Volume 7 (1973): Issue 2 (June 1973)

A gas chromatographic method was developed for the determination of N-nitrosodimethylamine (NDMA) in cigarette smoke. The NDMA in the smoke of 200 cigarettes was trapped in a solution of sodium hydroxide and separated from most of the smoke components by distillation from acidic and basic solutions. The aqueous solution was extracted for 8 hr. into ethyl ether in a Dean Stark apparatus. To concentrate the solution of NDMA, the ether was distilled until only 5 ml of the solution remained. An aliquot of this solution was analysed by means of a gas chromatograph equipped with a 200-ft. glass capillary column coated with Carbowax 20-M poly(ethylene glycoI). An alkali metal flame ionization detector with a selectivity of 10*/1 for nitrogen compounds to normal hydrocarbons was used. Small amounts (2 ng) of nitrosamines in the presence of large amounts of other compounds were easily detected. N-nitrosodimethylamine-C¹⁴ was used as an internal standard. The recovery of the method was 50 to 60%. The smoke of cigarettes made from a high-nitrate Burley tobacco and from a tobacco blend containing 8% NaNO₃ contained 50 ng and 95 ng of N-nitrosodimethylamine per cigarette, respectively. Positive identification of NDMA in the smoke of these cigarettes was made by mass spectrometry. No NDMA was found in the smoke of two popular domestic cigarettes or in a cigarette composed of a blend of Burley tobaccos. The limit of detection of the method is 10 ng/cigarette; therefore, if NDMA is present in the smoke of these domestic cigarettes, its concentration is less than 10 ng/cigarette.

A procedure for the determination of CO in the gas phase of cigarette smoke has been developed. Cigarettes are smoked on a 20-port syringe-type smoking machine and the gas phase is sampled by a specially designed sampling unit without appreciably affecting the level of particulate phase components or the operation of the smoker. After mixing, the concentration of CO in the gas phase is measured by nondispersive infra-red analysis. All operations are carried out automatically except the initial start-up, calibration of the analyser, and calculation of results. The procedure has been applied to a variety of cigarette brands with results ranging from 1.2 to 2.0 ml of CO per puff. The precision of the procedure was determined by repeated analysis of a standard cigarette over a four-month period. For a single value representing 100 cigarettes, the relative standard deviation was found to be 2.2 % for ml of CO per cigarette and 1.1 % for ml of CO per puff.

A procedure for the combined determination of total HCN and total gas phase aldehydes in cigarette smoke has been developed which is practical to use for the analysis of relatively large cigarette samples. The smoking system includes a Cambridge pad for collection of the particulate phase and a small tube of silica gel to trap gas phase components, with 5 cigarettes being smoked through each pad and trap. Following smoking, the Cambridge pad and silica gel trap are extracted; the silica gel extract is used for the determination of total gas phase aldehydes and the combined extracts for HCN analysis. Colourimetric procedures, automated through use of the Technicon AutoAnalyser as described previously, are employed for the analysis of the sample extracts. Evaluation of this combined procedure indicates that it yields reliable results for both total HCN and total gas phase aldehydes with greater speed and ease of operation than provided by the previously described methods. The procedure has been applied to various cigarette samples with the relative standard deviation for a single port of 5 cigarettes ranging from 1.9 to 5.5 % for gas phase aldehydes and from 2.6 to 8.4 % for total HCN.

Measurements have been made of the distribution of temperature and low molecular weight gases within a burning cigarette, using a sampling probe coupled directly to a mass spectrometer (or Bosch carbon monoxide meter). The interior of the combustion coal is effectively an oxygen-deficient pyrolytic region. The oxides of carbon are produced in two distinct regions: a high-temperature (about 400-800°C) combustion region and a low- temperature (about 150-400°C) pyrolysis region. In the high-temperature coal the carbonised tobacco acts very much as a classical oxidizing solid fuel bed of carbon to give the two carbon oxides (and water). In the low-temperature region behind the coaI tobacco decomposes to give a substantiaI proportion of the carbon oxides and a major proportion of the hydrocarbons found in mainstream smoke.

Tobacco samples containing 20 ppm p,p’-DDT, 200 ppm p,p’-DDT, 20 ppm p,p’-TDE and 200 ppm p,p’-TDE were smoked and their smoke condensates analysed for the p,p’-DDT and p,p’-TDE degradation products such as p,p’-DDT, p,p’-TDE, p,p’-DDE, p,p’-DDM, trans-dichlorostilbene, bis-(p-chlorophenyl)methane, and p,p’-dichlorobenzophenone. The degradation patterns and the amounts in which these degradation products were present in smoke condensates showed that 1. during the smoking process the more volatile compound has a better chance of not being destroyed; 2. the higher the amount of pesticide present in tobacco, the greater is the percentage loss of pesticide on smoking; 3. smoking being a pyrolytic reaction of a very short duration, “primary”, “secondary”, and “tertiary” reactions take place in the order given; 4. reactions, which have lesser energy requirements, occur more readily than those which have a higher requirement; 5. in the case of p,p’-TDE-treated tobacco smoke, some p,p’-DDE is also formed by the dehydrogenation of p,p’-TDE.

A simplified extraction and purification method is described for preparing tobacco samples for monitoring organochloride pesticides by GLC. For DDT, only one 8 hrs Soxhlet extraction in a paper thimble is necessary. Further extractions can be made for other pesticides such as dieldrin and endosulphan sulphate. No additional column chromatography purification is required before GLC analysis. The method gives at least as accurate and reproducible results as the other methods used for comparison, as well as having the following advantages:

1. Less chemicals are used and the cost per sample is reduced to one fifth of the cost of the old method.

2. The time required for each sample is greatly reduced including ancillary operations such as preparing and cleaning glassware. Thirty or more samples can be done per day including calculations.

3. Because fewer stages are involved in the new method less pesticide is lost from the samples during extraction and reproducibility and accuracy are improved.

The method developed by Peck (8) for observing smoke deposits on cigarette filters with the scanning electron microscope was extended to two techniques to determine how the particulate phase of smoke is deposited on celluIose acetate filters and on individual cellulose acetate fibers.

Technique A: Immediately after the smoke particles were deposited on the fibers, the filter was exposed to methyl 2-cyanoacrylate vapour; the methyl 2 cyanoacrylate monomer polymerized rapidly and formed a very thin film (0.05 µ thick) over the partially volatile particles so they could be examined in the vacuum chamber of the scanning electron microscope. This technique was used to observe smoke deposits on single fibers oriented either parallel or perpendicularly to the smoke stream.

Technique B: Methyl 2-cyanoacrylate vapour was drawn into a mixing chamber in front of the filter as each puff of smoke was taken. The monomer coated the particles and polymerized. The coated particles were subsequently trapped by the fibers and observed with the scanning electron microscope.

From techniques A and B, it was observed that single fibers oriented parallel to the smoke stream showed a heavy deposition of small particles (<< 0.1 µ in diameter). This observation qualitatively confirms the theory that diffusion is one of the predominant mechanisms of filtration. Relatively smalI numbers of large smoke particles (> 0.5 µ in diameter) were trapped by single fibers oriented perpendicularly to the smoke stream. These large particles were trapped by interception on fibers which were perpendicular to the smoke path. The edge of each Y-cross-section fiber, where interception is most likely to occur, was more heavily coated than other parts of the fiber. All of the large particles in a 28- × 45-µ area on a single fiber oriented perpendicularly to the smoke stream were counted. The total number of particles on the fiber were calculated and compared to the amount expected from the totaI number of particles per puff, the fraction of particles larger than 0.5 µ, and the single fiber efficiency. Good agreement between the experimental and calculated values was obtained.

With regard to the standardization of smoking parameters, the influences of different smoking conditions upon smoke yields have been investigated (type of smoking machine used: RM 20/71). The analytical results obtained show that none of the systems in question, i.e. piston pump - vacuum pump; rectangular puff profile - bell-shaped profile; Cambridge filter - electrostatic trap; “free” smoking - “restricted” smoking; is to be unambiguously preferred.

On the basis of an improved method of determining the combustibility of tobacco comprehensive investigations were carried out of factors which influence combustibility. On the strength of statistical studies of the correlations, the following conclusions were drawn:

1. The most precise unit of measurement of the combustion of tobacco leaf is the product of “duration of combustion” and “amount of substances burnt” (secs. × mg).

2. The mineral components influence combustibility variously, K2O positively, Cl-, Na₂O, Fe₂O₃ and P₂O₅ negatively, whilst SO₄- -, SiO₂, CaO and MgO have no influence.

3. The coefficients of combustibility suggested up till now, which are based solely on the composition of the mineral components, demonstrate only a very weak correlation.

4. Isolated groups of organic substances demonstrate no incisive correlation to combustibility.

5. The concentration of all organic substances in the volume unit, the “substance density”, shows a high, negative correlation to combustibility; the “porosity” of the leaf is also clearly correlated to combustibility.

6. Combustibility is influenced by the combined action of part of the inorganic components, the concentration of organic substances in the volume unit and the porosity of the leaf. The combustibility index (C.I.), which was ascertained on the strength of these combined actions, seems to be the most suitable indicator of combustibility.