General enquiries on this form should be made to:
General enquiries on this form should be made to:
Defra, Science Directorate, Management Support and Finance Team,
Telephone No.020 7238 1612E-mail:[email protected]
SID 5Research Project Final Report
Note
In line with the Freedom of Information Act 2000, Defra aims to place the results of its completed research projects in the public domain wherever possible. The SID 5 (Research Project Final Report) is designed to capture the information on the results and outputs of Defra-funded research in a format that is easily publishable through the Defra website. A SID 5 must be completed for all projects.
A SID 5A form must be completed where a project is paid on a monthly basis or against quarterly invoices. No SID 5A is required where payments are made at milestone points. When a SID 5A is required, no SID 5 form will be accepted without the accompanying SID 5A.
· This form is in Word format and the boxes may be expanded or reduced, as appropriate.
ACCESS TO INFORMATION
The information collected on this form will be stored electronically and may be sent to any part of Defra, or to individual researchers or organisations outside Defra for the purposes of reviewing the project. Defra may also disclose the information to any outside organisation acting as an agent authorised by Defra to process final research reports on its behalf. Defra intends to publish this form on its website, unless there are strong reasons not to, which fully comply with exemptions under the Environmental Information Regulations or the Freedom of Information Act 2000.
Defra may be required to release information, including personal data and commercial information, on request under the Environmental Information Regulations or the Freedom of Information Act 2000. However, Defra will not permit any unwarranted breach of confidentiality or act in contravention of its obligations under the Data Protection Act 1998. Defra or its appointed agents may use the name, address or other details on your form to contact you in connection with occasional customer research aimed at improving the processes through which Defra works with its contractors.
Project identification
1.Defra Project code
PS2521
2.Project title
The Development of NMR Spectroscopy for the Quantification of Active Ingredients in Pesticide Formulations
3.Contractororganisation(s)
CSL
Sand Hutton
York
YO41 1CZ
54.Total Defra project costs
£35000
5.Project:start date
01 October 2004
end date
01 April 2005
6.It is Defra’s intention to publish this form.
Please confirm your agreement to do so.YES FORMCHECKBOX NO FORMCHECKBOX
(a)When preparing SID 5s contractors should bear in mind that Defra intends that they be made public. They should be written in a clear and concise manner and represent a full account of the research project which someone not closely associated with the project can follow.
Defra recognises that in a small minority of cases there may be information, such as intellectual property or commercially confidential data, used in or generated by the research project, which should not be disclosed. In these cases, such information should be detailed in a separate annex (not to be published) so that the SID 5 can be placed in the public domain. Where it is impossible to complete the Final Report without including references to any sensitive or confidential data, the information should be included and section (b) completed. NB: only in exceptional circumstances will Defra expect contractors to give a "No" answer.
In all cases, reasons for withholding information must be fully in line with exemptions under the Environmental Information Regulations or the Freedom of Information Act 2000.
(b)If you have answered NO, please explain why the Final report should not be released into public domain
Executive Summary
7.The executive summary must not exceed 2 sides in total of A4 and should be understandable to the intelligent non-scientist. It should cover the main objectives, methods and findings of the research, together with any other significant events and options for new work.
· The major objective was to evaluate the use of quantitative nuclear magnetic resonance (QNMR) spectroscopy for quantifying active ingredients in commercial pesticide formulations, and to consider its potential to complement the current liquid chromatography (LC) and gas chromatography (GC) methods.
· Four active ingredients were investigated: dimethoate, malathion, dichlobenil and metazachlor. Experiments were performed on the active ingredient and various commercial formulations. A total of 34 formulations were available that had previously been analysed using chromatographic methods. Formulations were obtained in various physical states: non-viscous liquids, viscous liquids and solids. Methods were developed to extract the active ingredient in a form suitable for NMR analysis. These are transferable to similar systems.
· Key choices in method development were the deuterated solvent for the NMR lock signal, an appropriate reference for quantification, and an extraction. It was found that commercial NMR solvents did not specify the reference concentration precisely enough for QNMR; therefore reference solutions for quantitatation were made in-house. The requirement to produce specific reference solutions for QNMR of pesticide formulations significantly hindered the progress of the project as a whole and the resolution of this issue within this project will expedite any future investigations.
· The active ingredients of the pesticide formulations used contained both carbon and hydrogen atoms, and were studied by 1H (as detailed in the CSG7) and by 13C QNMR techniques. Robust NMR parameter sets for both 1H and 13C QNMR experiments were developed.
· Calibration curves were found to be linear with a high degree of correlation for the purified active ingredients.
· QNMR spectra were obtained for the active ingredients and all formulations. A total of six measurements were obtained for each sample: a single aliquot was analysed three times to provide repeatability data and three aliquots were separately analysed to provide reproducibility data.
· The repeatability of 13C QNMR measurements was found to be lower than that of 1H QNMR in a number of cases and this was largely due to the lower sensitivity of 13C NMR. However, when uncertainty due to sample preparation was included similar levels of precision were obtained for both methods. The QNMR experiments produced data with precision of the same order of magnitude as chromatographic methods, although the latter generally show a slight advantage.
· The accuracy of the QNMR techniques was found to be lower than that for existing chromatographic techniques when assessed against the declared active ingredient concentration. The accuracy of each technique was estimated as an average deviation from the percentage of the active ingredient in the formulation. These were 10% (1H NMR), 5% (13C NMR) and 1% (chromatography). The accuracy of the NMR techniques varied from formulation to formulation being within 0.6 and 23% of the declared values.
· Chromatographic measurement times ranged from 8 to 20 minutes, and measurement time was around 15 to 30 minutes for 1H and one to two hours for 13C QNMR methods. Development time can be considerable for both QNMR and chromatographic techniques.
· A major advantage of the NMR technique over existing methods was the greater information content of the data sets recorded. These could be interpreted to determine the presence and approximate concentration of co-formulants and impurities as well as active ingredients. This often only requires a single data set.
· These results indicate that QNMR is a complimentary technique to chromatographic methods if used solely for the quantification of active ingredient. The technique also provides information that is not currently available relating to impurities and co-formulants.
· A secondary objective was to evaluate the potential of NMR to detect contamination in pesticide formulations. Emphasis on this part of the project was significantly reduced at the request of the Pesticide Safety Directorate (PSD) who indicated that quantification of active ingredients in the formulations should be the major focus of the project. However, a wide range of compounds was detected in the pesticide formulations by NMR spectroscopy and the composition of formulations containing common active ingredients was found to vary widely. For this reason and given the limited availability of standard formulations, it was not possible to model the composition of the formulations in the manner originally proposed. It was noted however, that the composition of specific formulations could be characterised by NMR and therefore it is reasonable to assume that the identities of unknowns could be determined using the NMR profiling techniques. Detailed characterisation of these compounds was outside of the scope of this project. The ability of NMR spectroscopy to detect and characterise a wide range of compounds in complex mixtures such as pesticide formulations is well established and CSL have particular expertise in this area. It is anticipated that a study focussing on the presence of these components would provide PSD with an effective mechanism for the characterisation of unknowns in support of regulatory activities relating to the composition of specified formulation compositions.
· The NMR spectra collected in this project demonstrated that QNMR’s main advantage over traditional chromatographic techniques was the broad range of compounds that were simultaneously detected in 1H and 13C NMR spectroscopy. This enables many other species to be profiled and quantified at the same time as the active ingredient or indeed the quantification of multiple active ingredients. In applications where this is required (with the caveat that the quantification accuracy may not be as high as for traditional techniques), QNMR would be the method of choice. The benefit to PSD of employing this technique would therefore be that impurities and active ingredients could be determined simultaneously. The approach taken could be to use a target list to rapidly screen the composition of a formulation against its specification or to undertake a more detailed interpretation of the data from first principles. The latter would doubtless be the case for the determination of unidentified ingredients in formulations.
· In addition to further refinement of the methods presented here, further work should focus on the strengths of the NMR technique for the determination of formulation ingredients that cannot currently be easily detected by other methods. These include polymers such as polyvinylpyrrolidone (PVP). NMR spectroscopy is also well suited for the confirmation of the presence or absence of a particular ingredient within a formulation when quantification is not the main concern. Here rapid experiment times and minimal sample preparation could be exploited. This approach would also provide a ballpark figure for the concentration of a particular component in the formulation, anticipated from this study to have an error of not more than ± 25%.
· The major role for NMR spectroscopy in the analysis of pesticide formulations is likely to be the detection and characterisation of those compounds that have not been declared in the formulation specification, but can be clearly detected and characterised by NMR spectroscopy. Efforts should therefore be focussed on ensuring that appropriate techniques and databases are established in support of regulatory activities in this area and in support of legal cases.
Project Report to Defra
8.As a guide this report should be no longer than 20 sides of A4. This report is to provide Defra with details of the outputs of the research project for internal purposes; to meet the terms of the contract; and to allow Defra to publish details of the outputs to meet Environmental Information Regulation or Freedom of Information obligations. This short report to Defra does not preclude contractors from also seeking to publish a full, formal scientific report/paper in an appropriate scientific or other journal/publication. Indeed, Defra actively encourages such publications as part of the contract terms. The report to Defra should include:
the scientific objectives as set out in the contract;
the extent to which the objectives set out in the contract have been met;
details of methods used and the results obtained, including statistical analysis (if appropriate);
a discussion of the results and their reliability;
the main implications of the findings;
possible future work; and
any action resulting from the research (e.g. IP, Knowledge Transfer).
INTRODUCTION
The nuclear magnetic resonance (NMR) measurement is based on the differences in nuclear spin energy levels that occur when certain nuclei, including 1H and 13C, are placed in a magnetic field. The resulting NMR spectrum contains peaks from all examples of a specific nucleus in the sample, present above the limit of detection. Each compound has its own characteristic pattern of chemical shifts and coupling constants based on its chemical structure. NMR has long been employed for compound identification, both on pure compounds and for identifying components in mixtures. However, NMR may also be used as a quantitative method to determine accurately the absolute concentration of different components within a mixture. This application is known as quantitative-NMR (QNMR) and requires careful attention to experimental parameters to be successfully employed.
The principles behind QNMR have been established for over two decades (Cookson & Smith 1984), and today it is extensively employed in such areas as the quantitation of lignins (Xia et. al. 2001, Holtman et. al. 2004) and analysis of copolymers (Adrianensens et. al. 2003). New developments of the technique are also being investigated such as its use for isotope ratio determination at specific chemical sites (Tenailleau et. al. 2004), which utilise its high level of precision. Of direct relevance to the current project is a recent QNMR study on the agrochemicals 2,4-dichlorophenoxyacetic acid and 2,2-dichloropropionate (Wells et. al. 2002), in which the method was found to compare favourably with the more common chromatographic methods of quantification. However, unlike the current study, this paper concentrated on these two agrochemicals, and looked at the purity of the compounds themselves, rather than their presence in commercial formulations.
This project had two major objectives. The first was to evaluate the use of QNMR to compliment existing methodology for the determination of active ingredients in pesticide formulations. This formed the major part of the work, and involved a number of distinct stages. Suitable samples were first obtained for investigation, both of the pure active ingredient and as present in a number of formulations. Sample preparation methods were then developed for each compound, and initial NMR experiments performed to obtain an appropriate QNMR parameter set. A calibration series was then developed using the pure compound, and the amount of active ingredient in each formulation quantified using 1H and 13C QNMR methods and a chromatographic method for comparison. The results and methodology from QNMR and chromatographic measurements were then compared and contrasted.
The second major objective was to evaluate the potential of NMR as a rapid screen to detect the presence of contaminants in pesticide formulations. Two approaches could be taken. The first was to use a reliable database of variation within the range of pesticide formulations. If the composition of a pesticide formulation is sufficiently well defined, and a sufficiently large database of spectra are acquired from standard formulations, then various established multivariate and pattern recognition techniques can be employed to rapidly determine whether a formulation contains unacceptable contamination. The second approach utilises the high information content of the NMR data to determine the identity of the constituents of the formulation. Typically this can be achieved by interpretation of the NMR data (using for example J-couplings and chemical shifts) or by reference to a library of spectra from potential contaminants. Frequently these approaches are used in conjunction once a spectral library has been developed.
METHODS AND MATERIALS
Samples and materials
A total of four different active ingredients were investigated in a number of formulations: dimethoate, malathion, dichlobenil and metazachlor. For each pesticide type, the pure active ingredient was obtained from a reputable chemical supplier. A number of commercial preparations containing the active ingredient were also obtained. These came from various suppliers, and had been collected during previous surveillance exercises performed by CSL. The commercial preparations contained other ingredients in addition to the active ingredients, such as solvents, dyes and emulsifying agents. There were a total of 8 dimethoate, 3 malathion, 12 dichlobenil and 11 metazachlor formulations.
These pesticide formulations had previously been characterised by chromatographic methods, and a series of reports provided to PSD: Dichlobenil (FD 03/23); Metazachlor (FD 03/30); Dimethoate (FD 03/34); and Malathion (FD 03/35). The dichlobenil, dimethoate and metazachlor surveys were all analysed using CIPAC methods, which are the international referee methods for formulations analysis.
All other supplies for this project, including deuterated solvents, were obtained from reputable laboratory suppliers.
NMR sample preparation
Choice of solvent
The addition of a deuterated solvent was required to: a) provide a deuterium lock for the NMR signal; b) reduce sample viscosity thus ensuring good NMR line widths; and c) to contain the reference compound in a controlled amount. Pesticide formulations contain various dyes and emulsifying agents that cause solubility problems when dissolved in many common solvents. Therefore, for each active ingredient the solvent was carefully chosen following a number of preliminary tests and considerations:
· Determination of the miscibility of the solvent with the formulation. This is considered with reference to published solubility data, and assessed visually by checking for separation into distinct layers; the milky appearance of an emulsion; or precipitation of ingredients over time.
· Solubility of the internal standard in the prepared sample. This was an issue for Metazachlor formulations, where although the formulations were largely soluble in pure methanol, the internal standards TSP and trioxane precipitated from solution following mixing with the formulations.
· Chemical reaction between the solvent and the active ingredient. This is assessed by theoretical consideration of the species involved, and through careful observation during mixing and from interpretation of the resulting NMR data.
Choice of reference standard
An accurately determined concentration of reference standard was required to ensure full quantification of the active ingredient. In the initial stages of this project problems were identified with commercially supplied solvents spiked with the common NMR standard tetra-methyl-silane (TMS). Although the quoted concentrations were adequate for use as a chemical shift reference, they were not reliable for use in quantification. It was therefore necessary to make up a reference standard in-house. However, TMS is highly volatile, and could not be reproducibly prepared to an accurate concentration. A standard that was soluble in the chosen solvent, with a single, distinct peak occurring in a spectral region devoid of other compounds (as assessed in a preliminary spectrum of the formulation) was required. In the current study both trioxane (following Xia and Akim, 2001) and TSP (2,2,3,3-d(4)-3-(trimethylsilyl)propionic acid) were used as reference standards for different pesticides.
The concentration of the reference standard required was also an important consideration. If too much or too little was added, measurement inaccuracies would occur when comparing the peak areas of the reference material to that of the active ingredient. In the current study, the correct concentration was assessed by acquiring a preliminary test spectrum of the formulation to determine the intensity of the NMR signals. This allowed the appropriate concentration of the reference to be calculated.
Choice of preparation method
The final task of the method development was to determine the extraction protocol that would be used to obtain the active ingredient
The methods that were used in the current study are detailed below:
· Preparation of dimethoate formulations: 300 μl of dimethoate preparation was added to a 5mm NMR tube using an automatic pipette. To this was added 300 μl of DMSO containing 250 mM trioxane. The NMR tube was capped, and the solution thoroughly mixed using a vortex mixer prior to analysis.
· Preparation of Malathion formulations: 400 μl of deuterated acetone containing 266 mM trioxane was added to 200 mg of malathion formulation. The solution was thoroughly mixed, and the solution was added directly to a 5mm NMR tube for analysis. · Preparation of Dichlobenil formulations: The dichlobenil formulation consisted of approximately 6% active ingredient in an insoluble clay matrix. The matrix was first ground to a fine powder in a pestle and mortar, and homogenised by shaking. 200 mg of this powder was then mixed with 800 μl of deuterated acetone spiked with 81 mM trioxane. This was agitated for 30 minutes in a sonic bath to extract the active ingredient. The sample was then centrifuged to separate out the insoluble material, and the supernatant added to an NMR tube for analysis. Following 1H NMR analysis of the formulations chromium(III) acetylacetonate (2 mg in 20 μl of deuterated acetone) was added to each extract to reduce the relaxation delay for 13C QNMR analysis.
· Preparation of Metazachlor formulations: To 200 mg of the formulation was added 600 μl d6 DMSO spiked with 127 mM trioxane. Each solution was agitated for 15 minutes using a sonic bath, followed by centrifugation to remove any insoluble material. The supernatant was decanted and placed into an NMR tube for analysis.
NMR protocol
All NMR experiments were carried out on a Bruker ARX-500 spectrometer. The following procedure to determine the optimum parameter set for quantitative NMR was followed for each type of pesticide. This was conducted on a sample of the pure active ingredient in the chosen solvent.
13C NMR spectra were acquired using an inverse gated decoupling pulse sequence (Freeman et. al. 1972). 13C spectra were acquired with a central frequency of 125.773 MHz. 128k data points were acquired with a spectral width of 361 PPM, giving an acquisition time of 1.45 s. A total of four dummy scans and 64 acquisition scans were used. Decoupling was based on a soft pulse length of 32 μs. Values for the P1, D1 and total experimental time are given in Table 1.
A standard 1D pulse sequence was used for the acquisition of 1H NMR spectra. These were acquired at a central frequency of 500.140 MHz. 64k data points were acquired with a spectral width of 20.8 PPM, giving an acquisition time of 3.15 s. A total of four dummy scans and 128 acquisition scans were used. Values for the P1, D1 and total experimental time are given in Table 1.
P1/μs
D1/s
experiment time/mins
Pesticide
1H
13C
1H
13C
1H
13C
chrom.
Dimethoate
3.2
6.9
10
70
29
93
8
Malathion
3.2
7.5
12
80
34
81
20
Dichlobenil
3.1
7.6
12
40
34
40
10
Metazachlor
3.2
7.6
12
40
33
47
14
Table 1: Critical experimental parameters: D1 and P1 for QNMR, and total experiment time for all experimental methods.
The data from both 1H and 13C experiments were processed using Bruker’s XWIN-NMR software. The data were Fourier transformed, manually phase corrected and subjected to a polynomial baseline correction. A line broadening factor of 1 Hz was applied to all 13C NMR data before Fourier transformation.
Development of quantification methodology
Two quantification methods were evaluated: calculation of pesticide concentration based on the mathematical relationship between peak areas and the application of calibration curves. Both methods use the ratio between the areas of the reference standard and a peak resulting from the pesticide. For each pesticide a peak was selected for quantification using a number of criteria to ensure that the measured area would be as accurate as possible: i) where possible a singlet resonance was chosen; ii) the resonance was in a ‘clean’ area of the spectrum, well removed from other resonances from both the pesticide and other constituents of the formulations (as identified in preliminary investigations); and iii) the peak was of known assignment, so the number of atoms per molecule was available.
Two methods for measuring the peak areas were also investigated: the integration and deconvolution. The integration software sums spectral intensity over a range of frequencies. It can also allow a ‘slope’ and ‘bias’ correction to be performed over the region to compensate for any underlying baseline shift. Deconvolution of the NMR peaks aids the resolution of overlapping peaks and this technique has been used for quantification throughout. The definition of deconvolution of NMR spectra may differ from that used for other techniques. As the NMR signal is presented in the frequency domain the NMR signal follows a characteristic distribution of transitions about a mean value. This distribution gives the NMR signals their line-shape and this line shape follows a mixture of a Lorenztian and Gaussian distribution. Deconvolution in NMR terms therefore means fitting a mixed Lorentzian-Gaussian line to an NMR peak and using this fitting to determine the peak area from the equation that defines the fitted peak shape. The technique reduces deviations from the ideal lineshape due to, for example, neighbouring peak tails or low digital resolution of the NMR signal.
Quantification using a simple calculation based on the mathematical relationships between the peak areas and the compound concentrations was assessed. To calculate the pesticide concentration, CP, required the ratio of the pesticide to reference peak areas (AP/AR), the concentration of the reference standard, CR, and the number of atoms per molecule for both the reference, NR, (e.g. 3 carbon atoms, 6 hydrogen atoms for trioxane) and the pesticide, NP. These are related by the following equation:
P
R
R
P
R
P
*
*
N
N
A
A
C
C
=
Quantification using calibration relied on making an accurate series of pesticide solutions at known concentrations in a solvent containing a fixed level of reference standard. These were subject to the NMR experiment using the parameters chosen for QNMR. The ratio of peak areas was then calculated, and plotted against the known concentrations. These plots were then used to produce a fitted calibration curve, which was used to calculate the concentration from any peak ratio.
The reproducibility of these methods was assessed by application to fresh mixtures of pesticide and reference standards at known concentrations. The calibration curve method was found to be superior both in terms of reproducibility and in avoiding concentration dependent changes. Calibration was therefore adopted for all of the result presented in this report. Although it is preferable that the unknown concentration lies within the range used for calibration, if the curve is well characterised, extrapolation is also possible.
The calculation method had the advantage that it could be rapidly employed without the need for preliminary experiments. Although not as accurate as the calibration method, it is adequate for most applications. However, it should be noted that this method could only be used when an assignment of the pesticide peaks was available, so that the number of atoms representing each peak was known. The time taken to conduct an assignment could thus be more profitably spent on running a calibration series.
Quality control
The NMR spectra were subject to a series of quality control tests to ensure that the NMR spectrometer was operating efficiently. All spectra were monitored for line width and line shape. Those exhibiting poor line width/shape were rejected and the data re-acquired. Line width criteria used were 1.7 Hz maximum for 1H and 1.5 Hz for 13C measurements. Quality control was also applied during the quantification measurements. This was achieved by interspersing samples from the relevant calibration series before and after repeat measurements on the formulations. The newly calculated concentrations of these samples were compared to the original values to check for experimental drift or error and were found to be consistent throughout.
RESULTS AND DISCUSSION
Calibration data
Calibration series were obtained for all four of the pure pesticides. NMR spectra for the metazachlor calibration series are presented in Figure 1. These data show the increase in intensity and the decrease of the signal to noise ratio as the concentration of active ingredient was raised. The ratio of these peaks was plotted against the known concentration of active ingredient to form the calibration plots as shown in Figures 2 to 9.
Figure 1: Metazachlor 1H (A) and 13C (B) calibration series. Internal standard peaks are present at 5.2 ppm and 95 ppm in the 1H and 13C spectra respectively. The concentration of the pesticide is shown above and to the right of each spectrum.
The calibration curves were subjected to linear regression with the line restrained to pass through the origin. These lines are displayed on Figures 2 to 9. Table 2 presents the gradient of these lines for both 1H and 13C calibrations and all four pesticide standards. These were the values used to calculate concentrations from the ratio of NMR peak areas for all subsequent results. The measured ratio is multiplied by the relevant tabulated value to obtain the concentration of active ingredient.
compound
expt.
gradient/mMol
R2
SEc/mMol
Dimethoate
1H
540.7
0.9989
7.6
Dimethoate
13C
419.9
0.9986
14.9
Malathion
1H
212.8
0.9946
16.3
Malathion
13C
662.8
0.9943
16.9
Dichlobenil
1H
1152.2
0.9982
6.6
Dichlobenil
13C
118.7
0.9747
25.4
Metazachlor
1H
121.8
0.9962
14.0
Metazachlor
13C
390.9
0.9938
17.8
Table 2: Data on calibration curves: gradient, coefficient of determination (R2) and the standard error of the concentration (SEc).
Table 2 also presents the coefficient of determination (R2) and the standard error of the concentration (SEc) for each calibration. The R2 value compares estimated and actual y-values, and ranges from 0 for no correlation to 1 for full correlation. The SEc value gives information on the accuracy of an individual concentration calculation, based on the deviation of individual points from the trend line. The measured concentration should fall within ±2*SEc of the true concentration in 95% of cases.
The main observation from the calibration data was that the plots were all strongly linear, with R2 values close to 1. The SEc values were also very low, suggesting that the calibration would be accurate down to the lowest concentrations measured. Also notable was that the R2 values for both 1H and 13C calibration curves were comparable, suggesting that neither method had the advantage in this respect. As a consequence of this information, the calibration curves could be used with confidence in the quantification of active ingredients. However, some consideration must be given to the presence of components other than the active ingredients in the formulation as these may hinder quantification due to, for example, interactions and/or overlapping signals.
0
0.2
0.4
0.6
0.8
1
1.2
1.4
0
200
400
600
800
Concentration (Moles/Litre)
Ratio of Pesticide to Trioxane peak
areas
Figure 2: Dimethoate 1H calibration series.
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
0
200
400
600
800
Concentration (Moles/Litre)
Ratio of Pesticide to Trioxane peak
areas
Figure 3: Dimethoate 13C calibration series.
0
0.5
1
1.5
2
2.5
3
3.5
0
100
200
300
400
500
600
700
Concentration (Moles/Litre)
Ratio of Pesticide to Trioxane peak
areas
Figure 4: Malathion 1H calibration series.
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0
100
200
300
400
500
600
700
Concentration (Moles/Litre)
Ratio of Pesicide to Trioxane peak
areas
Figure 5: Malathion 13C calibration series.
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0
100
200
300
400
500
Concentration (Moles/Litre)
Ratio of Pesticide to Trioxane peak
areas
Figure 6: Dichlobenil 1H calibration series.
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
0
100
200
300
400
500
Concentration (Moles/Litre)
Ratio of Pesticide to Trioxane peak
areas
Figure 7: Dichlobenil 13C calibration series.
0
1
2
3
4
5
6
0
100
200
300
400
500
600
700
Concentration (Moles/Litre)
Ratio of Pesticide to Trioxane peak
areas
Figure 8: Metazachlor 1H calibration series.
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
0
100
200
300
400
500
600
700
Concentration (Moles/Litre)
Ratio of Pesticide to Trioxane peak
areas
Figure 9: Metazachlor 13C calibration series.
Appearance of NMR Spectra
Spectra fulfilling all quality assurance criteria were obtained for all standards and formulations. A complete assignment of the NMR spectra for each pesticide was not necessary for quantification and no assignment was necessary when a calibration curve was used for quantification. However, when an assignment was desired, a range of 2D NMR experimental techniques was used to determine chemical structure. Figure 10 presents the 2D HSQC experiment used in the assignment of dimethoate. The cross peaks in this spectrum identifies the connectivity between 13C and 1H atoms. This is the approach that would be taken to determine the identity of unknown compounds in pesticide formulations.
Figure 10: 13C-1H HSQC spectrum of dimethoate standard showing connectivity between 1H and 13C resonances. (Peak labels correspond to numbering on chemical structure.)
Figure 11 presents a representative NMR spectrum for the dimethoate formulation. It is immediately clear that both the active ingredient can be clearly identified, and that there are a large number of additional peaks. Regions containing many such low level impurities in both the 1H and 13C spectra are shown expanded in the figure. The number of additional peaks and their intensities vary with the formulations, indicating a great diversity in the composition of the commercial products. The majority of additional compounds were present at much lower concentrations than the active ingredient, although some intensities were comparable. Potential overlap with the long tails of neighbouring peaks of similar intensity may result in inaccurate quantification due to the combined contribution of both signals. This is likely to be more of an issue for 1H when compared to 13C NMR spectroscopy. 13C NMR spectroscopy offers greater spectral resolution and therefore the likely incidence of overlapping signals is significantly reduced. However, the reduced sensitivity of the 13C technique may offset resolution improvements as the reduced signal to noise ratio will result in longer acquisition times and/or higher spectral variability.
A large number of spectra were acquired during this study. A database of these spectra is available on request.
Figure 11: 1H (A) and 13C (B) spectra for a dimethoate formulation. Regions relating to low level impurities are expanded.
It was important for quantification that the peaks used for the calculation were in isolated regions of the spectra, away from the non-active ingredients. This was checked in a preliminary acquistion of the formulation before both the calibration series and the main quantitative experiments were performed. This would only present a practical problem for future quantification using these calibration curves if a new formulation contained a radically different additive. Even in this case however, it would be possible to reanalyse the calibration data to obtain ratios for a different NMR peak. The measurements could also be performed at a different temperature resulting in further resolution of the NMR resonances.
Calculated concentrations of pesticide formulations
For each pesticide formulation, a single aliquot was measured using the 1H and 13C QNMR methods on three separate occasions, allowing the repeatability of the NMR measurement to be determined. In addition, for each pesticide formulation a total of three separate aliquots were subject to analysis, in order to check the reproducibility of the method. The means, repeatability standard deviations, and reproducibility standard deviation for both 1H and 13C results were calculated for each formulation. This data is presented in Table 3. The GC results previously measured and reported by CSL are also given in this table, along with the declared concentrations of active ingredients as supplied by the manufacturers. The repeatability data for chromatography were obtained by duplicate injections of material from a single aliquot, whilst the reproducibility data were obtained from the measurement of multiple aliquots.
The resources and time allocated in this project were primarily focused on addressing the accuracy of the NMR measurement. It was found that the accuracy of the NMR method was significantly worse than the precision. Therefore, accurate determination of the precision was not thought worthwhile at this stage, as simply adjusting the acquisition time would almost certainly improve the precision of the method. The methods presented require further modification prior to application for regulatory measurements and once finalized these would undergo full validation. It was not thought useful to fully test a method (using 5 repeats) when the method would need to be further modified and therefore 3 repetitions were performed per sample.
The tabulated data allow a number of comparisons to be made. A comparison of the repeatability data to the reproducibility data shows consistently better repeatability than reproducibility for the 1H QNMR and the chromatographic methods using malathion and dimethoate, indicating that the precision of the analytical technique itself does not contribute greatly to the final variability in these cases. The results for the 13C QNMR method using malathion and dimethoate and all methods using dichlobenil and metazachlor are more varied. The high level of repeatability that is inherent in the NMR measurement suggests that it is unlikely that measurement is the source of this variation, but probably reflects the ability of NMR to capture information dynamically, and therefore not simply present a snap-shot at any given time-point. The different behaviours for different pesticide types is unsurprising due to the quite different preparation protocols employed. When lower repeatability for the 13C method is observed, it is most likely linked to both the lower sensitivity of 13C NMR, and the longer 13C relaxation times which permit fewer scans to be acquired in a given time (reducing the signal to noise ratio of the measurement when compared to 1H NMR).
The repeatability and reproducibly errors are given as %w/w of active ingredient in the formulations. They are not expressed as %RSD as the precision values for the NMR data approximate to fixed errors and therefore the %RSD decreases with increasing active concentration or run time. For 13C QNMR the limiting factor for the repeatability is the 13C NMR sensitivity and therefore the repeatability and the reproducibility are often approximately the same. For 1H NMR the repeatability is often significantly better than the reproducibility and thus the limiting factor here is the way in which the formulation has been prepared for analysis. Factors affecting this include interactions and errors due to the extraction procedure.
Reduced experiment times decrease the precision of the NMR methods but are not likely to affect the accuracy. To investigate the effect of reduced run times fully would require the analysis of a range of formulations with different concentrations of active ingredient. The NMR data acquisition time is related to the sensitivity of the NMR measurement and therefore also to the reproducibility. Very short acquisition times give larger errors, but if time is the limiting factor and not the precision of the measurement, then rapid acquisition times could be further investigated. The acquisition times chosen here for 1H NMR are roughly comparable with typical chromatographic run times. The exact relationship between the NMR sensitivity and the experiment time is that the signal to noise (S/N) ratio is proportional to the square root of the number of scans acquired. Each scan takes a finite time to acquire and this is determined during instrument optimisation. The worked example below gives an insight into the likely time reductions, and associated increase in error, that can be achieved. The calculation uses the errors associated with the determination of the dichlobenil content of a formulation with a declared content of 4%. The S/N ratio and the data acquisition times of the spectra associated with that analysis are then used to proportionate the S/N ratio and the acquisition times of the data from the determination of dimethoate (37.4%) and thus calculate the reduced experiment time. The result is a hypothesised error and time based on a minimum S/N ratio for the determination of dimethoate at a concentration of 37.4%. The calculated experiment time is based on a target signal to noise ratio of 1941.81, being the lowest value that was used for the determinations presented in this report (for sample number 42629). To achieve this requires the acquisition of multiple scans (NS). It is therefore necessary to insert an inter-scan delay of 10 seconds to ensure that the system has returned to equilibrium prior to the next acquisition. The acquisition time (AQ) per scan (13.145 s) therefore includes this delay. For single scan experiments this delay is not required and acquisition of quantitative data can be performed in 3.145 s using this experimental setup. However, the S/N ratio will be reduced and will thus lower the precision of the measurement. Further alteration of the acquisition parameters using a single scan experiment will result in acquisition times of less than 1 s. These are associated with low-resolution data and are only appropriate for formulations that do not contain significant impurities that have similar chemical structures to the active ingredient. As all of the formulations analysed were found to contain impurities at similar chemical shifts to the resonance positions of the active ingredients, high-resolution spectra were acquired throughout.
Dichlobenil (42629): S/N = 1941.81
Dimethoate (43390): S/N = 5526.81, NS = 128, AQ = 13.145 s
S/N = k*√NS
Therefore kdimethoate (proportionality constant) = S/N÷√NS = 5526.81÷√128 = 488.5
A S/N ratio of 1941.81 requires (1941.81÷488.5)2 = approximately 16 scans.
The total acquisition time is therefore AQ×NS=13.145×16 =210.32 seconds or 3 mins 30 secs.
The error for the dichlobenil determination can thus be imparted on this reduced acquisition time for the determination of dimethoate, so a signal to noise ratio of 1941.81 for a 37.4% dimethoate formulation should be achieved in 3 mins 30 secs with an associated error of 13% for repeatability. The reproducibly figure for the 37.4% dimethoate extraction is 1.8%. This would not be affected by the reduced run time, however the experimental repeatability would become the limiting factor using short run times and therefore an error of approximately 13% on repeated analyses would be anticipated.
A single scan experiment with an experiment time of 3.145 seconds would produce a S/N ratio of 488.5 for the active ingredient dimethoate present at 37.4%.
This calculation provides a theoretical value. In practise the experimental set-up for the formulation and the contents of the formulation that do not relate to the active ingredient will largely determine the reproducibility of the method. It should be noted that when peak overlap occurs in NMR spectra, it is usual to alter the experimental conditions and in particular the temperature and pH. This will shift the peak positions often separating overlapped peaks and this is analogous to altering the solvent or temperature gradient when using LC or GC respectively.
Comparing the QNMR data to the chromatographic data revealed that both the repeatability and reproducibility variation in chromatographic techniques was generally slightly less than that from QNMR methods. This was not exclusively the case however, and QNMR and chromatographic methods had precision with the same order of magnitude. On comparing the accuracy of calculated concentrations between methods and against declared values, it was found that there was greater consistency between the two different NMR methods than between NMR and chromatographic results for all cases except when analysing metazachlor. Metazachlor determination showed higher 1H values than both 13C and chromatography and this may have been related to the problems in finding a good preparation protocol. In general the chromatographic results were closer to the declared value than the NMR measurements for all cases except dichlobonil, where results were more varied. It is also worthwhile to note that the pesticide for which the preparation methods for QNMR and chromatography were most closely matched was dichlobenil. A comparison of the NMR based methods and current chromatographic techniques is given in table 4.
It is the author’s opinion that the 1H and 13C NMR techniques are not currently sufficiently accurate for routine regulatory implementation when assessed against the declared content of the active ingredient present in a pesticide formulation. Further effort would be required to produce a reliable protocol for the quantification of active ingredients in any given formulation by NMR spectroscopy in the presence of other formulation components. As with chromatographic methods, the physical form of the formulation (e.g. clay, liquid, solid) would need to be considered on a case by case basis to generate an NMR protocol which approaches the accuracy of current chromatographic methods. The accuracy of quantification by NMR spectroscopy is usually stated as approximately ± 1% for pure compounds and this was clearly supported by the calibration curves presented in Figure 3 to 9. However, the results generated here showed a significant deviation in the accuracy of the measurement from this value, with, in the most extreme case, an NMR generated value exceeding the declared content of a single formulation by 23% and in the most accurate case the NMR value being below the assigned value by 0.6%. The underlying reasons for this varying accuracy were likely to be primarily due to signal overlap and interactions between the quantification standards and the formulation. It may also be the case that the behaviour of the active ingredients in the formulations was different to that for the standards used to produce the calibration curves. It would be possible to resolve all of these issues given the resources, but it is not anticipated that NMR would be as accurate as current methods for the rather limited role of quantifying active ingredients. The results presented here clearly show that NMR spectroscopy was able to rapidly detect significant deviations from the formulation specification relating to declared active ingredients, however, much smaller deviations in the formulations relating to undeclared ingredients can be detected and characterised if not quantified to the high degree of accuracy required for regulatory purposes.
Development costs for NMR analysis are comparable with those associated with developing traditional methods of analysis (i.e. hours to weeks) and are heavily dependent on the physical and chemical characteristics of the formulations. Many general issues (such as the reactivity of common NMR standards/solvents and the formulations) would not pose the same degree of difficulties in subsequent studies as these were addressed here. Costs associated with routine NMR analysis are also dependant on the formulation and may be minimal when dissolution is all that is required. These costs would also mirror those required for conventional analysis methods.
Both NMR and traditional (chromatographic) methods can be automated and the time scales for analysis are therefore comparable. Until such time that the development work has proceeded to the point where agreement on the applicability of the NMR methods can be reached, some uncertainty will remain as to the overall cost of the analysis. It is not anticipated that the NMR methods will replace the routine regulatory tests, but will compliment these methods by providing additional and supporting information. In terms of analysis time, a ballpark concentration of the active ingredient could certainly be obtained by NMR more quickly than by traditional methods. However, further evaluation of the errors associated with the rapid methods would be required prior to a commitment on the time and costs associated with these analyses.
In addition to further refinement of the methods presented here, further work should focus on the strengths of the NMR technique for the determination of formulation ingredients that cannot currently be easily detected by other methods. These include polymers such as PVP. NMR spectroscopy is also well suited for the confirmation of the presence or absence of a particular ingredient within a formulation when quantification is not the main concern. Here rapid experiment times and minimal sample preparation could be exploited. This approach would also provide a ballpark figure for the concentration of a particular component in the formulation, anticipated from this study to have an error of not more than ± 25%.
The major role for NMR spectroscopy in the analysis of pesticide formulations is likely to be the detection and characterisation of those compounds that have not been declared in the formulation specification, but can be clearly detected and characterised by NMR spectroscopy. Efforts should therefore be focussed on ensuring that appropriate techniques and databases are established in support of regulatory activities in this area and in support of legal cases. Figure 11 clearly shows the presence of these impurities. The interpretable nature of the NMR data, as highlighted in table 4, ensures that it is an excellent tool for characterising these impurities. This concept was not further pursued here due to the reduced emphasis on this aspect of the proposal as directed by PSD at the onset of this work.
Declared
Mean value (%w/w)
% declared
Std_Repeatability
Std_Reproducibility
code
sample
ingredient
(%w/w)
1H
13C
Chrom.
1H
13C
Chrom.
1H
13C
Chrom.
1H
13C
Chrom.
45559
Rogor L40: C.
dimethoate
37.40
38.61
38.83
37.48
103.2
103.8
100.2
0.067
0.806
0.150
0.662
0.372
0.610
45561
BASF Dimethoate 40: B.
dimethoate
37.40
36.15
36.84
34.35
96.6
98.5
91.8
0.085
0.283
0.071
0.037
0.294
0.287
45866
Danadim: B.
dimethoate
38.00
42.86
43.18
38.63
112.8
113.6
101.6
0.024
0.620
0.206
1.008
0.362
0.206
45564
Rogor L40: D.
dimethoate
37.40
36.80
36.64
36.28
98.4
98.0
97.0
0.206
0.676
0.050
0.429
0.157
0.320
43390
Rogor L40: B.
dimethoate
37.40
38.85
39.36
37.15
103.9
105.2
99.3
0.093
0.811
0.071
0.701
0.408
0.287
42490
BASF Dimethoate 40: A.
dimethoate
37.40
40.37
40.50
38.03
108.0
108.3
101.7
0.166
0.289
0.050
0.749
0.361
0.112
42361
Danadim: A.
dimethoate
38.00
43.52
44.17
39.03
114.5
116.2
102.7
0.168
0.542
0.112
0.500
0.769
0.112
42351
Rogor L40: A.
dimethoate
37.40
38.27
39.15
37.33
102.3
104.7
99.8
0.034
1.060
0.112
0.637
0.353
0.260
47879
Fyfanon 440: B.
malathion
42.00
48.37
44.37
40.20
115.2
105.6
95.7
0.441
0.510
0.047
0.874
0.537
0.145
42486
Malathion 60.
malathion
56.60
63.80
62.90
55.72
112.7
111.1
98.5
0.084
0.728
0.044
1.440
0.626
0.067
42637
Fyfanon 440: A.
malathion
42.00
49.10
44.29
41.06
116.9
105.4
97.8
0.162
0.169
0.061
1.429
1.266
0.061
42355
Casoron G
dichlobenil
6.75
7.03
6.68
6.76
104.2
98.9
100.2
0.108
0.088
0.017
0.045
0.041
0.025
42484
Casoron G
dichlobenil
6.75
6.89
6.40
6.59
102.1
94.8
97.6
0.368
0.103
0.003
0.250
0.225
0.017
42561
Casoron G
dichlobenil
6.75
6.93
6.67
6.67
102.6
98.8
98.8
0.026
0.126
0.051
0.099
0.195
0.051
45563
Casoron G
dichlobenil
6.75
6.88
6.40
6.60
101.9
94.8
97.8
0.151
0.051
0.040
0.146
0.028
0.040
45862
Casoron G
dichlobenil
6.75
6.89
6.79
6.66
102.1
100.7
98.7
0.217
0.071
0.050
0.048
0.087
0.050
46128
Casoron G
dichlobenil
6.75
6.71
6.51
6.52
99.4
96.4
96.7
0.146
0.110
0.081
0.091
0.160
0.081
42649
Casoron G4 Weed Block
dichlobenil
4
3.95
3.88
3.81
98.8
97.0
95.3
0.021
0.081
0.035
0.101
0.156
0.046
42656
Casoron G4 Weed Block
dichlobenil
4
4.13
3.84
3.92
103.2
95.9
98.1
0.074
0.121
0.018
0.010
0.140
0.022
43419
Embargo G
dichlobenil
6.75
5.48
5.19
6.45
81.1
76.9
95.5
0.060
0.041
0.059
0.115
0.207
0.122
43416
Luxan Dichlobenil Granules
dichlobenil
6.75
5.60
5.44
7.40
83.0
80.6
109.6
0.172
0.111
0.042
0.414
0.351
0.078
42488
Luxan Dichlobenil Granules
dichlobenil
6.75
6.07
5.52
6.48
90.0
81.8
96.0
0.186
0.074
0.024
0.452
0.174
0.043
45864
Sierraron G
dichlobenil
6.75
7.30
7.13
7.20
108.1
105.7
106.7
0.058
0.216
0.049
0.070
0.372
0.049
Table 3: Tabulated statistical results for all pesticide formulations and experimental methods.
Mean value
percent of declared
Std_repeatability
Std_reproducibility
code
sample
ingredient
declared
1H
13C
Chrom.
1H
13C
Chrom.
1H
13C
Chrom.
1H
13C
Chrom.
45872
Alpha Metazachlor 50SC
Metazachlor
44.3
51.89
48.07
45.28
117.1
108.5
102.2
1.952
2.113
0.366
1.533
1.473
0.366
45867
Butisan S
Metazachlor
43.1
50.12
46.20
43.07
116.3
107.2
99.9
1.179
0.550
0.176
1.712
0.977
0.209
45868
Katamaran
Metazachlor
32.3
37.75
34.78
32.78
116.9
107.7
101.5
1.231
1.720
0.145
1.059
1.189
0.145
45568
Me2 Booty
Metazachlor
43.1
51.57
48.65
43.43
119.6
112.9
100.8
0.081
1.271
0.173
0.829
0.871
0.357
45870
Sultan 50 SC
Metazachlor
44.3
51.93
50.38
44.27
117.2
113.7
99.9
0.419
0.831
0.169
0.467
2.405
0.169
42738
Standon Metazachor 50
Metazachlor
43
51.56
46.80
43.34
119.9
108.8
100.8
0.083
0.898
0.175
1.324
1.690
0.259
43396
Alpha Metazachlor 50SC
Metazachlor
44.3
53.55
48.74
44.40
120.9
110.0
100.2
0.196
1.604
0.296
1.392
1.643
0.296
42357
Katamaran
Metazachlor
32.3
36.18
34.10
31.98
112.0
105.6
99.0
1.490
1.421
0.132
1.210
1.442
0.132
42359
Butisan S
Metazachlor
43.1
49.74
46.22
42.70
115.4
107.2
99.1
2.326
2.765
0.209
0.751
1.944
0.209
45566
Butisan S
Metazachlor
43.1
53.30
48.76
42.92
123.7
113.1
99.6
1.650
2.050
0.205
2.179
0.985
0.205
45783
Standon Metazachlor 50
Metazachlor
43
49.92
46.97
42.52
116.1
109.2
98.9
0.684
1.101
0.429
0.831
0.083
0.429
Table 3: Continued.
Accuracy
Precision
Development time
Cost
Information content
Interpretability
1H NMR
The mean deviation from the declared value for the four formulation types was ±10%.
On average this was 0.75 for repeatability and 1.35 for reproducibility.
This is difficult to estimate, as “start-up” time for this initial study should not be required for future work. Development time is comparable with that for chromatographic measurements.
Following development, cost per sample is determined from FEC rates based on sample numbers an associated staff/equipment usage.
Information about all protonated compounds in a mixture including the target and any impurities.
1H and 13C NMR are often used in combination for the determination of the chemical structure of unknowns. With appropriate expertise and facilities it is possible to determine the identity of unknowns in complex mixtures such as pesticide formulations
13C NMR
The mean deviation from the declared value for the four formulation types was ±5%.
On average this was 1.35 for repeatability and 1.35 for reproducibility.
See above. Development of 13C NMR is not in addition to development for 1H NMR.
See above.
Information about all organic compounds (above the limit of detection).
Chromatography
The mean deviation from the declared value for the four formulation types was ±1%.
On average this was 0.10 for repeatability and 0.17 for reproducibility.
Varies widely depending on the formulation type (hours to weeks).
The cost per sample for chromatography is generally higher than for NMR. However, this is in part offset by the high capital outlay and level of expertise required for NMR analysis.
Determined largely by the detection system used e.g. selected ion MS is very specific to the target analyte.
This is dependant on the detection system. Little information about compound ID is determined by UV detection. More by MS but often limited to the target analyte.
Comments
In the absence of a certified formulation the accuracy has been assessed by agreement with the declared value for the formulations.
The precision of each method is based on the standard errors for reproducibility and for repeatability. These are given as %w/w
Table 4: A comparison of the NMR and chromatographic methods
Evaluation of the potential of NMR for screening of contaminants
The NMR spectra of the different types of pesticide formulation were inspected in detail. The aim was to choose a particular formulation type from which a model of the genuine formulation could be developed. This model would contain information on the levels of variation occurring in the particular formulation type, and would allow adulteration with compounds outside the permitted range of active ingredients and legitimate additives to be rapidly detected. To be successful, this process requires a reasonable degree of specificity about the permitted make up of the particular formulation type.
It is apparent that NMR is a highly effective way of identifying the presence of components other than (but also including) the active ingredient. The sub-plots in Figure 11 present sections of both 13C and 1H NMR spectra of an example formulation showing the wealth of information on additional compounds. However, after a detailed inspection of the NMR spectral database acquired in this study, it was discovered that the definition for particular formulation types varied too greatly to make the construction of a model for any of the studied pesticides a viable proposition. Here we do not sugest that the composition of a formulation is unregulated, simply that the formulations for a particular active ingredient vary widely in their overall composition due to the presence of both impurities and specified additional content. It is therefore not possible to define a standard formulation for a particular active ingredient.
A more appropriate approach is to characterise the impurities and co-formulants present in each formulation on a case-by-case basis in support of legal or regulatory requirements. The NMR analysis provides information about all constituents of the formulation and not just the active ingredient, as is often the case for chromatographic methods. In addition, the NMR data is interpretable and therefore standard compounds are not always required to determine the identity of the components of the formulation. This is achieved by the interpretation of the peak positions and the J-coupling patterns of the 1D NMR data and also by recruiting 2D NMR techniques that link the 13C and the 1H nuclei. Indeed NMR is one of the principle tools for the determination of the chemical structure of unknown or newly synthesised compounds. However, if it were possible to use a target list approach, it would be a straightforward matter to ascertain the presence or absence of a list of compounds within the formulation, often achieved from a single NMR dataset. It would also be possible to quantify these components within the limitations of the precision data that is presented herein. Furthermore, this process could be automated using some of the tools that have been developed at CSL for the automatic detection of contaminants in foods using NMR spectroscopy.
CONCLUSIONS
Development of QNMR methods for pesticides
We have developed methods for the quantification of active ingredients in pesticide formulations by both 1H and 13C QNMR. The 1H method was more sensitive, and would be the only option if the target compound was present at very low concentrations, or acquisition time was limited. However, due to the lower spectral resolution, overlap between different 1H signals is more problematic, and could cause difficulties where the active ingredient signals were not the major component. In such cases the 1H technique would be better used to rapidly ‘fingerprint’ or ‘profile’ the formulation’s composition rather than to obtain accurate concentration values.
Using the 13C QNMR technique, separation between different NMR frequencies was much more pronounced. This assisted both in discriminating different compounds, and in obtaining discrete integrals from different spectral regions in highly crowded spectra. 13C QNMR is the better method where high resolution between formulation constituents is required, for example when trying to determine contamination by closely related molecules with only small frequency shifts between their spectra.
QNMR methods were compared with the more usual chromatographic techniques using a number of criteria. The precision of the techniques were found to be of the same order of magnitude, with the chromatographic measurements slightly more precise. One difference between the chromatographic and QNMR methods however, was that the experiment time for QNMR could be shortened at the expense of precision by reducing the number of scans of the data (a single NMR measurement is of the order of seconds, but successive scans are required to obtain improved signal to noise). Therefore, the QNMR method could be used with higher throughput when a less precise check on the quantity of active ingredient is required. Chemical methods can also be used to shorten the required delay between scans and these too could be used to reduce overall run time. These methods were not employed extensively in this study.
In terms of the accuracy of measurement, the chromatographic methods show greater agreement with declared values for the concentration of the active ingredients in the formulations. This is most probably due to overlapping peaks in the NMR spectrum. To address this a 13C QNMR method has been developed (in addition to the work proposed in the CSG7). This method had inherently greater resolution than the 1H QNMR method and thus the accuracy was improved with an error of approximately 10% for the 1H NMR method being reduced to approximately 5% for the 13C NMR method. This was therefore similar to the accuracy of current techniques, which were found to have an error of approximately 1% by the same criteria. The 13C QNMR method was of lower sensitivity than the 1H QNMR method and therefore longer acquisition times (approx 2hrs) were required. The remaining error was likely to be due to several factors including; residual overlapping resonances from for example isomers, interactions between formulation constituents and the NMR solvents, and the lower sensitivity of 13C NMR. These factors could be overcome with further investment. A proposed approach would be to use the 1H NMR method in the first instance. If resolution were found to be a problem then the experimental setup (temperature, pH etc.) would be changed. 13C QNMR would be used to provide better resolution if this did not resolve the issue.
For the similar levels of precision obtained for the determination of the concentration of the active ingredient, 1H QNMR methods were not notably quicker than the chromatograpgic methods. Chromatographic methods ranged from 8 to 20 minutes, whilst the measurement time was around 15 minutes for 1H and one to two hours for 13C QNMR methods. However, the NMR acquisition times can be substantially reduced with a commensurate reduction in the precision as previously discussed.
However, for a new formulation, method development time would dominate considerations for both QNMR and chromatographic methods. QNMR requires a solubility and stability analysis of the active ingredient and reference material, and time to construct a calibration curve. Chromatographic methods may also require extraction methods to be developed, require correct choice of column type and solvent, and investigation of potential derivatisation reactions when GC methods are employed. These considerations must be made on a case-by-case basis. It is clear that the QNMR methods form a useful compliment to the existing possibilities and add another level of capability when characterisation of the active ingredients or contaminants is required. Although not discussed here, NMR is particularly suitable for studying the interaction between compounds such as unanticipated reaction or degradation of the active ingredients and the kinetics thereof.
Detecting contaminants in pesticide formulations
A thorough evaluation of the available spectra was made to assess their potential for use as a database for detecting contaminants in pesticide formulations. However, it was found that the formulations were too diverse for statistical modelling using the number of spectra acquired in a study of this size. As a consequence, more project effort was placed into QNMR rather than elaborating on this area. Further development in this area would benefit from direct input from pesticide manufacturers to supply larger number of typical formulations, in order to define ranges of products found within these mixtures. A common practice in the pharmaceutical industry is that reputable companies store an analytical fingerprint of their product this can then used to verify downstream labelling claims by “matching” the product fingerprint against that collected at the point of origin. These fingerprints are obtained using holistic techniques such NMR spectroscopy. This is an exceptionally robust method for rapidly detecting fraud, non-compliance or significant product deterioration and should be considered if this is an issue in the agrochemical industry.
Information content advantages of QNMR
Whilst pursuing both the QNMR and contaminant detection areas it became clear that the significant advantage of the NMR technique over the chromatographic methods is in the area of information content. QNMR offers the ability to quantify numerous other compounds in the formulation at the same time as the active ingredient. This can allow for profiling of individual formulations, and the simultaneous detection of minor ingredients of interest. Data is obtained from all proton containing compounds in quantifiable amounts, and a single standard is used to quantify all compounds. By comparison GC-MS methods often only acquire data from two ions: one from the active ingredient, and one from a reference compound.
Despite the variability of the formulations it was apparent that NMR offers potential for use as a fingerprinting and profiling technique for the identification of constituents of pesticide formulations. During a QNMR experiment information is gained not only on the levels of active ingredients, but also on the wide range of additional compounds present in the formulation. This could be used to build up a profile of the formulation and which could be aided by hyphenating the NMR method with a separation technique such as liquid chromatography (i.e. LC-NMR). When an unknown is detected, NMR also offers a range of 2D techniques that can assist in compound identification.
In conclusion the NMR methods that have been evaluated here represent a realistic solution to the quantification of the active ingredient in pesticide formulations. The accuracy of NMR techniques are likely to be lower than for chromatographic methods, however, there is a substantial increase in the information content when using the NMR approach. Rapid NMR methods could be used if less precise determination is required. They may be used for example to verify that a formulation contains all of the specified ingredients using a single dataset that could potentially be acquired in less than a minute. NMR is also a powerful technique for the determination of unknowns and this could be applied to the characterisation of non-specified species in pesticide formulations.
Appendix 1: Details of the formulations used in this analysis are given below. The formulations were stored in CSLs dedicated pesticide store at ambient temperature at all times other than when submitted for analysis.
Code
Sample
Formulation type
Reg. No.
Batch No.
Manufacturer
Pack size (g)
Supplier
Date received
Label declared a.i. content
1H Data Acquisition date
13C Data Acquisition date
Original
Analysis date
45559
Rogor L40
EC
07611
UB11-06
Isagro
5000
Farmway Ltd
15/10/2003
400g/l (37.4% w/w) dimethoate
26-27/01/2005
28-30/01/2005
Jan 2004
45561
BASF Dimethoate 40
EC
00199
074 Date of Production 11/95
BASF
5000
Farmway Ltd
15/10/2003
400g/l (37.4% w/w) dimethoate
26-27/01/2005
28-30/01/2005
Jan 2004
45866
Danadim
EC
11550
3A6 30128 07
Cheminova
5000
BATA
03/11/2003
400g/l (38.0% w/w) dimethoate
26-27/01/2005
28-30/01/2005
Jan 2004
45564
Rogor L40
EC
07611
UF30.02
Isagro
5000
Harlow Agricultural Merchants
15/10/2003
400g/l (37.4% w/w) dimethoate
26-27/01/2005
28-30/01/2005
Jan 2004
43390
Rogor L40
EC
07611
C.20.02.SPC
Isagro
5000
Robertsons Crop Protection
14/07/2003
400g/l (37.4% w/w) dimethoate
26-27/01/2005
28-30/01/2005
Jan 2004
42490
BASF Dimethoate 40
EC
00199
37M 30118 06
BASF
5000
Cropwise Ltd
21/05/2003
400g/l (37.4% w/w) dimethoate
26-27/01/2005
28-30/01/2005
Jan 2004
42361
Danadim
EC
11550
3A6 30127 06
Cheminova
5000
Agrovista UK Ltd
15/05/2003
400g/l (38.0% w/w) dimethoate
26-27/01/2005
28-30/01/2005
Jan 2004
42351
Rogor L40
EC
07611
VC30.01
Isagro
5000
Hall Farm Merchants
15/05/2003
400g/l (37.4% w/w) dimethoate
26-27/01/2005
28-30/01/2005
Jan 2004
47879
Fyfanon 440
emulsion, oil in water, EW
11014
31R 20616 06
Cheminova
5000
East Riding Horticulture
27/02/2004
400g/l (42.0% w/w) malathion
24/02/2005
24/02/20025-03/03/2005
Mar 2004
42486
Malathion 60
emulsifiable concentrate, EC
08018
A2297
United Phosphorus Ltd
5000
CSC Crop Protection
21/05/2003
600g/l (56.6% w/w) malathion
24/02/2005
24/02/20025-03/03/2005
Mar 2004
42637
Fyfanon 440
emulsion, oil in water, EW
11014
31R 20618 06
Cheminova
5000
East Riding Horticulture
04/06/2003
400g/l (42.0% w/w) malathion
24/02/2005
24/02/20025-03/03/2005
Mar 2004
Code
Sample
Formulation type
Reg. No.
Batch No.
Manufacturer
Pack size (g)
Supplier
Date received
Label declared a.i. content
1H Data Acquisition date
13C Data Acquisition date
Original Analysis data
42355
Casoron G
GR
09022
AM3A21K000 2003.01.21
Crompton Europe Ltd
5000
Agrovista UK Ltd
15/05/2003
6.75% w/w dichlobenil
18-25/03/2005
18-25/03/2005
Nov 2003
42484
Casoron G
GR
09023
AM2K11K003
Nomix-Chipman
10000
Nomix-Chipman
21/05/2003
6.75% w/w dichlobenil
18-25/03/2005
18-25/03/2005
Nov 2003
42561
Casoron G
GR
10709
AM2K11Q002
The Scotts Company
5000
East Riding Horticulture
23/05/2003
6.75% w/w dichlobenil
18-25/03/2005
18-25/03/2005
Nov 2003
45563
Casoron G
GR
09022
AM1J06K007
Crompton Europe Ltd
25000
Farmway Ltd
15/10/2003
6.75% w/w dichlobenil
18-25/03/2005
18-25/03/2005
Nov 2003
45862
Casoron G
GR
09022
AM3A22K000 2003.01.22
Crompton Europe Ltd
15000
BATA
03/11/2003
6.75% w/w dichlobenil
18-25/03/2005
18-25/03/2005
Nov 2003
46128
Casoron G
GR
10709
AM2L15Q001
The Scotts Company
5000
East Riding Horticulture
17/11/2003
6.75% w/w dichlobenil
18-25/03/2005
18-25/03/2005
Nov 2003
42649
Casoron G4 Weed Block
GR
09371
270802
Vitax Ltd
250
Deans Garden Centre
04/06/2003
4% w/w dichlobenil
18-25/03/2005
18-25/03/2005
Nov 2003
42656
Casoron G4 Weed Block
GR
09371
1120500 14-50
Vitax Ltd
250
Deans Garden Centre
04/06/2003
4% w/w dichlobenil
18-25/03/2005
18-25/03/2005
Nov 2003
43419
Embargo G
GR
10367
Lotto 021203 (on piece of paper inside sealed box)
SumiAgro Ltd
25000
United Agri Products Ltd
14/07/2003
6.75% w/w dichlobenil
18-25/03/2005
18-25/03/2005
Nov 2003
43416
Luxan Dichlobenil Granules
GR
09250
2878066
Luxan Ltd
15000
Robertsons Crop Protection
14/07/2003
6.75% w/w dichlobenil
18-25/03/2005
18-25/03/2005
Nov 2003
42488
Luxan Dichlobenil Granules
GR
09250
1106 Manufacturing Date 02/1999
Luxan Ltd
1500
Cropwise Ltd
21/05/2003
6.75% w/w dichlobenil
18-25/03/2005
18-25/03/2005
Nov 2003
45864
Sierraron G
GR
09675
S0020806 14/05/02
The Scotts Company
25000
BATA
03/11/2003
6.75% w/w dichlobenil
18-25/03/2005
18-25/03/2005
Nov 2003
Code
Sample
Formulation type
Reg. No.
Batch No.
Manufacturer
Pack size (g)
Supplier
Date received
Label declared a.i. content
1H Data Acquisition date
13C Data Acquisition date
Original Analysis date
45872
Alpha Metazachlor 50 SC
SC
10669
03088030
Makhteshim-Agan
5000
BATA
03/11/2003
500g/l (44.3% w/w) metazachlor
06-11/05/2005
06-11/05/2005
Mar 2004
45867
Butisan S
SC
00357
Lot: 02482916KO Prod: 01-04
BASF
5000
BATA
03/11/2003
500g/l (43.1% w/w) metazachlor
06-11/05/2005
06-11/05/2005
Mar 2004
45868
Katamaran
SC
09049
Lot: 16626936WO Prod: 07-03
BASF
5000
BATA
03/11/2003
125g/l (10.8% w/w) quinmerac + 375g/l (32.3% w/w) metazachlor
06-11/05/2005
06-11/05/2005
Mar 2004
45568
Me2 Booty
SC
10659
308 10659 (given over the phone by supplier as not on container)
Me2 Crop Protection Ltd
5000
Harlow Agricultural Merchants
15/10/2003
500g/l (43.1% w/w) metazachlor
06-11/05/2005
06-11/05/2005
Mar 2004
45870
Sultan 50 SC
SC
10418
03078028
Makhteshim-Agan
5000
BATA
03/11/2003
500g/l (44.3% w/w) metazachlor
06-11/05/2005
06-11/05/2005
Mar 2004
42738
Standon Metazachlor 50
SC
05581
5121 JUL 02
Standon Chemical Co.
5000
Countrywide Farmers
10/06/2003
500g/l (43.0% w/w) metazachlor
06-11/05/2005
06-11/05/2005
Mar 2004
43396
Alpha Metazachlor 50SC
SC
*10418
02058009
Makhteshim-Agan
5000
Robertsons Crop Protection
14/07/2003
500g/l (44.3% w/w) metazachlor
06-11/05/2005
06-11/05/2005
Mar 2004
42357
Katamaran
SC
09049
Lot: 52868175LO Prod: 07-02
BASF
5000
Agrovista UK Ltd
15/05/2003
125g/l (10.8% w/w) quinmerac + 375g/l (32.3% w/w) metazachlor
06-11/05/2005
06-11/05/2005
Mar 2004
42359
Butisan S
SC
00357
Lot: 17529168EO Prod: 01-03
BASF
5000
Agrovista UK Ltd
15/05/2003
500g/l (43.1% w/w) metazachlor
06-11/05/2005
06-11/05/2005
Mar 2004
Code
Sample
Formulation type
Reg. No.
Batch No.
Manufacturer
Pack size (g)
Supplier
Date received
Label declared a.i. content
1H Data Acquisition date
13C Data Acquisition date
Original Analysis date
45566
Butisan S
SC
00357
Lot: 31718547G0 Prod: 12-02
BASF
5000
Harlow Agricultural Merchants
15/10/2003
500g/l (43.1% w/w) metazachlor
06-11/05/2005
06-11/05/2005
Mar 2004
45783
Standon Metazachlor 50
SC
05581
5122 JUL 03
Standon Chemical Co.
5000
Phoenix Agronomy Ltd
22/10/2003
500g/l (43% w/w) metazachlor
06-11/05/2005
06-11/05/2005
Mar 2004
*Incorrect Reg Number
References to published material
9.This section should be used to record links (hypertext links where possible) or references to other published material generated by, or relating to this project.
Adrianensens P.J., Karssenberg F.G., Gelan J.M., Mathot V.B.F 2003 Improved quantitative solution state 13C NMR analysis of ethylene-1-octene copolymers Polymer 44 3483-3489.
Cookson D.J., Smith B.E. 1984 Optimal conditions for obtaining quantitative 13C NMR data Journal of Magnetic Resonance 57 355-368.
Freeman R., Hill H.D.W., Kaptein R. 1972 Proton-Decoupled NMR Spectra of Carbon-13 with the Nuclear Overhauser Effect Suppressed J. Magn. Res. 7 327-329.
FD 03/23 Dichlobenil Formulations for Amateur and Professional Use: A Survey For Compliance With Specifications of Samples Purchased Between May and November 2003
FD 03/30 Metazachlor Formulations for Professional Use: A Survey For Compliance With Specifications of Samples Purchased Between May and November 2003
FD 03/34 Dimethoate Formulations for Professional Use: A Survey For Compliance With Specifications of Samples Purchased Between May and November 2003
FD 03/30 Malathion Formulations for Professional Use: A Survey For Compliance With Specifications of Samples Purchased Between May 2003 and February 2004
Holtman K.M., Chang H.M., Kadla J.F. 2004 Solution-state nuclear magnetic resonance study of the similarities between milled wood lignin and cellulolytic enzyme lignin J. Agric. Food Chem. 52 720-726.
Tenailleau E., Lancelin P., Robins R.J., Akoka S. 2004 NMR approach to the quantification of non-statistical 13C distribution in natural products: vanillin Analytical Chemistry 76 3818-3825.
Wells R.J., Hook J.M., Al-Deen T.S., Hibbert D.B., 2002 Quantitative Nuclear Magnetic Resonance (QNMR) Spectroscopy for assessing the purity of technical grade agrochemicals: 2,4-dichlorophenoxyacetic acid (2,4-D) and sodium 2,2-dichloropropionate (dalapon sodium) J. Agric. Food Chem. 50 3366-3374.
Xia Z, Akim L.G., Argyropoulos D.S., 2001 Quantitative 13C NMR analysis of lignins with internal standards J. Agric. Food Chem. 49 3573-3578.
�
A
B
B
A
13C
1H
SID 5 (2/05)Page 27 of 33
Chart1
298.59
0
597.17
154.36
398.12
77.18
497.64
1H calibration
75.02997.2751.2964986872298.59
0000
67.315194.1242.8838149001597.17
69.19140.9910.5924325418154.36
72.969136.9891.8773588784398.12
69.48724.170.347834846877.18
72.225168.6042.3344271374497.64
1H calibration
0
0
0
0
0
0
0
13C calibration
trioxanemalathionratioconcentration
12.855.410.4210116732298.59
0000
12.80511.670.9113627489597.17
13.0392.3880.1831428791154.36
12.6547.7210.6101627944398.12
13.0961.3260.101252290877.18
12.7029.6670.7610612502497.64
13C calibration
0
0
0
0
0
0
0
31P data
trioxmalaratiomMg/LdensitymassvolumeNMR volume% mala
standard34.7482.8330.081529872274.843908687620.0010893321
standard24.1042.0240.083969465677.0834400723
4248642.243.8320.09071969783.28010986790.86810208951.0607067138204.7217.126664311717.12666431130.27731014855
4248642.4183.8510.090786930183.341829435590.93544518091.0607067138204.7217.126664311717.12666431130.27731014856
standard15.3847.2130.4688637546430.41397068426.5486100.90619701488
standard15.6237.3150.4682199322429.8229457769426.5486100.76763721119
4248636.683.170.086423118979.335878255986.92039764451.0607067138203.5215.8538162544715.853816254430.153337364911
4248667.3925.40.080128205173.557187131680.58926296211.0607067138203.5215.8538162544715.853816254430.153337364912
standard34.6264.1960.1211806157111.243041238114
standard34.8634.2960.1232251958113.119952828215
424860001.0607067138200.4212.5656254417712.565625441729.83102437917
424860001.0607067138200.4212.5656254417712.565625441729.83102437918
standard0020
standard0021
66.1542857143
426.5485714286
132.3085714286
341.2457142857
255.9342857143
31P Calibration
TriphenMalaRatioConc
26.6652.240.084005250366.1542857143
15.2227.1260.468138221426.5485714286
16.892.1220.1256364713132.3085714286
16.3176.1290.3756205185341.2457142857
17.574.8190.2742743312255.9342857143
0000
31P Calibration
0
0
0
0
0
0
0