Document q3YbXqXBN5jvqRVYjg5JzNjYE

.1,__A > The Halide Meter -- The Myth and the Machine American Indn p- G. O. NELSON Lawrence Radiation Laboratory, University of California, Liiernujre, California 94551 (j No cridtncr has been found to support the belief that the halide meter operates t* on the copper or copper-halide excitation principle. Rather, it is a more subtle mechanism, involving enhancement of the nitrogen band spectra in the presence of a halogen. The electrode material used serves only to conduct electric current. A stain less-steel and platinum electrode system along with an ultraviolet transmitting filter has been proposed which increases the sensitivity fourfold and the longevity of calibration by 50 to 100 times. Ti data InfroditCiion HP HE SPARK-TYPE halide meter has been used fen more than a generation to mea sure the concentration of halides and halo- - * T * n'l ... - V - . hem u'd with good sucre-s to detect chlori nated hydrocarbons, but shows progiessivelv less response to biominated, iodinated, and fluorinnicd materials. The ability of the halide metei to respond in any halogen has always been based on the theory of halogen interaction with a roppei elcetiode and sub sequent copper excitation in a spark somee.' However, in spectra! studies which base been undei taken, the evidence roilected does not support tlai5 theory. E\jjt riments were conducted to ascertain the mission spectrum of tin- halide inetei spark in the visible and ultraviolet region'v-veral interesting and quite unexpected pliinomena were observed: (1 No detect able copper lines or copper halide bands weie noted at relatively high halogen coiiccntraiions ol 2000 ppm: '2' halide meti i read ings cnukl he obtained with a number of other metallic eicctrodi' which contained no detectable copoer; itui !".) the same ba-ic spectrum vva- always observed, no matter what electrode material was used or what uj * pi ia >ru d u mirr |h< ijvpif'i1 of llir l. S Alurti c i nr i C- fnjfns- i-n, halogen was sampled. Only the intensity of the bands changed when halogens were sampled. The purpose of this study was to identify the spectrum produced, and to piopuse a more appropriate mechanism responsible tor halide meler operation. A better understand ing of the operational parameters has led to several changes in the electrode and optical filter system. These alterations have made possible large increases in sensitivity, stability, and longevity of calibration. Experimental Procedure The spark spectrum produced in air and in air-halogen mixtures was measured and re corded graphicuilv bv the apparatus shown in Figiue 1. Ihe shutter and given filter were removed from die halide melt r and rvplaced with a quart'/, window. Light from the spark chamber passed through the window into a 0.20-meter Janeil-Ash Ebert monochromator (82 to HO A The monochroiT.atoi grating angle was changed with a dr liming motor powered by a !JU-voh variable power supply. The motor was also attached to a variable potentiometer (voltage divider which vouid alter the vt Itage from a 6-volt battery as it rotated. This voltage change was the sate.,! to the v-axis (wavelength) on an v-y ri voider Monoi liromatic light, exiting lrnm the mo nochromator, "as delected bv nr. ESH-RaihS f ( ;-ii'i*( I1111 | > i r f" Iu!,r 4 1!_! \ 1' hr` \\ s.`- viinpllm i imii.irter :h.s_ K-HC. \ 3 v , 11 - * m ; if*it`!is:r. r11 !' ` ! < order, \ :iioi j -t i v ff.l * i rJl- \s;i "ri r,rtr; ',n - 1 ` UCC 095883 American Industrial Hygiene Association journal 337 rushy of ns were identify ropose a isible for demandas led to ve made stability, sir and in 1 and reu s shown dter were 1 replaced the spark >w into a hromutor or grabng mg motor er Supply, a variable in h could ttery as it die signal i recorder, m the mo.MI-9336S Figure 2. Spertra ireneratcd bv halide meter spark in the presence of HCCk, HCBn, CFLI, SFi, and air. photomultiplier tube powered hy a John Fluke 412A high-voltage dc |>ower supply. The signal from the phototube (intensity i was amplified by a Keithley 410 miciomicroammeter and sent to the y-axis of a Vai ian F-80A t-y recorder. With this technique. light intensity versus wavelength (BLR) to BOO m^r) could be plotted simultaneously on the recorder.* A more precise spectral identification of the spectra was performed with a 3.4-meter Jarrell-Ash emission spectrograph. '1 he halide meter was used as a source excitation unit. Light from the spark chamber was collected on photographic plates aftei a 30-minute ex posure time. An iron pin was ignited with the regular Jarrell-Ash source unit. Since the iron emission lines are well known and accmatels tabulated, accurate positions of the experimental lines or bands could be found. The dispersion of this system, 5 A/inm. yielded accurate bandhead positions to 1 A. Spark Spectra Produced from Halide Meter Spectra were obtained from the halide meter spark souice hy using the usual plati num and copper electrode couple. Various halogenated and nonhalogenatcd hsdmcar- I UCC 095884 58(1 Sovcnt brr-Decern bir, 1968 'n f . J4 ^A -3 l j i**_ A--j ,,i j.i miiiirii nififcij. ! fc* - i 3300 I 3500 3400 WAVELENGTH - A ww- ^*4*'.^.* v ``./TNwiC; , *, ii 3000 3900 Ficcrf. 3. Photograph of halide meter spark spectra with various electrodes in air and airFreon 12 mixtures. Line A is the reference iron pin. Lines B and C are the spectra from platinum and copper electrodes in air and 1000 ppm Freon, respectively. Lines D and E are the spertra from platinum and stainless-steel electrodes in air and 1000 ppm Freon 12. Lines F and G are the spectra from two platinum electrodes in pure nitrogen and oxygen, respec tively. bom were sampled. Figure 2 shows the spectra obtained when air, 500 ppm HCCL, 500 ppm HCBr3) 600 ppm SF,,. and 2500 ppm CH3I wete sampled by the halide meter. Note that no new spertra have arisen, even though the halogen species sampled had changed. The onK process observed was the intensification or enhancement of ilit' air spectra. Other ma- 1 , 1J nn try 2000 ppm (XO-, XH,. XO, Hg, SO-, ace tone. benzene, and propane 1 but only S03 yielded a slight increase in spectral intensity. A number of other electrodes were sub stituted for the platinum and copper combi nation. and several of these ate shown in Tables I and II. All the electrodes used yielded the same basic spectra in approxi mately the same intensity as did the platinum and copper electrodes shown in Figure 2. Identification of Spark Spectra With the halide meter used as a source unit (except for the iron reference pin), accmate spectial photographs wete taken with the 3 4meler spectrograph. Air. air-gas mixoncs, niuogen, and oxygen wete photographed: a typical segment of the results ;s shown in Figmc 1 The top line (.4' is the reference emission spectrum ftom tils' iron pm. lanes B and C ate the spectia for air and I OOP ppm Ficon 12. tespertivelv, using a piatmum and eoppei electrode sotiple. Lull's D and A ate the lespeit-ne spertra ftom an and 1000 ppm Freon 12, using platinum and stainlesssteel electrodes. Spectra from pure nitrogen and pure oxygen are shown in lines F and G. Note that: (1) No new spectrum appears even with the introduction of relatively large amounts of Freon 12: (2i the same spectra are produced from the copper and stainlesssteel electrodes: (3) only band spectra are in wiitenre indicating that the observed phe nomena are primarily molecular rather than atomic; (4) all observed handheads were classified as arising from excited nitrogen or XV molecules1 and correspond in jiosition to those lines produced from pure nitrogen in line F; and (5) oxygen does not contribute significantly to the overall spectrum. Some cop|K'r (3247 A and 3274 A', and platinum (2650 A, 2702 A. and 2706 A) lines were observed, but they were outside the spectral range of the photodetector. Other photographs were taken using both gold and platinum electrodes, but again the only spec tra identified veie those mixing from nitro gen or ionizi-d molecular nitrogen. '1 he pi es sence of halogen only increased the intensity of the specti uin. Mechanism of Halide Meter Operation From the xpectrogruphir data obtained, one must suggest a mt < liamsm vastly diddent iiom the copper atom or copper halide molet ule excitation theoiv. No evidence has been seen to indicate that tlicre is sufficient exiitat.on of the copper or copper halide to Aru'iiitm / produce en; True, coppi sion s]H.'ctro: ever, under the halide n is not the u membered Cud, aie r stein test i. vastly dilTei and Cud 1 under spec: in disehan nev er been spectral re: The don is ionizatic molecules, which the reading, a gen spectr meter to > plavs, hov lv undefst One s- ;!iat the I me muwi used to /i w lien l !.'< sampled, acteri'tie wutei, O' tiate elt needed . transfer would 5 .iilditii'i' passing eiet tint to tile : p: ot s.i '1 ins i. s; leeti i. habile Tin '. l,"t lie U'g i 'i fa, t, p i mill sp-l-tl , i caw UCC 095885 r, l%8 *A i 11 O air- Liiim T arc Lines ^pec- 'tainlcssnitrogen and G. appears t-lv large spectra stainlessla are in , cd phe- i *t ds were rogen or isition to rogen in intribute A) and A) lines ride the . Other ;old and nly specm nitrohe presintensity in 'htained, y differi halide i nfe has sufficient talkie to A'Ht ruan Industrial Hy^ifnc Assot inlivn Journal produce enough radiation to be delected. True, copper lines are easily visible on emis sion spectrograph photographic plates. ITouecer, under the low-amperage conditions of the halide meter (18 mA), copper excitation is not the major process. It can also be re membered that copper halides, especially CuCl, are readily observable in flames (Bielstein test). However, spark spectra are often vastly different from flame emission spectra, and CuCl has been observed in a spark only under specialized conditions of high vacuum in discharge tubes.1 C'ttF and Culir have never been observed by any means over the spectral region in which the meter operates.' The dominant halide meter process present is ionization and excitation of the nitrogen molecules. It is the nitrogen spectrum upon which the meter is zeroed for background reading, and it is intensification of the nitro gen spectrum by the halide which causes the meter to respond. The exact role the halide plays, however, is mysterious and incomplete ly understood. One series of experiments has indicated that the halogen might directly interact with . .. . i , , , used to zero the instrument instead of air, and when 100 ppm chlorine gas in nitrogen was sampled, the meter responded in the char acteristic manner. This would indicate that water, oxygen, carbon or its oxides, and other trace elements normally present were not needed as intermediates in whatever energytransfer process might exist. From this it would seem possible that the halogen acquired additional energy from the energetic electrons passing through the air space between the electrodes. This energy was then passed on to the nitrogen molecule through collisional processes, and finally was lost as radiation. This radiation was the increased nitrogen spectrum observed by the phototube of the halide meter. There are other alternative explanations, but most of them are also lacking in suppos ing experimental evidence. Many theories, in fact, predict that quite the opposite phenom enon will be observed; that is, the nitrogen spectrum will be depressed rather than in creased in die presence of a halogen. For Taiile I Effect of Eire trodc MaOria! on Halide M*.lei Rpspone to 100 Ppm Freon 12 IJrctrPdc Upper Lower Kcsp ui'-r Si!?njl (iVj * \oixr (.?) Platinum Copper Platinum Stainless 5ire! Platinum 1 unuMrn** Platinum Crns^ Platinum Gold Platinum Platinum Platinum Platinum Platinum Copper Copper Stainless stt el Stamlrss steel 1 untptonc Tungsten BrasBrass Platinum Gold Aluminum** Carbon Tantalum* Molvbdf num Iron 35 45 C> 57 35 '50 55 44 35 34 28 30 -- 7 57 3 4 2 5 25 5 I5 y 2 > 4 6 _ 7 4 V,' \ 12 It 18 7 11 b 23 5 17 17 ; 5 -- 1 14 Reading oj> meter when 100 ppm Freon 12 sampUd, ^Ba^k^round fluctuations when air onlv 'aqs vamp'ed. 'Yellow powder (WQs or WO.-CL) fornird on t'liutsten electrode dL pper electrode ienitid and burned after about 3 minutes of sampling 'Aluminum electrode ignited and burned after 14 minutes of sampling. ^Tantalum electrode ignited immediately upon spaik ac tivation. ^Vlrtinlo if e It-nM 1-iSA'i'n tl-in t !n -s turo, the gas with the lowest ionization poten tial will ionize and exite the easiest. How ever, in the mixture of nitiogen (15.5 volts', oxygen (12.5 volts'), and chlorine f 15 2 volts), the amount of nitrogen ionization and excitation increases as the chlorine concen tration increases. This is just the opposite of what is predicted by the ionization potential data. Efftct of Electrode Materiel Twenty-one pairs of electrodes were sub stituted in the halide meter. The response to 100 ppm Freon 12, the background noise, and the ratio of these two values are tabu lated in Table I for fifteen pairs of electrodes. Note that all the electrodes, except those which ignited at the high spark temperatures (1100K), gave a meter reading. It can also be seen that the best signal-to-noise ratios were achieved when a relatively inert mate rial (platinum or gold) was used as an upper electrode. Platinum and gold greatly decrease the amount of oxide formation, which causes UCC 095886 590 Soi'cmber-Dtctniber, 1968 Table II Effect of Various Fitters on Hjtide Meter Response to 1(MI Ppm Freon 12" Elrrt.odrs t'ppiir LfWtT Platinum Platinum Vickt-'I Platinum Pl.t n ui ni Pidtinmn Stainless strtd Alunttl Xirlcel GuU PlaCjinirn C.opper T;r,- 5b A'* -V s/s 5 70 5 59 1 5 59 20 59 2.5 70 3 5 7j 10 47 39 10 21 20 75 15 14 17 21 15 12 4-76* .V S/S S 0 5 30 37 0 5 28 32 1.0 17 34 3 8 36 05 30 43 2 6 44 Filter 7-60* S S/S \ o 37 1 5 21 )-5 23 2 18 2.5 17 2 22 S 20 14 12 17 20 21 5-60 V S/S H M. F* S ,V S/S 0.5 40 47 2.5 ^ 19 0.5 2S 40 1 5 27 05 24 40 1.5 27 I 17 41 1.5 27 0.5 40 56 2.5 2n 3 7 51 4.0 13 x Rejdintp ^.'Berated bv an RGA phutotube exhibiiini( S4 type spectral response. *' Scbntt S Gen filter, 2 7 mm, maximum trutismenon at 325 idm (Ftfofl and 725 m/t (5094) wrtb leaa thus J.l'tj transmission from 425 to 680 mu Spectral tharacteristies sim'lar tj Corning filtrf No 7-54, e Corning hirer, 5 0 mm. nnximunt tun.inixsion at 480 mu 18064,), etreetise band width 375 to 5*5 mu tt Corning filter, 4,6 mm, maximum transmission at 3)0 mu (7064), ilTectise hand ss-dth 325 to 380 3 smaller band also appears with a maximum transmission at 740 mu (1094) and an eifective bund width from 720 t* 750 mu s Corning filter, 4.7 mm, maximum transmission at 420 ( 6594), rffectixe hand width (tolD 380 to 470 mu- * !Iii t!r merer filter sttpplted wirft the instrument, 3-0 mm, maximum traniuHssisan at 420 (7094). effective hand width frnm 380 to 460 mu. i J--.ignal from 100 ppm Freon 12; -V--background noise with on Frecn 12 being umpir'd, S/\--aigaal-tonn-se ratio. nnor electrical conduction and high back* ground noise levels. The copper electrode used in the halide meter is a rather poor choice. Under the ex treme conditions of electrical discharge, the halogens are present in a highly reactive state. They attack the copper electrode and form a coppei halide which coats the filter, slits, and electrodes. This changes the gap distance, which disrupts the longevity and ac curacy of the calibration curves. Comparative weight loss data for copper and stainless-steel electrodes were studied when 0,5% Freon 12 was sampled for 30 minutes. The copper electrode lost 6.3 mg and 0.9 mm of length, while a stainless-steel electrode lost only 0.08 mg with no discerni ble reduction in length. This would indicate a life of up to 50 to 100 times as long as that of the copper electrode. Effect of Various Filters Maximum sensitivity could be improved if those regions which showed the greatest en hancement could be better optically isolated from those areas in which few or no bands were present. The ultraviolet region (300 to 425 m^t) seemed to be one of the more prom ising areas to study. Four filters were tried in an attempt to improve me senxiuvny. x mu.j were substituted for the one supplied in the commercial instrument; the comparative re sults are shown in Table II. Meter response (signal) to 100 ppm Freon 12 and back ground (noise) were measured, and the signal-to-noise ratio was calculated for various electrode pairs. The UG-5 filter, in general, gave not only the best signal but the best signal-to-noise ratio as well. Effect of Both Filter and Electrode Material Table II also shows the effect of both elec trode and filter on meter response. The best combination is the platinum upper and stain less-steel (soft-tempered, 0.051-inch diame ter) lower electrodes with a Schott and Gen UG-5 (or Corning 7-54) filter. Typical re sponse curves are shown in Figures 4 and 5. Figure 4 is the response to 100 ppm Freon 12. Note that the backgiound noisa level is three times as high and the signal is about 30% lower when the platinum and copper electrodes are used with the regular filter. Figure 5 shows an even more striking com parison when methyl iodide is sampled while the platinum and copper electrodes are used. Ami titan I < 4. ti * 4 1 0|j;- { <l 1 17; * 5- > x 0 3 FlOt RE i ppm Fietitt time .4 to r pled at 6 . time C. V ground no and UGo h The meter tamimnt : to zero e\t sampled, fresh air i the appro minutes 1, t. i i ft f IU i jir.' _ n 11 Neil's. UCC 095887 Dec ern bt'r} 1968 I',* H. M F f .v/,v S y s/y 40 47 2 3 19 28 40 1 5 27 24 40 1.3 27 17 41 1.5 27 40 5fi 25 22 7 31 4 0 13 13 tiltan A smaller ru 720 to I mu. effective -*i%na!-t- tors were irird in .ensitivity. Filters e supplied in the comparatise rc. Meter response m 12 and backired, and the sib ilated for carious filter, in general, nal but the best tfrocU Material Ifert of both elecsponse. The best i upper and stain11 >51 -inch dianiea Srhott and Gen liter. Typical re- Figures 4 and 5. > 100 ppm Freon urid noise level is signal is about mum and copper regular filter, uoie striking com" is sampled while i. modes are used. American Industrial Hygiene Association Journal 591 Fcci re 4* Halide meter response curve to 100 ppm Freon 12. Air background is measured from time A to B, The Freon 12-air mixture is sam pled at B and reaches maximum scale reading at time C. Note greater response and lower back ground noise when the stainless-steel electrodes and L`G-5 filter are used. Figure 5. Halide meter response to 1000 ppm CHJ. Background is observed from time .4 to B. CHrl is sampled at B and reaches maximum meter reading at tune C. At time D pure air is sampled, and the original background observation is reached at E. Note erratic and inconsistent results when the platinum and copper electrodes are used with the halide meter filter. The meter reads about 4 /iA when the con taminant is introduced, but gradually drops to zero even while the methyl iodide is being sampled. The meter drops below zero when fresh air is again sampled, but it returns to the approximate original base line several minutes later. Clearly this system was not useful even at a comparatively high concentratton. 1 lie |Jltlmunt UIIU Mdimcjs-ucci vicetrode-UG-3 filter system, however, proved to be extremely stable and demonstrated useful response characteristics. Figure 6 gives an overall indication of the response gains when the stainless-steel elec* Fni'tc 6. Halidr metrr calibration for halides usinit carious conditions, tlraph was ob tained under normal conditions. Graph B shows the cain in response and diminution of back ground when a stainicw-stcel electrode and UG-5 filter are used. Graph shows the true sensitivity gain nser the conditions in because the noise leseis have heen equated. VCC 095888 592 Xovrmber-December, 1968 trode and a UG-5 filter are used to detect the \ annus halogens with the halide meter. Graph A shows a calibration curse under the standard instrument conditions. Graph B shows the increases in sensitivity obtained when a stainless steel electrode and UG-5 filter arc used. The same electrode and filter conditions are used in obtaining graph C, except that the slits were opened to approxi mate the noise conditions encountered when the platinum and copper electrodes were used with the halide meter filter. This gives a truer indication of response gains when elec trode and filter substitutions are made. Note that the new conditions increase the response fourfold for chloroform, bromoform, and sul fur hexafluoride. Methyl iodide was easily detected, -whereas platinum and copper elec trodes yielded eiratic results. Conclusions No evidence has been found to suppott the belief that the halide meter operates on the ''py-" nr f-nnner halide excitation pi maple. Rather, it is a more subtle mechanism in volving an incompletely understood halide in teraction with the nitrogen molecule. In the presence ot the halide, the. nitrogen spectrum becomes more intense, and the photodetector responds to this enhancement. The only im portant qualification of the electrode material is that it produce a smooth, stahle spark and be reasonably resistant to halogen gases. Experiments have shown that with a plati num and stainless-steel electrode couple and an ultraviolet filter system, a manyfold in crease in sensitivity is often realized. Some solvents, such as methyl iodide, could not be clearly measured with a platinum and copper electrode couple. However, with stainless steel, as an electrode material, such materials are easily and accurately detected. Some nonhalogenated materials such as sulfur dioxide and other sulfur derivatives could be mea sured as well. Stainless-steel electrodes have been found to increase the longevity of calibration by 50 to 100 times because of their resistance to spark decomposition products. These elec trode and filter changes have made it possi ble to measure all halogens more accurately and thus greatly increase the general reliabil ity of the instrument, r; -- 1, [>niis Hahfit Mettr lnstiuchun Manual, Bulletin No. J148 Oa\ii> Instrument Dimmou, Newark, Nen 2 PtMisf. R W B and A C Gavdon Thr lion of SUtltiuiar Structure, 3rd EH., Chapman and Hall Lui , London (1963) 3 R\o, P. R , and J K Brouy-. J, Ckem. Phyt 35: 776 (1961) Rrcnvrd M> 2*1. 1968 Dictionary of Mining and Mineral Industry Terms A comprchensw e dictionary of mineral science and technology, published by the Interior Department's Bureau of Mines, has been placed on sale by the Supei intendent of Documents. U. S. Government Printing Office. Titled A Dictionary of Mining, Mini rat, and Related Terms, the 1269-pagc volume contain'- approximately 150,000 definitions for some 55.000 separate entries. Included aie scientific and engineering terms, words common to patticnlar mineial-producing localities or regions, and nbbiev minns generally used in the technical literature on mineial technology published throughout the English-speaking countries of the world. Price of the new publication is $8.50. Up-to-date definitions reflect changing usage of older terms as well as addi tions to the vocabulary, but some older definitions, considered to be of \alue in histoiical research, ate also included. The authorin' or source for each definition is provided, giving an indication of the time and place in which the t< un was first used. The bulk of the entiies relate to metal and toal mining, quanting, geologv, metallurgy, glassmaking, ceramics and class, and minerals and rocks of economic significance. A Dietiiinary of Minina, Mintral. and Relah d Terms was compiled and edited bv Paul W, Thiush and the SlJT of the Bureau of Mines. A copy of the dictionary can be pun hawd from the .Superintendent of Documents, U. S. Government Piinting Office, Washington. DC 20402. It is not sold by the Bureau of Mines. Retentioi Admii R. G. T. Fission Product , tig Large doses peritoneally high levels of bt ,3`Cs increased, to those observe long-term comp, skeleton. Kinet of the individua Introduction A K OST MLDUi,1 EVA radionuclides in dosage levels. Distribi. terns obtained from e represent the `'nomu always been some dm he applicable if injm were used. For exar> that doses of "'I 1 ablation of the thvrn alter the normal frail tained in that org.tr believe that similar t other elements that k Strontium is know rema.n there for 1m urn lot ali/es prima1' rapidly excreted. 1 of these two eleiue'i difference in the w ated anil the lengtt peisists High ik1 injected singly, migh vv wk fwrfafim I >is!rr Dtp l b Mum.*. } nr'. tuumlaliun (or Mttfua i UCC 095889