Tobacco Documents Online

c.3

Technology Centre, British-American Tobacco Co. Ltd., SOUTHAMPTON CJPIBTM " THE SENSORY ENHANCEMENT OF THE INITIAL PUFFS OF LOW TAR PRODUCTS USING AN ALKALINE ADDITIVE P}AT-L027-93 c. 3 . REPORT NO: P.32 16 DECEMBER 1993 AUTHOR: CO-WORKER: C.J. SHEPPERD J.L. BEVEN ISSUED BY: A,G. STEPHENSON © 1993 BAT Co. Ltd. COPY NO.% Do not copy or show to unauthodsed persons

Technology Centre, British-American Tobacco Co. Ltd., SOUTHAMPTON CJPIBTM 16 DECEMBER 1993 THE SENSORY ENHANCEMENT OF THE INITIAL PUFFS OF LOW TAR PRODUCTS USING AN ALKALINE ADDITIVE PRJ-L027-93 c. 3 REPORT NO: P.32 SUMMARY: Reducing tar delivery will normally be expected to result in a reduction in the sensory impressions experienced by the smoker. In addition, the action of tobacco rod filtration means that the initial puffs are inevitably lower in delivery, and hence can be sensorically lacking. This characteristic is obviously a more prevalent feature of low and ultra-low delivery products where the sensory impression from the first few puffs could sometimes be described as minimal. However, the delivery per puff increases as the rod is consumed, with a corresponding increase in sensory levels, and therefore the middle and later puffs may well be at an acceptable level for the smoker of lower tar products. The sensory enhancement of low and ultra-low tar tobacco smoke, without increasing delivery, is seen as a potentially positive advantage. However, enhancement along the whole length of the tobacco rod might result in an excessive and unbalanced sensory impression in the middle and later puffs. Therefore, the study described in this report attempts to establish the viability of selective enhancement of the initial puffs only, thereby flattening the profile of sensory impression along the tobacco rod by lifting the front end. This involved modifying the smoke pH of the early puffs only, with alkaline additives, such that these initial puffs contained nicotine in the more sensorically potent free-base form. In this way, the additive (sodium carbonate) was applied during cigarette manufacture to the front end of the tobacco rod only, using a prototype pulsed spraying system, based on an existing device (the Arenco DOQ). The modification and operation of this device is described in this report. Cigarettes were produced that covered a range of deliveries (6.4 - 9.3mg) and additive application levels, along with appropriate water treated controls. These samples were subjected to both analytical and sensory testing. © 1993 BAT Co. Ltd. Do not copy or show to unauthodsed persons

PRJ-L027-93 ' c. 3 The analytical results showed that, whilst there was no major effect on the delivery of the products, treatment did result in an increase in the pH of the smoke of the first puffs of the treated cigarettes. This increase, as expected, was greater for samples containing higherloadings of sodium carbonate. Sensory testing demonstrated significantly increased impact and irritation for treated samples, as compared to controls, for the lower delivery products containing medium and high levels of additive on the front end of the rod. Further sensory testing on these particular products considered the sensory impression on the front, middle and back portions of the tobacco rod. It was demonstrated that the greatest differences between treated and control cigarettes existed on the front portions of the rod, and that as the rod was consumed, so the treated samples became more and more similar to the control sample. Thus the treatment had flattened the sensory profile along the cigarette by lifting the sensory impression of the first puffs. One problem with the technique, as it currently exists, is that the application of aqueous solutions to the forming tobacco rod, at a point so close to the paper, causes noticeable staining. Potential solutions to this problem currently under consideration include spraying at a location removed from the garniture (and paper), such as within the suction chamber, or injecting additive into the rod rather than onto its surface. Assuming this problem can be resolved, it is also necessary to establish the level of preference for treated product as compared to control cigarettes. This would be achieved, in the first instance, with a small-scale consumer study. Finally, the technology may well have applications in addition to that described in this report, such as the addition of non-volatile flavour materials or "flavour- enhancers" to the front or rear portions of the tobacco rod during manufacture. INDEX TERMS: Low Delivery Cigarettes Sensory Assessment Tobacco Additives Sodium Carbonate Sensory Quality Analysis Nicotine Impact Smoke pH © 1993 BAT Co. Ltd. Do not copy or show to unauthorised persons

~rv INTRODUCTION One problem frequently encountered with low and ultra low tar products is that whilst the middle and later puffs of the cigarette are adequately satisfying to the smoker, the first few puffs may offer insufficient sensations. This is due to the lower delivery of these initial puffs', brought about by the high level of filtration through the tobacco rod. As the cigarette is smoked, so the delivery per puff increases and consequently the sensory impression also increases. The last puffs may in fact offer slightly higher levels of sensations than the smoker desires, although this is easier for the smoker to control by modifications to his behaviour, such as reducing peak flow rates or puff duration. However, whilst the smoker can also increase puff volumes to compensate for the reduced delivery in the first few puffs, this increase may not be enough when low and ultra low tar products are smoked. It is assumed that smokers currently accept this limitation, as a consequence of selecting low delivery cigarettes. Incidentally, this problem can be more severe with 100mm products due to the longer rod tength. There have been a number of attempts to modify cigarette construction, such that the initial puffs are enhanced, and the later puffs reduced - thereby allowing generation of a flatter sensory profile along the cigarette. This has most frequently involved using two tobacco blends. The cigarette is thus constructed with 100% high nicotine/flavour tobacco at the front end of the tobacco rod which gradually decreases, yielding to 100% mild blend at the mouth end (1, 2). Logically, this approach should achieve the desired effect, and it is also possible that such a cigarette may generate the sensory impressions of a product from a higher delivery sector. However. constructing cigarettes of this dual blend type, at normal making speeds, is technically very difficult to achieve. One of the primary sensations available to the smoker is that of impact, and impact is generally considered to be synonymous with "strength". Therefore, increasing impact (provided the overall sensory balance is not disrupted) implies that the product is "stronger" and potentially more satisfying to the smoker. The impact sensation has been shown to correlate with the level of the more volatile "extractable" or free-base form of nicotine (3, 4). The proportion of the available nicotine in this free-base form is dependent on the pH of the smoke, i.e. at around pH 5 or less, the proportion of nicotine in the free-base form is close or equal to zero. As the environment becomes more acidic, so the nicotine is more likely to be protonated and simultaneously less volatile. Above pH 5, the amount of nicotine in the free-base form increases rapidly (Figure 1). Therefore, one way of increasing the impact sensation is to raise the pH of the smoke, thereby increasing the proportion of nicotine in the free- base form, thus making it more available to elicit the impact sensation, presumably as a result of an increase in volatility. However, if this principle is @ 1993 BAT Co. Ltd. Do not copy or show to unauthorised persons 570 676 7

4 applied along the whole length of the cigarette, the smoke rap~aly becomes harsh and "tiring" as the tobacco rod is consumed, with excessr, re ~mpact and irritation in the middle and later puffs. Sensory enhancement with alkaline additives Several compounds have been employed as alkaline additives at different locations within the cigarette, to raise the pH of the smoke and hence boost impact sensations. These include potassium carbonate, sodium bicarbonate, magnesium hydroxide and polyethyleneimine (5). However, the compound most frequently used in this way appears to be sodium carbonate - an additive that is approved, even at relatively high levels (max. 2% in manufactured cigarettes, according to the Forth Report of the Independant Scientific Committee on Smoking and Health (ISCSH) (6)), Js stable, non- toxic and previously has been shown to be effective as an additive to enhance impact by pH modification. Alkaline additives have been applied to the filter (7), and this has been shown to increase impact sensations, presumably by increasing the pH of the smoke as it passes through the filter. However, this too will boost impact (and irritation) along the whole length of the cigarette - and unless the levels of additive are kept sufficiently low, the smoke may again become harsh in the later puffs. In addition, alkaline materials can be incorporated into the tobacco rod as additives to the paper, and this has indeed been shown to enhance the sensory impression - mainly in the areas of impact and irritation (8). Again, this might be expected to result in a general increase in sensations along the length of the tobacco rod. Therefore, the alternative approach described in this report attempts to enhance the sensory character of the first few puffs only, making more efficient use of the nicotine already available by increasing the smoke pH of these initial puffs, through application of an alkaline additive (sodium carbonate) to the front portion of the tobacco rod only. This Concept has been examined in two stages: 1) A preliminary study that employed hand-made cigarettes and ad-hoc smoke testing, to establish the viability of the approach, followed by 2) A machine-making study involving the development of a method of applying additive to the front end of the tobacco rod only, during cigarette making, and subsequent small scale sample manufacture prior to more formal sensory panel testing. The initial stage is described briefly below, whilst the second approach, involving sample manufacture, forms the main body of this report. © 1993 BAT Co. Ltd. Do not copy or show to unauthorised persons 570; 6' 698

1. THE PRELIMINARY STUDY Alkaline additive on tobacco - the objective and sensory effects The preliminary experiments involved applying alkaline materials to tobacco rods, either by manual injection of aqueous solutions into the tobacco rods of ready made cigarettes, or by hand constructing cigarettes using AJJA 17 spills (manufactured by Gizeh, in West Germany) and tobacco presprayed with an aqueous solution of an alkaline material. The alkali chosen for this work was sodium carbonate, selected for the reasons listed previously. This compound can be included in cigarettes up to a level equivalent to 2% (dwb), as laid down by the Fourth report of the ISCSH (6). Initially, alkali was applied along the whole length of the tobacco rod. Standard low tar (9mg delivery), flue-cured cigarettes were hand injected with aqueous solutions to give 0% (control), and 0.5, 1.0 and 2.0% Ioadings (wwb) of sodium carbonate on the tobacco, and the samples allowed to condition for 24 hours prior to any testing. Five cigarettes of each type were smoked on a Borgwaldt RM20 smoking machine and the condensates trapped by electrostatic precipitation onto the inner surface of a glass tube. This condensate was washed from the glass surface with methanol (3ml) into a 50 ml conical flask, and the extract made up to 30 ml with distilled water. The pH of the resulting solution was determined using an Orion SA 720 pH meter fitted with a standard pH electrode, calibrated with appropriate buffer solutions. Each pH determination was performed in triplicate. The results of this experiment (summarised in Table 1 below) indicate that the application of sodium carbonate to the tobacco rod does increase the pH of the smoke condensate. The level of pH increase obtained might be expected to have an effect on the subjective levels of nicotine induced impact and irritation, due in part to an increase in the proportion of nicotine in the free-base form. Table 1 The effect of Sodium Carbonate addition on the smoke condensate pH of low tar flue-cured cigarettes Na2CO3 loading!/% wwb) 0 (Control) 0.5 Mean smoke condensate pH 6.2 6.6 1.0 6.6 2.0 6.8 © 1993 BAT Co. Ltd. Do not copy or show to unauthorised persons b'702676 9

Therefore, a sensory assessment of impact and irritation for these spiked samples was made. on an ad hoc basis, by a small group of smokers comparing a treated sample with the control. Any difference in the level of impact and irritation was simply noted as being present or absent, ie no particular score was attributed to the magnitude of the sensations. The direction of any change was, of course, also noted. In all cases, the treated samples were described as having higher levels of both impact and irritation than the controls, although differences between treated samples were less obvious. These results were not unexpected, and are consistent with previous work using alkaline additives. However it was noted that using high levels of sodium carbonate (ie >1%) appeared to offer no particular advantage, from a sensory point of view, in these particular cases. The second approach, designed to address this question, involved pretreating a standard flue-cured tobacco blend by hand spraying with either water, or sodium carbonate solution. In this way, three samples were generated to give 0% (control), 1 and 2% Ioadings (wwb) on the blend. These tobaccos were conditioned to consistent making moisture (approx. 13%) after spraying. A set of samples was made from these tobaccos, using a hand operated cigarette making device and AJJA 17 spills. ~'he first set of three cigarettes was made from each of the three tobaccos, prepared as above. An additional sample was generated where the cigarettes were constructed such that the front third of the tobacco rod (furthest from the filter end) comprised sodium carbonate treated tobacco (2% loading), whilst the remainder of the rod consisted of control tobacco only. These cigarettes were subjected to an ad hoc sensory test by a small group of smokers. In the first instance, fully treated cigarettes were compared with controls. The results confirmed the findings of the spiking experiments, in that the alkali treated cigarettes gave higher levels of both impact and irritation than the control sample. The "front-end" treated sample was considered separately. A panel of nine experienced smokers was supplied with the front-end treated sample and a control cigarette. Both cigarettes had been weight and pressure-drop selected, and then masked and coded. The tobacco rod of each cigarette was marked at two points along its length, thereby dividing it into three equal sections. The panellists were asked to assess impact and irritation at some point within each section, indicating on a supplied form whether the cigarettes were "equivalent", or whether one product exhibited "greater", or "much greater" impact and/or irritation than the other. The results of this test indicated that there were significantly greater levels of impact and irritation from the front-end treated sample than the control, for the front part of the rod, but that these differences became less marked in the © 1993 BAT Co. Ltd. Do not copy or show to unauthorised persons

middle and rear sections. In this way. it appeared as if the profile of impact and irritation had indeed been flattened along the tobacco rod. by raising the sensory impression of the front end of the cigarette. Therefore, justification to produce larger quantities of sample, and conduct more formal sensory panel testing, was established. This would require sample manufacture by cigarette making machine, and therefore a means to apply additive, in an intermittent manner, to the forming tobacco rod. Since this technology was not currently available, part of the project had now become to develop a pulsed spraying device to achieve these aims. The remainder of this report describes this development, the manufacture of samples and their subsequent sensory testing. 2. THE MACHINE-MAKING STUDY Pulsed application during manufacture A means has been sought to accurately apply a quantity of sodium carbonate to the front third only of each tobacco rod during cigarette manufacture, initially using a Molins Mk8 cigarette making machine. Tobacco rods are formed such that the eventual "front" ends of a pair of consecutive rods are adjacent. Thus, a single pulse spraying of additive solution could be applied to the front portion of two cigarettes. This gives the immediate advantage that any system that is employed to pulse spray onto the forming tobacco rod only needs to operate at half making speed. However, the device and its components should ultimately be capable of reliable operation at higher speeds eg at Molins Mk9 speeds or faster. In the development stage, however, a Molins Mk8 maker was to be used - at a speed of 1000 cigarettes/minute. In this case, since each spraying operation encompasses a pair of cigarette rods, the prototype device would be required to operate comfortably at a speed of 500 operations a minute. The chosen system would also have to be timed very accurately to apply only at the appropriate part of the tobacco rod, and there should be sufficient control, if deemed necessary, to enable the avoidance of "carry-over" onto the rear portion of the rod. A device that already exists to apply materials, primarily menthol or top- dressing flavours, onto the forming tobacco rod, is the Arenco DOQ. This device comprises a pressurised reservoir leading to a spray nozzle assembly (Figure 2). The operation of the unit is controlled using solenoid valves in both the air and liquid lines. The nozzle enables an atomised spray to be generated by mixing the flavour, or additive solution, with air, and this assembly is designed to fit immediately before the garniture on a Molins cigarette making machine. Flow rate is adjusted by means of a needle valve in the liquid line. In this way, material is sprayed vertically upwards onto the stream of tobacco passing above it and the application level is determined by the liquid flow-rate, concentration of solution sprayed, eventual rod density and also the making speed. © 1993 BAT Co. Ltd. Do not copy or show to unauthodsed persons

In order to achieve a pulsed, rather than continuous spray from this device, it has been necessary to incorporate two additional solenoid valves in the feed to the spray nozzle head; one in the liquid line and the second in the atomising air line. The valves operate simultaneously, switching both air and liquid on and off as required. The second valve was found to be essential, since without it the continuous air flow between pulses resulted in no clear cut.off, and some continuous bleeding of solution even when the liquid valve was off. The initial lab development of a suitably modified DOQ employed a small electronic device to switch the solenoid valves on and off with an appropriate frequency. However, once transferred to the manufacturing environment, the timing and frequency of pulsing must be referenced to the making machine. The site deemed most appropriate for this synchronisation was considered to be the cutting knife drive shaft. Full details of the attachmentJsynchronisation of the Arenco DOQ with the Molins Mk8 cigarette making machine are given in Appendix 1. Sample specification and manufacture With the aid of a cigarette design computer program ("CigDes"), three cigarette specifications were generated, to give a range of low delivery products with predicted deliveries of 9,1, 7 and 5 mg of tar. These specifications are included as Table 2. © 1993 BAT Co. Ltd. Do not copy or show to unauthorised persons