This article describes a device intended to produce replicas on filters by liquid deposition of anion or metal solutions. Schematically, the filters are housed in cassettes labelled automatically by means of a code. An automatic arm takes each cassette, reads the code, and deposits the amount of element required. Weighing before and after deposition allows the amount deposited to be accurately checked and determined. This automated system allows the production of replicas with high deposition regularity, replica dispersion for the most part being <1%. The samples produced can be used during proficiency tests where the assigned value is determined either by the participants or by the organizer.
The main aim of proficiency tests is to procure information for participating laboratories allowing them to demonstrate or improve the quality of their analyses. Exploiting the results of proficiency tests also allows the achievement of other objectives such as comparing different analytical methods, evaluating a new method compared to an existing standardized method, supporting accreditation bodies or public authorities, and demonstrating the analytical skills of laboratories to customers (Boley, 1998). For all these reasons, this type of test is tending to multiply, thereby requiring the preparation of an increasing number of samples.
For instance, in the field of occupational health, there exist a large number of tests organized by different bodies, for example American Industrial Hygiene Association in USA, Health and Safety Laboratories in UK, Institut National de Recherche et Sécurité in France, and Berufsgenossenshaftliches Institut für Arbeitsschutz in Germany (Schlecht et al., 1997; Breuer et al., 2005; Langlois et al., 2008; Stacey and Butler, 2008).
The constraints linked to preparing these samples depend partly on the analytical method to be employed. If this involves a direct analysis requiring prior calibration, the production method should be as close as possible to the method used to sample the airborne pollutant in the workplace. Hence, for the quantitative analysis of quartz by a direct method (X-ray diffraction or infrared spectroscopy), standards must be produced in the same way as any samples that may be taken at workplaces with the intention of measuring pollutant concentration (e.g. air sampling using a particle selector). This condition indeed guarantees identical particle deposition on the sampling support in both cases (standards and samples). On the other hand, if the analytical method is an indirect method, the morphology of the particles and the deposition surface in particular have less impact on the result of the analysis, leaving more flexibility in terms of preparation method. This is the case, for example, for the quantitative analysis of certain inorganic particulate aerosols (metals, metalloids, and acids), where the atomic spectrometry or ion chromatography analysis requires preliminary dissolution (ISO 15202-2, 2001; MétroPol 009, 2005; MétroPol 003, 2008; ISO/DIS 30011; ASTM Standard D7439, 2008; ISO/DIS 21438), and the main source of uncertainty would originate from sample preparation.
To test the proficiency of laboratories to carry out a quantitative analysis of this type of aerosol in accordance with the methods laid down in the different standards, the proficiency test, ALASCA (French acronym of Ability of Laboratories to Analyse Airborne Chemical Products), was developed for four metal elements (cadmium, chromium, nickel, and lead) and three anions (fluoride, chloride, and nitrate) (Langlois et al., 2008). This test focuses solely on analytical method. This is why the sampling support spiking is done by depositing metal or anion solutions, the application of which is easier than the generation of a homogeneous particle deposit. The aim of this article is to describe an automated system (SAGE: Automated sample generation system) that has been designed to deposit metal or anion solutions on filters that can be used in proficiency tests, in particular the ALASCA tests. In addition, recent work has highlighted the importance of taking wall deposits into account, particularly when sampling an aerosol by means of a closed cassette (Harper and Demange, 2007). This is the reason why this automated system has been designed specifically to allow the spiking of a sampling support previously fitted inside a cassette, the aim being to allow the use of a method that generates the solution directly in the sampling cassette (ISO 15202-2, 2001, Appendix D and MétroPol 003, 2008). This automated system enables the generation of large series of samples whose composition is accurately known, with very high reproducibility, in order to limit any bias caused by sample preparation.
The system has been designed to be upgraded easily. For instance, it could be fitted with various syringes (from 5 μl to 12.5 ml), cassettes (25 and 37 mm in diameter), and filter media. In the following, the generation of samples in the framework of the ALASCA proficiency test will be described, as it is the only one that has been presently validated.
The principle of producing samples from the SAGE automated system, which was manufactured by the company IMF (Maxéville, France), is the following:
The solutions to be deposited are made either from certified commercially available diluted solutions or from pure salts dissolution. The deposit solution concentration is calculated for a deposit varying from 25 to 500 μl per sample.
Cassettes containing new quartz fibre filters are labelled. The labels generated include a 2D code and the denomination of the sample. The cassettes are placed randomly on plates.
An automatic arm takes each cassette, the 2D code is read, and the volume of solution to be deposited, a function of the amount of element required, is deposited on the filter. Weighing prior and subsequent to deposition allows this to be checked. The weighing accuracy is 0.1 mg but could be improved to 0.01 mg.
The addition of n different elements to the filter is achieved by successive passages of the cassettes after changing the solution to be deposited.
All the data are archived on a computer in an ACCESS® database. The automated system is computer controlled with a programme developed in Windows XP®. The automation-system part of the programme was generated using CODESYS®. Figure 1 presents the different elements of the system.
Fig. 1.
Open in new tabDownload slidePhotograph of the system showing, from left to right: the cassette plate (1), the moving arm (2), the automatic syringe (3), the solution receptacle (4), the optical reader (5), and the balance (6).
The different sample production steps are detailed below.
The cassettes are labelled by means of a label printer (model CAB a3), which has a rotating mechanism fitted to it allowing the labels to be affixed automatically on the cassettes. The labels chosen are made of polypropylene, and printing is by thermal transfer. They are resistant to water, acids, and ultrasonic bath treatment. The label is positioned in the lower section of the cassettes.
The software sends the information for printing (sample number) from the ACCESS® database to the printer, which in turn generates a label that includes the name of the sample and its 2D code. Publishing is triggered by the insertion of the cassette in the rotating mechanism. Because of the reduced size of the 2D code, this can be printed twice on the label according to the sequence ‘2D Code–Sample number–2D Code’. This repetition speeds up reading of the 2D code in relation to the location of the cassette in front of the optical reader and increases reading reliability in the case where one of the two 2D codes may have been incorrectly printed.
The sample numbers are managed in ACCESS®; an indicator is flagged up in the database when the label is issued, which makes manual reprinting complicated and avoids any risk of double labelling.
A pneumatic cassette holding clamp, the solution container, and the automatic syringe are fitted to a mobile arm capable of movement in the x-, y-, and z-axes (Rexroth Cartesian Motion System).
The solution receptacles are flasks with a 500 ml capacity made of borosilicate glass. These flasks are easy to clean. The stopper is fitted with a gland allowing entry of the tubing.
The automatic syringe (Hamilton PSD3) has a volume of 25 μl. Deposits of between 25 and 500 μl are made by programming one or several full piston strokes.
This volume is programmable in 1-μl steps. Spiking system purging is also fully programmable.
The concentration of the deposit solution is determined according to the French occupational exposure limit values of each analyte, given that the volume to be spiked on the filter is <500 μl. For instance, the concentrations of analytes in deposit solutions used for the ALASCA tests are 50, 100, 500, and 1000 μg ml−1 for Cd, Pb, Cr, and Ni, respectively, and 2.5, 5 and 7.5 mg ml−1 for F−, NO3−, and Cl−, respectively.
The following are placed in the operational zone of the moving arm:
A plate comprising 48 emplacements allowing exact positioning of the cassettes. The plates are stored in a cabinet intended for this purpose.
A 2D code reader (Opticon) and a rinsing flask.
An analytical balance (Mettler Toledo AB54S/FACT). This balance has a range of 51 g and a display accuracy of 0.1 mg.
The pneumatic clamp takes hold of a cassette on the plate. The arm deposits the cassette on a revolving plate for automatic reading of the 2D code. This code, via the database, allows the volume of solution for deposit to be known as a function of the element selected. The cassette is transferred to an analytical balance. The mass is read after stabilization. The arm then positions the syringe at the centre of the filter and deposits the volume of liquid corresponding to the cassette number read by the 2D code reader. The mass is again read after spiking, following stabilization of the balance. The arm then takes the cassette, immerses the needle in the deposit solution to remove residual drops, and returns the cassette to its place on the plate. One syringe is used per analyte.
The spiking system is made of inert materials:
The flask containing the solution and the body of the syringe are made of borosilicate glass.
The tubing carrying the solution, the valve of the syringe, and the part of the piston in contact with the liquid are made of Teflon.
The needle of the syringe is made of polyetheretherketone.
The cassette management software uses an ACCESS® database, which manages the samples. Indicators are flagged up once a cassette has been labelled and spiking completed, making any sample duplication impossible without manual intervention in the database.
The spiking cycle can only take place if the cassette is properly positioned on the plate.
If the 2D code cannot be read, the cassette is returned to the plate with no deposit made. An indicator appears on the control panel to identify the cassette and to flag up the anomaly.
A synoptic table of the cassettes positioned on the plate allows any problems related to code non-reading or beyond-limit spiking to be viewed on the screen in real time. All the information available concerning a current plate sample can be displayed at any time.
A hood covering the entire system prevents any intervention during the spiking cycles. Opening one of the panels causes the immediate stoppage of SAGE. This stoppage must be acknowledged at computer level for spiking to be resumed.
The deposited solution volume V is given by the following relationship:with mthe mass of solution deposited, d the density of the solution deposited, Z the correction factor taking temperature into account in the mass per volume relationship ( ISO 8655-6, 2002 , Appendix A), and FE the evaporation factor which takes into account the amount of liquid having evaporated between the first and second weighing
Parameter d is measured at the time of producing the deposit solution.
Parameter md is measured at the time of depositing the solution.
Factor Z is calculated from the continuous temperature measurement.
Factor FE was determined experimentally by measuring the loss of mass due to evaporation of a distilled water deposit over a period equivalent to twice the deposit time. Whatever the volume deposited (from 25 to 500 μl), the evaporation factor FE is constant. It was equal to 0.54% for our experimental conditions.
The mass M of element deposited on a filter is obtained from the following relationship:with Cthe known element concentration of the deposited solution.
Series of 10-anion-spiked (fluoride, chloride, and nitrate) or metal-spiked (cadmium, chromium, nickel, and lead) replicas were produced with the aim of calculating deposit homogeneity.
The uncertainty of the mean deposited volume V, calculated from the mass md of solution deposited, is defined as being twice the random error sr. This random error corresponds to deposit dispersion and so to the standard deviation of repeatability (sr).
The two tables below recapitulate the results of these experiments (Table 1 for anions, Table 2 for metals). In total, 150 anion deposits and 200 metal deposits were made. The expanded uncertainty relative to the mean deposited volume Imd is equal to twice the random error (assuming k = 2).
Table 1.
Solution Set volume (μl) Mean deposited volume (μl) sr (%) Imd (%) Fluoride 25 25.05 0.19 0.39 50 50.12 0.14 0.28 100 100.05 0.26 0.53 200 200.34 0.31 0.62 400 400.56 0.13 0.26 Mean 0.21 0.41 Chloride 25 25.05 0.43 0.86 50 50.12 0.44 0.87 100 100.05 0.23 0.45 200 200.34 0.32 0.63 400 400.56 0.14 0.28 Mean 0.31 0.62 Nitrate 25 25.12 0.17 0.34 50 51.07 0.16 0.32 100 100.34 0.07 0.14 200 200.41 0.12 0.24 400 401.07 0.11 0.23 Mean 0.13 0.25 Solution Set volume (μl) Mean deposited volume (μl) sr (%) Imd (%) Fluoride 25 25.05 0.19 0.39 50 50.12 0.14 0.28 100 100.05 0.26 0.53 200 200.34 0.31 0.62 400 400.56 0.13 0.26 Mean 0.21 0.41 Chloride 25 25.05 0.43 0.86 50 50.12 0.44 0.87 100 100.05 0.23 0.45 200 200.34 0.32 0.63 400 400.56 0.14 0.28 Mean 0.31 0.62 Nitrate 25 25.12 0.17 0.34 50 51.07 0.16 0.32 100 100.34 0.07 0.14 200 200.41 0.12 0.24 400 401.07 0.11 0.23 Mean 0.13 0.25 Open in new tabTable 1.
Solution Set volume (μl) Mean deposited volume (μl) sr (%) Imd (%) Fluoride 25 25.05 0.19 0.39 50 50.12 0.14 0.28 100 100.05 0.26 0.53 200 200.34 0.31 0.62 400 400.56 0.13 0.26 Mean 0.21 0.41 Chloride 25 25.05 0.43 0.86 50 50.12 0.44 0.87 100 100.05 0.23 0.45 200 200.34 0.32 0.63 400 400.56 0.14 0.28 Mean 0.31 0.62 Nitrate 25 25.12 0.17 0.34 50 51.07 0.16 0.32 100 100.34 0.07 0.14 200 200.41 0.12 0.24 400 401.07 0.11 0.23 Mean 0.13 0.25 Solution Set volume (μl) Mean deposited volume (μl) sr (%) Imd (%) Fluoride 25 25.05 0.19 0.39 50 50.12 0.14 0.28 100 100.05 0.26 0.53 200 200.34 0.31 0.62 400 400.56 0.13 0.26 Mean 0.21 0.41 Chloride 25 25.05 0.43 0.86 50 50.12 0.44 0.87 100 100.05 0.23 0.45 200 200.34 0.32 0.63 400 400.56 0.14 0.28 Mean 0.31 0.62 Nitrate 25 25.12 0.17 0.34 50 51.07 0.16 0.32 100 100.34 0.07 0.14 200 200.41 0.12 0.24 400 401.07 0.11 0.23 Mean 0.13 0.25 Open in new tabTable 2.
Solution Set volume (μl) Mean volume deposited (μl) sr (%) Imd (%) Cadmium 25 25.10 0.48 0.96 50 50.14 0.16 0.32 100 100.24 0.13 0.27 200 200.31 0.16 0.32 400 401.07 0.10 0.20 Mean 0.21 0.41 Chromium 25 25.03 0.23 0.45 50 50.18 0.17 0.34 100 100.35 0.10 0.21 200 200.28 0.23 0.45 400 400.89 0.18 0.36 Mean 0.18 0.36 Nickel 25 24.48 2.24 4.48 50 50.11 0.17 0.34 100 99.02 0.60 1.20 200 199.96 0.11 0.21 390 390.07 0.15 0.30 Mean 0.65 1.31 Lead 25 24.81 0.86 1.72 50 49.41 1.05 2.10 100 99.50 0.59 1.17 200 200.04 0.12 0.25 400 399.76 0.13 0.26 Mean 0.55 1.10 Solution Set volume (μl) Mean volume deposited (μl) sr (%) Imd (%) Cadmium 25 25.10 0.48 0.96 50 50.14 0.16 0.32 100 100.24 0.13 0.27 200 200.31 0.16 0.32 400 401.07 0.10 0.20 Mean 0.21 0.41 Chromium 25 25.03 0.23 0.45 50 50.18 0.17 0.34 100 100.35 0.10 0.21 200 200.28 0.23 0.45 400 400.89 0.18 0.36 Mean 0.18 0.36 Nickel 25 24.48 2.24 4.48 50 50.11 0.17 0.34 100 99.02 0.60 1.20 200 199.96 0.11 0.21 390 390.07 0.15 0.30 Mean 0.65 1.31 Lead 25 24.81 0.86 1.72 50 49.41 1.05 2.10 100 99.50 0.59 1.17 200 200.04 0.12 0.25 400 399.76 0.13 0.26 Mean 0.55 1.10 Open in new tabTable 2.
Solution Set volume (μl) Mean volume deposited (μl) sr (%) Imd (%) Cadmium 25 25.10 0.48 0.96 50 50.14 0.16 0.32 100 100.24 0.13 0.27 200 200.31 0.16 0.32 400 401.07 0.10 0.20 Mean 0.21 0.41 Chromium 25 25.03 0.23 0.45 50 50.18 0.17 0.34 100 100.35 0.10 0.21 200 200.28 0.23 0.45 400 400.89 0.18 0.36 Mean 0.18 0.36 Nickel 25 24.48 2.24 4.48 50 50.11 0.17 0.34 100 99.02 0.60 1.20 200 199.96 0.11 0.21 390 390.07 0.15 0.30 Mean 0.65 1.31 Lead 25 24.81 0.86 1.72 50 49.41 1.05 2.10 100 99.50 0.59 1.17 200 200.04 0.12 0.25 400 399.76 0.13 0.26 Mean 0.55 1.10 Solution Set volume (μl) Mean volume deposited (μl) sr (%) Imd (%) Cadmium 25 25.10 0.48 0.96 50 50.14 0.16 0.32 100 100.24 0.13 0.27 200 200.31 0.16 0.32 400 401.07 0.10 0.20 Mean 0.21 0.41 Chromium 25 25.03 0.23 0.45 50 50.18 0.17 0.34 100 100.35 0.10 0.21 200 200.28 0.23 0.45 400 400.89 0.18 0.36 Mean 0.18 0.36 Nickel 25 24.48 2.24 4.48 50 50.11 0.17 0.34 100 99.02 0.60 1.20 200 199.96 0.11 0.21 390 390.07 0.15 0.30 Mean 0.65 1.31 Lead 25 24.81 0.86 1.72 50 49.41 1.05 2.10 100 99.50 0.59 1.17 200 200.04 0.12 0.25 400 399.76 0.13 0.26 Mean 0.55 1.10 Open in new tabThese data are represented in Fig. 2 in the form of histogram of distribution in percentage of the relative errors between the deposited volume and the mean volume for the different series studied.
Fig. 2.
Open in new tabDownload slideHistogram of the distribution of the relative errors between the deposited volume and the mean volume for the different series studied.
The samples sent out within the framework of these proficiency tests are not replicas. Each sample has a unique assigned value (M) per element determined by weighing the deposit.
The uncertainty of the assigned value was calculated from the quadratic sum of the uncertainties of different parameters (CSD, d, Z, FE, and md). Standard NF EN ISO 8655-6 (2002) served as the reference to determine these uncertainties. The results are given in Table 3.
Table 3.
Parameter md (%) d (%) Z (%) FE (%) CSD (%) Relative uncertainty 0.1 0.01 0.1 0.5 Parameter md (%) d (%) Z (%) FE (%) CSD (%) Relative uncertainty 0.1 0.01 0.1 0.5 Open in new tabTable 3.
Parameter md (%) d (%) Z (%) FE (%) CSD (%) Relative uncertainty 0.1 0.01 0.1 0.5 Parameter md (%) d (%) Z (%) FE (%) CSD (%) Relative uncertainty 0.1 0.01 0.1 0.5 Open in new tabThe relative uncertainty of the deposited solution concentration is equivalent whether this is prepared from a certified commercially available solution or from pure salts.
Combining all these uncertainties enables the calculation of the total uncertainty of the assigned value of an element of a sample.
Typically, the relative uncertainty of the mass of element deposited is <1%.
For the 2008 ALASCA campaign, 1200 samples were prepared: 720 samples for the metals campaign and 480 samples for the anions campaign.
To produce these samples, 2880 deposits of metal-containing solutions and 1440 deposits of anion-containing solutions were made by the SAGE automated system. The mean difference between the set volumes and the real volumes was 0.06% with a standard deviation of 1.19% (Fig. 3).
Fig. 3.
Open in new tabDownload slideHistogram of the distribution in percentage of the relative errors between the set volume and the deposited volume for the samples distributed in the 2008 ALASCA campaign.
These samples were distributed and analysed by the participants in the 2008 metals and anions campaigns.
S: standardized result, R: result of the participant, and VA: assigned value of the element on the sample.For each result handed in, a standardized result was calculated:with R: standardized result, R: result of the participant, and VA: assigned value of the element on the sample.
Aberrant results (RS < 50 or RS > 150) were deleted before calculating the statistics presented in Table 4.
Table 4.
Element Fluoride Chloride Nitrate Cadmium Chromium Nickel Lead Mean 97.00 98.03 99.26 95.32 95.79 97.30 97.00 Median 97.11 98.27 99.65 96.50 96.45 98.31 97.18 Standard deviation 10.00 9.09 7.88 7.80 6.13 6.89 8.28 Element Fluoride Chloride Nitrate Cadmium Chromium Nickel Lead Mean 97.00 98.03 99.26 95.32 95.79 97.30 97.00 Median 97.11 98.27 99.65 96.50 96.45 98.31 97.18 Standard deviation 10.00 9.09 7.88 7.80 6.13 6.89 8.28 Open in new tabTable 4.
Element Fluoride Chloride Nitrate Cadmium Chromium Nickel Lead Mean 97.00 98.03 99.26 95.32 95.79 97.30 97.00 Median 97.11 98.27 99.65 96.50 96.45 98.31 97.18 Standard deviation 10.00 9.09 7.88 7.80 6.13 6.89 8.28 Element Fluoride Chloride Nitrate Cadmium Chromium Nickel Lead Mean 97.00 98.03 99.26 95.32 95.79 97.30 97.00 Median 97.11 98.27 99.65 96.50 96.45 98.31 97.18 Standard deviation 10.00 9.09 7.88 7.80 6.13 6.89 8.28 Open in new tabThe results detailed in this article show that the SAGE automated system allows a very precise production of replicas; this in turn allows them to be supplied for proficiency tests where the assigned value is determined from the results of the participants.
The data presented in the histogram of Fig. 2 show that 95% of the deposits are located at ±1% of the set volume. The highest divergences were noted when depositing the nickel and lead solutions. Phenomena of liquid re-entering the syringe due to capillary action after a certain period of use would appear to be the likely cause of the increase in the spiking uncertainties for this series. Improving the maintenance schedule of the deposition needle should allow an improvement in the uncertainty related to mean deposited volume. Whatever the case, since the deposited volume is measured, it is still possible to eliminate replicas to obtain the variability required.
Knowledge of the concentrations of the mother solutions used and the dilutions made allows the calculation of an assigned value per sample produced with a very high accuracy. This is the choice made within the framework of the ALASCA tests. In this case, the bias and the dispersion of the laboratory results can be calculated. As an example, the results presented in Table 4 show that the mean bias of the laboratories for the 2008 campaign was <5%.
The device presented here to produce replicas is complementary to other devices described in the literature where the particles to be deposited are generated in the air (Baron and Deye, 1987; Anglov et al., 1993; Dyg et al., 1994; Skogstad et al., 1996; Stacey and Butler, 2008). This automated method to generate replicas for proficiency tests presents many advantages:
Low assigned value uncertainty: in this respect, the assigned value can be set by the organizer, allowing the extraction of other information related to the results of a campaign (comparison of techniques, bias of laboratories, etc.).
Very good reproducibility: the production of replicas by solution deposition guarantees higher replica accuracy (on the order of 1%) compared to replicas produced by air sampling (in general a few percent).
Traceability of the deposits made: management of cassette identification and amount deposited in the cassette by a computerized application guarantees this traceability and reduces the risk of spiking error.
Random distribution of replicas to different participants: although the different laboratories participating in an annual campaign analyse the same replicas in terms of amount of material deposited per element, these are distributed randomly, by means of a computer programme, during the three tests comprising a campaign. This limits the risk of collusion between laboratories and therefore guarantees test fairness between the different participants.
The permanent presence of an operator is unnecessary. The deposition phase requires very few manual interventions (changing the deposition solution and changing the plate), the remainder being fully automated. Moreover, the control panel allows an at-a-glance view of the smooth execution or otherwise of a deposition session.
Possibility of modifying the support: within the bounds of the impregnation ability of a support, it is possible to use other sampling supports (mixed cellulose ester membranes, for example).
In conclusion, this automated system, on account of the regularity and traceability of the operations, provides a valuable aid to the organizers of proficiency tests in relation to compliance with certain requirements of quality assurance reference systems.
Acknowledgements—The authors would like to thank Mr Bruno Lacroix for his participation in the design of the device.
Anglov
JTB
Holst
E
Dyg
S
Iron, manganese, copper and titanium in welding fume dust on filters for internal and external quality assurance
,Fresenius J Anal Chem
,1993
, vol.245
(pg.335
-9
), vol.(pg.
ASTM Standard D7439
,Standard test method for determination of elements in airborne particulate matter by inductively coupled plasma–mass spectrometry
,2008
West Conshohocken, PA
ASTM International
pg.24
pg.
Baron
PA
Deye
GJ
Generation of replicate asbestos aerosol samples for quality assurance
,Appl Ind Hyg
,1987
, vol.2
(pg.114
-8
), vol.(pg.
Boley
N
Benefits arising from proficiency testing schemes: the comparison of analytical methods
,Accred Qual Assur
,1998
, vol.3
(pg.459
-61
), vol.(pg.
Breuer
D
Maybaum
B
Gusbeth
K
Ringversuch mit Probennahme zur Bestimmung von flüchtigen anorganischen Säuren (HCl und HNO3)
,Gefahrstoffe–Reinhaltung der Luft
,2005
, vol.65
(pg.122
-8
), vol.(pg.
Dyg
S
Anglov
T
Christensen
JM
Preparation of filters loaded with welding dust. A homogeneity and stability study of hexavalent chromium
,Anal Chim Acta
,1994
, vol.286
(pg.273
-82
), vol.(pg.
Harper
M
Demange
M
Concerning sampler wall deposits in the chemical analysis of airborne metals
,J Occup Environ Hyg
,2007
, vol.4
(pg.81
-6
), vol.(pg.
ISO 15202-2
,Workplace air—determination of metals and metalloids in airborne particulate matter by inductively coupled plasma atomic emission spectrometry—Part 2: sample preparation
,2001
Geneva, Switzerland
International Organization for Standardization
pg.38
pg.
ISO 8655-6
,Piston-operated volumetric apparatus—Part 6: gravimetric methods for the determination of measurement error
,2002
Geneva, Switzerland
International Organization for Standardization
pg.14
pg.
ISO/DIS 21438
,Workplace atmospheres—determination of inorganic acids by ion chromatography (3 parts)
Geneva, Switzerland
International Organization for Standardization
ISO/DIS 30011
,Workplace air—determination of metals and metalloids in airborne particulate matter by inductively coupled plasma mass spectrometry
Geneva, Switzerland
International Organization for Standardization
pg.32
pg.
Langlois
E
Boulet
A
Kauffer
E
ALASCA proficiency testing scheme for occupational hygiene laboratories
,J Environ Monit
,2008
, vol.10
(pg.1460
-6
), vol.(pg.
MétroPol 003
,Métaux, Métalloïdes
,2008
Vandœuvre les Nancy, France
Institut National de Recherche et Sécurité
pg.20
pg.
MétroPol 009
,Anions minéraux
,2005
Vandœuvre les Nancy, France
Institut National de Recherche et Sécurité
pg.30
pg.
Schlecht
PC
Song
R
Groff
JH
Interlaboratory and intralaboratory variabilities in the Environmental Lead Proficiency Analytical Testing (ELPAT) Program
,Am Ind Hyg Assoc J
,1997
, vol.58
(pg.779
-86
), vol.(pg.
Skogstad
A
Eduard
W
Huser
PO
A laboratory method for generation of replicate filter samples of asbestos fibers in air
,Am Ind Hyg Assoc J
,1996
, vol.57
(pg.741
-5
), vol.(pg.
Stacey
P
Butler
O
Performance of laboratories analysing welding fume on filter samples: results from the WASP proficiency testing scheme
,Ann Occup Hyg
,2008
, vol.52
(pg.287
-95
), vol.(pg.
© The Author 2009. Published by Oxford University Press on behalf of the British Occupational Hygiene Society
With high quality products and considerate service, we will work together with you to enhance your business and improve the efficiency. Please don't hesitate to contact us to get more details of Automatic Sample Tester.