Tobacco Documents Online

OF FYnR0CF'\' claNTn~ l-CIe.;: Er, E s':~::E Robert H. Cundiff Research Department, R. J. Reynolds Tobacco Company Winston-Sa1em, North Carolina The co::?aratively recent emphasis on the gas phase of cigarette smoke has i<<:reased the need for specific methods for determination of individual co;~.iponents of this phase. One such compound whose presence in smoke is well confirmed is hydrogen cyanide. Osborne, Adamek, and Hobbs (1956) and Philippe and Hobbs (1956) list representative references on smoke constituents, and describe infrared methods for identification and determination of'several constituents, including hydrogen cyanide. Kensler and Battista (1963) also report values for hydrogen cyanide, and Norman, Newsome, and Keith (1963) mention hydrogen cyanide as one of the vapor phase constituents determined in their work. The present method for determination of hydrogen cyanide was developed so that the efficiency of filter cigarettes could be more readily compared. Bark and Higson (1963) provide an excellent review of the methods available for determination of hydrogen cyanide. The method of Schilt (1958) was selected because of its reported specificity and sensitivity and because of the stability of the various solutions employed. The procedure is based upon the formation and extraction of the neutral dicyano-bis(1,10-phenanthro- line)-iron (II) complex produced through the exchange reaction between tris- (1,10-phenanthroline)-iron (II) and cyanide ions. r

L Aan;:ratus and Reacents S;oking machine. Standard analytical type, adjusted to take one 35-m1. puff of two-second duration per minute. Smoke receiver. 800-m1. Kjeldahl flask modified by attaching a 25 x i00-r.:a. tube and a No. 2 straight-bore stopcock to the bottom of the flask. The receiver is illustrated in Figure 1. An 800-m1. Kjeldahl flask with- out modifications may be substituted for this receiver. Spectrophotometer. Beckman Model B, or equivalent. 1,10-Phenanthroline monohydrate. G. Frederick Smith Chemical Company. Ferrous ammonium sulfate, hexahydrate. Analytical reagent grade. Ferroin reagent solution. Dissolve 1.96 g. ferrous atmnonium sulfate and 3.17 g. 1,10-phenanthroline in water and dilute to 1 liter. Potassium cyanide, reagent grade. Determine purity by argentimetric titration. Disodium hydrogen phosphate. Saturated aqueous solution. Hydroxylammonium chloride, 107.. Dissolve 100 g. hydroxylammonium chloride in 1 liter of water. Sodium hydroxide, 1 N. Dissolve 40 g. sodium hydroxide in water and dilute to 1 liter. Buffer solution, PH 5.5. Mix five volumes of saturated disodium hydrogen phosphate with one volume of 10% hydroxylammonium chloride. Adjust the pH of the solution to 5.5 with glacial acetic acid. ~ Thymol blue indicator. 0.1% aqueous solution. Chloroform. Water-saturated.

3 Place 50 r..l. of oH 5.5 buffer solution in the smoke receiver. Attach to sr:oking r.:.chine and smoke 50 mm. of each of five previously selected cioarettes cirectly into the buffer solution, taking a 35-m1. puff of two- second duration each minute. Record the number of puffs required to smoke the five cigarettes. After smoking, allow the smoke solution to stand 10 minutes, then drain the solution into a 100-m1. volumetric flask. Rinse the entry tube and receiver with additional portions of buffer solution, draining the rinses into the flask. Finally dilute to volume with the buffer solution. Preparation of Standard Calibration Curve Accurately weigh 0.2410 g. potassium cyanide (100.0% KCN basis) and transfer to a 100-m1. volumetric flask. Dilute to volume with water and mix. Dilute 10.0 ml. of this solution to 100 ml. with water. Dilute 2, 4, 6, 8, 10 ml., respectively, of the dilute standard to 100 ml. with pH 5.5 buffer solution. Pipet 10 ml. of each of the final diluted solutions into each of five 25 x 200-mm. test tubes. This corresponds to 20, 40, 60, 80, 100 micrograms, respectively, of hydrogen cyanide. Add two drops thymol blue indicator solution, and add 1 N sodium hydroxide until the solution turns blue. Add 5 ml. of ferroin reagent solution, mix, and place the tubes in a boiling water bath for 10 minutes. Cool to room temperature and transfer with water to 60-m1. separatory funnels. Extract with 10 ml., then three 5-ml. portions of water-saturated chloroform. Combine the extracts in 25-m1. volumetric flasks and dilute to volume with chloroform. Determine the percent transmittance at 597 r~j in a 1-cm. cell using water- saturated chloroform as the reference solution. Plot micrograms of hydrogen cyanide versus %T on semi-log graph paper to prepare the standard calibration curve.

, . - '::: S O_ SOIUt1oP.s Pipet a 5- or 10-m1. aliquot of the smoke solution into a 25 x 200-mm. test tube. Proceed as described above in preparation of the standard calibration curve, starting with addition of the thymol blue indicator. Determine the micrograms of hydrogen cyanide-from the standard calibration curve. Calcu?ations Five cigarettes smoked, 10-m1. aliquot taken for color development: Micrograms hydrogen cyanide: Per cigarette = micrograms from curve x 2 Per puff = micrograms from curve x 10/No. of puffs Five-milliliter aliquot taken for color development: Micrograms hydrogen cyanide: Per cigarette = micrograms from curve x 4 Per puff = micrograms from curve x 20/No. of puffs. EXP ERIAENTAL Stability of Smoke Solutions Standardized smoking parameters were utilized as nearly as possible in establishing this procedure. These included smoking more than one cigarette to obtain the smoke sample and adsorbing the smoke at ambient temperatures. In this manner better precision should be realized. Initially, a major hindrance to using standardized smoking conditions was the rapid decomposition and/or side reaction of the cyanide after adsorption of the smoke in each of several solutions. Solutions investigated included 0.1 A; trisodium phosphate, 1 Pi disodium phosphate, 0.1 N sodium :1ydroxide, silver nitrate solutions, and buffered ferroin solutions. None

.5 o: .'-.,:!se was sa:.isfactory, because a decre::se in cyanide content was obs,2rved in each of the solutions even on short standing periods. The silver nitrate and ferroin solutions were also somewhat difficult to handle because of precipitation problems. Unsatisfactory cyanide recovery was realized in these media as a result. A drop in PH after smoking was noticed in each of the basic solutions, indicating the reactivity of the acidic corpounds present in the smoke. Only when buffer solutions of pH 7.4 or less were employed, did the PH of the smoke solution remain unchanged. Increased cyanide stability was concomitant with increased acidity of the smoke solutions. When a buffer solution of yH 5.5, prepared as described above, was employed, the cyanide analysis of smoke solutions remained unchanged after standing for 4-6 hours. Other evidences of sufficient stability of cyanide under these conditions included the fact that replicate results on a per cigarette basis were realized regardless of the number of cigarettes smoked. Also, closely similar results to those obtained when smoking into the buffer solution were realized if cigarettes.were smoked into a dry flask at -80° C., and the flask residue then analyzed spectrophotometrically. Interference Studies Schilt (1958) established the specificity of his procedure by study of the effect of diverse ions on color development. The procedure is par- ticularly applicable to analysis of a complex mixture, because the violet- colored reaction product can be separated quantitatively from the aqueous ferroin solution with chloroform, leaving many potential interferences in t1he aqueous layer.

:'.o net::i:. i Et , aGC: :i0:1fi1 possiI.:i: 1ntCrfering subst::nces, iCnown to bte present ir, tobacco s::oke, were studied. These included nicotine and nicotine salts, acetonitrile (methyl cyanide), and acrylonitrile (vinyl cyanide). None of these r„aterials yielded a violet color, nor interfered in quanti- tative discernment of hydrogen cyanide. Cellulose rods, equivalent in size, weight, and draft resistance to standard 85-r.r^. cigarettes, were smoked in a comparable fashion, and the smoke solutions analyzed for cyanide content. No violet color was noted, indicating the lack of interference from the combustion products of cellulose. Cigarettes were smoked with and without Cambridge filters interposed between the buffer solution and the cigarette. In these tests approximately one-half of the hydrogen cyanide was either retained by the filter of reacted with the particulate phase. In analysis for hydrogen cyanide, it is not feasible to utilize this type of filtration to remove possible interfering materials. Recovery of Known Amounts of Cyanide Recovery studies for cyanide were made by adding known amounts of potassium cyanide to the buffer solution in the receiver of the smoking apparatus, then smoking five cigarettes in the regular manner. After cor- rection for the amount of cyanide present in the cigarette smoke, as determined in separate analyses, the amount of cyanide recovered from the smoke solutions could be calculated. Table I summarizes the results of these recovery studies using regular 85-mm. filter cigarettes as the control. Precision of the Total Procedure The precision was checked by analyzing a single production lot of standard 85-rn. filter cigarettes. The cigarettes were conditioned at 25° C.

7 and 65% relative humidity but were selected at random without weighing or checking draft resistance. The results expressed in micrograms hydrogen cyanide per cigarette from 12 analyses are listed in Table II. • RE SULTS of a J Tables III, IV,•and V list the hydrogen cyanide content of the smoke ` ~ . . number of domestic cigare'ttes with results expressed both in micrograms, per cigarette and average micrograms per puff. Most values are results of replicate determinations with the cigarettes being selected at random for analysis. Fifty millimeters of each of five cigarettes were smoked except s • for brand "Q" where 45 mm. were smoked. Results for this brand are extra- polated to a 50 mm, basis for comparative purposes. Untipped cigarettes fo 70-mm. length were prepared from (1) a flue- cured tobacco blend, (2) a noncased burley tobacco blend, and (3) a Turkish tobacco blend. The same tobaccos were cased, then made into blended ciga- rettes with and without bonded charcoal filters. The hydrogen cyanide content of the smoke from each of these five cigarettes-was determined with the results as listed in Table VI. The rate of cyanide delivery into the mainstream smoke on a per puff basis was checked by smoking 20 cigarettes, collecting the first-two puffs of each cigarette in the first receiver, the third and fourth puffs in a second receiver, the fi-fth and sixth puffs in a third receiver, the seventh and eighth puffs in a fourth receiver, and, wherever possible, the ninth and tenth puffs in a fifth receiver. Five cigarettes of the same brand,.were - smoked separately, taking either 8 or 10 puffs, whichever was applicable, to obtain the average per puff value for the whole cigarette. These.data for representative cigarettes are given'in Table VII, and Figure 2 illus- trates the cyanide delivery graphically.

8 DISCUSSIO\ Ten-milliliter aliquots were taken for color development in analysis of the low yield charcoal filter cigarettes, such as brarids "P"; "Q", "R", and "S"; five-jnilliliter aliquots were taken with the other•cigarettes. As indicated in'Tables III, IV L and V, the highest cyanide yield is . f • ~ . . from the 70-mmd unfiltered cigarettes. Some regular filter cigarettes a apparently remov,e•small amounts of•cyanide, although others yield approxi- • ~ mately the same amounts as the king=sized unfiltered cigarettes. Only the ` filter cigarettes with sufficient activated charcoal incorporated into the filter are efficacious•sin appreciably lowering the cyanide delivery. The data in Table VI indicate that the total hydrogen cyanide yield per cigarette is less from the Turkish tobacco than from the other types. The yield is considerably lower on a per puff -basis because of the slower burning rate of these cigarettes. It is problematical whether or not there is a significant difference between the tobacco types and the blended cigarette in this regard, but the effectiveness of the charcoal filter in reducing the cyanide yield is readily apparent. In this particular study the filter reduces the cyanide delivery nearly 75%. Table VII and Figure 2 emphasize the effectiveness of some of the charcoal filters for removal of cyanide. It seems significant that the cyanide deliver-y'increases only gradually with the more effective filters (brands "P", "S", and "R") at least through the eighth puff, whereas the increase is much more rapid with the other brands.• As indicated in Figure '2 some of the media seem to lose their retention capacity after only 4 puffs .j with a marked increase in cyanide delivery noted thereafter.

J. Si.:.:i.'_D3.l~:1 for r ep: r_ .'Z2o.^. S 2.::.. to X. ii. V ~...:o}' =or E::Cellc?nt .'.JcRn=cF.1 ass:s t c':nC2. 10 LIT-P'~ TG"RE CIicD 1. Bark, L. S. , and H. G. Higson, "A Review of the rfethods Available for Detection and Deter;aination of Small Amounts of Cyanide." ANALYST, 88, 751-60 (1963). 2. Kensler, C. J., and S. P. Battista, "Components of Cigarette Smo':ce with Ciliary-Depressant Activity. Their Selective Removal by Filters Containing Activated Charcoal." RTEiJ ENG. J. MED., 269, 1161-1166 (1963). 3. Norman, V., J. R. Newsome, and C. H. Keith, "Vapor Phase Analysis of Tobacco Smoke." Abstracts of Papers, 145th Meeting American Chemical Society, New York, September, 1963. p. 23A. 4. Osborne, J. S., S. Adamek, and M E. Hobbs, "Some Components of Gas Phase of Cigarette S,oke." ANAL. C'rIEaTM1., 28, 211-215 (1956). 5. Philippe, R. J. , and M. E. Hobbs, "Some Components of the Gas Phase of Cigarette Smoke." ANAL. CHEM., 2S, 2002-2006 (1956). 6. Sc:ilt, A. A., "Colori:n:tric Determination 30, 1!:09-11 (1958). of Cyanide." ANAL. C:iEM. ,

11 TABLE I RECOVERY OF CYANIDE AFTER ADDITION TO BUFFER SOLUTIONS BEFORE SMOKING ~ Cyanide Added , Mg. Cyanide Recovered, Mg . Recovery, % 0.40 0.40 100.0 0.50 0.52 104.0 0.60 0.59 98.3 1.00 0.98 98.0 1.00 1.00 100.0 1.00 1.00 100.0 1.00 1.04 i 104.0