Document LgmaKE3pXK8Jnbv6v3Y8Ox2z3

LABORATORY NOTE: A STUDY IN LABORATORY AUTOMATION Analysis of PCBs in trans former oils 8y $.1. Lormon, H. Gordon, and J.P. Hendricks OLYCHLOftfNATSO biphenyls ling thousands of sample* per year P(PCBs) have been commercial* had to be developed. ly manufactured since 1929. The system that was finally devel They were, until recently, wideloyped has been in use for more than used throughout industry because of three yean. More than 1000 samples their chemical and thermal stability, have been assayed with it. The three- low or nonflammability, and good part method consists of gas chroma electrical insulating properties. But tography. sample preparation, and the very properties that have made automation, both instrumental and PCBs so useful have resulted in a ban data reduction. These were inte on their manufacture and strict reg grated into one efficient system for ulation of their disposal by the Envi speedy handling of many samples ronmental Protection Agency and rapid analysis of data. (EPA). PCBs are so chemically inert that they persist in nature long after many other organic chemicals have Sample tracking broken down. Most of the biosphere Sample preparation consists of Is contaminnted, although no natural four discrete tasks, allowing fa as source of PCBs exists. Not only are sembly-line work and minimum crit PCBs environmental pollutants, but ical path planning. The first task was they are suspected carcinogens as to log in an identification number for well. each sample. Next, each sample was When the EPA announced the reg prepared and its weight entered into ulations on disposal of PCBs In the log. The sample was then loaded 1971, the technical personnel at Con into the autosnmpler for injection solidated Edison were faced with a into the gas chromatograph. The last major problem. It was necessary to task was to eater the sample's i.d. determine which of the 23,000 trans and weight into the computer. Two formers in the Consolidated Edison technician-days were required for 50 systems contained PCB-comami- samples, at a rate of about three nated transformer oil. and at what samples per technician hour. levels. The oils were almost certainly contaminated because at one time fuinollititsd Edison had used PCBs Oas chromatography as transformer oil. When the switch Two Perkia-Elmer Sigma stria was made, the new oil introduced be gas chromatographs were used. Each came contaminated by PCB residua. was equipped with a Perkia-Elmer This meant that method for hand- autosampler. The Sigma 3 gat chro matograph used is a single-channel Mr. Leman It Senior Stuff Chemist, Mr. Cordon Is Croup Lender of the Special unit, equipped with an NlM electron capture detector and capable of tem Analysis Division, and Mr. Hendricks is perature programming. The Sigma 4 Senior Technician ofthe Specie/ Analysis Division, ConeoUdated Edison Company of Now York. The authors wish to ac knowledge Meryl Mendhfor drafting this ankle. This paper is adaptedfrom a prosentetfon made at the S7nd Pittsburgh Conference on Analytical Chemistry and Applied Spoctraecopy, Atlantic City, New ferny, March mi. is a dual channel, isothermal GC. The chromatography was carried out under the following conditions: column packing, Chrpmosorb W HP; column toad, 4% OV-22S; mesh sin, 60-00; column material, glass; column sise: 1.1 m x 4 nun; carrier gas, argon: methane, 95:5; carrier flow. 30 mL/min; oven temperature, 225 *C; auxiliary temperature, 373 *C; standing current of etactran capture detector. 0.5. No data reduction wee pafprmed on the gas chromatographs:Thtout put signals were tai'-dkmQt hno e computer. Sample cleanup and peaparution Rapid analysis was repaired for the large numbers of templet. For the first analyea, sample cleanup was performed on a packed Plororil column. But, because using Florastl packing is a very laborious proce dure, at this time only dirty or high samples (greater than 500 ppm of PCB) or samptas giving poor chro matograms were cleaned. This proce dure war changed for later PCB aampta. Ail inter analytes were paformed using Sep-Paks (Waters Associates) which are FlorocB units in small disposable packets, This eliminated the need for cohuna prep aration. The preparation procadnre was a follows: 100 jiL of sample were dis pensed on e previously weighed ScpPak. The loaded Sep-Pak wee than reweighed. Next the sample was diued with 25 mL of hexane, using a syringe, directly into a 25-mL vol umetric flask. Tablet H ftnnovnry IraniBgpPtii ml Heatawe 6 10 18 20 20 %h aBgMl m 4.0 0.7 0.3 0.2 TM 034 10.7 B4 10.7 90.0 ^ 176 : FEBRUARY 1962 MGNS 003138 I 21,OM 14 Si Pttrarch for organoaillcon pscialtist I organoslIIcon 1 loompounds --| ' now Rogiotor [ A Rovlow providos data on ovor 1000 products Edition S-S ol the Potrarch Organoaillcon Raglatar A Review hat |uat baan publithad. Thit uaaful documant pratants structural, physical and toxicological data on ovar 1000 compounds. An absolute neeeaaity for the research chemist -- and equally indispensable for the technologist and engineer engaged in matarlals development, eurfaoe modification, compoeite structures, rheological studiee and other applications. Seven authoritative chapters discuss the uea of organotllanaa in tyntheaia, tailoring surfaces and other topics of current Interact. For your free copy of thte uaaful catalog, contact Petrarch today. i Flpurv 1 PCB tarnp/a before ofeemrp. pctrarch nam MC. Bertram Road Bristol. PA 19007 215: 781-9255 in tits Far East: Chtaao Corp., Tokyo Chela Bsadai Bandas Card Net 97 17B : FEBRUARY 1982 _ 2 Samp* la Fpur* 1 after cleanup. Mod paaSa and We oftromafopranTf are Ifle him, Ii PMIarlal from th cfmn-ufi procedure. mare la no toe* ol *19 no Mea of PCS The recovery was found to be ex cellent (rM* /). A sample known to be clean gave tha same chromato gram bosh before and after deenup. It wu concluded, therefore, that no PCBe were loet. A known dirty sam ple save a well-reeolved ebromatogram aher dsonup (Ffgtrcw 1 ~4j. Stmdtrds Originally the transformer oils were identified by trade asms, nab* as standards Askasal, Insrteea. and Pyranol, which era dUfaeeac sUaturea of PCBa km all centals sham 30*h trlchlorobmeane. Later, far HONS 003139 I Figure 3 A very poor cnromttogram batora tampla ciaanup. Figure 4 Sampta f/t Figure 9 attar efaanup. Nota tha hfgtify Improvad raaoludon, and much cfaanar ohromatogram at a raault of tfia ciaanup procedure. Now, improved Mark III ZJppatts diepaneare are highly roaletant to breakage. Tha boroaiiieata giaaa syringe ia ancaaad In protective polypropylene. Oeltvery velvet ere positioned within the reservoir for added sefety. The delivery tip ie secured by a rigid platform to eliminate dislodging by accident. Dispensing range fully adluatebie throughout syringe capacity. Five sizes available: 0-0-5 ml, 0-9 mi, 0-10 ml, 0-30 ml, 0-50 ml. Fully autociavabie without disassembly. Available with low profile, sntMtp, safety reservoir of amber glees, or one gallon amber plastic coaled bottle for the 0-30 ml and 040 ml sizes. Screw threaded adapters for use with common laboratory con tainers are available. consistency in reporting, the Aro* don 1242, 1254, and 1240 ware used. Recently Arodort 1016 end 1221 were added. Now commercial Arodor standard# of 1000 ppm in isoctane art used. The standard# were prepared by diluting to 2 ppm PCS in hexane. This is equivalent to taking 0.1 g of oil containing 500 ppm PCB, and di luting to 25 tnL with hexane. Five mkroUters of the standard, contain ing 20 ng PCB, were injected into the chromatograph. Standardization was performed, and the standard usually hcM with very little drift. For complete information, sand for the Wheaton Zlppette brochure. Tel: 900425-1400, extension 2232. WHEATON INSTRUMENTS 1301 N. 10th at.. MMt Nw Jnq 00338 ciwi, nmcmifcn AMERICAN LABORATORY : 17* RONS 003140 FCBe continue* Datt reduction An HP 3354 LAS computer was used for data reduction. The com puter it a real-time on-line system with instrument and autosampler control and It compatible with any lvolt analog instrument output. In (hit cam, it received data directly Tram the gae chromatographs. Hie computer was programmable in BASIC, allowing for individual data analysis. The system-level lan guage was a user-oriented conversa tional language for peak-integration analysis. Calculation and Arocior selection were done as follows. The computer searched for 5 key standard peaks for each Arocior of interest. It also calculated a concentration for each Arocior based on total peak area. Each standard peak was calibrated at total concentration, ail 3 peaks averaged, and a relative standard de viation (BSD) calculated. An error factor was calculated using both the BSD and the total-area value. The Arocior with the lowest error factor was choeen. The computer then looked at the % BSD. If it was below 40%, the mean value was reported. If It was above 40%, the total-area value was reported (figuns 3 and 6, reOfrJ). There are several reasons why this software is so significant. The pro- 0 cgr rine .17*44 t*.74>>3 apoclop u<0 0 PT T[M 1.5 ftf '5CL0S i:e 2 pph :o.*n7 mean -t?.o?r: flE"ii eppop pet U"E 1,4*63? 2.*47; 2.60l?> 3.13204 4. 7'365 2 pen 5>> ?*><9 T-XJ 3*0. 500 AP0CI.OP ME All I500 P1 ET TINE ai;e4ociivP ?*> 5m0e0an EPPOP ;r: .'#E 0 0 PE f Tirrc :-.i?204 4,r;.5 r. 04-544 *.17044 10.1*443 APOO.0P 1254 2 pcm 1:45.4 51.0,:*5 44.4445 23.3112 MEAH 2*7.514 PET Tint 1.3 AP0CL0P 1234 PPM 2i:.-';5 MEAN 213.- 33 hame *12**-l i2<0-2 peak-: > l24'-jl 1242-4 1242-5 PEA*? 5 n1am2e42-r PEAk 3 i-*m| *1254-1 41234-2 *1254-3 11234-4 1254-3 PEAK? 3 1254-T PEAk i . #50 4.47';4C- Pj0 . P'itl 33.044 Tablet AralIw alfrtlhm Areatar 1MB All peaks? Vaa naan 37J %mo 10M Total man. V M*.2 area *ohm si.a (% HB0) x 4638 IBM Yba SU 7M m2 11.4 0.1 rag 1MB Yea 36.4 28.4 8&4 28 aa.i M3 BmaptaieAroaiorim ijec mean vetue ., (Mena. - ip* anay(mmw * seeai. ' . total EPP0P 3C0*E 24.34U for 1949, aa expected, gram is not complex. It is merely simple statistical process written in BASIC. The computer program wee United to identifying rocognlxabie Arocior chromatograms, t.g., trans former oils. The data were still re viewed by a technician, who used the printouts to write reports and trou bleshoot the instrumentation. The results of all analyses were stored in e data file, and specific runs (standards and pertinent runs) could be stored la tbrir entirety. Numerous utility programs were also available. For example, files could be mwiipulated for system operatloa aad tr- When this project began, PCBs were classified into two ripnfleam concentration level*. Samples wbh poenl, stmpt for milsnnmnel m entry. Any eamplm greninr than SB ppm wart dmriftad ee PCImaterial* 1B0 : PUmjARY 19B2 MOMS 003141 t KIT TIWI AROCLOR 1261 9 RET TINE l.S AROCLOR i2se 2 PPM hern e 2 PPM 138.431 MEAN 139.431 NAME PEAKS e NAME 1260-T PEAKS l RSD 9 < RSD tee ERROR SCORE 9 0 RET TIME 1.3430* 2.K024 3.39462 5.05479 1242 0 RET TIME l AROCLOR 1242 2 PPM 15.3950 9.99*14 14.3320 13.9122 46.2421 MEAN 19.7176 2 PPM 193.; MEAN 103. ERROR SCORE 54.9765 NAME 1242-1 1242-2 1242-3 1242-4 1242-5 PEAKS 5 NAME 1242-T PEAKS 1 \ RSD 69.2299 >, RSD 80.5755 REAGENTGRADE SOLVENTS 0 RET TIME 3.33462 5.95479 9.93997 19.6215 AROCLOR 1254 0 RET TIME 1.5 AROCLOR 1254 2 PPM 36.5069 49.1013 64.9035 59.1655 MEAN 47.9143 2 PPM 147.577 MEAN 147.977 ERROR SCORE 11.6913 NAME 41254-1 1254-3 1254-4 1254-8 4 PEAKS 4 NAME 1254-T PEAKS 1 :. RSD 22.9693 '< RSD 51.9997 Ptfure t Printout foe sompto. oils with over SO ppm PCS, a third class of concsntratfcm had to bs oh ktlBt.^1 tMiaa. Mfbk MUMMtpa. done greater than 500 ppm were re ferred to as PCB material. Samples that were between SO and S00 ppm are now labeled PCO-contamlnaied. In terms of disposal, this dassl- flcatlon of the tamplee meant that noo-PCB material was dlspoeedof as required severs rsstrictions. Conclusion Analytical systems designed for very large numbers of temples must be carefully designed. AH aspects of analysis, sample preparation, Instru* meat automation, and data reduc tion must be considered, both sep arately and as a system. Even where Users oI Corco chromatographic, nine- tronic and spoctrophomatrie reagent- grade solvents Know they can count on Corco for highest quality. Reliability, And delivery. In pints, gallons, frvo-gaaon*. or drums. At Corco, we've bean specialists in reagent-grade solvents since 1969- end have bmlt our reputation on meeting the requirements and speeifteabont of our customers. ACS and ASTM Corco can also provide reagent- grade acids, bases, and specialty chem icals. Write or call regarding your specific solvent requirements and our complete solvent listing in Bulletin 10. Or circle the number. ' rial required some restrictions regardlag disposal, and PCB material sary, the savings in labor alone make such a system profitable. Ctwile Eseder tarries CardWe. tie HONS 003142 AMERICAN LABORATORY : IB* r