Solvent

This measurement method is suitable for monitoring the Occupational Exposure Limits (OELs) of solvent components at workplaces. Sampling is performed by drawing air through a silica gel tube using a suitable sampling pump. The adsorbed substances are desorbed with a ternary mixture consisting of dichloromethane, methanol and water. The sample solution is analysed by means of gas chromatography using two capillary columns of di ﬀ erent polarity with the help of flame ionisation detectors (FID). Quantitative evaluation is based on multiple-point calibrations using internal standards.


1G eneral principles
In this analytical method sampling is performed by drawing adefined volume of air through asilica gel tube using asuitable sampling pump. The adsorbed substances are desorbed with at ernary mixture consisting of dichloromethane, methanol and water. The sample solution is analysed by means of gas chromatography using two capillary columns of different polarity with the help of flame ionisation detectors (FID). Quantitative evaluation is based on multiple-point calibrations using internal standards.

Chemicals
The substances to be determined should have the highest possible purity. • Helium 5.0 (carrier gas) • Hydrogen 5.0 • Synthetic air (free of hydrocarbons)

Calibration standards
Calibration standards are necessary for obtaining the required calibration functions.
The minimum concentration range they should cover is at least one tenth to twice the respective Occupational Exposure Limits or MAK values. If this is not possible for reasons of linearity, then calibration is carried out in the linear range and the sample is subsequently diluted accordingly. 50 μLe ach of the solvent components (see Table 1) are added to a5mL volumetric flask, into which approx. 2.5 mL of desorption solution have already been placed. The flask is then filled to the mark with desorption solution and shaken. Details for the preparation of the stock solution are shown in Table 1. The respec-2172 1) Determination of the content was 85%: The exact content of 2,6-Dimethyl-4-heptanone has to be determined by gas chromatography for every new delivery.
tive purities of the reference substances must be taken into account when calculating the concentrations.

Note:
The stock solution is stable in the refrigerator at approx. 4°Cfor up to four months. The date of preparation must be noted on the flask.
Seven calibration solutions are prepared from the stock solution in 5m Lv olumetric flasks. The volumes of stock solution listed in Table 2are added to the volumetric flasks, into which approx. 2.5 mL of desorption solution have already been placed in each case. The flasks are then filled to the mark with desorption solution and shaken. 3S ampling and sample preparation

Sampling
As ilica gel tube (BIA type) is opened and connected to af low-regulated pump for sampling. Then, workplace air is drawn through the silica gel tube at aflow rate of 5L /h (0.083 L/min). The recommended sampling duration is two hours, during which an air sample volume of 10 Ls hould not be exceeded. The pump and the tube are either worn by ap erson during working hour or stationary sampling is carried out. After sampling, the silica gel tube is sealed with the plastic caps that are supplied with it and unambiguously labelled.

Sample preparation
The loaded silica gel tube is opened in the laboratory and the content is transferred into a2 0m Ls crew-capped vial. Then the silica gel is covered with 5m Lo fd esorption solution and the vial is sealed tightly. After as tanding of 18 hours (overnight) 2.5 μLo ft he internal standard (IS) are added to the sample and shaken briefly. The sample solution is filtered through adisposable filter into an autosampler vial and analysed.

4O perating conditions for gas chromatography
Each sample is analysed in parallel on two capillary columns of different polarity. 2-Methylheptane is used as internal standard for the non-polar column and n-undecane for the polar column.

5A nalytical determination
For the analytical determination of the samples prepared according to Section 3.2 1 μLo ft he sample solution is injected in each case into the gas chromatograph using the autosampler and divided between two columns of different polarity. The mean value of the measurement data from both columns is used for the evaluation in each case. If the deviation of the data from both separation columns is less than 20%, the mean value is used for the evaluation. In the case of agreater deviation, it must be determined on the basis of different criteria, which value should be used for the evaluation, e.g.: • If the signal in achromatogram of asubstance to be determined is overlapped by that of another substance, then the result from the other chromatogram is used for the evaluation.
• If the signal of the internal standard in ac hromatogram is overlapped by that of another substance, then the results from the other chromatogram are used for the evaluation.
If the measured concentrations are above the calibration range, then asuitable dilution must be prepared and the analysis must be carried out again.

6C alibration
The internal standard method is used for calibration of the method. In order to obtain the calibration functions, each calibration standard (see Section 2.4) is injected into the gas chromatograph three times and analysed in the same manner as the sample solution (determination in triplicate). The calibration must be performed anew if the analytical conditions change or the quality control results indicate that this is necessary.

7C alculation of the analytical result
On the basis of the resulting peak areas the respective mass X in μgo fasolvent component can be obtained from the relevant calibration function. The corresponding mass concentration (ρ)o fac omponent is calculated according to Equation (1) as follows: Equation (2) enables calculation of the value at 20°C and 1013 hPa: where: ρ is the mass concentration of asolvent component in mg/m 3 ρ 0 is the mass concentration of the component in mg/m 3 at 20°C and 1013 hPa X is the mass of the component in the analytical sample in μg V is the air sample volume in litres η is the recovery t a is the temperature during sampling in°C p a is the air pressure during sampling in hPa

8R eliability of the method
The characteristics of the method were calculated as stipulated in EN 482 [1], EN 1076 [2] and DIN 32645 [3]. The experiments to determine the characteristics of the methods were -if not otherwise described -carried out in ad ynamic test gas facility. Test gases were generated in the concentration range of approx. one tenth up to twice the OELs or MAK values of the investigated substances at ar elative humidity of between 20 and 80%. Samples were taken from each of these concentrations at room temperature (air sample volume 10 mL), prepared according to Sections 3.2, 4and 5and analysed.

Precision
The precision of the method was determined at three concentrations for each substance. For this purpose six silica gel tubes per concentration were loaded simultaneously in adynamic test gas facility. These experiments were carried out at arelative humidity of approx. 50%. The calculated parameters are shown in Table 3.

Recovery
The recoveries of the investigated solvent components were evaluated in the course of the determination of the precision in the lowest measurement range. For this purpose, six silica gel tubes were loaded for each experiment. The recoveries were constant throughout the selected measurement range. The results can be found in Table 3.

Limit of quantification
The limits of quantification for the selected substances were determined from 10point calibrations in ac oncentration range of approx. 2t o2 0mg/m 3 as stipulated in DIN 32645 [3]. For all substances alimit of quantification of approx. 20 μg/5 mL (absolute 4ng) was calculated which, on the basis of an air sample volume of 10 L, asample solution of 5mLand an injection volume of 1 μL, is equivalent to alimit of quantification of approx. 2mg/m 3 .

Expanded uncertainty
The expanded uncertainties for the determination of the individual solvent was estimated taking all relevant influencing factors into consideration as stipulated in EN 482 [1] and EN 1076 [2] and calculated according to [4]. The expanded uncertainties are between 11 and 13% for the selected substances. The individual results are shown in Table 3.

Storage stability
Studies on the storage stability of the loaded silica gel tubes were carried out at two different concentrations over ap eriod of four weeks. For this purpose twelve silica gel tubes were each spiked with concentrations of 10 mg/m 3 as well as with concentrations that are equivalent to once the OELs or MAK values. The silica gel tubes were spiked at arelative humidity of approx. 50% and afterwards sealed tightly and stored at room temperature. Three samples each were analysed after one day and then after 7, 14 and 28 days. No significant losses could be ascertained over the investigated time period of four weeks.

Influence of the air humidity
The influence of humidity was investigated for all the selected solvents at aconcentration of 80 mg/m 3 at three different humidity levels (approx. 5, 30 and 70%). Additionally, the influence of humidity on the sampling of acetone at arelative humidity of approx. 70% and ac oncentration of acetone of 500 mg/m 3 was tested. There was no evidence that the air humidity had any influence.

Note:
The extracting agent contains water, as the recovery for some of the substances at a relative humidity of <30% are distinctly reduced.

Capacity of the adsorbent
Experiments on the breakthrough behaviour of the selected substances were carried out at ar elative air humidity of 80%. The concentrations used in this case were equivalent to twice the OELs of the investigated solvents. At asampling duration of two hours and av olumetric flow rate of 5L /h (0.083 L/min) none of the tested substances showed abreakthrough.

Interferences
Interferences from overlapping signals should be corrected by the use of two separation columns of different polarity arranged in parallel. Experience of using this equipment has shown that, apart from afew exceptions, at least one chromatogram without the signal being influenced is obtained and can be evaluated. The chromatograms of the different capillary columns used are shown in Figure 1.