, remains a challenging task but is highly desired for shifting the regulation of high-risk candidates to polar compounds.

1.2.3 Analytical methods

Another challenge is the analysis of polar compounds. Gas chromatography (GC) and liquid chromatography (LC) based methods are techniques that can analyse analytes in complex matrices. However, (very) polar or permanently charged compounds are required to be volatile for GC analysis, and are minimally retained within a reversed phase LC (RPLC) method. RPLC methods are based on nonpolar interactions, which fails to retain polar or ionic compounds3_. Chromatographic methods that are amenable for the analysis of polar compounds include hydrophilic interaction LC (HILIC) in combination with mass spectrometry (MS) detection10,19_, mixed-mode LC – high-resolution MS (MMLC-HRMS)20_{, supercritical fluid chromatography} (SFC)19_{, ion chromatography (IC)}21_{or capillary electrophoresis (CE)}22_{. In Chapter 2.}

Analysing polar compounds

, analytical methods used for polar compounds are extensively explained and discussed.

1.3. Aim of the study

This literature review aims to suggest multiple analytical methods for analysing polar compounds in water to obtain an overview of commercial purposes' most suitable methods. An objective thoroughness and quality analysis of the primary methodologies used for polar compounds is necessary to obtain more knowledge on existing and recently developed research techniques. Also, insight of possible methods available to implement in commercial labs are being investigated and discussed based on their advantages and limitations. The centre of attention are polar compounds; thus, nonpolar compounds, together with detailed information on existing or recently developed treatment processes, will not be covered in this review.

1.4. Outline literature review

(13)

Consequently, most of these compounds remain unknown due to the lack of methods available to detect and quantify them. Many polar compounds are of serious concern due to their ability to spread along partially closed water cycles, reach surface waters, and eventually drinking water. In this study, an overview of existing literature for the analysis of polar compounds in water is reported to obtain an overview of commercial purposes' most suitable methods. This is achieved by explaining the definition and scope of polar compounds (section 2.1. Theoretical background and 2.2. Defining the scope of polar compounds) and investigate currently used analytical techniques (section 2.3. Analytical methods). Various sample preparation methods (section 2.3.1. Sample preparation methods) will be examined and compared, together with alternative extraction methods (section 2.3.2. Alternative sample preparation methods). Currently used and alternative chromatographic or electro migrative separation methods will be discussed and criticised based on their commercial applicability (section 2.3.3. Chromatographic separation and 2.3.4. Alternative separation techniques), and detection with MS is being explored (section 2.4. Mass spectrometry detection). Knowledge is acquired by exploring literature not older than twenty years, focusing on the most recently published literature of the last decade. Keywords such as 'polar compounds', 'water', 'PMT compounds' combined with a sample pre-treatment method or chromatographic technique are used for literature search. By comparing and summarising all the described methods, together with their advantages and disadvantages, conclusions are drawn (chapter 3.

Conclusion and Future

remarks

. By reflecting on a few future remarks (chapter 3.

Conclusion and Future

remarks

, this review will be closed.

(14)

2. Analysing polar compounds

2.1. Theoretical background

It is important to gain more knowledge on the specific definition of polarity, and to assess how polar a compound is in order to analyse polar compounds. Polar bonds can be related to the electronegativity of an element which is the attraction of atoms on their electrons23,24_{. Not every} atom attracts electrons with the same force meaning the higher the pull of electrons, the higher its electronegativity. Consequently, the atoms will share their electrons unequally as the electrons will be closer to the atom with the highest electronegativity. If both atoms' electronegativity is identical, the difference will be zero, and a nonpolar bond occurs. If the electronegativity difference is significant, an atom can take an electron from the other atom. These bonds are also known as ionic bonds and are the most completely polar bonds25_.

Nonpolar, polar and ionic bonds are insufficient descriptors to assess how polar a compound is. To obtain a less abstract understanding of this, the physical-chemical property octanol/water partitioning coefficient Kow or P, or the distribution coefficient Dow is used to describe lipophilicity. The Kow is a measure of the differential solubility of a compound in two immiscible solvents, mostly octanol/water, assuming equal quantities. The ratio between those concentrations makes it a unitless parameter and is usually expressed in logarithmic terms. Log Kow values are an efficient tool to approximate the polarity of compounds as a lower value corresponds with a higher mobility and hydrophilicity3_{. However, this parameter can only be} used for neutral compounds or where compounds exist in a single form. Therefore, for compounds with multiple charge-forms, the pH-dependant logarithmic octanol/water distribution coefficient log Dow or log D is used. Log D is described as the sum of the concentrations of all forms of the compound in the nonaquatic/octanol phase divided by the sum of the concentrations of all forms in the aqueous or buffer phase. In most literature, log D values are calculated at a pH of 7.4, as at a pH of 7.4, most polar compounds will have a charge number of 0.5 or higher22_{. For consistency and comparing compounds, this study also focusses} on log D values at pH 7.4. A very recently published paper from Knoll et al. (2020) investigated the correlation between log P and their corresponding log D values, as seen in Figure 3.

(15)

Figure 3. Scatterplot of log P and corresponding log D values at pH 7.4 for a list of compounds. Colour codes indicate different charge numbers. Numbers are for selected compounds; (1) gadobutrol, (2) neomycin, (3) tobramycin, (4) gadopentetic acid, (5) cephalosporin C, (6) gentamicin, (7) glyphosate, (8) cephalozin, (9) 3-methylphosphinic acid, (10) itaconate, and (11) 3,4-dihydroxyphenylacetic acid (adapted

from ref [13]).

From Figure 3, it is stated the lower the log P value, the lower the log D value and the higher its polarity22_{. However, it must be said that despite these calculations, challenges exist in} establishing models for the more complex compounds with multiple functional groups instead of having one single functional group, as the behaviour of these compounds is different26_{. Yet,} as the log D sufficiently describe the polarity of a charged compound, this review uses log D values to assess a compounds' polarity.

2.2. Defining the scope of polar compounds

Based on log Kow and corresponding log D values at pH 7.4, it is possible to identify compounds as 'nonpolar', 'polar' and 'very polar'. From literature, many papers classify polarity based on log D values10,9,19-22_{. Hence, in this review only the log D values are used to assess polarity. To} compare analytical methods from literature and divide compounds in multiple groups, this study classifies compounds as follow; nonpolar is defined as a log D value larger than 2 (log D > +2), polar compounds consist of log D values between -2.5 and 2 (log D -2.5 to log D +2), and very polar compounds will cover the range of log D values below -2.5 (log D < -2.5). Very polar compounds with log D values below -2.5 are restricted to other separation techniques relative to RPLC while both nonpolar and polar compounds can be separated or monitored in one run by using one RPLC column19_.

(16)

To get an overview of very polar compounds and their negative log D values at pH 7.4, Table 1 shows 30 out of 237 compounds with the lowest log D values reported by Knoll et al. (2020)22_. However, it must be noted that this list is only a small subset of all possible very polar compounds and is shown to give a little insight into potential emerging compounds and their function. From this small subset, many compounds are well-known in the water quality and treatment plants as they appear in raw water sources and are only partially removed by treatment plants20,27_{. A few examples are the herbicides diquat and glyphosate, the antidiabetic compound} metformin or the food additive fumarate, which all originate from excessive agriculture, human medicines usage or industry20,27_{. Challenges in removal of these polar compounds indicates the} desire for proper analysis and detection, which is discussed in upcoming chapters 2.3. Analytical methods and 2.4. Mass spectrometry detection

(17)

Table 1. List of 30 very polar compounds with log D values below -2.5 at pH 7.4. Information adapted from ref [22] and log D values were calculated with Chemicalize provided by ChemAxon.

Compound log D Compound class

1_Gd-DTPA _-17.90 _{Contrast agent}

2_Gd-BOPTA _-16.03 _{Contrast agent}

Neomycin -15.47 Aminoglycoside antibiotics

Amikacin -15.10 Aminoglycoside antibiotics

Citrate -9.47 Organic acid

Ristocetin A -9.16 Glycopeptide antibiotic

Tartate -7.89 Organic acid

Cephalosporin C -7.64 Beta-lactam antibiotic

Glyphosate -7.26 Herbicide

Diquat -7.03 Herbicide

Fumarate -6.51 Food additive

Maltotriose -6.47 Carbohydrate

3_GMP _-6.36 _Nucleotide

Itaconate -6.25 Intermediate compound

3-methylphosphinicopropionic acid -6.23 Pesticide metabolite

4_AMP _-5.75 _Nucleotide

Metformin -5.62 Antidiabetic

Threonate -5.55 Sugar

Cephazolin -5.01 Beta-lactam antibiotic

Oxytetracycline -4.92 Tetracycline antibiotic

Vancomycin -4.85 Glycopeptide antibiotic

Ascorbate -4.81 Vitamin

Maltose -4.70 Carbohydrate

Choline -4.66 Essential nutrient

Ethyl glucuronide -4.54 Human metabolite

Glycolic acid -4.41 Alpha-hydroxy acid

5_cCMP _-4.33 _{Second messenger}

Emtricitabine carboxylate -4.00 Drug metabolite

Sorbitol -3.73 Nutritive sweetener

Trifluoroacetic acid -2.62 Disinfection byproduct

1_{Gd-DTPA - Gadopentetic acid;}2_{Gd-BOPTA - Gadobenic acid;}3_{GMP - guanosine monophosphate;}4_AMP

(18)

2.3. Analytical methods

As mentioned in 1.2. Challenges in studying polar compounds, many challenges exist in analysing polar compounds relative to nonpolar compounds in sample pre-treatment and analysis. Various sample pre-treatment methods will be discussed in upcoming section 2.3.1. Sample preparation methods, and analytical methods currently being used for analysis are extensively explained in section 2.3.3. Chromatographic separation. Additionally, the applicability of implementing methods on a commercial scale, together with possible drawbacks or benefits of techniques, will be considered.

2.3.1. Sample preparation methods

Sample pre-treatment is often needed to preconcentrate an analyte, increase the analytes’ recovery, and remove matrix compounds28_{. The separation of an analyte from a matrix is based} on physical-chemical differences between the analyte and the matrix, such as mobility or polarity. For water matrices, this might be complex as the polarity between analytes and matrices can be similar. Currently, the three most used techniques to enrich an analyte from an aqueous matrix are liquid-liquid extraction (LLE), solid-phase extraction (SPE) or evaporative concentration (EC)22_{. LLE utilises the analyte's distribution between an aqueous matrix and a} water-immiscible solvent, SPE relies on adsorption of the analyte to solid material, and EC evaporates the aqueous matrix14_{. In}_{Figure 4}_{, an overview is shown with possible sample} preparation methods and their log D ranges. QuEchERS, an abbreviation of quick, easy, cheap, effective, rugged, and safe, is an extraction method for nonpolar pesticides. Although it is a green chemistry method and has many food analysis applications, extraction of polar compounds in water has not been reported yet, and QuEchERS analysis will not be discussed in this review29_.

Figure 4. Overview of sample preparation methods with their range of log P/log D values which can be covered. QuEchERS, quick, easy, cheap, effective, rugged, and safe; DLLME, dispersive liquid-liquid microextraction; EC, evaporative concentration; EME, electro membrane concentration; CNT, carbon

(19)

2.3.1.1. Liquid-liquid extraction (LLE)

The basic principle of LLE is the transfer of target analytes from a liquid matrix into another immiscible liquid matrix, according to solubility differences28_{. Analytes from aqueous} environmental samples are extracted into a nonpolar or less polar organic solvent. It is one of the oldest and standard extraction methods and is used in many commercial laboratories30_. However, LLE requires large volumes of (toxic) organic solvent, is time-consuming, has a lot of waste, and is not environmentally friendly30_{. Besides, automation is complicated, and LLE} cannot extract polar compounds. Therefore, LLE is of little interest in this review. A more rapid, low cost and more straightforward method is dispersive liquid-liquid microextraction (DLLME). It is a LLE mode that involves the dispersion of fine droplets of extraction solvent in an aqueous sample31_{. A syringe rapidly injects a mixture of extraction solvent and disperser} solvent into the aqueous sample, whereby a cloudy solution is formed30_{. After centrifugation,} the extracting solvent's fine particles containing the target analytes are separated from the aqueous phase32_{. Because of the microemulsion formed, the contact surface is enhanced, and} equilibrium is formed quickly33_{. A DLLME workflow is depicted in}_{Figure 5}_.

Figure 5. A schematic workflow of dispersive liquid-liquid microextraction (adapted from ref [33]).

Many papers describe the successful use of DLLME for pesticides or pharmaceuticals in water samples32,34,31_{. However, the focus is more on nonpolar compounds such as polyaromatic} hydrocarbons (PAHs) and polychlorinated biphenyls (PCBs). Also, Figure 4 shows that DLLME is more beneficial for nonpolar analytes28_{. Proper analysis with (D)LL(M)E is not suitable for} polar compounds as they will stay in the aqueous solution, which is unavailable as there are no polar organic solvents which are not miscible with water. Besides, research on the use of (D)LL(M)E on very polar compounds could not be found and therefore, (D)LL(M)E will not be discussed further.

(20)

2.3.1.2. Solid-phase extraction (SPE)

Offline or online SPE allows isolation and enriching of analytes from environmental samples such as water. The same basic principle of chromatography applies to SPE. Analytes of interest are immobilized on a bed consisting of polymeric or silica-based stationary phase and are retained and separated from the matrix35_{. After conditioning of the stationary phase, samples} are loaded on the cartridge. Unwanted material is rinsed, and the analytes of interest are eluted and collected. The whole SPE process is conducted under vacuum conditions. A schematic overview of SPE is depicted in Figure 6. The analyte is being eluted using a solvent from the solid-phase, and enrichment, solvent exchange, and sample clean-up are achieved14_{. Thus,} interactions with analytes and sorbent material result in retention.

Figure 6. A schematic overview of the four basic steps in solid-phase extraction (adapted from ref[35_]).

Depending on the polarity of compounds, a wide range of SPE sorbents are available that can obtain high recoveries and enrichment factors of analytes before separation22_{. The Oasis} hydrophilic and lipophilic balanced (HLB) sorbent is the most common sorbent for environmental samples14,36,37_{. Other SPE sorbents are Strata, Bond-Elute and Isoelute ENV+}38_. HLB sorbents are stable for the whole pH range and have hydrophilic and lipophilic properties suitable for many compounds. They are commercially used, have a high sample capacity, and help remove interferents in complex matrices. However, from Figure 4, it is shown that HLB and C18 sorbents can only be applied to a small range of log D values and limitations are observed for polar compounds. A paper by Köke et al. (2018) used HLB-SPE cartridges to analyse 26 very polar compounds with log D values below zero and concluded that only 30% of the compounds could be retained with HLB39_{. Comparison with different preconcentration,} e.g., multi-layered SPE (mlSPE) or evaporation methods, showed that HLB sorbents could not

(21)

enrich analytes determined with mlSPE or evaporation. Retention of analytes with HILIC showed decreased capability of HLB compared with mlSPE and evaporation methods.

As many polar compounds are permanently charged or can be ionised, ion exchange functionalised SPE materials might be an excellent alternative to C18 or HLB sorbents14_. Examples are weak or strong anion exchangers (WAX/SAX), weak or strong cation exchangers (WCX/SCX) or mix-mode anion/cation exchangers (MAX/MCX). Scheurer et al. (2012) effectively used a WCX sorbent before HILIC-MS for the detection of metformin (log D -5.62) and its metabolite guanylurea (log D -3.82)22,27_{. Björnsdotter et al. (2019) detected ultra-short} PFAS, e.g. trifluoracetic acid (TFA, log D -2.62)22_{, using a WAX sorbent prior to SFC-MS/MS} analysis40_{. Although many papers report methods for very polar compounds or classes, a} significant drawback of ion exchange materials is that it is limited to ionisable compounds and only one charge state at a time, which makes it ineffective for screening and multi-targeted methods. However, target and suspect screening is still possible, and ion exchangers might remain a good option for sample pre-treatment.

Another alternative to SPE sorbents is the mixed-bed cartridge, where multiple sorbents are present in one cartridge. Combinations of WAX, WCX, ENV+, and HLB sorbents can be used to analyse a broader range of analytes. Organic micropollutants41_{, industrial products}12_{, and} polar disinfect by-products42_{, all with log D values below -2.5, were detected using mixed-bed} cartridges. Kern et al. (2009) used a combination of WAX, WCX, ENV+, and HLB sorbents to achieve sufficient enrichment of 19 TPs, including the pesticide metabolite atrazine-desethyl41_. Schultze et al. (2019) and Zahn et al. (2019) combined both WAX with WCX sorbents for the determination of industrial chemicals and halomethane sulfonic acids (HMSAs), including methyl sulphate (log D -2.84) and chloromethane-sulfonic acid in water matrices, respectively12,42_{. The extensive review on SPE variants by Zahn et al. (2019) concluded that} mix-bed sorbents are the most promising sample preparation methods for compounds with a log D value equal to or below zero14_{. However, by using a combination of anion and cation} exchangers, salt formation might occur. In addition, many mix-bed sorbents are not commercially available and, therefore, limited for routine laboratories that analyse high numbers of samples. But mixed-bed cartridges are continuously being improved, and it is expected that these may be viable in future prospect.

(22)

2.3.1.3. Evaporation concentration (EC)

EC extraction methods cover the most comprehensive range (from log D -8 to 8) to enrich very polar compounds based on log D values, together with electro membrane concentration (EME) and carbon nanotubes (CNT) methods. The basic principle for evaporation sample preparation methods is that the aqueous matrix is evaporated, the solvent is exchanged, and analyte enrichment occurs14_{. The use of evaporation for the enrichment of polar analytes is possible for} non-volatile compounds, which are often very mobile. Three variants of EC exist where; 1. the sample is frozen as solid, and the matrix is sublimated at low pressure (lyophilisation), 2. the matrix is evaporated at low pressure and elevated temperatures (vacuum-assisted) or 3. the sample is diluted with acetonitrile (ACN), and evaporation of the resulting azeotrope is achieved with high temperatures (azeotropic evaporation)14_.

An example of vacuum-assisted evaporation (VEC) showing it effectiveness compared to mlSPE and HLB, is from Köke et al. (2018). They investigated several pre-treatment methods for 26 polar compounds with log D values below zero in groundwater samples, analysed with HILIC-HRMS39_{. It was shown that the VEC method showed the highest recoveries relative to} mlSPE or HLB methods together with the detection of analytes that could not be detected by mlSPE or HLB. Recoveries are depicted in Figure 7. From Figure 7, it is shown that for the HLB material, 8 of the 26 (30%) compounds could be identified, and for evaporation (19/26, 73%) and mlSPE method (17/26, 65%), percentages were higher. Mean recoveries increased in the order or HLB, mlSPE and evaporation. In addition, with VEC, solvent exchange was possible, which is necessary for HILIC analysis. However, high electrospray ionisation (ESI) matrix effects and lower enrichment factors were observed, which is undesirable prior to analysis.

(23)

Figure 7. Recoveries from various analytes in tap water for the mlSPE, evaporation, and HLB methods, analysed with HILIC-MS/MS. Mean recoveries of all analytes amenable for each enrichment method are given by dashed lines in their respective colours. The optimal range of recoveries (80 to 120%) is shaded in

light green (adapted from ref [39]).

Another example of VEC showing its broad applicability range in environmental waters is from Mechelke et al. (2019). They used VEC to study 590 organic compounds with log D values between -14 to 8 in river and wastewater samples43_{. The analysis was performed with a polar} RPLC column coupled to high-resolution tandem mass spectrometry (HRMS/MS) via an ESI interface. Enrichment of polar compounds was achieved, and some very polar compounds could be detected with VEC, which are lost with SPE, e.g., tromethamine (TRIS, log D < -4.6). The VEC method showed other advantages over SPE, such as accurate and precise quantification, low cost, and low sample volumes. However, the VEC-enriched samples showed more signal suppression which is the disadvantage of EC as all analytes are being enriched, also matrix components.

Schulze et al. (2019) developed various analytical screening methods for 57 PMOCs with a log D varying from -5.6 to 3.4 in surface and drinking water samples12_{. Eight different sample} preparation techniques were investigated, including seven methods based on SPE and one evaporation method where the sample was evaporated to dryness. It was concluded that the evaporation method prior to HILIC analysis was one of the enrichment methods capturing most PMOCs. Montes et al. (2017) used freeze-drying as evaporation technique onto various real

(24)

water matrices such as surface water and drinking water20_{. After reconstitution of the sample,} the sample was filtered and evaporated to dryness. It turned out that 21 PMOCs with log D values varying between -6 to 4 such as metformin (log D -5.62)22_{were identified. One} advantage of this method is that discrimination of PMOCs is being avoided, which could be produced with SPE. However, this method focused more on identifying compounds and not on quantification, which limited the investigation of possible matrix effects.

2.3.2. Alternative sample preparation methods

Alternative options exist in the analysis of polar compounds. The discussed sample preparation methods in 2.3.1. Sample preparation methods, such as EC or mixed-bed SPE, cover a broad range of (very) polar compounds. As these methods are great options for polar compounds but not a single method can cover all those compounds. Therefore, it is important to get an overview of other possible pre-treatment methods. Although the commercial availability is necessary and probably is not being solved short-term, a few other techniques deserve attention as research of polar compounds in aqueous matrices is continuously developing. In this chapter, sample preparation methods that are not frequently used but are of interest for future developments are explained and discussed.

2.3.2.1. Electro membrane extraction (EME)

Another extraction technique which is gaining popularity since 2006 is EME. EME is a hollow fibre liquid-phase microextraction (HF-LPME) method for the extraction of charged compounds44_{. It is an electric driven sample preparation technique that uses an electric field to} accelerate the mass transfer of ionic analytes during extraction22_{. EME is based on applying an} electrical field between a sample and an acceptor compartment, separated by a polymeric supported liquid membrane (SLM) impregnated by an organic solvent33_{. By applying a voltage,} ionic compounds migrate under the SLM's electrical field and are concentrated into the acceptor compartment. An EME set-up is depicted in Figure 8. Advantages of EME are that it has effective sample clean-up for complex matrices, good selectivity and analyte enrichment33_. Besides, it is a one-step sample preparation method, and because interferents such as proteins, salts or phospholipids do not cross the SLM, low matrix effects are observed44,45_.

(25)

Figure 8. Schematic overview of the set-up for electro membrane extraction (adapted from ref [44]).

An example of EME being used on polar compounds can be seen at Koruni et al. (2014). They showed the analysis of five basic and three acidic drug compounds, including enalapril (log D -1.06)22_{in a single run with a broad polarity range using EME-capillary electrophoresis} (CE)/Ultra-violet (UV)46_{. Extraction conditions were investigated in standard aqueous solutions} and applied to urine samples. Two different SLMs with different conditions were tested, and low limit of quantification (LOQ) values (15-45 ng/L) were observed. The method showed its applicability on urine samples but not on water matrices. However, they explained its potential to extract compounds of an extensive polarity range. The EME ability to be used in hyphenated systems can be seen in the work of Oedit et al. (2016) as they tested the compatibility of EME hyphenated with LC, GC and CE47_{. They concluded that EME is an excellent sample} pre-treatment technique with high prospects for hyphenation to analytical methods with high matrix tolerance. However, for hydrophilic compounds, extractions are not that exhaustive. Despite this, further research in EME could make the extraction of polar compounds in water feasible.

Although EME has not been used on water samples, it is effective on urine samples which was observed in the work of Fernandez et al. (2017). They reported for the first time highly polar drug substances in urine samples48_{. Low matrix effects were observed as most matrix} compounds were unable to cross the SLM. EME its potential is not studied on polar compounds in water, but Zahn et al. (2020) expect increased potential for water analysis14_{. However,} arguments to support this statement are absent, and it remains debatable if this technique is superior to mixed-bed SPE or evaporation extraction. Even though EME is environmentally friendly, has the potential for automatization, and has a straightforward setup which makes it attractive for commercial routine-laboratories, it is not expected that EME will leaves research

(26)

institutes soon, as strong evidence has to be collected first that EME is applicable for the extraction of polar compounds in water matrices.

2.3.2.2. Carbon nanotubes (CNT)

New kinds of sorbents might be desirable so that the enrichment of analytes can be improved. Examples of alternative SPE sorbents are carbon nanotubes (CNT). This type of sorbents combines polar and nonpolar interactions and makes it possible to extract all kind of analytes14,22_{. CNT consists of a graphite sheet rolled up in a nano-scale tube. The electronic} properties allow interacting strongly with inorganic molecules via hydrogen bonding or other non-covalent forces49_{. Especially multiwalled CNT (MWCNT) are of interest as the multiple} layers of graphene causes increased loadability and a large surface to volume ratio. A comparison of enrichment efficiencies with MWCNT and conventional SPE sorbents was investigated by Zhou et al. (2007), who analysed five sulfonylurea herbicides, including thifensulfuron-methyl (TSM, log D -2.03)22_{in environmental waters}50_{. Simple} matrices such as tap water showed no significant difference between MWCNT and C18 SPE, but for complex waters such as seawater and well-water, much better extraction efficiencies and recoveries were observed for MWCNT, which is shown in Figure 9. They concluded that MWCNT is a valuable alternative to SPE sorbents.

Figure 9. The chromatogram of five sulfonylurea herbicides in spiked well water enriched by CNT and C18. Peak identification: (1) nicosulfuron; (2) thifensulfuron-methyl; (3) triasulfuron; (4) metsulfuron-methyl;

(5) bensulfuron-methyl, (adapted from ref [50]).

Other published papers examined the use of CNT sorbents for different compound classes, polar and nonpolar, in water such as pesticides, phenoxy alkanoic acids or triazines49,51,52,53_{. Li et al.} (2009) investigated six polar organophosphorus pesticides in seawater, including the insecticide

(27)

Oasis HLB sorbents53_{. Pyrzynska et al. (2011) also concluded that for various polar} organophosphorus pesticides, Oasis HLB sorbents could be supplemented by CNT49_{. All in all,} CNT have great potential in the preconcentration of large volumes of water samples as they have a strong adsorption affinity for a wide variety of organic compounds. However, the costs of CNT materials limit the commercial availability. Besides, the toxicological hazard of CNT is not widely investigated and require further research49_.

2.3.2.3. Other methods

Other extraction techniques or sorbents that are under development in the analysis of polar compounds are graphitised carbon black (GCB), porous graphitic carbon (PGC) sorbents, or ionic liquids (IL). GCB is made of heating carbon black and is provided with various ionisable groups on the surface, making it possible to exchange ions and bind to acidic analytes. Schultze et al. (2019)12_{showed that GCB could be used for polar compounds with log D values ranging} from -6.7 to 3.6. However, not all compounds could be eluted, and it was concluded that retention strongly depended on the interactions of an analyte with the surface of GCB. PGC does not contain ionisable compounds, and as it has hydrophobic and electronic interactions, a wide polarity range can be retained. For dipole-interactions, it seemed that PGC is more effective than HLB sorbents9_{. The same as GCB was observed for PGC by Schultze et al.} (2019)12_{, and PGC turned out to be insufficient and could not retain all polar compounds.}

IL are solvents that are liquid at room temperature and consist of non-molecular solvents. Their low melting point and volatility makes them suitable as extraction media54_{. Various} combinations of cations and anions make IL have multiple properties and suitable for a wide range of applications, for example, in GC analysis55_{. However, emulsions can be formed, which} might cause reduced recoveries, and for synthesis, high costs are required. IL it is a green alternative and has many advantages, including high extraction efficiencies and selectivity56_. Zahn et al (2020) stated that the use of IL and its endless variability, which allows tailoring them to specific problems, may be helpful for polar compounds14_{. Besides, Zhao et al. (2006)} concluded that implementing IL might be feasible for various industry applications54_{. However,} they did not mention the use of IL in aquatic environmental samples. Despite the encouraging words from Zahn et al., from 2006 until now, not a single paper is found on using IL for polar compounds in aqueous environmental samples, which limits the applicability, especially for commercial purposes.

(28)

2.3.2.4. Summary sample preparation methods

From al discussed sample preparation methods, an overview is made in Table 2. All discussed literature is summarized to obtain a clear overview of all possible extraction methods. From Table 2, it is shown that of all ten papers discussed for sample preparation, eight articles (80%) implemented a SPE sample pre-treatment step independent of the selected separation method. SPE combined with multiple sorbents added to a cartridge was proven to be effective in analysing polar compounds relative to SPE, where only one sorbent was used. Also, it is shown that many papers used a combination of multiple ion exchangers or various extraction methods for comparison. It is expected that the problem of limited commercial availability will be reduced as research is strongly developing. Evaporation techniques, however, showed even better recoveries and broader polarity coverage relative to (ml)SPE. Besides, EC allows evaporation of the organic elution solvent, which is essential before HILIC analysis, and lower solvent consumption is necessary. However, from experiments with EC, higher matrix effects and lower enrichment factors were observed.

Alternative extraction methods such as EME might have potential as high selectivity and high matrix tolerance is achieved. However, EME is tested on several matrices but not yet on water, limiting the applicability of polar compounds. Extraction with CNT or other alternatives such as IL is still under development and are not expected to leave research institutes in the short-term. However, research in alternative extraction methods will likely grow as developments in other extraction methods is investigated. Nevertheless, alternative methods remain at a very early stage and will not be implemented for routine laboratories yet.

(29)

Table 2. Pre-treatment, separation and detection methods together with analytes and matrices studied for the analysis of polar compounds in water from all discussed papers.

Analytes Matrix Pre-treatment Analysis MS analyser Reference

26 polar compounds (log D < 0)

Tap and surface water, wastewater effluent

mlSPE, (HLB, GCB, WCX, and WAX)

Evaporation

HILIC-HRMS HESI-Orbitrap Köke et al.39

Metformin and guanyl urea Surface and influenced raw waters

SPE (WCX) HILIC-MS ESI(+)-QQQ Scheurer et al.27,57

29 PFAAs Groundwater, stormwater, surface and drinking water

SPE (WAX) SFC-MS/MS ESI(-)-QQQ Björnsdotter et al.40

TPs of pesticides, biocides and pharmaceuticals

Surface water, sewage effluent

SPE (WAX, WCX, ENV+ and Oasis HLB)

LC-MS/MS ESI-Linear Trap Quadrupole orbitrap

Kern et al.41

64 target analytes (of which 44 log D < 0)

Surface water, groundwater bank filtrate SPE (WAX, WCX) mlSPE (GCB, WCX and WAX) SPE (MCX + ENV+) Evaporation MMLC-MS/MS HILIC-MS/MS RPLC-MS/MS SFC-qTOF

ESI(+/-)-QQQ Schulze et al.12

HMSAs Surface, drinking and tap water

mlSPE (WAX, WCX and ENV+)

Evaporation

HILIC-HRMS NMR

HESI-orbitrap Zahn et al.42

590 substances (log D -14 to 8)

Wastewater, surface water VEC

mlSPE (Oasis HLB, WAX, WCX, ENV+)

LC-HRMS/MS ESI(+/-)-orbitrap Mechelke et al.43

45 target analytes (log D -6 to 1.8)

Surface, ground and drinking water, effluent wastewater

Evaporation (freeze-drying) MMLC-HRMS ESI-QQQ Montes et al.20

17 PMOCs (log D − 3.06 to 1.23)

Surface and drinking water Azeotrope evaporation SFC-HRMS RPLC-MS/MS

ESI-qTOF ESI-QQQ

(30)

2.3.3. Chromatographic separation

Many chromatographic separation techniques rely on C18 material, where retention is based on nonpolar interactions between the solid-phase and the analyte. For RPLC, a hydrophobic stationary phase is used, and hydrophobic or nonpolar compounds will have increased affinity to the stationary phase, making separation possible. Polar compounds are of high interest as they are minimally retained on conventional RPLC columns and therefore are difficult for separation. They interact weakly with the solid-phase and elute early or even in the dead volume, which is the volume of a LC system between the point of injection to the point of detection, excluding the column14_{. Yet, for increased knowledge and occurrence of polar} compounds, it is necessary to analyse them. By investigating other types of chromatographic separation, identification and detection of polar compounds is possible, and therefore various separation techniques will be discussed in section 2.3.3. Chromatographic separation

With polar compounds being mostly non-volatile, making them challenging to evaporate, GC is not suitable for polar compounds14_{. In GC separation, volatile compounds are passed through} a capillary column, and analytes are separated based on their volatility or boiling point. Although polar compounds can be volatile, compounds with a log D value below zero have low vapour pressures, which indicates that evaporation is challenging for these compounds. Although derivatisation might circumvent this and reduce the polarity of a compound, which was achieved by Steinborn et al. (2016) for glyphosate (log D -7.26)58_{, it introduces} inaccuracies and is a time consuming and labour-intensive process14_{. Therefore, the} applicability of GC on polar compounds is limited to specific nonpolar compounds with log D values higher than one and will not be discussed in this review.

2.3.3.2. Hydrophilic interaction liquid chromatography (HILIC)

HILIC can be described as a variant to normal phase LC (NPLC), which both uses hydrophilic stationary phases with reversed phase type eluents. Where NPLC uses completely organic mobile phases, HILIC uses water-miscible organic mobile phases. HILIC is a suitable technique for separating polar compounds, and retention involves the hydrophilic partitioning of the analytes between the mobile phase and the water layer adsorbed onto the stationary phase9_. Stationary phases often consist of bare silica or a polar modified solid support. Mobile phases consist of a combination of strong solvents such as water and a water-miscible weak organic

(31)

solvent such as acetonitrile (ACN)14_{. As a result, polar compounds will elute later than nonpolar} ones, and opposite elution order to RPLC is achieved.

HILIC has many advantages over NPLC or RPLC. Retention of polar compounds that are too water soluble for RPLC is possible with HILIC as polar analytes show good solubility in the aqueous mobile phase and show retention with the polar stationary phase. Besides, HILIC is favoured over RPLC prior to MS-detection, as for ionization and desolvation high organic solvent is required. This is achieved by using high volumes of organic solvent at the beginning of the HILIC separation, and is in contrast with RPLC, where high contents of water are needed to retain polar analytes37_{. As a result, increased sensitivity is achieved in MS analysis which} means lower detection of the analyte’s concentration. However, in HILIC, undesired secondary electrostatic or hydrogen bonding interactions might occur, and to stabilise the water layer, long equilibration times are necessary to obtain reproducible retention times59_{. HILIC, in} combination with aqueous samples, might seem impossible as high elution strength of water is achieved by using HILIC60_{. Nevertheless, many papers successfully reported the analysis of} polar compounds in water, which are interesting to discuss together with the possibility of implementing HILIC in commercial laboratories.

From an extensive review by Salas et al. (2017), it was stated that bare silica, zwitterionic, amide and diol stationary phases are mainly used for HILIC separation in environmental analysis37_{. Scheurer et al. (2009) and a follow-up study (2012) developed a HILIC-MS/MS} method using zwitterionic phases to determine the antidiabetic pharmaceutical metformin (log D -5.62) in surface water27,57_{. A Strata-WCX was used for sample clean-up, and isotopically} labelled standard metformin-d6 was used for quantification. This method was proven worthwhile for metformin and guanylurea, with low detection limits (10 ng/L) and high recoveries (90%). Also, both papers published data on the occurrence of metformin and guanylurea in sewage and surface water of Germany for the first time. Their environmental fate and effective removal in treatment plants was also studied. However, this paper lacks information on matrix effects or ionisation suppression which affect the reliability of the method.

A comparison between HILIC and RPLC can be seen in the work of Bisceglia et al. (2010) which studied thirteen different LC columns to separate 23 drugs compounds in wastewater,

(32)

including the polar drug metabolite benzoylecgonine (log D -0.60)22,61_{. The thirteen stationary} phases tested are depicted in Figure 10.

Figure 10. Summary of all tested RPLC and NPLC/HILIC columns, (adapted from ref [61]).

They compared HILIC and RPLC phases prior to tandem MS analysis. They observed enhanced retention for polar analytes with cross-linked diol, zwitterionic, and embedded polar group stationary phases. However, low resolution i.e., the separation power of the complete chromatographic system relative to the components of a mixture, was observed and could not be improved by adjusting pH, temperature, solvent, or gradient conditions. Analyte recoveries were investigated on six different SPE sorbents and Strata XC provided the best recoveries for most analytes. However, the recovery of cocaine remained dependent on its esterification and varied from very low (19%, ecgonine) to high (87%, benzoyl ester-containing metabolites), which is not very robust in terms of reproducibility of a method.

An example of HILIC used with SPE as pre-treatment method can be seen in Echeverría et al. (2013). They developed a SPE coupled to HILIC-MS/MS method for five (very) polar iodinated X-ray contrast media in sewage, including iohexol (log D -1.95)22,62_{. RP and HILIC stationary} phases were compared, and after optimisation of separation parameters, the zwitterionic

Column Name Manufacturer Modea _{Phase Type}

Particl e Size (µm) Pore Size (Å) Column Dimensions (mm)

Zorbax Eclipse XDB Agilent Technologies

RP C18 end-capped 5 4.6 ´ 250

Atlantis T3 Waters Corp. RP C18 end-capped 5 3.0 ´ 150

Gemini NX Phenomenex RP C18 end-capped 3 2.1 ´ 150

Polaris C18-Ether Varian RP EPGb_-alkoxy ₃

3.0 ´ 150

Polaris RP-Amide Varian RP EPG-amide 5 3.0 ´ 150

Ascentis RP-Amide Sigma-Aldrich (Supelco)

RP EPG-amide 5 3.0 ´ 150

Allure Basix Restek RP Propyl cyano 5 4.6 ´ 150

Ultra IBD Restek RP EPG proprietary 3 2.1 ´ 150

Ultra PFP Restek RP PFPc ₃ ₁₀₀ _{2.1 ´ 150}

Allure PFPP Restek RP PFP-propyl 5 60 4.6 ´ 150

Viva PFPP Restek RP PFP-propyl 5 300 2.1 ´ 150

Luna HILIC Phenomenex NP EPG-diol 5 4.6 ´ 250

Obelisc-N SIELC NP EPG proprietary

(zwitterionic)

5 3.2 ´ 150

Ultra IBD Restek NP EPG proprietary 3 2.1 ´ 150

a_{Mode of separation: reversed-phase (RP) versus hydrophilic interaction or normal-phase (NP)} b _{Embedded polar group (EPG) within an extended (typically C18) alkylsilane chain}

(33)

RPLC. However, high matrix effects affected the sensitivity of the method. Although this method seems promising, further work on the clean-up of the sample during pre-treatment is needed.

Another example of HILIC being successfully used to separate polar compounds is with Zahn et al. (2019) who recently discovered a few polar disinfection by-products HMSAs in surface and tap water with HILIC-MS/MS42_{. Four HMSAs, including bromomethane sulfonic acid (log} D -2.38)22_{, were first synthesised as they are not commercially available and quantitative} nuclear magnetic resonance (qNMR) was used for concentration determination. mlSPE pre-treatment was used to enrich the polar compounds, and an acquity amide column was used for HILIC- scheduled multiple reaction monitoring (sMRM) analysis. However, quantification was limited to only HMSAs, and this method indicated high matrix effects. Although this method requires more investigation, this method extended the existing methods on HMSAs analysis and a successful synthetization of four HMSAs which is favourable when establishing a method for monitoring HMSAs in water treatment plants without the availability of reference standards.

Another example of HILIC being efficiently used to separate polar compounds in waste water was by Sordet et al. (2018) who described a HILIC-MS/MS/MS method with triple stage fragmentation for the detection of seven polar iodinated contrast media (ICMs) in sewage waters, including iopamidol (log D -0.74)22,63_{. Only a dilution step was performed as sample} pre-treatment, and a WAX mixed-mode column was selected for analysis after optimisation. Water, ammonium acetate and ACN were used for the mobile phase, and MRM3_parameters were optimised using ICM standards. The method was validated and applied to freshwater samples, and six out of seven ICM could be detected in at least one sample. This study reported a method with MRM triple-stage fragmentation which allowed high selectivity and sensitivity for contrast agents. In addition, low ion suppression (5%) and ion enhancement (9%) was observed, and low limit of quantification (LOQ) values of 0.5 µg/L for all the ICMs were evaluated. ESI, in combination with MRM3_{showed a better signal to noise ratio compared to} conventional MRM modes. The method was proven to be quick, simple and has low matrix effects due to the absence of sample pre-treatment. Although little research is done on triple stage fragmentation, it proved to be a good alternative for detecting polar ICMs. However, triple stage fragmentation is not widely applied in commercial labs and requires special expertise, limiting its applicability for routine laboratories.

(34)

HILIC can also be used to separate herbicides in freshwater, which was demonstrated by Fauvelle et al. (2015) who developed a HILIC-MS/MS method to detect nineteen acidic herbicides, including bentazon (log D -0.19)22,64_{. A design of experiment (DoE) was performed} for SPE optimisation to study the parameters extraction pH, the proportion of methanol in the washing fraction and the washings fraction volume. This resulted in selecting an Oasis HLB cartridge, which first seemed ineffective for polar compounds. In addition, to minimise matrix effects and maximise extraction, recoveries extraction was performed. A zwitterionic silica-based stationary phase was used for analysis, together with ammonium acetate, ACN and methanol as a mobile phase. The method was validated and showed recoveries above 64% and LOQ values between 5-22 ng/L. River samples were used to test the method's practicability and showed the wide presence of acidic herbicides and their metabolites. This robust method presented the usefulness of HILIC in polar herbicide analysis. Also, attention was paid to limiting interferences which are appreciated in developing analytical methods. However, this method is limited to acidic herbicides and does not cover other classes of polar compounds. Yet, the method might be suitable for targeted analysis in commercial laboratories due to its straightforward analysis and reliable results.

Only recently HILIC was combined with nontargeted screening approaches, which was demonstrated by Kolkman et al. (2021), who quantified 32 very polar compounds, including melamine (log D -2.34)22_{in surface and drinking water with HILIC-HRMS}65_{. In addition, data} from HRMS analysis could be screened for unknown compounds. A simple evaporation and reconstitution step was performed before analysis. For HILIC analysis, a high-purity silica column was selected. The method was validated, and an average recovery of 89% was obtained for surface water. For compounds that are often difficult to analyse, e.g. gemcitabine (log D -1.47)22_{, this method could be used. The method was applied on real surface, and drinking water} samples and twelve out of 32 compounds could be detected. Matrix effects were studied by monitoring the peak areas of used internal standards. The method proved to be suitable for polar compounds. However, more internal standards could improve the method's strength, although this remains challenging in HILIC analysis as those are limited in availability. In addition, strong acids could not be detected due to poor retention and implied unsuitable for this method.

By addressing many HILIC methods used for environmental analysis, many acidic or basic analytes show proper retention, especially for compounds with amines, carboxylic acids and

(35)

the solvation of the ions in the mobile phase. HILIC can be used commercially as it can be used on conventional LC instruments, and reliable results have been reported on polar compound analysis. However, developing only one method covering the whole range of (very) polar compounds remains challenging.

2.3.3.3. Supercritical fluid chromatography (SFC)

SFC is a separation technique similar to GC and LC but uses a supercritical fluid as the mobile phase. If a liquid or gas operates above its critical temperature (Tcrit) or pressure (Pcrit), it changes in a supercritical fluid66_{. SFC was introduced in the 1960s}67_{, but many technical} challenges were encountered and gained popularity only in the last two decades as commercial instruments became available, and industry could develop robust SFC methods22,68_{. Carbon} dioxide (CO2) is a commonly used supercritical fluid since its critical temperature is just 31.1 °C and its critical pressure is 7.38 MPa. Other mobile phases are ethane, n-butane, nitrous oxide, dichlorodifluoromethane, diethyl ether, ammonia or tetrahydrofuran. Retention mechanisms are predominantly determined by the stationary phase and include nonpolar interactions such as hydrogen or van der Waals interactions14_{. By using polar stationary phases, NP/HILIC retention} behaviour is observed, and thus, polar analytes obtain increased retention. Modifiers such as methanol and additives such as acids, bases, or salts are often added to the mobile phase to increase its polarity and the elution strength of polar analytes from polar columns14_.

Supercritical CO2 is the most favourable mobile phase as its critical temperature and pressure are easy to reach but also as it has a low viscosity, low cost and is inert against many compounds. Therefore, high separation efficiencies and narrow chromatographic peaks can be achieved14_{. SFC brings many advantages as it can be used for complex samples to separate} small molecules69_{. Also, due to the low viscosity of the mobile phase higher flow rates can be} obtained and reduced analysis time is achieved. Low solvent is required which makes SFC a green technique as the amount of hazardous wasted is reduced. However, many polar compounds are not soluble in critical CO2 which makes them poor candidates for separating them by CO2 SFC70_{. Besides, a pressure gradient is along the column is required to maintain} the solvent in supercritical state. This can change the density, elution strength and thermal conductivity of the method. If the pressure is increased, higher density and thus higher elution strengths are observed. Therefore, a minimal pressure drop is desired to obtain high method stability14_{. Besides, SFC cannot be performed on a conventional LC system and the CO2 tanks} required for SFC analysis are big and heavy which makes it both unattractive for commercial

(36)

laboratories with limited lab space. As SFC is primarily used in pharmaceutical analysis to separate enantiomers, and not for separating polar analytes, it is debatable if SFC is a suitable technique for polar compound analysis. Yet, a few papers reported the analysis of polar compounds in water with SFC which showed interesting results for discussion.

SFC has helped to develop a method for the analysis of fifteen polar compounds between a log D range of -3.06 to 1.23, including melamine (log D -2.34, see Table 1) in various matrices by using SFC-HRMS7_{, which was demonstrated by Schulze et al. (2020). For enrichment of} samples, azeotrope evaporation (section 2.3.1.3. Evaporation concentration (EC)) was used, and the application of the method was tested on six water samples, including surface, ground and drinking water. A BEH (a hybrid stationary phase with reversed and normal phase characteristics) and a Diol normal phase column showed the best performance for compounds in positive ionization mode or negative ionization mode, respectively. A mobile phase consisting of methanol/water and ammonium hydroxide was selected as modifier. Adding methanol allowed improvements of the solubility and dissolution of the compounds. The method was validated and compared with a RPLC-MS/MS method. The SFC method was superior to RPLC in terms of peak shapes and retention times, which is depicted in Figure 11.

Also, low LOQ values in the single- to double-digit ng/L range were determined. Of the fifteen compounds, nine were detected in at least one of the samples, including acesulfame (log D -1.49)22_{. Still, high signal suppression was observed, which probably is due to the evaporation} step. However, this sample clean-up step is very generic and quick as the only pre-treatment step was to evaporate the mixture (sample with acetonitrile) to dryness at 40 °C and could cover

Figure 11. (a) SFC-HRMS extracted mass chromatograms of a standard mixture of the target PM substances sorted by log D; (c) RPLC-MS/MS-extracted MRM chromatograms of the

A critical review of analytical methods for the analysis of polar compounds in water

Literature Thesis