A Sampling and Analytical Method for Simultaneously Assessing Multiple Organic Solvent Exposures for Plastic Material Printing Industry Workers

The aim of this study is to develop methodologies suitable for conducting multiple organic solvent exposure assessments for workers in the plastic material printing industries. By reviewing the existing sampling and analytical methods for the seven organic solvents used in the industry (including isopropyl alcohol, 2-butanone, ethyl acetate, methyl isobutyl ketone, toluene, n-butyl acetate, and cyclohexanone), coconut shell charcoal is suggested as an appropriate sorbent media for collecting samples, and GC/FID is chosen as the instrument for chemical analyses. Both CS2 + 5% (v/v) 1-butanol and CS2 + 5% (v/v) isobutanol are proposed as candidate desorption reagents. Since the recovery rates obtained from the latter (83.47%–99.84%) are higher than those of the former (80.12%–98.00%), the latter is chosen in the present study. The linearities of all the resulting calibration curves are r ≥ 0.995, with recovery rates (R) of ≥ 75%, and the corresponding coefficients of variations (CV) ≤ 7% for all target organic solvents. The results also suggest that the collected samples should be stored in a –10°C environment and be analyzed within 30 days. Finally, suggestions are made for amending the storage temperatures and storage days currently promulgated in related NIOSH methods based on data obtained from present study. The proposed methodology would be beneficial to the plastic material printing industry, enabling the simultaneous assessment of workers’ exposure to multiple organic solvents.


INTRODUCTION
According to 2012 statistics, the global revenue for the printing industry was ~422 billion US dollar (IBISWorld 2012).Among various types of printing industry, the plastic material (including the plastic membrane, plastic bag, plastic sheet, and leather-like plastic) printing industry accounts for ~46% of the total market.The products of plastic material printing industry are widely used in food, pharmacy, and cosmetic industries for various packing purposes.According to governmental statistics in Taiwan, there were 36,707 workers employed by 2,245 printing and reproducing industries with annual market ~2.87 billion US dollar in 2009.In order to increase the production rate through fast coating inks on the printed materials, many organic solvents with characteristics of low boiling point and high volatility usually contain in inks used in plastic material printing industries.In principle, seven organic solvents, including isopropyl alcohol (IPA), 2-butanone (MEK), ethyl acetate (EA), methyl isobutyl ketone (MIBK), toluene, n-butyl acetate (n-BA), and cyclohexanone (CHA) can be found in plastic material printing industries.Though many air monitoring results obtained from printing industries have indicated that the time-weighted-average (TWA) exposures to individual organic solvent might not exceed occupational permissible exposure limits (PELs), overexposure in the combined mixture of printing solvents was frequently found for printing industry workers (Samimi, 1982;Verhoeff, 1988;White et al., 1995;Horstman et al., 2001;NIOSH, 2001;Leung, 2005;Brueck, 2007).The above results clearly indicate that the importance on conducting multiple organic solvent samplings for assessing the exposures of printing industry workers.
In fact, multiple chemical exposures are commonly found in indoor, outdoor and workplace environments (Cao et al., 2011;Hsu et al., 2012;Kim et al., 2012).For workplace environments, the currently developed sampling and analytic methods are only good for assessing the exposure of a single organic solvent or solvents with similar characteristics.Table 1 summarizes sampling methods promulgated by National Institute of Occupation and Safety and Health (NIOSH) for the seven organic solvents (mentioned above) commonly found in the plastic material printing industry (NIOSH, 2008).It can be seen that beaded carbon is the required sorbent media only for MEK, whereas the others consistently use coconut shell charcoal as the sorbent media.In addition, inconsistencies could also be found among various organic solvents in their sample stabilities.Among them, storage days for four solvents are labeled from 6 to at least 90 days; however, the other three are listed as unknown.Furthermore, requirements of sample storage environment are also inconsistently listed.Storage temperatures of four chemical compounds range from -5 to 5°C, but the other three are labeled as refrigerated, stored in freezer, and unspecified, respectively.Table 2 further summarizes analytical methods promulgated by NIOSH for the above seven target organic solvents (NIOSH, 2008).As shown in Table 2, the gas chromatography/flame ionization detector (GC/FID) are consistently suggested for analyzing the above mentioned target organic solvents, however, five different GC columns were selected.There are two solvents recommended for sample desorption, including 1% 2-butanol in CS 2 and 100% CS 2 .Based on the above information, therefore, it can be expected that conducting multiple organic solvent exposure assessments would be both labor and cost intensive considering the number of sorbent tubes should be used during sampling periods, the differences in the environmental temperatures and storage time for sample stability after samplings, the intrinsic differences in desorption reagents for sample desorption, and the use of different columns in GC/FID analyses.
In this study, the currently available sampling and analytical methods for each individual organic solvent were assessed.Then, methodologies suitable for conducting multiple organic solvent exposure assessments in plastic material printing industries were proposed, and finally tested by examining the linearity of the calibration curves, sample recovery rates and corresponding coefficient of variations, and sample storage stabilities.The methodology proposed by the present study would be beneficial to the plastic material printing industry for assessing workers' inhalatory exposures to multiple organic solvents simultaneously.

Selection of Sampling and Analytical Methodologies
As shown in Table 1, six target organic solvents consistently use the coconut shell charcoal as their sorbent media, with the exception of MEK using the beaded carbon as its sorbent media.Yet, it is true that the beaded carbon, a carbon molecular sieve material, has the advantages in both water vapor adsorption (typically a factor of 4 lower than the activated carbon), and chemical adsorption (4 folds increase in the 222Rn adsorption coefficient) (Scarpitta, 1996).Here, it should be noted that both coconut shell charcoal and beaded carbon share same adsorption principles, and most importantly the former is cheaper and more widely used in the field than the latter.Therefore, the former is suggested as the sorbent media in the present study.This selected glass charcoal tube is known with a dimension of 7 cm × 4 mm (ID) and 6mm (OD) with both ends being flame sealed.The tube contains two sections of charcoal (coconut shell burned at 600°C) with grain size of 20 to 40 mesh.The front and rear section respectively contain 100 and 50 mg of charcoal.The above two sections are separated by a 2 mm-PU sponge and positioned by glass wool at the front section and PU sponge at the rear section.
According to the technique listed in Table 2, the gas chromatography and flame ionization detector (GC/FID) was selected for analyzing all target organic solvents.In this study, the concentrations of all organic solvents were determined using a Hewlett-Packard (HP) gas chromatograph (GC) (HP 5890A; Hewlett-Packard, Wilmington, DE, USA), a flame ionization detector (FID) (HP 5972; Hewlett-Packard, Wilmington, DE, USA), and a chemstation (Aspire C500; Acer, Taipei, Taiwan).This GC/FID was equipped with a separation column of crosslinked methyl silicone (HP-1, 30 m × 0.53 mm × 2.65 μm), and an automatic sampler (HP-7673A).The injected volume of each sample was 2 μL and the flow rate of carrier gas (N 2 ) was set at 10 mL/min.The temperature was initially set at 40°C and held for 3.00 minutes, increased to 60°C at 70 °C/min and held for 1.20 minutes, and to 95°C at 20 °C/min, then immediately

Proposing Sample Desorption Reagents
As shown in Table 2, two reagents were suggested for sample desorption, including 1% 2-butanol in CS 2 and 100% CS 2 .A researcher successfully used CS 2 mixed with 5% 2-butanol as a reagent for the desorption of alcohol from charcoal (Yu, 1990).Chen reported that using CS 2 mixed with 3% IPA as a desorption-assisted reagent would appreciably increase the desorption efficiency of n-butanol and MIBK from sorbent tube samples collected from paint manufacturing industries (Chen, 1993).Based on the above information and considering both 1-butanol and isobutanol are less interference to the retention time on gas chromatograph, this study proposed two separate solutions, including CS 2 mixed with 5% (v/v) 1-butanol and CS 2 mixed with 5% (v/v) isobutanol, as the candidate reagents for sample desorption tests.

Selecting Sample Desorption Reagent
In this study, twelve charcoal tubes were spiked with seven target organic solvents of IPA, MEK, EA, MIBK, toluene, n-BA, and CHA with the spiked amount for each compound equivalent to a charcoal tube collecting 3 L air volume (i.e., the maximal collected air volume of the charcoal tube) under a 1-PEL concentration environment.In principle, the sampling flow rate for a charcoal tube is usually set at 40 ± 5 mL/min for assessing personal exposures to organic solvents.Under the maximal collected air volume situation, the calculated sampling time of every charcoal tube could be 75 minutes or above.Therefore, a two-hour standard spiking would be comparable to in-situ sampling situations, and was adopted in the present study.After a two-hour spiking, the samples were refrigerated over night.In this study, desorption efficiencies were evaluated by testing against six spiked charcoal tubes for each of the two proposed desorption reagents.To test desorption efficiencies of each desorption reagent, the spiked charcoal was removed from the tube and placed in a 1.8 mL flask, and subsequently added with 1 mL of desorption reagent.The flask was then capped and put on a shaker for one hour prior to GC analysis.Finally, the resultant desorption efficiencies of the two proposed sample desorption reagents were compared, and the one with better desorption efficiencies would be chosen.

Evaluating Desorption Efficiencies of the Selected Sample Desorption Reagent and the Linearity of the Calibration Curve
Eighteen charcoal tubes were used in this part of study, and were classified into three groups (i.e., six charcoal tubes for each group).Charcoal tubes in each of the three groups were spiked with the seven target organic solvents with the spiked amounts for each compound equivalent to a charcoal tube collecting 3 L air volume under the 0.5-, 1-, and 2-PEL concentration environments, respectively.After a two-hour spiking, the samples were refrigerated over night.Finally, the resultant desorption efficiencies of the selected desorption reagent under three spiked conditions  were determined, and further compared with the performance criteria, including (1) the coefficient of variance (CV) for desorption efficiencies of the three different spiked conditions must be less than 7%; and (2) the difference between the maximum and minimum mean desorption efficiencies should be less than 7% in each of the three spiked conditions.
The linearity of the calibration curve was also tested in this part study.Here, the calibration curve is obtained by injecting the solution into GC column at concentration equivalent to 1/16 to 2 times of PEL with the samples air volume of 3 L.The calibration curve was required to meet the criteria of (1) the relative prediction error must be less than 7% within the pre-determined range of target concentrations, and (2) the linearity of calibration curve should be found with a correlation coefficient (r 2 ) greater than 0.995.

Assessing Sample Stability
As shown in Table 1, inconsistent sample storage conditions can be seen for the seven target organic solvents.Among them, four chemical compounds are found with storage temperatures ranging from -5 to 5°C, and the other three are labeled as refrigerated, stored in freezer, and unspecified, respectively.Table 1 also shows the inconsistency in the storage days.Among them, four chemical compounds are labeled from 4 to at least 90 days; however, the other three are not specified.Based on the above information, the storage temperatures of -10°C and room temperature (25°C), and the storage day from 0 to 30days were chosen in this study.A total of 36 charcoal tubes were used and each is spiked with each individual target organic solvent with the amount equivalent to the sample collected with the air volume of 3 L under a 1-PEL concentration environment.Among them, 6 charcoal tube samples were analyzed immediately after the two-hour spiking (i.e., representing the stability of the 0 th day).The remaining charcoal tubes were then divided into two groups (each group contains 15 charcoal tube samples) and stored at -10°C and room temperature, respectively.For each group of charcoal tube samples, the recovery efficiencies of three charcoal tubes were analyzed at 3 rd , 7 th , 11 th , 20 th , and 30 th days, respectively.

Desorption Efficiencies of the Two Proposed Desorption Reagents
Fig. 1 shows the two chromatographs respectively obtained from using reagents of (A) CS 2 + 5% (v/v) 1-Butanol and (B) CS 2 + 5% (v/v) isobutanol for the desorption of charcoal tube samples.Here, the concentrations of seven targeted solvents were prepared at levels of 0.5 PEL on the basis of the sampling volume specified at 3 L.As such, concentrations of IPA = 1246.0μg/mL, MEK = 958.3μg/mL, EA = 1775.8μg/mL, MIBK = 317.5 μg/mL, toluene = 688.1 μg/mL, n-BA = 991.7 μg/mL, and CHA = 160.3μg/mL were used in the present study.Both chromatographs show good separation among various compounds (including the seven target organic solvents, two compounds containing in the desorption reagent, and the internal standard of n-heptane) suggesting that both desorption reagents would not interfere with subsequent chemical analyses.However, it should be noted that when a sample contains benzene, the retention time of benzene could be seriously interfered by that of 1-butanol on the chromatograph.Therefore, the use of CS 2 + 5% (v/v) 1-butanol as the desorption reagent might not be feasible for samples containing benzene.
Table 3 shows desorption efficiencies of the seven targeted organic solvents using the two candidate desorption reagents for comparison.For CS 2 + 5% (v/v) 1-butanol, the mean recovery rates of the seven targeted organic solvents ranged from 80.12% to 98.00%, and the corresponding CVs ranged from 0.42% to 1.17%, respectively.For CS 2 + 5% (v/v) isobutanol, the mean recovery rates and their corresponding CVs ranged from 83.47% to 99.84 and 0.91% to 2.55%, respectively.All mean recovery rates and their corresponding CVs complied with the criteria respectively set at ≥ 75% and ≤ 7% suggesting that both desorption reagents could effectively remove the seven target organic solvents from charcoal tubes.However, the recovery rates obtained from the desorption reagent of CS 2 + 5% (v/v) isobutanol for six target organic solvents were found to be higher than those of CS 2 + 5% (v/v) 1-butanol (t-test, p  0.1) with the exception for IPA (i.e., no significant difference).The above results indicated that the former desorption reagent was in general better than the latter.

Desorption Efficiencies of the Selected Sample Desorption Reagent and the Linearity of the Calibration Curve
Though both CS 2 + 5% (v/v) 1-butanol and CS 2 + 5% (v/v) isobutanol were feasible for the desorption of the seven target organic solvents and subsequent analyses, the recovery rates obtained from the latter were higher than those of the former.Therefore, the latter was selected as the sample desorption reagent in this study.We conducted a study in a PET plastic membrane printing industry and discovered that some solvent mixtures used for plastic printing contained methanol, and hence the compound was included in this part of study.For assessing the linearity of calibration curves, 100 ppm n-heptane was used as an internal standard during analyses.In the present study, the linearity of a calibration curve was assessed by its correlation coefficient (r 2 ).This study yielded r 2 of 1.0000 for methanol, MEK, toluene, BA, and EA, 0.9999 for MIBK, 0.9998 for IPA, and 0.9997 for CHA.Though slopes of calibration curves reveal that the analytical signals of methanol and CHA were relatively weak among the eight target solvents, but the linearity of all calibration curves complied with the criterion of r 2 ≥ 0.995.
Table 4 shows desorption efficiencies of the eight target organic solvents for samples spiked with 0.5-, 1.0-, and 2.0-PEL concentrations using CS 2 + 5% (v/v) isobutanol as the desorption reagent.Results show that recovery rates of seven organic solvents were well above the criterion set at ≥ 75%, with the exception for methanol at 0.5-PEL concentration (= 73.63%).In addition, all resultant coefficient of variations (CV) complied with the criterion of ≤ 7%, and the relative difference (RD) between the maximum Table 3. Desorption efficiencies of the seven target organic solvents for the two desorption reagents of CS 2 + 5% (v/v) 1-Butanol and CS 2 ＋5% (v/v) isobutanol (n = 6).and minimum mean desorption efficiencies for all target organic solvents also complied with the criterion of ≤ 7%.
The above results suggest that the mixture of CS 2 + 5% (v/v) isobutanol would be feasible for the desorption of samples collected from the plastic material printing industry for conducting multiple organic solvent exposure assessments.

Stabilities of Samples
Fig. 2 shows stabilities of all target organic solvents containing in charcoal tubes during a 30-day test period and being stored at the room temperature (25°C) and -10°C environments, respectively.The prepared concentrations for the seven target solvents were the same as those used in Fig. 1. Results show that no significant difference in recovery rates could be found for EA, toluene, and n-BA while samples stored at both the room temperature (25°C) and -10°C environments.The resultant coefficient of variation (CV) of recovery rates for the above three solvents during the 30-day test period were consistently less than 1%, with the exception for BA on the 30 th day while stored at room temperature (= 3.19%; could be possibly explained by the random error from of laboratory).Nevertheless, all these values still complied with the criterion set at 7%.The above results suggest that storage temperature does not have a significant effect on recovery rates for the above three compounds.Comparing with the storage durations and storage temperatures specified in NIOSH methods 1457, 1501, and 1450 respectively for EA (6 days @ 5°C), toluene (30 days @ 5°C), and n-BA (30 days @ 4°C), our results suggest a broader storage range in both storage day and temperature would be feasible for the above three compounds.
Though recovery rates of IPA and MIBK were significantly higher for samples stored at -10°C environment than those at room temperature, all resultant recovery rates were still greater than 75%.The resultant CV of recovery rates for the above two solvents during the 30-day test period were consistently less than 1% with the exception for MIBK on the 20 th day (= 4.53%).The above results also suggest that both storage temperatures were suitable for both compounds.Comparing with the storage durations and storage temperatures specified in NIOSH methods 1400, and 1300 respectively for IPA (unknown, store in freezer), and MIBK (unknown, refrigerated), our results suggest a clearer storage period and a broader temperature range would be suitable for the above three compounds.
For methanol, MEK, and CHA, all resultant recovery rates were less than the criteria of 75% for those charcoal tube samples stored at room temperature after 1-day storage.In particular, the resultant CV of the recovery rate for the above three compounds obviously exceed the criterion of 7% after 3-, 11-, and 20-day storages.The above results clearly indicate that the room temperature environment was not suitable for storing charcoal tube samples while the above three were included in the analyzed target compounds.On the other hand, while charcoal tube sample were stored at the -10°Cenvironment, all resultant recovery rates for the above three compound were higher than 75%, and corresponding CVs were lower than 7% during the 30-day test period.The above results indicate that the storage of charcoal tube samples at the -10°Cenvironment would be necessarily.Finally, as we examine the storage durations and storage temperatures specified in NIOSH methods 2000, 2500, and 1300 respectively for methanol (at least 30 days at 5°C), MEK (at least 90 days @ -5°C), and CHA (unknown), our results were somewhat comparable with the former two, but can provide clearer information for the last one.
Based on the information obtained from this part of study, it is concluded that charcoal tube samples collected from the plastic material printing industries should be stored at the -10°Cenvironment and be analyzed within 30 days.

Conducting Exposure Assessment in a Plastic Material Printing Industry
To test the feasibility of the developed sampling and analysis method, an exposure assessment study was conducted on 169 workers in the plastic material printing industry.Results show that the eight targeted solvents were detectable from collected samples.Among them, the four solvents of the IPA, MEK, EA and toluene were found with relatively high exposure concentrations.However, neither of them exceeded their PELs.Nevertheless, judging from the cumulative exposure of the eight analyzed solvents (defined as the summation of the exposure concentration divided by its PEL of all analyzed solvents), ~14.8% of workers exceeded permissible exposure level (i.e., 1.0), ~34.9% were above 0.5 permissible exposure level, and particularly ~1.2% of workers reached 2.0 permissible exposure level.The above results clearly indicate the importance for conducting multiple organic solvent samplings for assessing the exposures of printing industry workers, and the suitability for using the sampling and analysis method developed in the present study.

CONCLUSIONS
In this study, the coconut shell charcoal was suggested as the sorbent media for collecting samples from the plastic material printing industry for multiple organic solvent exposure assessments.The collected samples were suggested subsequently using GC/FID for analyzing all target organic solvents.Though results of chromatographs using the two proposed desorption reagents showed good separation for all target organic solvents, the recovery rates obtained from the reagent of CS 2 + 5% (v/v) isobutanol were higher than those of CS 2 + 5% (v/v) 1-butanol.The linearity of all resultant calibration curves complied with the criterion of r 2 ≥ 0.995.For samples being spiked with 0.5-, 1.0-, and 2.0-PEL concentrations, all resultant recovery rates for all target organic solvents were well above the criterion set at 75%, with the exception for the methanol at 0.5-PEL concentration.Considering all resultant coefficient of variance (CV) comply with the criterion of ≤ 7% and the relative deviations (RD) between maximum and minimum desorption efficiencies were also below the criterion of ≤ 7%, the reagent of CS 2 + 5% (v/v) isobutanol was chosen for sample desorptions.The recovery rates and their corresponding CVs for EA, toluene, n-BA, IPA and MIBK for charcoal tube samples while stored at both the room temperature (25°C) and -10°C environments during the 30-day test period were consistently greater than 75% and less than 1%, respectively.But for methanol, MEK, and CHA, only while charcoal tube samples stored at the -10°C environment would result in recovery rates and their corresponding CVs complying with the above criteria during the 30-day test period.For conducting multiple organic solvent exposure assessments in plastic printing industries, it is recommended that the collected charcoal tube samples should be stored at the -10°C environment and be analyzed within 30 days.

Fig. 2 .
Fig. 2. Stabilities of targeted organic solvents containing in charcoal tube samples stored at (A) room temperature (25°C) and (B) -10°C environments during a 30-day test period.

Table 1 .
NIOSH sampling method for the seven target organic solvents.
to 120°C at 70 °C/min and held for 0.71 minutes.Therefore, the sample analysis could be completed in ~7.30 minutes.

Table 2 .
NIOSH analytical method for the seven target organic solvents.