Determination of Pesticides in Cannabis

(1)

Determination of Pesticides in

Cannabis

Graduation report

Monique van den Kieboom, Bilthoven, June 2011 Version 1

(2)

Determination of Pesticides in Cannabis i

Determination of Pesticides in

Cannabis

Graduation report

Version 1: June 2011

Candidate: Monique van den Kieboom

2009837

maw.vandenkieboom@student.avans.nl Graduation place: National Institute for Public Health and

Environment (RIVM)

Department: Laboratory for Health Protection Research (GBO) RIVM

Antonie van Leeuwenhoeklaan 9 3721 MA Bilthoven

Supervisor: Dr. Jan van Amsterdam

Jan.van.Amsterdam@rivm.nl

Tutor: Hans Cremers Hans.Cremers@rivm.nl

Educational institute: Avans University Breda

Academic for Technology for Health and Environment

Education: Forensic Laboratory Science

Specialization: Forensic Analytical Chemistry

Supervisor educational institute: Ir. Hans van Amelsvoort jaga.vanamelsvoort@avans.nl

(3)

Determination of Pesticides in Cannabis ii

Foreword

For the graduation of the training “Forensic Laboratory Science” at the Avans University in Breda, a research was performed with mainly chromatographic techniques. The investigation took place from January 24, 2011 to June 10, 2011 at the National Institute for Public Health and Environment (RIVM) in Bilthoven, the Netherlands. A part of the investigation took also place at Avans University in Breda.

During this graduation period I have learned a lot about chromatographic and mass spectrometric techniques, mainly gas chromatography and ion trap mass spectrometry. I would like to thank my supervisor and tutor Dr. Jan van Amsterdam and Dhr. Hans Cremers, from the RIVM, for their accompaniment during my internship. Furthermore I would like to thank everyone from the section “Gezondheidsbevolkingsonderzoek” for the questions I could ask, and the good time I have had during my graduation internship. Also, I would like to thank the police forces who collected the samples for my investigation.

I would also like to thank from the Avans University Breda, Dr. Ben de Rooij, for the opportunity to analyze my samples on the LC-MS, and Ing. Fred Geenen and Ing. Edward Knaven, for their guidance with the LC-MS during the days I was analyzing my samples at the Avans University.

At last, I would like to thank Ir. Hans van Amelsvoort for his pleasant accompaniment from school, during my graduation internship.

6 June 2011, Bilhoven Monique van den Kieboom

(4)

Determination of Pesticides in Cannabis iii

Summary

The National Institute for Public Health and Environment (RIVM) has been requested by the Ministry of Health, Welfare and Sports, to investigate whether Dutch Cannabis contains pesticides. To answer this question, RIVM requested the police to collect hemp plants from illegal hemp plantations. The yield of these plantations is sold often to coffee shops for distribution. The aim of this investigation is to determine whether and if yes, which pesticides are used on hemp plants. The other goal is to quantify the amount of used pesticide. In a questionnaire sent to police forces it was reported that occasionally jerry cans have been found by hemp plantations that may contain pesticides. Previous studies have reported that in the past abamectin, dichlorvos and fenbutatin were used in hemp plantations,. This investigation does not only include some legal substances, like abamectin, imidacloprid, tebuconazole and triadimenol but also three groups of illegal substances, like organochlorides, organophosphates and poly chlorinated biphenyls. The most samples were delivered by police forces from the regions “Haaglanden” and Flevoland. In total there were 114 samples, from illegal plantations. 88 samples came from 44 plants, and 26 samples were already packed in plastic bags. Besides these samples eight samples were obtained from coffee shops and three samples of medicinal hemp from the Trimbos institute. For the sample preparation two methods were used for comparison. The first method was developed for a previous study (determination of pesticides in tobacco) and had a preparation time of two days. The second method was the DisQuE kit from Water, which has a preparation time of 1-2 hour. The samples were analyzed according to two analytical methods on the gas chromatograph to determine which analytical method was more adequate. The first method was also from the previous study, and the second developed from the paper entitled “determination of multiclass pesticides in food commodities by pressurized liquid extraction using a GC-MS/Ms and LC-MS/MS”. 48 samples, including six coffee shop samples were also analyzed on abamectin with the LC-MS at the Avans University.

The analysis of GC-MS resulted in the fact that one of the 114 samples was contaminated with substance. The library of the MS identified this as pronamide, but further investigation with the pronamide standard proved this not to be. The analysis of abamectin showed that six samples out of the 48 were contaminated with abamectin, which is one of the substances regularly found by the police.

The main disadvantage of this investigation was, that there was no knowledge beforehand about the pesticides that are used on hemp plantations, other than the reported ones in the previous study. There was a large amount of samples, but no samples with certain use of pesticides, or clean ones. These were the main reason to develop a screening test and not a specific test. At the end of the internship the positive sample was detected with the GC-MS, but the suspected standard was not available. When the standard was available only an identification analysis could be performed. This investigation needs further optimization with more types of pesticides to create a more reliable screening test. Also an LC-MS analysis should be performed standard in this investigation. Some of the pesticides can only be determined with the LC-MS.

(5)

Determination of Pesticides in Cannabis iv

Table of Content

1 Introduction... 1 2 Theoretical background ... 2 2.1 Pesticides... 2 2.1.1 Abamectin ... 2 2.1.2 Imidacloprid... 3 2.1.3 Tebuconazole ... 4 2.1.4 Triadimenol... 4 2.1.5 Pyrethrins ... 5 2.1.6 Tebufenpyrad ... 5 2.1.7 Piperonylbutoxide ... 6 2.2 Sample preparation ... 6

2.2.1 Solid phase extraction ... 6

2.2.2 DisQuE sample preparation ... 7

2.3 Analytical research... 8

2.3.1 Varian GC column ... 8

2.3.2 Scherzo SM-C18 column for the HPLC ... 9

2.3.3 MS coupled to the GC... 9

2.3.4 MS coupled to a HPLC ... 10

3 Materials and methods ... 13

3.1 Preparation of hemp plants ... 13

3.2 Standards for the analytical research ... 13

3.3 Standards for the GC-MS analysis... 13

3.4 Standard for the LC-MS analysis... 14

3.5 Sample preparation GC-MS analysis... 14

3.5.1 Sample preparation with the first method ... 15

3.5.2 Sample preparation with the second method ... 16

3.6 Sample preparation for LC-MS analysis... 17

4 Results... 18

4.1 Preparation of hemp plants ... 18

4.2 Standards on the GC-MS ... 19 4.3 Standard on the LC-MS ... 21 4.4 Samples on the GC-MS ... 22 4.5 Samples on the LC-MS... 24 5 Discussion ... 27 5.1 Sample preparation ... 27

5.2 Comparison of the two GC-MS methods... 28

5.3 Sample analysis... 28 6 Conclusion ... 30 7 Recommendations... 31 7.1 Sample preparation ... 31 7.2 Sample analysis... 32 Literature... 33 Annex... 36 Annex I... 36

(6)

Determination of Pesticides in Cannabis v Annex II ... 37 Annex III... 39 Annex IV... 42 Annex V ... 46 Annex VI... 48

(7)

Determination of Pesticides in Cannabis 1

1 Introduction

The National Institute for Public Health and Environment (RIVM) has been requested by the Ministry of Health, Welfare and Sports (VWS), to investigate whether Dutch cannabis, sold in coffee shops, contains pesticides. To answer this question, RIVM requested the police to collect hemp plants from illegal hemp plantations for investigation. The yield of these plantations is sold often to coffee shops for distribution. The aim for this investigation was to determine whether and if yes, which pesticides are used on hemp plants. Another goal was also to quantify the amount of used pesticides. The police has found, during dismantling of hemp plantations, several unlabeled jerry cans, which may contain pesticides. In a questionnaire of RIVM sent to police forces, it was reported that occasionally jerry cans have been found, which contain legal substances. The expectation was that if there was any evidence of pesticides use, that it would be a legal substance.

Previous studies have shown that abamectin, dichlorvos and fenbutatin were used in hemp plantations, which are mainly legal substances. This investigation does not only include the legal substances, but also the illegal substances, like organophosphates, organochlorides and poly chlorinated biphenyls.

The RIVM is a knowledge and research institute promoting public health and a healthy and safe living environment. The main task of RIVM is to support the government policy, in national and international context. The other tasks includes, national coordination, knowledge development and research, provision of information to professionals and the general public, and many more. RIVM is responsible for providing reliable information in the fields of health care, medication, the environment and nutrition and safety. It is an official research facility of the ministries of VWS and Housing, Spatial Planning and the Environment. RIVM works also for the ministries of Agriculture, Nature Management and Food Quality; Transport, Public Works and Water Management, and for the Dutch Food and Consumer Product Safety Authority (nVWA).

The internship took place on the Laboratory for Health Protection Research (GBO). This laboratory identifies, analyzes and quantifies the risk and health impact associated with chemical substances, biological products, and food and lifestyle factors. This can be the effect of exposure to toxic substances, or the health consequences of smoking and alcohol consumption. The tasks of GBO includes assessing potential health risks and effects, providing scientifically based recommendations to government and providing scientifically based evidence to support other work at RIVM.

This graduation report is divided in several chapters. Chapter two provides the theoretical background that was necessary for this investigation. Chapter three reflects the materials and methods for the analytical research for this investigation. The results of the investigation are found in chapter four, and the discussion and conclusion of these results in chapter five and six respectively. Chapter seven reflects the recommendations for this investigation. This report closes with a list of consulted literature and several annexes.

(8)

2 Theoretical background

This chapter provides the theoretical background of pesticides, sample preparation methods and the analytical methods that will be used during this project.

2.1 Pesticides

Pesticides is a collective noun for all the substances that are developed for the extermination of vermin and pests. The different types of pesticides are also named after the organism that they exterminate, like insect are destroyed with insecticides and mites with miticides. The European Plant Protection Product Directive 91/414/EEC, reports in Annex I of the directive, which substances under which conditions are allowed for use in the EU. In countries outside the EU, some of the hazardous substances, like DDT, are still allowed for use, although they are hazardous for the health and the environment. The results of the questionnaire of RIVM sent to police forces reported that occasionally jerry cans have been found, which contain legal substances. Legal substances are substances that are allowed according to annex I of the directive 91/414/EEC. However, the way of use can be illegal. For example, if a substance is only allowed in greenhouses but is used in a garden, the use of the substance is illegal. Previous studies from Rikilt and Avans University have determined that abamectin, dichlorvos and fenbutatin were used in the past on hemp plants. The following substances are quite regularly found by the police in hemp plantations: abamectine, imidacloprid, tebuconazole, triadimenol, pyrethrins, tebufenpyrad and piperonylbutoxide. The next section will give more information about these substances.

2.1.1 Abamectin

Abamectin is a fermentation product of the bacterium Streptomyces Avermitillis and is used as an insecticide and miticide to exterminate insects and mites. Abamectin is also used as seed protectant to destroy nematodes. It acts on pests by interfering with the nervous systems. The pests will become paralyzed after intoxication with abamectin[1]. Abamectin is a mixture of avermectin B1, containing equal to or more than 80% avermectin B1a and equal to or less than 20% avermectin B1b. Figure 2-1 contains the structural formula of abamectin. The R in the structure represents the group that distinguishes avermectin B1a from avermectin B1b.

(9)

Figure 2-1: structural formula of abamectin

Following ingestion of high dose, abamectin is toxic to humans. Intoxication can cause hypotension and subsequent aspiration, and even life-threatening coma. A dose of 100 mg/kg or higher of ivermectin plus abamectin is toxic. During the production or use of the product, one may also be exposed to the substance through dermal contact. Protective clothing reduces this exposure. Abamectin is expected to be mobile in soil but, according to its Henry’s law constant, vapor pressure and water solubility, it is not volatile. The half-life of abamectin in soil is two weeks to two months. Under the influence of light, abamectin will degrade rapidly with a half-life of approximately two to six hours[2].

2.1.2 Imidacloprid

Imidacloprid is a systematic neonicotinoid insecticide used mainly on food crops. Neonicotinoids are derivates of nicotine, the active substance in tobacco leaves. Nicotine is also an insecticide. Imidacloprid inhibits the nicotine acetylcholine receptors which disrupts nervous system signaling. Imidacloprid is persistent in the environment which increases the probability of exposure to non-target organisms. Imidacloprid potentially poses an acute and chronic risk to small mammals and fresh water and estuarin/marin invertebrates. Figure 2-2 reflects the structural formula of imidacloprid[3].

Figure 2-2: Structural formula of imidacloprid

In humans, imidacloprid has a low toxicity because neonicotinoid insecticides bind less tight to human cholinergic receptors compared with insects. They also do not pass readily the human blood-brain barrier. Because of this poor penetration, imidacloprid gives no toxic effects if the exposure is low, though through dermal contact at workplaces occupational hazard and harm may occur. Imidacloprid has a low mobility in soil. The amine group in the structural formula indicates that this compound primarily exists in the cation form in the environment. In an environment with vegetation imidacloprid has a half-life of 48 days. In an environment without vegetation, the half-life is 190 days. This

(10)

Determination of Pesticides in Cannabis 4 means that imidacloprid degrades more rapidly in an environment with vegetation than without[3].

2.1.3 Tebuconazole

Tebuconazole is used as fungicide to exterminate fungi in food crops. Because of its use as fungicide, tebuconazole will be directly released to the environment. Figure 2-3 reflects the structural formula of tebuconazole.

Figure 2-3: Structural formula of tebuconazole

Tebuconazole is a non-volatile substance. Tebuconazole is persistent in soil but the mobility of the substance increases as the soil organic matter decreases[4]. According to the US Environmental Protection Agency (EPA) tebuconazole is a possible human carcinogen, but it is not classified as a mutagenic or genotoxic substance[5]. The half-life for tebuconazole in soil is under aerobic conditions some 600 days. Under anaerobic conditions the half-life is 1260 days[6].

2.1.4 Triadimenol

Triadimenol is the major metabolite of triadimefon. Triadimenol and other triazole fungicides metabolizes in animals and plants to form compounds containing the triazole moiety. Figure 2-4 depicts the structural formula of triadimenol.

Figure 2-4: Structural formula of triadimenol

Triadimenol and triadimefon have a common mode of action as they both block the re-uptake of the neurotransmitter dopamine. More specifically, the substances act as an indirect dopamine agonist by binding to the dopamine transporter. Animal studies have shown that both substances are able to induce hyperactivity. Used as fungicide, triadimenol will be directly released into the environment. Compared to other triazole compounds triadimenol is persistent in soil. The half-life of triadimenol is determined as more than one year. Some studies indicate a half-life of 500 days. Triadimefon degrades

(11)

Determination of Pesticides in Cannabis 5 with a half-life of 15 days to form the metabolite triadimenol. Triadimenol has a low toxicity following oral, dermal and inhalation exposure[7].

2.1.5 Pyrethrins

Pyrethrins exist in two forms; pyrethrin I and pyrethrin II. Figure 2-5 and 2-6 shows the structural formula of both substances.

Figure 2-5: Structural formula Figure 2-6: structural formula of pyrethrin II of pyrethrin I

Pyrethrins are not classified as possible human carcinogens, but there is suggestive evidence of carcinogenicity. Ingestion of 5-50 mg pyrethrins causes no toxic effects, but higher doses can induce symptoms which resemble DDT intoxication, like tremors, and hypothermia and even death. Persons with chronic respiratory diseases can be oversensitive for pyrethrins. Natural pyrethrins are moderately toxic following oral exposure but highly toxic if exposed via the parental route. Synthetic pyrethrins act on sodium channels. Normally these channels are closed off after depolarization, but the pyrethroid molecules hold these channels in an open position. Pyrethrins act also on the central nervous system. Used as contact insecticides, pyrethrins will be directly released into the environment. Pyrethrin I may exist in vapor phase, but it will degrade easily by a reaction with hydroxyl radicals in 1.3 hours[8]. The half-life of pyrethrin II is not available, but based on its structure it will probably degrade rapidly in the environment, like pyrethrin I[9].

2.1.6 Tebufenpyrad

Tebufenpyrad is a pyrazole insecticide. There is suggestive evidence of carcinogenicity, but is not classified as a human carcinogen. Figure 2-7 reflects the structural formula of tebufenpyrad.

(12)

Determination of Pesticides in Cannabis 6 The exposure to the environment is limited, because tebufenpyrad only is allowed for use in green houses[10]. Tebufenpyrad acts as an inhibitor on complex I in the mitochondrial respiratory chain. The substance is immobile in soil when it is released into the environment. The half-life of tebufenpyrad is more than 28 days[11].

2.1.7 Piperonylbutoxide

Piperonylbutoxide is, in comparison with the previous substances, not a pesticide but a synergist for pesticides. Piperonylbutoxide is mainly used in combination with pyrethrins. A synergist is a chemical with no pesticidal effects on its own, but it enhances the pesticidal properties of other chemicals[12]. Figure 2-8 depicts the structural formula of piperonylbutoxide.

Figure 2-8: Structural formula of piperonylbutoxide

Piperonylbutoxide inhibits the metabolism of pesticides by insects. Without such synergist insects may be able to degrade pesticides before any effects occur. With a synergist the amount of pesticides can be reduced for an effective result. The substance is classified as a possible human carcinogen. In the environment piperonylbutoxide behaves as a vapor. The half-life of piperonylbutoxide in air and soil is 3.6 hours and 14 days, respectively[13].

2.2 Sample preparation

Before chemical analysis of pesticides is possible, the tops and leaves of the hemp plants have to be prepared for the extraction of pesticides. Different kinds of sample preparation to extract pesticides are available. Two kinds of the available sample preparation methods, the solid phase extraction and the DisQuE dispersive sample preparation kit from Waters, will be discussed in the following sections.

2.2.1 Solid phase extraction

Solid phase extraction (SPE) is a useful sample preparation technique, because SPE prevents some of the problems associated with liquid/liquid extraction such as incomplete phase separation, low recoveries and disposal of large quantities of organic solvents. SPE is mainly used to prepare liquid samples and the extraction of semi volatile or non-volatile analytes. There are four phase types of columns; reversed phase SPE, normal phase SPE, ion exchange SPE and adsorption SPE. Reversed phase separation contains a polar liquid phase and a nonpolar modified solid phase. It is based on hydrophobic

(13)

Determination of Pesticides in Cannabis 7 interactions such as nonpolar-nonpolar interaction, van der Waals forces or dispersion forces. The analyte of interest is typical mid- to nonpolar. Normal phase separation contains a mid- to nonpolar liquid phase and a polar modified solid phase. This method is based on hydrophilic interactions such as polar-polar interaction, dipole-dipole interaction, dipole-induced dipole interaction, hydrogen bonding or pi-pi interaction. Normal phase SPE has two possible solid phases that can be used; polar functionalized bonded silicas and polar adsorptions. The analyte of interest, depending on the investigation, is mainly a polar compound. Ion exchange SPE is based on electrostatic attraction of charged groups on the sorbent surface to charged groups on the compound. Ion exchange SPE has two types; anion exchange SPE and cation exchange SPE. The silica in the anion exchange column has an aliphatic quaternary amino group. The cation exchange column contains silica with aliphatic sulfuric acid bonded to the surface. The last phase type discussed here is adsorption SPE. Adsorption is based on the interaction of compounds with the solid phase. Hydrophobic or hydrophilic interaction may apply, depending on which solid phase is used, and depending on which type of analyte is of interest[14].

The column type applied in the present study is a Bond Elut Si column obtained from Varian. This is a normal phase column based on adsorption. The column contains silica gel with no bonded phase and is suitable for the extraction of compounds, like pesticides. The underivatized silica is commonly used as sample clean-up. The silica in the column is extremely hydrophilic and must be kept dry. This means that the samples must be completely water free. The adsorption of polar compounds of the matrix of the samples occurs by the functional groups of the free hydroxyl groups on the surface of the silica particles. Silica columns may be used for the adsorption of polar compounds from nonpolar matrices with subsequent elution of the compounds in an organic solvent. The Bond Elut Si column is used as an adsorption media, when an organic extract is applied to the silica bed, the analyte of interest passes through unretained and the unwanted matrix compounds adsorb on the silica. The column is then discarded[14].

For the analysis on the LC-MS system, also a Bond Elut C-18 column is used. This is a reversed phase column, to extract mid- to nonpolar pesticides. In this research this column was used to extract abamectin from the plants.

2.2.2 DisQuE sample preparation

Another type of sample preparation is the sample preparation with the DisQuE dispersive sample preparation kit from Waters. The DisQuE dispersive sample preparation kits are based on the QuEChERS method. QuEChERS is an acronym for Quick, Easy, Cheap, Effective, Rugged and Safe. The two most commonly used methods with the DisQuE kits described are (1) the European Committee for Standardization (CEN) method 15662 and (2) the Association Of Analytical Communities (AOAC) official method 2007-01. Both methods use buffers in the extraction step. The CEN method uses a citrate buffer, whereas the AOAC method uses acetic acid as buffer[15].

(14)

Determination of Pesticides in Cannabis 8 The QuEChERS extraction is designed for multi residue pesticides analysis of fruits and vegetables with high water content matrices (85-90%). For samples with a low water content, water is added before the extraction step. During this investigation, the CEN method is used to extract the pesticides from hemp plants. The CEN method contains two tubes. Tube 1 is a 50 ml extraction tube containing 4 g of magnesium sulfate, 1 g of sodium chloride, 1 g of trisodium citrate dehydrate and 0.5 g of disodium hydrogen citrate sesquihydrate. Tube 2 is a 2 ml clean-up tube containing 150 mg magnesium sulfate, 25 mg primary-secondary amine (PSA) and 25 mg C-18 sorbent. Magnesium sulfate is used because of its ability to bind large amounts of water so that the pesticides are forced to move into the organic layer. PSA is used to remove polar interferences such as sugars and organic acids. The C-18 sorbent removes fats and other very nonpolar interferences[15]. The most commonly used solvents are acetonitrile, acetone and ethyl acetate. Obviously, each solvent has its advantages and disadvantages. Chosen is for acetonitrile as solvent. Acetonitrile is highly compatible with the most gas chromatographic and liquid chromatographic methods. Compared to the other solvents, acetonitrile has a lower volatility and does not extract very much lipofilic materials. On the other hand the residual water can be removed better by drying agents in acetonitrile than in other solvents[16].

2.3 Analytical research

The analytical research will be performed on two locations. The first is on a gas chromatograph (GC) CP 3800 coupled to a mass spectrometer (MS) 225, by the RIVM in Bilthoven. Both devices are purchased from Varian. The second research will be performed on the high performance liquid chromatograph 1200 series (HPLC) coupled to a 6320 MS at the Avans University in Breda, purchased from Agilent. Both MS’s are ion trap MS’s. In this section, the columns of the GC and the HPLC and the ion trap MS will be discussed. Annex I contains the pictures of the analytical equipment that was used in this study.

2.3.1 Varian GC column

The column that was used for the GC analysis is a Varian VF- 5m factor four column. This column has a low column bleed, which means that the column gives a little background noise. Figure 2-9 contains the structure from the column.

Figure 2-9: structure of the column

This column is highly inert with the following low bleed phase, 5% phenyl and 95% dimethylpolysiloxane. The highest temperature with low bleed is 325°C. The column is

(15)

Determination of Pesticides in Cannabis 9 selective for aromatic compounds. This selectivity combined with the high inertness of the column makes this column applicable for semi-polar and polar compounds. The typical application area are the following substances, alcohols, amines, drugs, pesticides, poly chlorinated biphenyls, sugars and many more. The substances that will be analyzed are poly chlorinated biphenyls and other pesticides[17].

2.3.2 Scherzo SM-C18 column for the HPLC

This column is a multimode C-18 column (further described as ODS column). It does not only contains the standard ODS ligands, but also cation and anion ligands which results in the fact that this column can be used to separate highly polar compounds. The Scherzo column has three types of separation, reversed phase, anion exchange and cation exchange. These interactions enable the separation of water-soluble substances without the use of ion-pairing reagents. Because of the ligands the column is also able to separate anionic and cationic compounds, using one column and one method. This makes it also able for the identification of pesticides[18].

2.3.3 MS coupled to the GC

The MS 225 from Varian is an ion trap MS. This MS has two methods of sample ionization; electron ionization and chemical ionization. In the electron ionization (EI) mode, electrons pass through the ion trap cavity during the ionization period. The electrons collide with the sample molecules present in the ion trap cavity. During the collision, part of the energy from the electrons will be transferred to the molecules. These energetic molecules decompose through different reactions and produces ions and neutral fragments. The ions that were produced through ionization create a unique mass spectrum. The other ionization technique is the chemical ionization (CI) mode. CI ionizes the sample molecules in two steps. In the first step the reagent gas ions are formed as the reagent gas is ionized by interaction with electrons emitted by the filament. The next step is that the reagent gas ions react with the sample molecules to form sample ions. The following four reactions can occur between the reagent gas ions (R) and the sample molecules (M) to create sample ions (M+)[19].

1. Proton transfer: (RH)+ + M  (MH)+ + R 2. Hydride abstraction: R+ + M [M-H]+ + RH 3. Association: R+ + M (MR)+

4. Charge transfer: R+ + M  M+ + R

2.3.3.1 Selected Ion Storage

There are different kinds of detection possibilities in the ion trap MS. One of these detection possibilities is the Selected Ion Storage (SIS). SIS enriches the sample ions relative to unwanted matrix ions by ejecting those unwanted ions from the trap with resonant ion ejection. The trapped sample ions have a characteristic frequency, which depends on the amplitude of the storage radio frequency field and the mass of an ion. The

(16)

Determination of Pesticides in Cannabis 10 characteristic frequency of an ion increases when the storage field increases. The frequency of an ion decreases when the mass of an ion increases[19].

2.3.3.2 Tandem mass spectrometry

Tandem mass spectrometry (MS/MS) is another possibility to detect samples in an ion trap. In an ion trap MS/MS, EI has three basic operations;

 ion formation and matrix ion ejection  precursor ion isolation

 product ion formation an product ion mass scanning

Ion formation and matrix ion ejection

EI ionizes the sample and matrix molecules. During the ionization, a multi frequency waveform is applied to the ion trap to eject ions below the specified precursor ion mass. It removes the unwanted ions from the trap, which improves the mass resolution. After ionization, a second waveform is applied to the trap, which removes ions with a higher mass than the specified precursor ion mass. As a result ions with the desired m/z value remain in the trap along with a few ions with an m/z value slightly below or above the desired m/z value[19].

Precursor ion isolation

The precursor ions can be isolated completely from the matrix ions in two steps. The first step ejects ions with masses lower than the precursor ions mass by ramping up the radio frequency field applied to the trap. The ions that have been ejected have masses up to and just below the precursor ion mass. The second step is to eject the all the ions with masses above the precursor ion mass. A second waveform will be applied to the trap to eject these unwanted ions. Isolation occurs at frequency levels with an optimal mass resolution[19].

Product ion formation and mass scanning

The product ions are formed from the precursor ions mainly by collision with helium gas. If the energy, released from the collision, is high enough one or more chemical bonds may break. After breaking of these chemical bonds, ions with a lower m/z than the precursor ion will be formed. When the product ions are formed, a single radio frequency is applied to the trap. This frequency creates a field which guides the ions through the trap into the electron multiplier[19].

2.3.4 MS coupled to a HPLC

The MS coupled to the HPLC is also an ion trap MS. This MS works with a different mechanism than the MS coupled to the GC. Figure 2-10 depicts a schematic figure of an ion trap spectrometer that can be coupled to a HPLC[20].

(17)

Figure 2-10: Schematic figure of the ion trap MS

The ion source that is used during this research is electrospray ionization (ESI). The electrospray interface generates the ions in the spray chamber (figure 2-11 and 2-12). In this chamber ESI follows four steps 1) formation of ions, 2) nebulization, 3) desolvation and 4) ion evaporation[20].

Figure 2-11: the spray chamber on the outside Figure 2-12: the electrospray chamber

2.3.4.1 Electrospray chamber

The ion formation occurs through different mechanisms. Ions can be generated in solution before nebulization, only when every aspect is correct, like the chemistry of the analyte, use of buffers. This can give a high analyte ion concentration. But it is not necessary to ionize the samples in solution before analysis. The nebulization creates an aerosol of the sample. The nebulizing gas enters the spray chamber through a tube that surrounds the needle, where the solution with sample comes from. The combination of

(18)

Determination of Pesticides in Cannabis 12 the gas and the voltage of the mesh electrode and the end plate breaks the solution into droplets. These droplets disperse into ions of one polarity, which results into a fine spray of charge droplets. If the droplets are allowed to enter the vacuum system, they will create a noise. This is prevented by aiming the stream orthogonally to the vacuum entrance. Before the ions can be analyzed, the drying gas evaporates the solvent by heating. This creates smaller droplets of sample that readily evaporate in the system [20].

2.3.4.2 Ion transport and focusing

The ions pass through the capillary from the sample chamber to the first vacuum space. The capillary has two primary functions, 1) the transfer of the ions from the spray chamber to the vacuum region and 2) a barrier that separates the atmospheric pressure region of the source from the vacuum region of the MS. The ends of the capillary help the ions to migrate towards the capillary entrance, and then pushed through the capillary. When the ions leave the capillary they are drawn to the skimmer and guided then through the octopoles and lenses. The skimmer, visible in figure 2-10, removes the drying gas. The ions pass then into the octopoles that focused and transports the ions to the ion trap. The two octopoles have different voltages to block out unwanted interference. The ions will then be guided into the ion trap for detection[20].

2.3.4.3 Detection

For the analysis of abamectin, the ion trap will be set on MS/MS. This means that only ions with a specified target m/z will be isolated from the sample, along with ions that have an m/z ratio slightly below or above the specified target m/z. For abamectin is the target mass, according to literature, 895 m/z. The mass of the abamectin molecule is approximately 872 m/z. The difference of the 23 m/z ratio corresponds to the mass charge ratio of sodium. Literature say that abamectin form easily an adduct with sodium. However, it is unknown if all the abamectin molecules present in a sample form a sodium-adduct, or only a part of the present molecules. This means that the 895 m/z corresponds to abamectin ionized by addition of sodium. To form more adducts of abamectin and sodium, a higher sodium concentration is the samples are needed. However, sodium is a non-volatile element that can block the sampling cone by precipitation. Therefore, sodium is a substance that can not be used continuously in the mobile phase[21]. To ascertain the identification of abamectin, MS/MS can be used to fragment the target m/z. If these fragments, as result of the fragmentation of the target m/z, occur on the same retention time as the target m/z, then the conclusion can be drawn that it is abamectin.

The theoretical background that was described here was necessary to conduct the investigation. The following chapters will describe the steps that were taken to conduct this investigation.

(19)

3 Materials and methods

This chapter provides the information about the materials and methods that have been used during this investigation. The pipettes that were used during this project were 100 µl, 200 µl and 5000 µl pipettes.

3.1 Preparation of hemp plants

The RIVM has received several hemp plants, which were obtained from illegal hemp plantations, from the police. The date of receipt of each plant was written in a log. The leaves and flowering tops were then cut from the plant with scissors and packed separately in small plastic bags. After sealing, the bags were labeled with a unique tag number of identification. The bags were then weighed on a Mettler Toledo Classic balance and stored at -20°C. The weight of each bag with sample was noted in the laboratory journal.

Besides these hemp plants, there were also three bags of medicinal hemp obtained from the Trimbos Institute, under the names of Bedrocan, Bedrobinol and Bediol. There were also eight coffee shop samples for determination of pesticides.

3.2 Standards for the analytical research

The standards for this investigation were 46 substances of pesticides. Four of these substances were substances regularly found by the police during dismantling of plantations. The other substances could be divided in three groups of illegal substances. The first four substances were abamectin, tebuconazole, imidacloprid and triadimenol. The three illegal groups of pesticides were organochlorides, organophosphates and poly chlorinated biphenyls. Abamectin was obtained from Sigma Aldrich, analytical standard HPLC grade 97.6%, CAS number 71751-41-2 and with batch number SZE923OX. Abamectin was obtained in powder form, and was only used for the LC-MS analysis. The rest of the standards were obtained from Accustandard (USA). The concentration, CAS number and batch number for each standard are described in Annex II. The standards from Accustandard were dissolved in methanol and delivered in glass ampoules.

3.3 Standards for the GC-MS analysis

For the analysis of the pesticides two methods were used. Chosen is for two methods, for comparison. The first method was developed in a previous study, determination of pesticides in tobacco. The second method was based on the paper, entitled “determination of multiclass pesticides in food commodities by pressurized liquid extraction using a GC-MS/Ms and LC-MS/MS”. The standards, with exception of abamectin were delivered in glass ampoules. After breaking of these glass ampoules, the pesticide standards were transferred into their respective transfer vials. To ascertain the applicability of the current GC-MS method (determination of pesticides in tobacco), the standards were diluted 1:50 with hexane (Biosolve, HPLC grade, batch number 63-33-1, CAS number 110-54-3) to a

(20)

Determination of Pesticides in Cannabis 14 final volume of 5 ml. After dilution, 100 µl of each diluted standard was transferred into a new vial to create a pesticide mix. From each standard and the mix 100 µl was transferred separately into GC-vials. All the GC-vials were placed in the autosampler of the GC-MS and analyzed by two methods. Annex III contains the settings for both methods. These methods are used for screening tests on the samples. When a sample was positive a calibration curve was made with different concentrations, for an indication of the amount of used pesticide.

3.4 Standard for the LC-MS analysis

For the LC-MS analysis only abamectin was used as standard, because this was the pesticide of interest. For the optimization of the MS, an undetermined amount of abamectin was dissolved in 10 ml acetonitril. The MS was optimized by using direct perfusion of abamectin in the ion trap with a syringe. The mass charge ratio of 895 m/z was strongly present in the solution, this means that this m/z is the target mass for optimizing of the MS. Table 3-1 reflects the optimized settings of the ion trap MS.

Table 3-1: The optimized settings of the ion trap MS

Part of optimization Value Unit

Capillary -3500 V

End plate off set -500 V

Nebulizer 15.0 Psi

Dry gas 10.0 l/min

Dry temperature 325 °C Skimmer 35.0 V Cap exit 360.0 V Octopole DC 1 13.88 V Octopole DC 2 1.73 V Trap drive 79.34 Octopole RF 300.00 Vpp Lens 1 -5.0 V Lens 2 -77.5 V

The eluent that was used for the analysis of abamectin was 80:20 acetonitrile with 0.1% formic acid and water with 0.1% with formic acid. the eluent was made with 1000 ml acetonitrile (Biosolve, France, HPLC-S, CAS number 75-05-8, lot number 757091) and 1 ml formic acid (Fluka, Puriss p.a. 98% eluent additive LC-MS, CAS number 64-18-6, lot number 1343727 22207026) and 1000 ml water (Biosolve, LC-MS, CAS-number 7732-18-5, Lot number 775841) with 1 ml of formic acid.

3.5 Sample preparation GC-MS analysis

Two methods were used to prepare the samples for the GC-MS analysis. The first sample preparation method was developed for the previous study of the determination of

(21)

Determination of Pesticides in Cannabis 15 pesticides in tobacco. The second method was with the DisQuE dispersive sample preparation kit from Waters, according to the QuEChERS method.

3.5.1 Sample preparation with the first method

From the samples 500 mg was weighed in and cut, with scissors, in little pieces, and transferred in a glass screw-cap tube. Due to the small amount of sample material, samples 75-78 were not weighed in for this investigation. These samples were coming from the same plantation as sample number 79 and 80, and are suspected of abamectin contamination. 10 ml of acetone (Merck Germany, for analysis, lot number K40096814 922, CAS number 67-64-1) was added to the screw-cap tubes with a glass graduated cylinder. The screw-cap tubes were mixed on a Coulter mixer (Coulter electronics Ltd, England), for 15 minutes and allowed to stand overnight. On the next day the mixing was repeated for 15 minutes and centrifuged in the Hettich Rotixa_P centrifuged (Depex B.V. de Bilt, Netherlands) at 23°C for 10 minutes at 3500 rpm. The supernatant of each sample was transferred to a new screw-cap tube using a funnel and circular filter paper. The acetone fraction was evaporated at 45°C with nitrogen to a volume of 0.5 ml in a sample concentrator from Techne. Subsequently, 3 ml of hexane and 4 ml of Milli-Q were added to the tubes. The tubes were shaken and centrifuged again, at 3500 rpm for 10 minutes at 23°C. After centrifugation the hexane fraction was collected and evaporated to a volume of 1 ml, figure 3-1 shows a photograph of a tube with the hexane layer on top and the water layer under it. The samples were now ready for clean-up with SPE.

(22)

Determination of Pesticides in Cannabis 16 For the clean-up, the Bond Elut Si 500 mg columns were conditioned with 2.5 ml of hexane. The columns contain a silica solid phase, also called silica bed. After the hexane reached the silica bed, and the top layer of hexane is equal to the top layer of the silica bed, the sample was brought onto the column. Figure 3-2 shows a photograph of the SPE columns with the samples in it.

Figure 3-2: SPE columns with the samples in it

When the level of a sample in the column was equal to the level of the silica bed, the hexane for the conditioning of the column was discarded. Then the possible pesticides were eluted from the column with 5 ml of hexane. The eluted samples were collected in new tubes and evaporated to dryness in the sample concentrator at 45°C with nitrogen. After evaporation 600 µl of hexane was added to dissolve the dry extract, to create more concentrated samples. The dissolved samples were stored in sealed 2 ml vials at -20°C.

3.5.2 Sample preparation with the second method

The second sample preparation method was performed with the DisQuE dispersive sample preparation kit from Waters. From the samples 10 gram was weighed in and cut, with scissors, in little pieces. To tube 1 (a 50 ml tube) 10 ml of acetonitrile (Biosolve, HPLC supra-gradient, CAS number 75-05-8, lot number 766421) was added. The tubes were shaken manually to create a suspension with the magnesium sulfate in the tube. The little pieces of hemp were then transferred to tube 1. The tubes were shaken and centrifuged for 3 minutes at 1500 rcf. 1 ml of the top layer of the samples was transferred to tube 2. These tubes were shaken for 30 seconds and centrifuged for 5 minutes at 10,000 rpm with the micro centrifuge (Centrifuge 5417R, Eppendorf). From every sample five GC-vials were filled with 100 µl of the top layer of tube 2. The samples were then ready to be analyzed.

(23)

3.6 Sample preparation for LC-MS analysis

The sample preparation for the LC-MS analysis was different than the preparation for the GC-MS analysis. Sample number 75-80 was suspected of abamectin contamination. Beside these samples, the coffee shops samples were also saved for the LC-MS analysis, due to the small amount of sample material. From the suspected samples, 1 gram of the samples was weighed and cut in to little pieces. These little pieces were transferred in a screw-cap tube. The extraction happened by adding 2 ml of acetonitrile (Biosolve, HPLC supra-gradient, CAS number 75-05-8, lot number 766421) to the screw-cap tubes. The tubes were mixed on the mixer (described in 3.5.1) and then centrifuged for five minutes at 2500 rpm with the centrifuged (described in 3.5.1). After centrifugation the supernatant of the samples was transferred separately in new screw-cap tubes. This extraction step was repeated, which made the final volume of supernatant 3-4 ml, for each sample. To the tubes with the supernatant, 10 µl of triethylamine (ICN, CAS number 121-44-8, lot number 8213A) and 6 ml of water were added. For the clean-up, the Bond Elute C-18 columns were used. These columns were conditioned with 3 ml acetonitrile and 3 ml mixture of (acetonitrile/triethylamine/water). This mixture was made by adding 96 ml of water to 64 ml of acetonitril and 160 µl of triethylamine, to a final volume of 160 ml. The samples were added to the column after conditioning, and then the columns were dried by vacuum. With 6 ml of acetonitrile the columns were eluted, and the eluted samples were collected and then evaporated to dryness with nitrogen at 50°C. After evaporation the samples were dissolved in 600 µl acetonitrile, and then ready for analysis.

(24)

4 Results

This chapter provides the results of the investigation of the determination of pesticides in samples, obtained from hemp plantations and coffee shops.

4.1 Preparation of hemp plants

The received hemp plants were prepared by cutting of the leaves and flowering tops with scissors. Figure 4-1 shows a paper bag with a hemp plant, the way the police delivered most of the samples. Some of the samples were already packed in plastic zip-lock bags.

Figure 4-1: a photograph of a hemp plant delivered in a paper bag

Figure 4-2 and 4-3 reflects how the plants were stripped of the tops and leaves and how they were packed. The packing occurred by putting the tops and leaves separately in small plastic bags which were then sealed with tape. These bags were labeled with a unique identification number that is reported in the laboratory journal.

(25)

Figure 4-2: The hemp tops were cut from the Figure 4-3: plastic bag with hemp tops. plant with scissors

In total there were 114 samples. 88 samples came from 44 plants and 26 samples were already packed in plastic bags. Annex IV contains the sample number and the weight of each bag. Annex IV contains also the weight for each sample for the sample preparation for analysis on the GC-MS. From the Trimbos institute, three bags of medicinal hemp were received, under the names Bedrocan, Bedrobinol en Bediol. These samples were also prepared for GC-MS analysis. The last samples were eight samples from coffee shops. These samples were saved for the LC-MS analysis of abamectine, because of the small amount of samples.

4.2 Standards on the GC-MS

From the standards 25 out of 45 can be determined on the GC-MS with the method developed from the previous study, determination of pesticides in tobacco. These standards are only a couple of several hundreds, and most of them belong to the group’s organophosphates, organochlorides and poly chlorinated biphenyls. Some of the standards, like the several Aroclors, were not compatible with the GC-MS method. This can be because this is a mixture of carbon atoms and chlorine atoms, with different proportions. With these determined standards a compound table was developed for the screening method. Table 4-1 reflects the compound table, with the specific retention time and mass-charge ratio for each compound.

(26)

Table 4-1: The compound table for the method, developed from the previous study determination of pesticides in tobacco. The table reflects the retention time and mass-charge ratio’s for each compound

Retention time Name compound Specific mass-charge ratio

8.158 Dichlorvos 109.0-185.0-187.0 9.955 Mevinphos 127.1-192.9-192.2 10.660 Imidacloprid 191.2-192.2-206.1 12.022 Naled 109.0-145.0-85.0-300.9 12.356 Phorate 231.0-120.9-97.0-259.8 12.544 BHC 183.0-81.2-218.8-216.9 12.758 Dimethoate 87.1-93.0-125.0-212.0 13.251 Lindane 183.0-181.2-218.8-216.9 13.273 Diazinon 305.2-304.1-137.3-276.1 13.626 Disulfoton 89.0-88.1-97.0-244.9 14.781 Methyl parathion 263.0-109.0-125.1-247.2 15.090 Heptachlor 272.2-274.1-100.2-336.7 15.704 Malathion 173.0-127.0-125.2-93.2 16.233 Aldrin 292.9-91.3-263.3-291.2 17.638 Heptachlor epoxide 353.1-354.8-356.7-351.3 18.055 Triadimenol 168.0-128.1-130.0-127.3 18.670 DDE 318.0-246.8-248.3-316.2 20.428 Dieldrin 277.1-279.0-281.0-263.3 21.461 Endrin 231.1-97.1-199.2-233.0 22.324 DDD 235.3-236.9-165.2-236.1 22.326 TDE 235.3-236.9-165.2-236.1 24.460 DDT 165.0-235.3-237.0-165.2 25.345 Tebuconazole 249.9-251.9-163.0-251.0 28.114 Methoxychlor 227.2-228.0-274.1-239.0 29.862 Phosalone 182.2-183.9-366.7-121.2 30.095 Azinphos-methyl 320.0-159.8-104.1-93.1

With the second method, developed according to the paper entitled “determination of multiclass pesticides in food commodities by pressurized liquid extraction using a GC-MS/Ms and LC-MS/MS”, 28 of the 45 standards can be determined. Table 4-2 shows the compound table for this method, with the specific retention time and mass charge ratio’s for each compound. Also with this method some of the standards were not compatible with the GC-MS method.

(27)

Table 4-2: The compound table for the method, developed according to the paper “determination of multiclass pesticides in food commodities by pressurized liquid extraction using a GC-MS/Ms and LC-MS/MS”. The table reflects the retention time and mass-charge ratio’s for each compound Retention time Name compound Specific mass-charge ratio

6.196 Dichlorvos 109.0-184.9-186.8-127.0 7.848 Mevinphos 127.0-191.9-192.7-164.0 8.843 Imidacloprid 191.1-192.0-205.8-163.1 11.076 Naled 109.0-145.0-184.9-147.0 11.726 Phorate 230.9-120.8-97.9-259.5 11.950 BHC 183.0-181.2-218.8-216.9 12.392 Dimethoate 87.1-93.0-125.1-212.0 13.260 Lindane 183.0-181.2-218.8-216.9 13.658 Diazinon 304.1-305.0-137.3-179.3 14.153 Disulfoton 88.1-89.0-97.0-141.8 16.482 Methyl parathion 262-.8-109.0-125.1-247.2 16.888 Heptachlor 272.2-274.1-100.1-270.3 19.012 Malathion 172.8-127.0-125.1-93.2 19.338 Aldrin 292.9-263.2-91.3-291.0 22.668 Heptachlor epoxide 352.9-354.8-351.2-356.7 24.215 Triadimenol 168.0-112.0-128.1-99.0 25.502 DDE 246.5-318.0-248.2-316.1 29.273 Dieldrin 277.0-278.8-263.0-280.9 31.878 Endrin 281.0-279.1-316.8-344.8 35.287 DDD 235.2-236.9-165.2-236.1 35.288 TDE 235.2-236.9-165.2-236.1 35.810 Ethion 230.9-97.1-232.8-125.1 40.859 DDT 235.3-237.1-165.2-236.2 43.014 Tebuconazole 249.9-125.1-252.0-127.0 49.099 Methoxychlor 227.3-228.0-274.-238.0 51.309 Phosalone 182.0-183.9-121.0 51.337 Azinphos-methyl 132.0-159.6-103.9-93.0 56.443 Coumaphos 3692-226.0-333.8-363.7

4.3 Standard on the LC-MS

Abamectin could only be determined on a LC system. Chosen is for a LC-system coupled to an ion trap MS. Based on the molecule the expected mass charge ratio was 872 m/z. But according to literature and analysis it occurs to be 895 m/z, shown in figure 4-4.

(28)

Figure 4-4: abamectin peak on 5.2 minutes. The spectrum underneath the peak is the mass spectrum of abamectin with sodium

The difference of the 23 m/z between the expected m/z and the analyzed m/z corresponds to the mass charge ratio of sodium. It is unknown of all of the abamectin molecules form an adduct with sodium or a part of the present abamectin molecules. This makes it difficult to determine the concentration. But for identification, this method is suitable.

4.4 Samples on the GC-MS

All the 114 samples, except for the samples number 75-78, have been prepared according to the sample preparation of the previous study, determination of pesticides in tobacco. Annex IV contains the precise weight of each analyzed sample. Sample number 56 contained after evaporation a white powder, which easily dissolved in hexane. This sample is also the only sample which has a peak that does not corresponds to a standard in the compound table of the two GC-MS methods. This peak also does not correspond to a peak of the plants. In the most chromatograms several peaks are on the same retention time, which indicates that these peaks are from the plant. The unknown peak occurred at a retention time of 18 minutes visible in figure 4-5. The other samples did not have an unidentified peak, or a peak that corresponds to the standards in the compound tables.

(29)

Figure 4-5: sample number 56, which had an unidentified peak

According to the library this peak was identified as pronamide/propyzamide. This standard is later injected on the GC-MS but this has a retention time of 13 minutes. This means that the sample has an unidentified peak.

(30)

Figure 4-6: the picture on the left reflects the positive sample. The picture on the right reflects the standard pronamide. The retention time of each peak is 18 minutes and 13 minutes respectively

It is necessary to conduct further investigation because there is a chance that several pesticides were not detected with this method. This will be further discussed in chapter 5. Two samples were prepared with the DisQuE dispersive sample preparation kit from Waters. The chromatograms of these samples, did not show good peaks, it was more a bulk of peaks with no separation between the peaks. There was also a lot of carry over of THC out of the samples. After analyzing of the sample, a pure diluted standard was analyzed, in this chromatogram a THC peak occurred. It appeared that there was some THC residue left in the liner and syringe. But after these components were replaced there was still a THC peak noticeable in chromatograms of standards. In total 40 blanco (hexane) runs were needed to clean the GC-MS. Based on this, the decision was made to prepare the plants only with the first preparation method.

4.5 Samples on the LC-MS

From the 114 samples from illegal plantations, 42 samples were prepared and analyzed on the LC-MS. The coffee shop samples were also prepared, but two out of the eight were lost during the preparation. Annex V contains the precise weight of each sample analyzed on the LC-MS. In total 48 samples were prepared and analyzed on the LC-MS. Six out of 48 are suspected positive for abamectine. These samples did have a peak at a retention time of 5 minutes and with a mass charge ratio of 895 m/z. The suspected samples were analyzed again, but then with fragmentation.

Figure 4-5 shows sample 60, which contains a peak at the 5 minutes. This peak has an m/z of 895 m/z which can corresponds to abamectin.

(31)

Figure 4-7: sample 60, with an isolation of 895 m/z. The peak is detected with a retention time of 5 minutes

For further investigation the m/z of interest was, after fragmentation of the 895 m/z, 751 m/z. Figure 4-6 reflects the chromatogram of this m/z in sample number 60. Because this is on the same retention time, it can be said that it is abamectin.

Figure 4-8: sample 60, with fragmentation of 895 m/z and isolation of 751 m/z. This shows a peak on a retention time of 5 minutes

To be certain of the presence of abamectin, the fragment 751 were fragmented to smaller m/z. The m/z of interest was 607 m/z. This fragment was also detected on a retention time of 5 minutes, show in figure 4-7.

(32)

Figure 4-9: sample 60, with fragmentation of 751 m/z and isolation of 607 m/z. This shows a peak on a retention time of 5 minutes

Figure 4-8, contains all the three results of sample 60 in one figure. The arrow indicates the sample with isolation of 607 m/z.

Figure 4-10: All three result of sample number 60. The arrow indicates the result of 607 m/z isolation

Because all three peaks, the target m/z and the fragmented m/z, occur on the same retention time, it is certain that this is abamectin. Abamectin was detected in six samples. Annex VI contains the results of the other five positive samples.

(33)

5 Discussion

This chapter provides the discussion of this investigation, the problems that have occurred and the solution.

The aim of this investigation was to determine whether pesticides were used in hemp plantation and, if yes which pesticides. The other goal was to quantify the amount of pesticides. The expectation was that if there was evidence of pesticides use, that it would be al legal substance.

Because there was no knowledge which kind of pesticides were often used in hemp plantations, several standards were obtained from three main pesticides groups, organophosphates, organochlorides and polychlorinated biphenyls. Some of these standards, like the several Aroclors, were not compatible with the GC-MS with the two methods. This can be because this is a mixture of carbon atoms and chlorine atoms, in different proportions. Because of this specific property of the substances, it was not possible to identify Aroclors with a screening run. The choice for a screening run was made based on the amount of samples. The expectation was that there would be 8-10 samples, but this proved to be 114 in total. Because of this large amount, a screening run is performed, and the expectation was to optimize the method when there was a positive sample. For this investigation a screening run for 25 good detectable pesticides was developed. Other problems that occurred will be discussed in the next sections

5.1 Sample preparation

Two types of sample preparation were applied during this investigation. The first method was developed in the study of pesticides in tobacco; the second was the DisQuE dispersive sample preparation kit from waters. The last one was discussed in several articles.

The first method was very useful for hemp plants. To determine if the method was applicable some of the samples were spiked with DDT, and this pesticide could be extracted from the hemp plants. The main disadvantage was the preparation of the samples, according to this method, lasted two days. During the first day the samples were weighed, acetone was added to the samples, mixed for 15 minutes and then leaved to stand overnight. The next day the rest of the preparation occurred. This method is based on hair samples, that is the reason why the samples were leaved to stand overnight in acetone, to destroy the hairs. Because of this preparation time the choice is made to look to other sample preparation. The most promising one was the DisQuE dispersive sample preparation kit from Waters. This method may be very useful for the extraction of pesticides in fruit and vegetables, for not for extraction in hemp plants. After the preparation, the sample extracts were a clear solution with a green color. After analysis with the GC-MS, lots of peaks were detectable, but with bulks of peaks or overlap of the peaks overlap of the peaks. Besides this, there was also carry over with THC. The standards were analyzed after the samples, and in this run, THC was identified in

(34)

Determination of Pesticides in Cannabis 28 standard. This was unexpected because it were pure standards, diluted 1:50 in hexane. The next step was to determine, how the THC peak could occur in a pure standard. At first several blanco’s (hexane) were analyzed, the result was that the THC peak was still detectable but smaller with each blanco. This meant that the THC either was present in the injection syringe, liner or column. The second step was to change the injection syringe and the liner from the GC-MS. Both were contaminated with THC. After this, 40 blanco’s were analyzed to reduce the THC to a minimum, and then the GC-MS were leaved on bake-out mode for the night.

Because of the many steps it takes to clean the GC-MS after analysis of the samples prepared with the DisQuE method, it is not worth it to use this method. The time it takes to clean the GC-MS is almost equal to the time it takes to prepare the sample according to the first sample preparation. To reduce the time, a cleaning step should be put between the DisQuE method and the GC-MS analysis. This method need to be tested thoroughly, with spiked samples with a pesticide with a known concentration, to analyze the recovery of the pesticide.

5.2 Comparison of the two GC-MS methods

Both methods are applicable as screening test, but the method from the previous study is in time a lot quicker. This run takes 43 minutes, while the other run takes 74 minutes. For a screening run the first method is very useful, although there could have been substances which were not detected. From the 500-600 pesticides that exist, only 45 have been analyzed by this method, whereas 25 were identifiable by this screening run.

The other method did identify 28 out of the 45 standard, but the runtime is a lot longer. Also with this method there is a change that there were substances that were not detected, only 45 standards were analyzed with this method.

However, both methods are screening methods. For an exact determination of the concentration of the used pesticides, both methods need optimizing. The aim of this investigation was also to determine the concentration, but this was not successful due to the large amount of samples. In total there were 114 samples, while the expectation was for 8-10 samples. With this amount only a screening test could be done for the determination, also the one positive sample was detected a few weeks for the end of the internship period.

5.3 Sample analysis

All the samples were analyzed according to first GC-MS method. From the 114 samples, one was positive for a substance that does not belong in the hemp plant. The library of the MS identified this as pronamide, but further investigation with the pronamide standard proved this not to be. A large amount of the samples were also analyze according to the second method. This method is also very useful but it takes al lot of time to analyze a

(35)

Determination of Pesticides in Cannabis 29 sample or a standard. The choice was then made to perform the quicker analysis of 43 minutes. This method did not give any other result than the first method.

For the analysis of abamectin only 48 samples were analyzed. Six out of these 48 were positive for abamectin, but there could be more samples contaminated with abamectin, which were not analyzed. Previous studies reported that abamectin has been used in hemp plantation. One study was performed by Rikilt, and the other one on the Avans University with coffee shop samples.

(36)

6 Conclusion

The aim of this investigation was to determine whether and if yes, which pesticides are used on cannabis in coffee shops. The other goal was to quantify the amount of used pesticides. The analytical research was performed on a GC-MS at RIVM in Bilthoven and on a LC-MS at Avans University in Breda.

With this investigation it is determined that there have been pesticides used on cannabis in illegal hemp plantations, but not the amount of pesticides. The GC_MS analysis has determined that one sample out of the 114 samples, from illegal plantations was contaminated with an unidentified substances. The library of the MS identified this as pronamide, but further investigation with the pronamide standard proved this not to be. The LC-MS analysis showed that six out of the 48 samples (42 from illegal plantations and six from coffee shops) were contaminated with abamectin. This was confirmed by fragmentation and isolation of the pesticide.

The method that has been developed for the study determination of pesticides in tobacco showed to be very useful as screening method during this investigation. However, for quantification of the amount of used pesticides, which was another part of the aim, the method needs further optimization.

Due to the large amount of samples, it was not possible to optimize the method for quantification. In total there were 114 samples from illegal hemp plants, eight samples from coffee shops and three samples of medicinal hemp from the Trimbos Institute. It was also difficult to optimize the method because there was no knowledge beforehand whether the samples were contaminated. When the one positive sample was detected with the GC-MS, the suspected standard was not available. When the standard was available only an identification analysis could be performed.

(37)

7 Recommendations

This investigation showed that many aspects need to be improved of adjusted for an optimal method and result. The goal of this investigation was to identify and quantify pesticides on hemp with an optimized method. This was not entirely successful, because of the amount of samples. To achieve this goal, various aspects of the investigation should be evaluated for improvement. The following sections describe possible modifications and improvements.

The main disadvantage of this investigation was that there was no knowledge beforehand which pesticides can be and are used for the grow of hemp plants. This was also the main reason to develop a screening test and when there was a positive result a specific test could be preformed. Only one sample out of a 114 was positive. The leaves of the plants were contaminated with pronamide according to the library of the MS. Further investigation with the pronamide standard proved that is was not pronamide

Six out of the 48 analyzed samples were contaminated with abamectin, determined with the LC-MS method. This can mean that there are more positive samples, so this method should be used more often for identification.

It would be useful to analyze a blanco hemp sample, where the certainty is that on this sample no pesticides were used. This is rather difficult to achieve because the samples come from illegal plantations. After analyzing a few samples some of the peak could be identified as specifically from the plants.

7.1 Sample preparation

Two different methods were used for the sample preparation, the method from the study determination of pesticides in tobacco and the DisQuE dispersive sample preparation kit from Waters. The main disadvantage of the first method is the preparation time. With this method the preparation time is two days. To determine if this can be shortened is to spike the samples with a known concentration, and to leave to stand over various times and not only overnight, and then measure the recovery of concentration of the method. With the recovery the minimum time for the extraction can be determined.

The DisQuE method can be evaluated to build in an extra clean-up step with SPE columns between the DisQuE method and the analysis on the GC. When the samples were directly analyzed after the DisQuE method, the GC-MS was very contaminated. It is possible that this can be prevented with an extra clean-up step. This improvement can also be determined with samples with a known concentration and then determine the recovery of the samples.

(38)

7.2 Sample analysis

There were two methods of GC-MS analysis, which could identify 25-28 pesticide standards from the 45 available standards. There is a chance that some pesticides were in the samples, which could not be determined by this method. One of the ways to determine the method on it ability to detected pesticides is to use other standard from different groups and analyze these standard. If they are detectable with the current method, then these standards need to be added to the current compound table. This will create a more complete compound table for a screening run. There are about 500-600 pesticides available, so to achieve this goal, these pesticides should be divided in different groups, and from every group there need to be two standards analyzed for the screening run.

The LC-MS analysis for abamectin is very useful for identification, but not for quantification. Abamectin forms easily a sodium adduct, but it is unknown how many of the abamectin molecules from a sodium adduct, and how many do not. Some of the paper written about abamectin, claims that the eluent need a salt to ensure that abamectin does not form a sodium adduct. However, salt in the ion trap mass spectrometer is not good for the equipment. Salts in the mobile phase or samples need to be investigated thoroughly. A good start for the investigation is the method described in the paper “Determination of abamectin and azadirachtin residues in orange samples by liquid chromatography-electrospray tandem mass spectrometry”. This article described a method for analysis of abamectin, where the samples obtain a sodium concentration of 60 mM. The extra sodium in the samples create more adducts of abamectin and sodium.

It can also be useful to include a LC-MS analysis for the determination of pesticides, because there are also pesticides that can be analyzed only on a LC-MS system. If this is not possible, there is a chance that there are some pesticides not detected, while they are present in the samples.