Document 9J8B54LNqg4x4OnLXD9RLV70R

I COMMENTS OH SAMPLING AND ANALYTICAL METHODS Minor comments are on the original NIOSH draft. General comments: In using trade names, the document should be consistent. Any shorthand such as dropping registration notation should be documented at the earliest opportunity and then a standard format adopted thereafter e.g. ^roc]or R Florlsll R Environmental Sampling and Analytical Methods 3714 ....evaluated under laboratory conditions using aerosol generation tech niques. Field studies were also performed*In a PCB environment containing aerosol and vapor components* The statement as It stands leaves open the question of aerosol or vapor studies. The document should be completed by Inclusion of this Informa tion In a manner similar to the above amended statement, and expressing the true nature of the exercise. Both aerosol and vapor concentrations must be Included. 3724 -3726 t Earlier In the text, 2 solvents are mentioned along with the use of single and multiple Implngers. Which solvent was used should be clarified (toluene or ethylene glycol) and the number of Impingers. The conclusions reached Imply a study using an aerosol of PCBs or an equilibrated vapor environment. Is this true? 3737 to use when personal sampling for PCBs. 3738 This circumvents other techniques such as area monitoring for calculating effective exposures. NIOSli has not tested all available solid sorbents (e.g. porous polymers). Hence, this line should be reworded to reflect only those methods tested. 3752 Electron capture gas chromatography Is the most widely used method, but not necessarily the "preferred" method. For example, the wide variance In response factors referred to In Table I Is a distinct disadvantage. 3001 -3003 In this day, it is not possible to eliminate a method based on the lach or expense of a computer. In many small companies, the OSHA requirements will rapidly force the purchase of such a unit for many applications. 3009 -3010 3824 One might prefer NBS as a repository for "standard" samples. Percblorination should definitely not be used to quantitate PCBs. lie have confirmed that not only biphenyl, but also substituted biphenyls, cause large positive errors. At least one of the replacement fluids for PCBs in capacitors is a substituted biphenyl (Chemical and Engineering News, p. 25, Nov. 15, 1976). Hence this interference is already present In the manufacturing facility and precludes use of perchlorination. Sec attached copy of letter from J. Coleman Weber to Dr. I. E. Wallen. HONS 056083 l 2- 3834 -3635 3843 3645 See coonents on alternate standardization procedure, line 5354. See above comments on perchlorlnation. APPENDIX I 5248 Sables collected must be representative of the personal exposure of Individual workers. 5252 "5253 Records should Include: Pump model & Serial no. Sampling Tube type and no. i.e. sampling equipment details 525B *5200 Calibration data should also be recorded. 5266 -526B The aim of sampling is to permit determination of the personal exposure level primarily. This leads to a decision on compliance/ non*compllance situations. It cannot lead to the "lowest feasible level" without other major input beyond the scope of sampling and analysis. 52/3 Airflow through the pump shall be controllable within 5* of the desired rate during the entire sample perl537 ~ 5291 A more frequent requirement for pump calibration is required. At least monthly and preferably in the method for a dally setup caTIbratlon, 5292 "5293 Spot-checked Is an open statement and means nothing. For digital readout pumps specify parity check between expected stroke count versus actual. For pump with rotameters specify before and after readings. We would prefer a calibration after flow stabilization and again before pump Is turned off, for each sample. 5303 The sorbent tube should not be vertical. Pointing down - loses"gTass wool and Florlsll Pointing up - dust from overhead, hard hat, etc. enters. Suggest a close to horizontal location, attached to the collar of the worker. , 5304 5308 5310 5320 Air being sampled should pass directly Into the open inlet of the sorbent tube This precludes filters. The recommended sample volume for this method Is 50 liters (U.S. spelling) The sorbent tubes should be labelled. At this point it should reiterate checking label and I.D. Number. A bulk air sample Is of no value unless the restraints required to obtain a good and recoverable sample are fully documented. MOMS 056054 5324 5346 5354 3* * Throughout Appendix II nomenclature for PCS and PCBs Is confused. In general, PCBs should be used as a noun describing a PCB mixture, while PCS Is an adjective. This reads more smoothly If written, "This would correspond to a detection limit of 40 ng oinf 5 ml of desorbent". This alternate standardization procedure Is of highly questionable utility. As pointed out In line 535B, It has not been evaluated In NIOSH laboratories. As correctly stated In the draft, the reason an alternate procedure Is needed are two-fold. First, most of the GC peaks In a chromatogram of a PCB product contain more than one component. Secondly, electron capture detector response differs significantly for different PCB Isomers (See Table I of the draft). These two factors Indicate that the absolute quantity of PCB represented by a given peak containing unresolved components varies depending on the relative amounts of the components. The relative composition of a given peak can vary for the Yollowlng reasons: 1. PCBs with a given degree of chlorination (l.e. 423!) from different manufacturers contain different Isomer ratios. 2. Component ratios vary for products with different degrees of chlorination. For example, from Table XII - 6, the ratio of dlchloro- to trlchiorobfphenyls for the peak with RRT 28 In Aroclor 1221 Is 85:15. From Table XII - 8, the same peak In Aroclor 1242 has a ratio of 25:75. Similarly, for peak RRT 70 In Aroclor 1242 and 1254, the tetrachloro-to pentachloroblphenyl ratios are 90:10 and 25:75. 3. Vapor pressure differences alter the composition of vaporized PCBs relative to the liquid. (See attached figure showing chromatograms of liquid and vapors from Aroclor 1016). 4. Interfering components may be present which distort the sample chromatogram. This Is likely to be a major problem In capacitor and transformer manufacturing facilities because several of the PCB replacement products (l.e. phthalate esters and chlorobutyldlphenyl ether) have similar GC retention times to PCDs and give an electron capture response. The weight factor technique of Webb and McCall does not adequately account for these difficulties. For those situations where the sample chromatogram closely matches the chromatogram of a reference PCB mixture, the Standard Analysis described In 5554 ff should be followed. However, if the chromatograms do not match, we believe the only satisfactory analytical method which will yield accurate results is gas chromatography/mass spectrometry (GC/MS) using selected ion monitoring (SIM). The detection limit of this technique for Aroclor 1016 is about one nanogram of PCD Injected Into the , instrument. Considering the environmental limit of 0.05 mg/m reconmended in this document. It might be possible to reach this lower detection limit for the total procedure by sampling as outlined in the draft, evaporating the hexane desorbent and analyzing the concentrate by GC/MS. HONS 056085 ' S3B1 5396 5410 5460 5515 5533 5590 611B -* Alternatively, thermal desorption from a suitable solid sorbent directly Into the GC/HS can be used. This well docunented technique analyses the entire collected sample at once, rather than only an aliquot at a time. Response factors vary only slightly for different PC8 Isomers and most interferences are eliminated. The bulk sample must not be shipped In the same container as the sample tubes. The volume of air sampled can be measured to within at least i5X. Delete next sentence. Delete loot and conclude sentence with *.... recovered with a relative standariTHevlatlon of 4.4t for 27 spiked samples". Oelete next sentence. , Type of electron capture detector Is not Important. rinsing with pesticide grade acetone and hexane should follow the water rinses. Argon/methane Is also a suitable GC carrier gas. The flow rate for either carrier gas should be only about 30 ml/mln for a 2 rnn Id column. Actual retention times should be used In calculations If available. If not available, then the measurements described here can be used. The vapor pressures In Table XIII - 5 appear to be too high. What Is the source of these data, and has their accuracy been verified? 1 HONS 056086