Document KJo2Q5kVwRO6dVmVkkv09MNV6

Instrumentation for Dust Measurement JOINT EUROPEAN INVESTGATIONS OF NEW GENERATIONS OF DUST SAMPLING INSTRUMENT J. H. VINCENT Institute of Occupational Medicine, 8 Roxburgh Place Edinburgh EH8 9SU, Scotland, UK INTRODUCTION Considerable success has been achieved in understanding the predominant relationships between the risk of coalworkers' pneumoconiosis and exposure to fine particles of airborne coalmine dust. This has led to the setting of meaningful stan dards and, in turn, substantial reductions in die incidence of disease by improved dust suppression. However there is still die need for further improvement to deal with a number of important dust-related health problems which remain. In order to make progress in these areas, new research ques tions are posed requiring more detailed information about the properties ofthe airborne dust which cannot be obtained just by the use of instruments like those employed in much of die previous research. A need is therefore identified for a new generation of dust sampling instrument. Various new instruments have emerged in recent years, and it is timely to critically assess some of them in relation to die current research needs. To this end a Joint Project, involving six laboratories from five European Member States, has been carried out under the auspices of the Commission of Euro pean Communities (CEC). The participating laboratories were: Bergbau-Forschunginstitut GbmH, Essen, West Ger many (BF) Silikose-Forschunginstitut, Bochum, West Germany (SF) Centre d'Etudes et Recherche des Charbonnages de France, Verneuil en Halatte, France (CERCHAR) Insdtut d'Hygiene des Mines, Hasselt, Belgium (IHM) Institute di Medicina del Lavoro, Milan, Italy (IML) Institute of Occupational Medicine, Edinburgh, UK (IOM) This paper describes the project coordinator's preliminary assessment of what was achieved. RATIONALE The new generation of dust sampling instrument includes a range of particle size-selective devices from which informa tion about health-related fractions of airborne dust may be obtained. The rationale against which to compare and evaluate these instruments was based primarily on the 1983 recommendations of the International Standards Organiza tion (ISO),1 updated where appropriate in the light of more recent experimental evidence. The dust fractions in question are inspirable (the fraction that enters through die 618 nose and/or mouth during breathing), thoracic (that penetrates below toe larynx) and respirable (that penetrates to the alveoli). Of these, it is the inspirable fraction which, when it is referred to below, has been updated from die 1983 ver sion, thus bringing it in to line with the definition contained in the 1985 recommendations of die American Conference of Governmental Industrial Hygienists (ACGIH).2 In addi tion, since it forms die basis of present sampling in some European countries, an alveolar fraction was also included, describing a fine fraction which takes account of die fact that--in actual human exposure--the finest inhaled particles remain airborne for long enough to be exhaled during the exhalation phase of the breathing cycle.3 The important philosophy embodied in die ISO recommendations is that all the dust fractions which deposit in the regions of the respiratory tract are subfractions of the inspirable fraction. This is consistent with what happens during actual human exposure. Ideally, it should also be reflected in sampler per formance; namely that die efficiency with which particles enter the sampler in the first place should match die inspirable fraction. The information provided by the various instruments includes details not only about airborne mass concentrations within defined fractions but also about mineralogical composition (and, possibly, physical properties such as particle shape). The instruments themselves fall into two categories. In die first, dedicated instruments sample to a given, single criterion (e.g., respirable dust), although some can also provide in formation about `total' dust. In the second, more versatile instruments--broadly referred to as spectrometers--can pro vide a wider range ofinformation. These operate cm die prin ciple that, if a defined fraction of airborne dust can be aspirated and its particle aerodynamic size distribution subse quently obtained, then all the information is available to allow determination of the particle aerodynamic size distribution and airborne mass concentration of any health-related sub fraction which can be defined numerically. If die dust thus classified can be recovered in sufficient quantities for analysis, then the mineralogical composition of such subfrac tions can also be determined. THE INSTRUMENTS During a Workshop which took place in Edinburgh early on in die project, involving all die participants, the following instruments were identified for inclusion in the study:  The French 10 1/min CIP10 personal sampler for respirable dust, also capable of providing a measure of `total' dust. Its pre-selector operates on die principle of filtration by porous foam filtration media. The Italian 2 1/min modified-Zurlo (M-Z) personal sampler for respirable dust. Its pre-selector operates on the principle of virtual impaction. The Italian 3.5 1/min personal sampler for `total* dust (TD) The British 3 1/min static inspirable dust sampler dOMID). The Italian 0.41/min static dust spectrometer (INSPEQ. It operates on the principle of inertial classification. The Italian 2 1/min personal dust spectrometer (PERSPEC). This too operates on the principle of iner tial classification. The German 40 1/min static dust spectrometer (PCI). It operates on the cascade impactor principle. The British 101/min static inspirable dust spectrometer (SIDS), also operating on the cascade impactor principle. The British 21/min personal inspirable dust spectrometer (PIDS), also operating on the cascade impactor principle. In addition to these, various instruments from the previous generation were also included for the purpose of comparison. These were: The German 50 1/min cyclone-based static sampler for alveolar dust (TBF50), also capable of providing a measure of `total* dust. The German 461/min horizontal elutriator-based static sampler for respirable dust (MPGII). The Italian 2.4 1/min cyclone-based personal sampler for respirable dust (CYCLO). The French 50 1/min cyclone-based static sampler for alveolar dust (CPM3). The British 2.5 1/min horizontal elutriator-based static sampler for respirable dust (MRE). The British 1.91/min cyclone-based personal sampler for respirable dust (SIMPEDS). The Belgian 17 1/min static sampler for total dust (STASER). The mean features ofdie above instruments are summarized in Table I. THE PROGRAMME OF WORK The research was carried out during the 3-year period 1985 to 1988. During die Edinburgh Workshop, it was agreed that, as far as possible, each of the principal instruments iden tified for inclusion in the trial should be evaluated by more than one laboratory and that each laboratory should evaluate more than one instrument. Thus assessment of a given in strument is less likely to be biased by die findings of, say, just one laboratory. In addition, it was agreed that two of die instruments--namely the CIP10 and PCI--would be evaluated by all six participating laboratories. The project called for a programme ofcomparative trials both in the laboratory and underground in mines (pyrites for Italy, coal everywhere else). Each laboratory developed its own Instrumentation for Dust Measurement experimental strategy, determined by the resources and ex pertise at its disposal. Thus the emphasis varied considerably between laboratories. At IOM, for example, die main em phasis was placed on the laboratory aspect, based on the ex tensive facilities available (notably the large wind tunnel) and associated expertise. Elsewhere, greater importance was given to the underground trials. In some, greater stress was placed on the abilities of the instruments to provide infor mation about mineralogical composition; and, in others, on the basic performance characteristics (notably with respect to particle size-selectivity) ofdie individual devices. The net effect of all the complementary contributions has been to pro vide an overall, balanced programme of work, as sum marized in Table n. The experimental inquiry fell into three broad areas: Experiments to assess basic performance characteristics (e.g., aspiration efficiency, particle size selectivity). Measurements of concentrations of health-related dust fractions and subtractions. Measurements of mineralogical composition. RESULTS In this paper, only concise, largely qualitative summaries of the results available at the time ofwriting are given. Whilst most are based on information obtained directly during the Joint Project itself, some information obtained during other studies has also been taken into account in some cases. The full experimental and statistical details of the individual studies are given in the final reports ofthe six individual com ponent projects, while die combined analysis and overall con clusions will appear in the synthesis report which is still in preparation. Basic Performance Characteristics This aspect of the work was conducted in the laboratory. One area of interest is the efficiency with which particles enter the sampler initially. For ideal health-related sampling, this should match the inspirable fraction, since any samplers for which this is true are consistent with the ISO rationale re ferred to above. Experiments to assess entry efficiency were performed with this in mind, mostly in the large wind tun nel at IOM. All devices intended for use as personal samplers were tested in that mode, mounted on the torso of a tailor's mannequin which, during sampling, was rotated step-wise through 360 degrees (to eliminate preferred-orientation ef fects). In the case of the CIP10, it was also tested as a static sampler (since it is used by some workers in this mode). The results are summarized in Table HI where, here and in the following tables, the quantitative experimental information reported in the original investigations has been reduced to the qualitative form shown. At this stage, until further analysis of the data is carried out, it is possible only to place the instruments into arbitrarily-chosen broad performance categories, without reflecting the degree to which each either conforms or fails to conform. In Table m, therefore, `YES* indicates acceptance, with more than 50% of the available data points falling within 10 percentage points of the definition in question (inspirable or true total dust). *NO* indicates non-acceptance, with less than 50% ofthe available data points lying within the same band. In certain cases, the 619 Instrumentation for Dust Measurement Table 1 The Instruments Tested and Their Main Features 1---------------------Sampler Type Flowrate l/min Principle of Nominal size selection fraction Other fractions C1P10 M-Z TD IOMID INSPEC PERSPEC PCI SIDS PIDS Dedicated Dedicated Dedicated Dedicated Spectrometer Spectrometer Spectrometer 10 2 15 3 0.4 2 40 Spectrometer 10 Spectrometer 2 Porous foam filtration Virtual impaction Aspiration Aspiration Inertial separation Inertial separation Cascade impactor Cascade impactor Cascade impactor Respirable/ Total' alveolar Respirable - Total' Inspirable - - - -- True total + subfractions Inspirable subfractions Inspirable + subfractions " _ | 1 I TBF50 MPGII CYCLO CPM3 MRE SIMPEDS STASER Dedicated Dedicated Dedicated Dedicated Dedicated Dedicated Dedicated 50 Cyclone Alveolar/ Total' respirable 46 Horizontal Respirable - elutriator 14 Cyclone Respirable Total' 50 Cyclone Alveolar/ respirable - 15 Horizontal Respirable Total' elutriator 1.9 Cyclone Respirable Total' 17 Aspiration True total - judgement may be influenced also by any obvious contradic tory trends present in the data. The table shows that the PCI provides a fair sample of true total dust (not surprisingly, since sampling with this instrument is arranged to take place almost isokinetically by virtue of the choice of number of entry nozzles). So too (for similar reasons) should the STASER (although this has not been investigated experiment ally). The M-Z, IOMID, SIDS and PIDS all match the inspirability criterion quite well. So too does the CIP10 in its personal mode, but nof as a static sampler. The basic selectivities of the two new samplers dedicated (nominally) to the respirable dust fraction (CIP10 and M-Z respectively) were also assessed at some laboratories. For the CIP10, selectivity matches die BMRC-definition (as a subfraction of the inspirable fraction) quite well except at small particle sizes where the finest particles are not collected by the porous foam final collection stage of the instrument and so are lost. However, it is noted that the proportion of the mass carried by particles lost in this way may be expected always to be very small in most practical situations. In any case, it may be argued that die dust which is lost in this way 620 is roughly equivalent to that which is exhaled. Therefore die CIP10 selection curve has features in common with both the BMRC respirable dust definition and that for die alveolar fraction (although it matches neither perfecdy). For the M-Z, agreement with the BMRC-definition is fair. For die earliergeneration instruments, selectivity is available from previous ly published information. For these devices, it is worth noting that the TBF50 and the CPM3 both exhibit selection characteristics which more closely reflect true alveolar deposition. The MPGII and MRE both conform closely to the BMRC-definition. The INSPEC and PERSPEC require special comment. The performance of the first was found to exhibit effects associated in part with its low sampling flowrate; namely, biased entry characteristics (depending on the type of entry piece attached), high particle losses between die entry and die sensing region, and collected mass too small to allow gravimetric assessment. The first two effects are more pro nounced the larger die particle size. Furthermore, in its pres ent mains-powered version, INSPEC does not satisfy intrinsic safety criteria which would allow its use underground Table n Outline of Programme of Work Carried Out Instrumentation for Dust Measurement Sampler CIP10 M-Z TD IOMID INSPEC PERSPEC PCI SIDS PIDS LABORATORY BF SF CERCHAR IHM IML P L.U L.U.P.M U,M U,P U U.M L,U L L U.P.M U.M U IOM L,P L,P L,P U.P P P L.U.P L.U.P L,P TBF50 UU MPGII uu CYCUO CPM3 MRE S1MPEDS STASER L,U L.U.P - U.P L,P U,M - L,U L U,M - L = Comparative trials in the laboratory U = Comparative trials underground R = Evaluation of basic performance characteristics M = Evaluation of instrument's ability to provide mineralogical data in coalmines. In its present form, this instrument would seem to be more suited to fine-particle aerosol studies in the laboratory or in less ardous workplace conditions. PERSPEC, with its higher sampling flowrate does permit the collection of larger dust deposits. However the recovery of fractions classified according to particle size is difficult in present versions ofthe instrument since it requires precise dissection of the collection filter. For such reasons, these two instruments did not feature significantly in the com parative studies that subsequently formed the bulk ofthe pro ject. It is understood that both are undergoing farther development to improve performance and practical applicability. Comparative Performances in Relation to Health-Related Dust Fractions Large numbers of comparative trials were carried out, both in die laboratory and underground in mines. In each in dividual run, an instrument was identified which provided a reference for the fraction of interest. For example, for true total dust the reference was usually a thin-walled probe fac ing into the wind and aspirating isokinetically. For the inspirable fraction, it was the IOMID, SIDS or PIDS. For respirable dust, it was the MRE (or an equivalent horizontal elutriator-based sampler such as the MPGII), and for the alveolar fraction the TBF50 (or CPM3). For the thoracic fraction, no suitable reference sampler was available. For this, therefore, it was decided to use the thoracic sample ob tained from the PCI. For the dust spectrometers (i.e., PCI, SIDS, PIDS), the determination of the dust concentration in each fraction was carried out by first determining the particle aerodynamic size distribution for the sampled dust, and then numerically cal- 621 Instrumentation for Dust Measurement Table m Summary of the Entry Characteristics of the Instruments Tested Sampler Entry efficiency True total Inspirable CIP10 (personal) CIP10 (static) M-Z TD IOM1D INSPEC PERSPEC PC! SIDS PIDS NO NO NO NO NO NO NO YES NO NO YES NO YES NO YES NO YES NO YES YES TBF50 MPGII CYCLO CPM3 MRE SIMPEDS STASER NO * * NO NO * YES NO * NO NO * * YES - unqualified acceptance * - no information NO - not appropriate dilating die size (frequency) distribution of die fraction of interest. The area under this new curve gives the mass sam pled in die fraction of interest, and hence its airborne concentration. From the large body ofdata available from all die trials that were carried out. Table IV aimmariras qualitatively how well die various instruments provide information relevant to die various health-related dust fractions. Here, as in the previous table, a fairly bland assessment of die relative performance is given. It is based on examination of combinations of the various instrument comparisons against suitable reference samplers and information about their selection characteristics. Where there are inconsistencies, die judgement is made by inspecting the total information available and, where ap propriate, a qualified acceptance (`OK') is indicated. It should be noted that although die first two columns appear to be die same as those in Table m, die ratings now take into ac count the accessibility ofdie sampled dust in those fractions. Hence, for example, although the CIP10 actually aspirates the inspirable fraction quite satisfactorily, it is not so easy to recover it for gravimetric assessment. Therefore a `YES' in Table m becomes 'OK' in Table IV. 622 Later, when all the results have been combined and ana lyzed in greater detail and have been discussed by all die participants in the Joint Project, a more detailed picture will become available. Mineralogical Assessment of Sampled Dust One important aspect of sampler performance is the ability to collect dust samples within desired fractions or classified size ranges in a form (i.e., quantity, accessibility) suitable for mineralogical assessment. This was studied, again both in the laboratory and in die field trials. The conclusions are summarized in Table V for typical coalmine dusts. In these studies, the main emphasis was placed on the quartz contentreflecting the general interest in health-effects associated with quartz-containing dusts. The methods which were used for mineralogical assessment included infrared spectrophotometry and X-ray diffractometry. For both, the greater the amount of dust which is available for die assay, the better. Obviously, however, die minimum amount of dust required to carry out a satisfac tory analysis depends greatly on die particular analytical in strumentation available. Within the laboratories participating in the Joint Project, such capability varied appreciably. For present purposes, the simple 'rule-of-thumb' was adopted that a minimum mass of 0.1 mg of mixed mine dust should be available in order to enable assessment for quartz con tent. Table V therefore indicates judgements made on die basis of estimates of amounts collected--in the various parts of each instrument as appropriate--for typical dust concen trations over typical (up to 8-hour) sampling shifts. Of the spectrometers, the 401/min PCI comes out particularly well since amounts of dust are provided at each of the impactor stages more than adequate for determination of the mineralogical content of the dust throughout die particle aerodynamic size distribution. The same can be achieved us ing the lower-flowrate SIDS and PIDS but with less sensitivi ty due to the smaller amounts of dust available for analysis. CONCLUDING REMARKS hi general, die dedicated samplers were found to be generally easier and more convenient to use. Some ofthe ones intend ed for respirable (or alveolar) dust may also provide--with some additional effort--a reasonable measure of total' (or, in some case, inspirable) dust. By contrast, the spectrometertype devices require more skill on the part of die operator. This is the price of the greater versatility required in some of the expected research applications. During the Joint Project which has been described, a number of dust samplers, originating from a number of European countries, have been tested and (heir performances compared. From the results, it should be possible to judge the relative strengths and weaknesses ofeach in relation to each proposed new application and to choose the instrument appropriate to die task accordingly. Although there is no single instrument which emerges as die universal 'best', it is clear that certain of die instruments are not appropriate for certain tasks. It is recommended that, in designing new studies to further understanding of die health-related properties of airborne dusts in mines, sampling instrumentation should be chosen after careful consideration ofthe results ofthis Joint Project. Instrumentation for Dust Measurement Table IV Summary of the Performances of the Instruments Tested in Relation to the Various Health-Related Dust Fractions Sampler True total Fraction Inspirable Thoracic Respirable Alveolar CIPIO (personal) CIPIO (static) M-Z TD IOMID INSPEC PERSPEC PCI SIDS PIDS NO NO NO NO NO NO NO YES NO NO OK NO OK NO YES NO YES YES YES YES NO NO NO NO NO * OK YES YES YES OK OK OK NO NO * OK YES YES YES OK OK NO NO NO * OK YES YES YES TBF50 MPGII CYCLQ CPM3 MRE SIMPEDS STASER NO * NO NO YES NO * NO NO NO NO NO NO NO NO NO NO YES OK OK YES YES NO YES NO NO YES NO NO NO YES - unqualified acceptance OK - qualified acceptance NO - not appropriate * - no information 623 Instrumentation for Dust Measurement Table V Summary of die Performance Characteristics of the Instruments Tested in Relation to Their Abilities to Provide Information About Mineralogical Content of the Dust During a Typical Sampling Shift in a Mine Sampler Coarse fractions Dust accessible Sufficient dust for analysis? for analysis? Fine fractions Dust accessible Sufficient dust for analysis? for analysis? CIP10 M-Z TD IOMID INSPEC PERSPEC PCI SIDS PIDS OK OK YES YES OK OK YES OK OK YES YES YES YES NO YES YES YES YES OK YES NO NO YES YES YES OK OK YES YES NO NO NO YES YES YES OK TBF50 MPGII CYCLO CPM3 MRE SIMPEDS STASER OK * * OK OK OK YES YES * * YES YES YES YES YES YES YES OK YES YES NO YES YES YES YES YES YES NO YES - unqualified acceptance OK - qualified acceptance NO - not appropriate * -no information REFERENCES 1. International Standards Organization (ISO): Air Quality--Particle Size Fraction Definitions for Health-Related Sampling. Technical Report No. ISO/TR 7708-1983. Geneva, ISO (1983). 2. American Conference ofGovernmental Industrial Hygienists (ACGIH): Particle Size-Selective Sampling in the Workplace. ReportoftheACGIH Technical on Air Sampling Procedures. Cincinnati (OH), ACGIH (1985). 3. Stuke, J., Emmerichs, M.: Das gravimetrische Staubprobennahmegerat TBF50. Ergebnisse von Untersucbungen aufdem Gebeit derStaub und Silikosebekampfung im Steinkohlenbergbau. 9:47-51 (1973). Acknowledgements: The author wishes to thank the Commission of Euro pean Communities (CEQ and the European Coal Industry for their finan cial support of this Joint Project. In addition, particular thanks go to foe individual Project Leaders in foe participating laboratories (Dr. L. Annbruster, Dr. H. Bauer, Mr. P. Courbon, Prof. G.F. Peruzzo and Mr. B. Preat) and to our numerous supporting colleagues. 624 Instrumentation for Dust Measurement COMPARATIVE MEASUREMENTS WITH VARIOUS INSTRUMENTS: PROBLEMS IN THE EVALUATION OF DUST EXPOSURES IN THE HARD COAL MINING INDUSTRY H.-Dl BAUER* K. Robockf Silicosis Research Institute, Bochum, FRG fK. Robock Mining Research Institute, Essen, FRG Dust measurements with the tyndalloscope for the evalua tion of dust exposure in the mining industry were performed from the middle of the fifties until 1973. Threshold values based on this measuring procedure. For the introduction of measurements, the tyndalloscope was at first the suitable device since it obtains data per minute about time-referred concentrations. Measuring values could therefore be allocated to a defined working process, thus indicating priorities of dust development and introducing measures ofdust suppres sion. The disadvantages of the tyndalloscope were the dependence of scattering light intensity not only on concen tration but also on particle size. Realizing that the evaluation of dust conditions according to mass concentrations of fine dust results in a more suitable risk evaluation than other measuring parameters had, in con nection with the establishment of mass-referred maximum workplace values for the whole mining industry, die conse quence of converting die whole measuring and evaluating system. Thus the tyndalloscope was no longer suitable for die general occupational medical assessment of workplaces since die allocation of intensity values could not be realized in individual cases (Figure 1). Due to conversion to gravimetry, partly very different, evaluations of dust condi tions in comparison with tyndalloscope assessments could be observed. Including tyndalloscope measuring data in epidemiological studies raised therefore many uncertainties about earlier critical dust conditions. In the FRG, maximum workplace concentrations are derived from the Johannesburg Convention fine dust definition. The MPG II equipped with a horizontal elutriator (Figure 2) theoretically meets this defined fractionation. In die follow ing time, it served as reference instrument in the German mining industry. When using devices with other separating functions comparative measurements with the MPG II are obligatory to determine whether conversion relations with a sufficient statistical significance are present. The TBF 50 (Figure 3), a double-cyclone instrument, used in routine measurements without follow-up filter behind the second cyclone, was tested in 180 comparative measurements. Con centration levels are compared in Figure 4. Because of the formerly supposed global connection between ash propor tion and dust particle size, a correction of TBF 50 values with regard to ash proportion was tried. Based on these calculations, a corrective diagram was made (Figure 5) which was referred to for the indication of concentration-equivalent values for the MPG II. Obviously, alterations ofmining and support techniques as well as of mine layout have increas ingly blurred die connection between ash proportion and dust particle size. Thus, a correction via ash proportion is ques tionable at present. According to recent research findings, the conversion factor between these two instruments is in dependent of this parameter (Figure 6). The individual registered measuring positions represent mean values of 3 to 4 measurements at die same positions over a whole shift. Any position reflects various faces with different coal types and different mining techniques. Relating die results obtained by the TBF 50 instrument with and without filter, a dependency of dust retained in the sec ond cyclone on dust particle size is still present (Figure 7). However, distributions are also enormous when referring to dust particle size so that further influences are supposed to play a role. They might occur due to concentrations consider ing that cyclone efficiencies in addition to the particle size of dust to be collected also depend on concentrations. Fur thermore, aggregates of suspended dust may be destroyed in cyclones. This factor has also an effect cm dust masses separated in individual stages, thus being able to falsify the reference to primary conditions ofairborne suspended dust. Dust may show various aggregation degrees which could not be correlated to defined workplace atmosphere parameters up to now. The dependence of cyclone efficiencies on suspended dust uptake is also revealed in a comparison with the personal dust sampler Simpeds 70 MK n. In case of high concentra tions, the throughput decreased which had the consequence that less dust was separated on the follow-up fine dust filter (Figure 8). In addition, varying flow velocities had a substan tial effect on the collecting capacity ofthe intake. When per forming alternating measurements with both instrument types, these parameters should be considered for concentra tion determinations. Different conversion factors had to be taken into account, too, in a comparative test using the French device CIP 10 (Figure 9). Applied at the same position and at the same time hitherto obtained test results of instruments show a de- 625 Instrumentation for Dust Measurement IffiMkmtrrtftm SlwMrt Figure 1. Comparison between tyndalloscopic and gravimetric measuring values. Figure 2. MPG II. 626 A Figure 3. TBF 50 Instrumentation for Dust Measurement pendency of the conversion factor level on mined coal type as well as on mining method (Figure 10). Outgoing from gas-flame coal, a low rank coal, and ending with high rank coal, the conversion factor increases, especial ly for plough mining. This tendency is less distinct when min ing is performed with shearer-loaders. For this type of min ing, uncorrected results of almost equal concentrations for MPG n and CIP 10 instruments can be based in general. The different reaction of both types during dust measure ments in various mines applying different mining methods and the mining of coal with varying ranks is the result of different coarse dust pre-extraction in connection with vary ing particle size distributions of suspended dust. In one case for example, a change of mining methods from stripping to cutting resulted in the reduction of average particle size diameters by about 26 per cent. Likewise decreased the con version factor from 1,4 to 1,1. The outcome of these comparative measurements indicates the difficulty to use instruments with a deviating fractiona tion when referring threshold limits to a specific fine dust definition. Usually, general conversion factors cannot be ap plied; allocation has to be face-specific. Due to the conversion to gravimetric methods, it is nearly impossible to recognize individual emitters and to propor tionate them according to mining methods. Therefore, a new handy instrument, measuring on tyndallometric basis, was Figure 4. Concentration comparison between Ctof and Cmpg- 627 Instrumentation for Dust Measurement cTBFSOoF cMPGII V10 060/97* Figure 6. Ash proportion obtained by MPG II in mass %. 628 CTBFSOwithout Httf cTBF SO with tutor Instrumentation for Dust Measurement Figure 7. Conversion factors for TBF 50 with and without filter. CSJH CMPG Figure 8. Conversion factors for Simpeds and MPG II. 629 Instrumentation for Dust Measurement Figure 9. CIP 10. developed: die TM digital jiP which indicates single and average values for random measuring periods (Figure 11). Measuring sensibility due to particle size was diminished in this equipment. The advantages were mainly attained by measuring scattered light at an angle of 70 compared to 30 for the former tyndalloscope, and using monochromatic primary light of a wave length of0,94 pm instead of visible light. Although the primary objective for using the TM digital jiP was dust measurement for technical purposes it was also designed as supplementary or auxiliary device for occupa tional medical surveillance under specific operational con ditions. At first, comparative measurements with die MPG n did not yield encouraging perspectives (Figure 12). The wide distribution of comparative values seemed to exclude an acceptable allocation of scattering light values to gravimetric concentrations. Classifying values according to specific characteristics of mining did not result in a substan tial improvement, either. However, face-referred evaluations and limitation to areas of low exposures obtained good cor relations between MPG II and tyndallometer (Figure 13). It is true that conversion factors vary widely from face to free; a linear relationship to comparative values is obtained, however, if a specific face is referred to. At first, this assessment had die only objective to find out which tyndallometric measuring values have to be determined for the "worst case'' in an area of low dust make and thus a low health risk in order to abstain from time-consuming gravimetric measurements. The linear correlations shown in figure 13, however. To prove this, comparative measure ments during about 100 subsequent shifts were performed Figure 10. Conversion factors CIP : MPG n. 630 Figure 11. TM dig. |iP. in various mines with the TM digital |iP, the MPG D, the TBF 50 and with a fine dust measuring device developed for permanent measurements with remote transmission of values on the basis of the tyndallometer digital pP. These devices were placed in a frame to maintain the same arrange ment of instruments even when positions in mines changed (Figure 14). Instrumentation for Dust Measurement To avoid dust deposits in the measuring chamber of the tyn dallometric dust measuring device, clean air flows through a small fan at the inside of the measuring chamber. Speed Geometric mean value ytM Figure 12. Relation between gravimetric and tyndallometric intensity values of respirable dust concentrations. Figure 13. Conversion relation for different faces (low concentrations). 1 PMA (measuring head) 2 FMA (electronic element) 3 MPG I U TBF 50 5 TM digital pP Arrangement of dust measuring instruments in comparative measurements v io i7i ms Figure 14. Arrangement of dust measuring instruments in comparative measurements. 631 Instrumentation ftr Dust Measurement can be adapted to environmental velocity. In order to il lustrate die comparison of results achieved by the tyndallometric fine dust measuring instruments with those of other instruments a field test typical for instrument reaction is described. The comparison with the MPG II showed a linear correlation over the whole sphere of concentrations (Figure IS). This also includes the other field tests which show in some cases a varying increase of the balancing straight line. However, die positions of balancing straight lines are typical for each face. Their rise remains nearly un changed during varying operational processes in die same face. Thus, a face-specific allocation to die MPG II is feasi ble but also necessary. In this case, the conversion factor is not rally valid for short-term but also for long tram periods. Regarding the TBF 50, correlations are less distinct (Figure 16). The distribution is higher. In another case, an alloca tion was even impossible (Figure 17). Since measuring values of the tyndallometer can be face-specifically allocated to gravimetric measuring values ofdie MPG II as the basic in strument, the following consequences can be drawn: 1. It is suitable to give a review on dust conditions be tween two gravimetric measurements. 2. It can help to decide for which shifts ofoperational pro cedures gravimetric measurements are required and fra* which shifts separate measurements should be carried out. 3. It is apt to indicate whether normal dust conditions were prevailing in the time of gravimetric measurements in order to exclude positive or negative extreme situations which may influence long-term classifications of the face. 4. hi particular cases, die interval between two gravimetric measurements could be extended under the condition that the frequency of tyndallometric measurements in creases in the meantime or a permanent tyndallometric surveillance is provided. 5. Basically, tyndallometric measuring instruments offer die possibility to allocate short-term concentration changes to specific operational procedures, to introduce dust suppression measures and to check die efficiency of them. However, the tyndallometric measurement will not be able to replace the gravimetric measurement. As far as test methods for the direct determination of specific fibrogeneity of a dust collective are not at disposal substance quan tities obtained by gravimetric samplers will have to be classified more extensively than hitherto to acquire a better knowledge on changing proportions of individual com ponents, their particle size distributions and information about their potential interactions in dust mixtures with regard to fibrogenic tissue reactions. Primarily, single particle analyses on homogeneous and heterogeneous compositions including element analyses by electron microscopy and lamma spec trometry are required. As in many other countries, the evaluation ofquartz-including fine dust mixtures in die FRG is carried out according to total fine dust concentration considering die quartz propor tion ofthis dust mixture. Quartz as individual mineral serves as reference value (Figure 18, middle-line). Thus, the ap proved fine dust concentration in case of quartz quantities < 100% is an operand only. In die FRG, however, this method is not consequently applied in cases of low quartz l(FMA) X : xJ" xHx x.xXX* : X Jt "x..* X X* XX x* _ X -------------- ---------------1 ( i ;i :I i c (MPG II) in mg/m3 '> 6 V10169/88 Figure 15. Comparison tyndallometer (FMA) and MPG n. 632 l(FMA) : -X * X* X x* xx X X jr* `x X . X X X * ic* *X XX X X X X Instrumentation for Dust Measurement 0 2 4 6 8 10 12 14 16 c(TBF50) in mg/m3 V10169/88 Figure 16. Comparison tyndallometer (FMA) and TBF 50. I(FMA) X X X 1 X X* * XX m X * *x XX - ,,'** *X a, * " * ** : "X X X _____ 8 _____ X Xx 0 2 4 6 8 10 c (TBF 50) in mg/m^ V1017Q/88 Figure 17. Comparison tyndallometer (FMA) and TBF 50. 633 Instrumentation for Dust Measurement quartz i mss % quartz M. - % 7.5 7.0 6.5 6.0 5.5 5.0 Scheme of bands for quartz concentrations in case of modified quartz evaluation Figure 18. Scheme of bands for quartz concentrations in case of modified quartz evaluation. proportions. In the presence of low quartz proportions, a defined fine dust threshold value was established which is not to be exceeded during a long-term assessment period. In case of small proportions, quartz is believed not to be die decisive biological parameter. At least in die hard coal min ing industry it is doubted that quartz has the same fibrogenic power under any petrographic condition. Due to different developments, quartz as single component might show vary ing activities, or, referring to its harmfulness, it could be modified on account of interaction with other mineral com ponents during or after deposit formation. For example, in spite of high quartz dust concentrations in a mine of a sedimentary hydroxide iron ore over a long-term exposure period did not provoke lung damages in exposed miners. At that time, this outcome was attributed to insoluble quartz sur face masking. In other hard coal mines, too, the risk to disease obviously cannot be directly and generally related to quartz fine dust concentrations. For example: the quartz proportion of dust originating from high rank coal is essentially lower than that ofyounger strata (Figure 19). Hie number ofdiseases is con trasting, however. These hints and findings raise two ques tions 1. Can occupational medical evaluation he based on a stan dardized quartz definition? When referring to pure quartz wouldn't it be preferrable to make modifications considering the different fibrogeneity and to fix specific limit values? 4.5 4.0 3.5 _____m_______ high rank coal ------------------------- low rank coal Figure 19. Quartz contents in respirable dust (measurement results obtained between 1980 and 1988). 2. Is it justified to restrict approved total fine dust con centration mathematically only by taking into account toe respective quartz proportion without considering the components in the dust mixture interacting with quartz? Several countries might have based their limit values on reference values for quartz of different origin which have different effects, therefore. This could explain the partly widely varying approved mass concentrations. A well-known fact is that free crystalline silica has not only structural dif ferences but also varying biological effects. The results of animal experiments after using quartz of different genesis but also with cristobalite, tridymhe, coesite, sdshovhe as well as amorphous silica confirm these findings. This means that toe conditions ofquartz formation and oftoe growth ofquartz crystals in the plutonic development from early release out ofliquid magma up to the telethermal phase in the hydrother mal sphere can vary widely. Therefore, deviant effective potentials oftoe mineral which is generally regarded as quartz should be taken into account. The Si04 tetrahedron arrange ment determining SiOj modifications does not seem to be decisive, but rather toe undisturbed or disturbed formation ofindividual tetrahedrons, for example substitution of Si ions by aluminium or phosphor. Under the condition that a specific limit concentration of for example 4 mg/m3 for respirable dust including quartz must not be exceeded toe curve progression of toe approved fine dust concentration considering toe proportion of the modified 634 quartz component would change (Figure 20). However, the potentially inhibitory effect of substances in die dust mix tures would not be taken into account when applying this purely mathematical procedure. A more reliable assessment might be possible if chemical, physical and mineralogical characteristics could be determined for the specific nocuousness of a total respirable dust collective. This assess ment cannot be realized yet for the hard-coal mining industry. Subject of present discussions is a model to better adapt the occupational medical assessment ofdust uptake in workplace atmospheres in case of exposure to dust originating from various stratigraphic horizons by means of correction fac tors. In case of an uncorrected reference to the quartz pro portion, this mineral component in the dust of seams with low rank coal was the decisive parameter of the approved total fine dust concentration since the limit of 5 mass per cent was essentially exceeded in general. However, disease frequency in the presence of seams with high rank coal and substantially lower quartz components was much higher. These different findings are intended to be harmonized by correction factors for exposure evaluation. This means that the quartz proportion in mass percent of seam strata with low rank coal has to be converted into an "`effective quartz proportion" referring to die conditions in layers with a higher rank coal. Irrespective of this parameter, the fine dust con centration of 4 mg/m3 represents the maximum limit con- quartz mss % Scheme of bands for quartz concentrations in case of modified quartz evaluation Figure 20. Scheme of bands for quartz concentrations in case of modified quartz evaluation. Instrumentation for Dust Measurement 0.7 Figure 21. Approved respirable dust concentration in dependence cm quartz proportion and the applica tion of correction factors. centration in an assessed period (Figure 21). In die practice, the consequences would be as follows: Applying factor 1, valid for seam strata with high rank coal, quartz evaluation would continue to begin for a quartz proportion of 5 mass percent in the German hard-coal mining industry. The fur ther progression of approved respirable dust concentrations will ensue from the orientation to the pure respirable quartz dust concentration of 0.2 mg/m3. Using factor 0.S, quartz evaluation would start for a quartz proportion of 10 mass percent only, i.e., that the calculated respirable quartz fine dust concentration of 0.4 mg/m3 analytically determined via the quartz proportion in mass per cent would be converted into an effective concentration of 0.2 mg/m3. It is certainly not yetjustified to provide as many categories for seam strata as shown in figure 21. A relatively rough differentiation in to 2 or 3 groups of factors would be preferrable. In our opin ion, the observed risk variations could be better taken into account by such a procedure, even when evaluating dust up take in various stratigraphic horizons. Such a convention demands that the adaptation to an evaluation is restricted to modified factors of the quartz component until general systems to evaluate the specific nocuousness of the whole dust collective will be developed. 635 Instrumentation for Dust Measurement MEETING DUST ASSESSMENT NEEDS OF AN AUTOMATED MINING INDUSTRY KENNETH L. WILLIAMS U.S. Bureau of Mines, Pittsburgh Research Center Pittsburgh, PA, USA INTRODUCTION The Bureau of Mines is vigorously conducting research to automate mining processes in an effort to keep U.S. coal min ing competitive in the world market. However, just as die industrial revolution and its aggressive push for productivi ty exposed increased numbers of workers to serious injury, will "high-tech" mining also mean high-risk mining? The answer is "no," or at least "not necessarily." In fact, one potential benefit of automation is to remove humans from die most hazardous underground tasks. This concept is cer tainly not new. Tethered or radio-operated remote control miners signaled die very beginnings of automation. Miner operators could now work under well-supported roof, away from potential methane ignitions and high dust levels in the face area. The Bureau is now conducting the next logical step in automation research. A continuous miner has already been outfitted with a suite of sensors and a computer to interpret sensor data and control movement ofthe machine. With the push of a button, die mining machine can execute a com plete sump-shear-load cycle with the machine head position controlled to within 1 cm. Navigation research is well under way, so eventually the miner will be able to mine coal within the seam with very litde human intervention. Several schemes to detect the interface between the coal and surrounding strata are being researched. Sophisticated laser, acoustic, inertial, and magnetic guidance systems will soon become feasible. The ultimate goal, of course, is completely automated opera tion that requires no human involvement. Does this mean that our worries about pneumoconiosis, silicosis, and related dust-induced diseases are over? Not for many years. While the objectives ofautomation and robotics are admirable, humans will still be going underground well into the next century. Individual exposure to dust may be reduced in many cases, but certainly not eliminated. For the foreseeable future, robotic miners will require Human super vision; and like today's technologically advanced automo biles, tomorrow's mining systems will still require mainte nance. Maintenance will be a service that highly trained humans will continue to provide, and those humans will be exposed to dust. Machines designed to mine more coal will likely liberate more dust, unless dust control research keeps pace. In addition to health concerns, increased dust levels may pose problems for optical or laser guidance systems and 636 other types of sensors, as well as increase the requirement for rock dusting to prevent dust explosions. A critical element of any control system is monitoring. In formation about the contaminant must be gathered so con trol efforts can be assessed and adjusted as required. This paper provides a brief overview ofa Bureau project that ad dresses improved monitoring and analysis of hazardous coal mine dusts. Since the project is a recent initiative, the intent of the paper is not to provide extensive technical detail, but only to introduce the reader to the work being conducted. REAL-TIME DUST LEVEL ASSESSMENT Since the respirable coal mine dust exposure standard in the United States is expressed as a mass concentration (2 mg/m3), gravimetric dust sampling techniques are ap propriate and acceptable for compliance monitoring if con ducted properly. The Bureau recognized several years ago, however, that a real-time method for assessing dust levels was needed to locate dust sources and evaluate dust control systems efficiently. The long sampling time required to col lect filter samples and the delay involved in weighing the filter make gravimetric techniques too time-consuming and labor-intensive for such purposes. This realization brought about the development of several light-scattering dust monitors, including the widely used RAM-I1 and the more recent MINIRAM. Other private sector instruments were developed without Bureau sponsorship. The advantages of these devices are almost instantaneous indication of dust levels, portability made possible by small size and batterypowered operation, and relative mechanical simplicity. Many researchers have evaluated the performance of these and other light-scattering dust monitors. The conclusion com mon to almost all of these works is that the response of photometers is not directly related to the mass concentra tion of the dust. Particle characteristics such as size, index of refraction, and shape all affect the response. A special concern when sampling near water sprays used for dust abate ment is that water droplets entering the instrument sensing chamber can scatter light and cause falsely high readings. The water droplet problem is minimiraH with instruments like the RAM-1 that use a cyclone preseparator. In that case, the cyclone captures most droplets larger than a few micrometers. In passive, open-chamber instruments like the MINIRAM, however, the problem can be severe unless a cyclone adaptor is used. Such uncertainty in light-scattering measurements makes them unsuitable for compliance measurements, but is generally acceptable for relative ``before-and-after'' measurements associated with evaluation ofcontrol systems. Even here, however, results can be very misleading if the size distribution of the dust cloud is dramatically altered by the dust control system. The Bureau is conducting basic research to develop a light scattering dust monitor that accurately measures the mass concentration ofdust, even in the presence of water droplets. The Mie theory of scattering of electromagnetic radiation is often applicable to the scattering oflight by respirable dust particles. The detailed mathematics are quite complex, but in general, the intensity of light scattered by a particle is a function of detection angle, intensity and wavelength of the source light, and particle size, index of refraction, shape, and surface properties. The Bureau is using computer models of Mie scattering to study the implications of varying instru ment configurations and particle characteristics. The theory and computer models deal with ideal spherical particles. Although particle irregularity will introduce unknown changes into the model predictions, the model can still pro vide general guidance regarding the selection of important instrument parameters. As an example. Figure 1 shows a two-dimensional diagram of die intensity of light scattered by a spherical particle as a function of angle for a given set of conditions. The value o, called the particle size parameter, is the ratio of the par ticle diameter to the source light wavelength. Figure 2 shows an intensity diagram for a somewhat larger particle, all other parameters remaining the same. This analysis indicates that each particle will have a scattering signature that may be unique to its physical characteristics. A novel experimental apparatus, called DAWN-A, has been obtained by the Bureau's Pittsburgh Research Center that will allow direct three-dimensional measurement ofdie intensity of light scat tered by a particle as a function of angle. As shown in Figure 3, the device consists of a sphere upon which are mounted several photodetectors. As a particle passes through the sphere, laser light is scattered to the detectors in a pattern associated with that particle. Intensity information is pro cessed by a computer. Research during the remainder of the project will examine the scattering signatures for a wide variety of particles likely to be found in coal mines. Once these signatures are known, a photometer may be designed that uses only those parts of the scattering signature needed to discriminate between liquid and solid particles, and to com pensate for particle size, shape, and index of refraction ef fects. The eventual design might need to include more than one source and detector in order to gather enough informa tion to complete the analysis. Long-term research might even lead to limited dust component analysis using the scatteredlight signature. The anticipated result of the research will be a dust monitor that can continuously and accurately measure the real-time mass concentration ofdust particles in a coal mine. A monitor with such capabilities will find applications in dust control research, and perhaps even in compliance monitoring. The real value of such a device, however, lies in automated dust Instrumentation for Dust Measurement MIE SCATTERING a = 1.0 Figure 1. Scattered light intensity as a function of angle a = 1. control systems. Dust control research has identified many viable methods to control respirable dust, but operating parameters must often be adjusted to fit the situation at hand. Water spray pressure or ventilation rate may need to be changed, for example. To automate the adjustment of dust control parameters requires that information about dust levels be fed back to a control unit that can decide what change to make in the operating parameters. The improved photometer could serve as that critical feedback mechanism. Requiring worker presence to adjust dust control system operating parameters manually would largely defeat one pur pose of automated mining, that is, to remove personnel from hazardous areas. Automated dust controls would greatly reduce the need for human presence. In addition, they would address the other main reason for automated mining, competitiveness. Controlling dust is not free. Power for fans, scrubbers, and water pumps is an ex pense that must ultimately be reflected in the cost of coal. Along with reducing labor costs to monitor and adjust dust 637 Instrumentation for Dust Measurement MIE SCATTERING a- 4 Figure 2. Scattered light intensity as a function of angle a * 4. control operating parameters, automated systems could pre vent unnecessary costs incurred by using overly restrictive dust control methods. DOST COMPONENT ANALYSIS While real-time knowledge of airborne mass concentrations of respirable coal mine dust is important for control purposes, health specialists know dial lung diseases, especially silicosis, are not correlated simply to levels of coal mine dust. The individual components of the dust have an important bear ing on the likelihood ofcontracting disease. This realization is reflected in die practice of reducing exposure standards in coal mines when quartz levels exceed 5 pet. Some Euro pean data suggest that silicosis is not directly related to the percentage of quartz alone. Other minerals such as kaolin and mica have some fibrogenic capacity of their own. On the other hand, minerals such as feldspar, calcite, calcium sulphate, siderite, hematite, pyrites, etc., exist in high quan tities in coal mine dust samples and may reduce the toxicity of the quartz present. All of this research points to die im 638 portance ofbeing able to determine die amount of quartz and other components in respirable coal mine dust samples accurately. Real-time, in situ component analysis of airborne coal mine dust remains a researcher's dream, but significant progress has been made in spectroscopic analysis techniques. The Bureau has purchased a Fourier transform infrared (FTIR) spectrometer to assess its capabilities. Already, die instru ment has demonstrated an order of magnitude greater sen sitivity to quartz than dispersive infrared techniques. These results were obtained by the manufacturer during courtesy analyses of Bureau-prepared filter samples. Dispersive infrared spectroscopy has served as a mainstay analysis technique for quartz for many years. It has a work ing measurement range of25 to 250 pg ofquartz with a preci sion of 13 to 22 pet. Figure 4 is a diagram depicting die operation of a typical dispersive infrared spectrometer. By a system of mirrors and lenses, die source beam is split and follows two separate paths to die detector. Synchronized beam choppers (C] and C2) allow die beams to alternately pass through a sample and a reference cell to the detector. The reference cell measurement allows compensation for such things as variation in source light intensity, temperature, pressure, etc. Infrared light from die source is viewed in discrete wavelength intervals throughout the range ofinterest, and the transmitted intensity is measured at the detector at each wavelength interval. These wavelength intervals can be referred to as "resolution elements." According to Skoog and West,1 "The quality of die spectrum--that is, the amount of spectral detail--increases as die number of resolu tion elements become larger or as the frequency intervals between measurements become smaller." For dispersive infrared spectroscopy, then, increased spectral quality in volves two costs. The first is the increased time required to measure transmittance at a greater number of resolution elements. The second is diminished sensitivity. This results because as the resolution interval gets smaller, the signal available to die detector is smaller. Figure 5 is a diagram of a typical FTIR spectrometer. Here as well, die beam is split, but there are no choppers to alter nate beam paths. Halfthe beam is reflected from a fixed mir ror, through the sample to die detector. The other half of the beam is reflected from a mirror that moves at a welldefined rate, changing die path length ofhalf die beam. The recombination of die two beams results in an optical in terference that strengthens or diminishes die signal at die detector. In fact, the rate of change of signal strength is directly proportional to die movement of the oscillating mir ror. By applying a mathematical Fourier transform to the function that describes die detected signal intensity as a func tion of mirror position in time, a function describing the in tensity as a function of wavelength, that is, the absorption spectrum, can be obtained. The advantage of the FTIR is that all resolution dements for a spectrum are measured simultaneously. Separate measurements need not be taken for each wavelength as is the case in the dispersive infrared system. Since quartz is recognized as a major health hazard and receives special emphasis under the respirable coal mine dust Instrumentation for Dust Measurement Figure 3. DAWN-A exposure standard, the project is directing substantial effort to improving the analysis methods for quartz. The Mine Safe ty and Health Administration (MSHA) is already consider ing the use of an FT1R in its Method P7 for routine coal mine dust sample analysis. Although users will enjoy the benefits of improved sensitivity, they must conduct die somewhat laborious sample preparation required by Method P7. Preparations include low-temperature ashing and sample redeposition. One objective of the Bureau project is to develop a valid, convenient method for direct on-filter FTIR analysis for quartz. As discussed above, other minerals appear to either enhance or diminish the toxicity of quartz, or even cause damage 639 Instrumentation for Dust Measurement Fixed mirror on their own. Thus, to understand completely occupational ly related lung diseases, the capability of measuring the other components in die dust sample will be essential. Multi component analysis of dust samples is, therefore, another of the many long-range objectives of the project. SUMMARY The project reviewed in this report has two primary goals. The first is to provide accurate real-time measurement ofthe mass concentration of airborne respirable coal mine dust. Such capability is needed to provide feedback regarding dust levels to future automated dust control systems. The DAWNA, a unique experimental apparatus for studying light scat tered by dust particles, will be used to design an improved photometer. The second goal is to provide improved capabilities for respirable dust sample component analysis. Such information will be needed to understand occupational lung disease and to better assess die hazards of workplaces. Fourier transform infrared spectroscopy has been selected as a promising technique to accomplish that goal. Just as the Bureau of Mines is applying high technology solutions to problems of production and competitiveness in die interna tional mineral industry market, it is also applying state-ofthe-art technology to the measurement and analysis of respirable dust. The project tasks are in their early stages, but initial work points to an exciting and fruitful future. REFERENCES 1.Skoog, D. A., and D. M. West. Saunders College. Principles ofIn strumental Analysis, p. 241. (1980). * Use of trade names is for identification only and does not imply endorsement by the Bureau of Mines. 640 Instrumentation for Dust Measurement ASSESSMENT OF PERSONAL DUST EXPOSURE WITH THE CIP10 FOR A BETTER MEDICAL MANAGEMENT OF THE PNEUMOCONIOSIS RISK IN COAL WORKERS M. ZITTER B. Mahteu E. De Surgy G. Auburtin A. Mas Houilleres du Bassin de Lorraine (HBL), Freyming-Merlebach, France ABSTRACT According to French regulations, level of coal dust exposure in each underground working must be measured by static sampling. In collieries of Lorraine a single sampling site, in the return air, is selected for each working. Each miner is assigned to one working in accordance with his fitness for work as determined by the occupational physician. A new individual dust sampler (CIP 10) developed by the CERCHAR has been used in a national survey in which more than 5000 measurements in 194 jobs were carried out. That sampler is now at the occupa tional physician's disposal for a better prevention of pneumoconiosis. So far it has been possible to: look after the placement of pneumoconiotic miners still occupied underground. A survey (207 measurements) showed that those workers were in average exposed to 0,64 mg/m3 respirable dust TWA; check dust exposure of miners with a profusion of 0/1 level (263 measurements, mean -- 0,89 mg/m3); document the exposures associated to some job suspected by the physician to be specially at risk. Some over exposure situations have been already detected. They offer a possible explanation for recent cases of particular pneumoconiosis. A strategy for the use ofCIP 10 is proposed, based on 5 successive days ofmeasurements, eventually repeated following the results dispersion and their extreme values. See Table of Contents, Part n, for Paper. 641 Instrumentation for Dust Measurement CORRELATION OF TESTS FOR MATERIAL DUSTINESS WITH WORKER EXPOSURE FROM THE BAGGING OF POWDERS WILLIAM A. HEITBRINK William F. Todd Thomas J. Fischbach National Institute for Occupational Safety and Health, Division of Physical Sciences and Engineering Engineering Control Technology Branch, 4676 Columbia Parkway Cincinnati, Ohio 45226, USA INTRODUCTION Laboratory dustiness tests have been devised1 to provide a quick and convenient means of estimating a material's relative dustiness. These tests are empirical in that they do not measure a fundamental property or response of the material being tested. In using these dustiness tests, one assumes that the dust generation in the test simulates die dust generation in an actual powder handling operation. In order to be useful, the results ofthese tests must be correlated with personal dust exposures. Because this correlation has not been evaluated, NIOSH researchers conducted a study to evaluate die correlation between worker dust exposure and die results oftwo dustiness tests. The two dustiness test devices are the Heubach Dust Measurement Appliance and the Midwest Research Institute (MRI) tester.1*2 This study was conducted in the packaging room for a powdered acrylic resin production line. The plant produced a variety of resins which differ in bulk density, particle size, moisture content, and observed dustiness. The resin powders were auger fed into tuck-in valve bags. The bags were filled with 50 pounds of powder, they were sealed and dropped onto a conveyor belt which transported foe bags to a palletiz ing operation. The operator tended a number of bag pack ing machines. Several workers rotated between foe bagging equipment and foe palletizing equipment in an adjacent storage area. EXPERIMENTAL PROCEDURES For six different resins, the workers' dust exposures were measured and dustiness tests were conducted on bulk samples of foe material to determine if the dust exposures and the dustiness test results were correlated. For each material packaged, exposures to total dust were measured using NIOSH Method 0500.3 Air samples were collected using personal pumps operated at 3.7 liters per minute. Separate sets of measurements were taken for different workers who rotated through foe bagging machine operations. Usually, 4-6 measurements were taken for each powder. The Heubach unit, depicted in Figure 1, consists ofa horizon tal rotating drum with internal baffles that produces a repeated dust fall through a regulated airstream. Airborne dust from foe drum enters a settling chamber and is then collected on a preweighed glass fiber filter (50 mm, Schleicher and Schull GmbH). The test parameters (mass ofmaterial, airflow rate, and total flow) for foe Heubach dustiness tester are not unique; they are set for each type of powder tested so that a desirable quantity of dust is collected on the filter. A sam ple of about 20 grams, a flow rate of 4 liters/minute and a sampling time of 5 minutes were selected as appropriate test conditions for this study site. In foe MRI tester shown in Figure 2, powder is poured out of a metal beaker in an enclosed space and foe resulting air- 642 Figure 2. MR1 dustiness tester. borne dust is collected on a preweighed filter (47 mm glass fiber Gelman type AE) at a rate of 10.8 liters per minute. The cup was rotated at a constant speed to dump the powder. A vibrator mounted to the cup shaft helps to dislodge the dust. The sample pump was run for 10 minutes after the rota tion of the cup was initiated. The MRI dustiness index was computed from the following formula: Dustiness Index = Dust collected (mg)/((Sample Weight [Kg])(Flow rate [1 pm])). RESULTS The personal dust exposure data and the dustiness test in dices were fit to a regression model of the following form: in (X) = a + b (Y). In this model, the terms "a'* and "b" Instrumentation for Dust Measurement are the regression coefficients, the term "X" is the individual dust exposure, and the term "Y" is the average dustiness index for a material. For both die MRI and Heubach dustiness test indices, a significant correlation was found between MRI and Heubach dustiness test results and worker dust ex posures. Statistical results for the analyses are listed in Table I. In Figures 3 and 4, the exposure data, the predicted worker dust exposure, and the 95 % prediction intervals for individual dust exposures are plotted as a function of dustiness test results. The prediction intervals include 95% ofthe exposures which would be predicted from the regression model.4 The prediction interval width is proportional to the standard er ror of estimate (S^, which is essentially die standard devia tion about the regression line. It is die result of two sources of error: (1) the lack of fit of die model to the data; and (2) the sampling error in measuring the dust exposure. The significance of die 1st source of error was evaluated using the method described by Mendenhall.4 This method tests whether the error caused by the lack of fit is larger than the sampling error. The significance of this difference is stated as "the-significance level for lack of fit" in Table I. This indicates that the correlation between the MRI dustiness test and the worker dust exposure involves a significant lack of fit. Apparently, this source of error causes the wider predic tion intervals for the MRI dustiness tester. For the Heubach dustiness test, the lack of fit was not significant. This means that the width of the prediction interval is caused by the variability in the workers' exposure data. Thus, the predic tion intervals in Figure 3 cannot become much smaller. DISCUSSION AND CONCLUSION The preceding regression analysis shows that dustiness test results were correlated with worker dust exposure and can be used to predict worker dust exposure to within an order of magnitude. Hie width ofthe prediction interval about the regression lines was largely caused by the variability in the worker dust exposures and the width of this prediction can not become much smaller. The correlations between worker dust exposure and dustiness test results are totally empirical and the results of the regression analysis must be used care- Table I Evaluation of Exposure Models Statistical terms Heubach MRI intercept (a) slope (b) Probability of a larger F R2 se significance level for lack of fit test (Probability of a larger F) -0.5 10 <0.0001 0.59 0.75 0.28 -0.1 0.09 <0.0001 0.45 0.86 0.013 643 Instrumentation for Dust Measurement dust exposure (mg/cublc meter) * heubach dustiness index ((MM (ration toot tn poroont) Figure 3. Predicted dust exposure, and prediction intervals plotted as a function of weight % lost, Heubach test. mri dustiness Index Figure 4. Predicted dust exposure, and prediction intervals as a function of MRI Dustiness Index. fully. The regression equations present in this paper are useful only to the extent that conditions at this plant at die time of this study are duplicated. If conditions at the plant change, the correlation will change. The feet that a significant correlation between dust exposures and dustiness test results was observed in an actual plant shows that addressing material dustiness is important in predicting and controlling worker dust exposure. It also sug gests that significant correlations may be present at other plants and other processes. As a result of this, dustiness testers can presently be used to do predictive industrial hygiene (the estimation ofexposures before they occur). For example, suppose a new product is being considered for pro duction in a process or an operation where two or more dif ferent materials are being used. For this process or opera tion, one can develop a correlation between dustiness tester results and dust exposure. The correlation and dustiness test results from a small sample of this new material could be used to predict the dust exposures to within an order of magnitude. This could allow one to make dust control recom mendations before the new product is produced or used on an industrial scale. Presently, dustiness testers are empirical tests which are used to simulate the formation of airborne dust during powder handling operations. Unfortunately, the mechanism of aerosol generation during operations such as bag dumping is not well understood in terms of the identity and magnitude ofdie forces which affect dust generation. An improved fun damental understanding of airborne dust generation by powder handling operations would allow one to select and devise dustiness tests which closely simulate the actual pro cess which generates the airborne dust. REFERENCES 1. Midwest Research Institute. 1986. Exposure toParticulate When Han dling Small Volumes. EPA Prime Contract No 68-02-4252. U.S. En vironmental Protection Agency, Office of Pesticides and Toxic Substances. Washington, D.C. 2. Heubach Inc. Heubach DustMeasuring Appliance. Heubach Avenue, Newark, New Jersey. 3. National Institute for Occupational Safety and Health. 1985. Nuisance Dust, Total Method 0500. In: NIOSHManual ofAnalytical Methods, 3 ed. NIOSH Publication 84-100. Cincinnati, Ohio. 4. Mendenhall W. 1968. Introduction to Linear Models and the Design and Analysis ofExperiments. Dusbury Press. Belmont California. Note: A more complete version of this paper has been submitted to Ap plied Industrial Hygiene. 644