Routine Analytical Chemistry Sub-Group. Technical Report

Transcription

1 Routine Analytical Chemistry Sub-Group Technical Report 2017 Collaborative Study of CORESTA Monitor 8 (CM8) for the Determination of Test Piece Weight, TPM, Water, Nicotine, NFDPM, Carbon Monoxide and Puff Count Obtained under Mainstream ISO and Intense Smoking Regimes November 2017 Coordinator: Linda A. Crumpler, Cerulean, USA Authors: Guy Jaccard and Donatien Tafin Djoko Philip Morris International, Switzerland

2 Table of Contents 1. Introduction Organisation Participants Protocol Raw Data Statistical Analysis Exclusion of outliers Numerical outlier technique: Cochran & Grubbs test Repeatability and Reproducibility estimation Data Representation ISO smoking regime Intense smoking regime Comparison Comparisons of Linear with Rotary smoking machines for test piece CM Comparison with previous studies Conclusion APPENDIX A List of Participating Laboratories APPENDIX B Experimental Protocol APPENDIX C Departures from Experimental Protocol APPENDIX D Raw Data Set APPENDIX E IUPAC 1994 Harmonized Statistical Procedure APPENDIX F Linear Rotary Comparisons APPENDIX G Rotary Smoking s... 74

3 1. Introduction The CORESTA Routine Analytical Chemistry Sub-Group has been given the responsibility to organize the annual testing of the CORESTA Monitor test piece. The 2017 study was designed: to measure mainstream ISO and Intense smoke yields of nicotine-free dry particulate matter (NFDPM), nicotine (NIC) and carbon monoxide (CO) to verify the current monitor test piece CM8 to determine intra- and inter-laboratory variability for the measured ISO and Intense smoke yields for the CM8 to verify the conditioned weight for the CM8 2. Organisation 2.1 Participants In total 32 laboratories participated in the 2017 study. Appendix A lists the participating laboratories in alphabetical order. Note: the mix of labs between ISO and Intense differed. 32 laboratories delivered data for ISO mainstream smoke for the CM8 test piece using 45 smoking machines (24 rotary smoking machines and 21 linear smoking machines). 26 laboratories delivered data for Intense mainstream smoke for the CM8 test piece using 34 smoking machines (16 rotary smoking machines and 18 linear smoking machines). One laboratory provided an incomplete data set (conditioned weight data missing). Table 1 summarises the number and type of smoking machines used. Table 1: Number of smoking machines Class ISO Intense Linear smoking machines LX-20 type 2 2 SM450 type LM20X type Hawktech 2000 type Rotary smoking machines RM20 A type 1 1 RM20D type 2 2 RM20H type RM 20M type RM 20 type RM200 A type 13 9 Total RAC-134-CTR CM8 Coll. Study 2017 November /103

4 Table 2 summarises the number of datasets provided for the reported parameters. Table 2: Number of datasets obtained Physical Parameter WEIGHT 220 Smoking Parameter ISO Intense TPM WATER NIC NFDPM CO PUFF Total individual data obtained Total individual data expected A code number has been assigned in confidence by the study co-ordinator to each of the individual smoking machines used in the study. 2.2 Protocol Participants were requested to follow the protocol Annual Study of the CORESTA Monitor Test Piece CM Experimental Protocol (provided in Appendix B) to analyse the product CM8 and to report the seven parameters listed in Table 3. Table 3: Parameters to be reported Parameter Number of replicates Unit Code Conditioned Weight 5 mg/test piece (t.p.) WEIGHT Total Particulate Matter 5 mg/test piece (t.p.) TPM Water content (smoke) 5 mg/test piece (t.p.) WATER Smoke Nicotine 5 mg/test piece (t.p.) NIC Nicotine Free Dry Particulate Matter 5 mg/test piece (t.p.) NFDPM Carbon Monoxide 5 mg/test piece (t.p.) CO Puff count 5 puff/test piece (t.p.) PUFF CM8 test pieces should have been smoked under two different mainstream smoking regimes with a fixed butt length of 33 mm as defined in the following table: Table 4: Mainstream ISO and Intense smoking regime parameters Smoking Regime Puff Volume (ml) Puff Frequency (s) Puff Duration (s) Vent. Blocking (%) ISO Intense RAC-134-CTR CM8 Coll. Study 2017 November /103

5 The CM8 test piece was purchased by the participants individually via Borgwaldt KC or Cerulean. 3. Raw Data Table 5 provides the results for conditioned weight, while Tables 6 to 11 list the results for the reported parameters for ISO smoking regime. Tables 12 to 17 list the results for intense smoking regime. Basic statistics (mean and standard deviation) were applied to summarize the raw data for each smoking machine used and parameters reported for CM8 test piece. At this stage no outlier statistics were applied. Appendix D contains the raw data (single values) for every measured parameter reported by smoking machine included in the study. RAC-134-CTR CM8 Coll. Study 2017 November /103

6 Table 5: WEIGHT mean and standard deviation per laboratory code Lab Code Mean SD CV [%] Sample size 1B L B L A L L L A L B L C L L L L L L L L L L L L L L A R A R A R B R A R B R C R A R B R R R B R R A R B R RAC-134-CTR CM8 Coll. Study 2017 November /103

7 Lab Code Mean SD CV [%] Sample size 14 R R D R A R B R R R R R RAC-134-CTR CM8 Coll. Study 2017 November /103

8 Table 6: ISO smoking regime. TPM mean and standard deviation per laboratory code Lab Code Type Mean (mg/t.p.) SD CV [%] Sample size 1B L B L A L L L A L B L C L L L L L L L L L L L L L L A R A R A R B R A R B R C R A R B R R R B R R A R B R RAC-134-CTR CM8 Coll. Study 2017 November /103

9 Lab Code Type Mean (mg/t.p.) SD CV [%] Sample size 14 R R D R A R B R R R R R RAC-134-CTR CM8 Coll. Study 2017 November /103

10 Table 7: ISO smoking regime. WATER mean and standard deviation per laboratory code Lab Code Type Mean (mg/t.p.) SD CV [%] Sample size 1B L B L A L L L A L B L C L L L L L L L L L L L L L L A R A R A R B R A R B R C R A R B R R R B R R A R B R RAC-134-CTR CM8 Coll. Study 2017 November /103

11 Lab Code Type Mean (mg/t.p.) SD CV [%] Sample size 14 R R D R A R B R R R R R RAC-134-CTR CM8 Coll. Study 2017 November /103

12 Table 8: ISO smoking regime. NICOTINE mean and standard deviation per laboratory code Lab Code Type Mean (mg/t.p.) SD CV [%] Sample size 1B L B L A L L L A L B L C L L L L L L L L L L L L L L A R A R A R B R A R B R C R A R B R R R B R R A R B R RAC-134-CTR CM8 Coll. Study 2017 November /103

13 Lab Code Type Mean (mg/t.p.) SD CV [%] Sample size 14 R R D R A R B R R R R R RAC-134-CTR CM8 Coll. Study 2017 November /103

14 Table 9: ISO smoking regime. NFDPM mean and standard deviation per laboratory code Lab Code Type Mean (mg/t.p.) SD CV [%] Sample size 1B L B L A L L L A L B L C L L L L L L L L L L L L L L A R A R A R B R A R B R C R A R B R R R B R R A R B R RAC-134-CTR CM8 Coll. Study 2017 November /103

15 Lab Code Type Mean (mg/t.p.) SD CV [%] Sample size 14 R R D R A R B R R R R R RAC-134-CTR CM8 Coll. Study 2017 November /103

16 Table 10: ISO smoking regime. CO mean and standard deviation per laboratory code Lab Code Type Mean (mg/t.p.) SD CV [%] Sample size 1B L B L A L L L A L B L C L L L L L L L L L L L L L L A R A R A R B R A R B R C R A R B R R R B R R A R B R RAC-134-CTR CM8 Coll. Study 2017 November /103

17 Lab Code Type Mean (mg/t.p.) SD CV [%] Sample size 14 R R D R A R B R R R R R RAC-134-CTR CM8 Coll. Study 2017 November /103

18 Table 11: ISO smoking regime. PUFF COUNT mean and standard deviation per laboratory code Lab Code Type Mean (puff/t.p.) SD CV [%] Sample size 1B L B L A L L L A L B L C L L L L L L L L L L L L L L A R A R A R B R A R B R C R A R B R R R B R R A R B R RAC-134-CTR CM8 Coll. Study 2017 November /103

19 Lab Code Type Mean (puff/t.p.) SD CV [%] Sample size 14 R R D R A R B R R R R R RAC-134-CTR CM8 Coll. Study 2017 November /103

20 Table 12: Intense smoking regime. TPM mean and standard deviation per laboratory code Lab Code Type Mean (mg/t.p.) SD CV [%] Sample size 1B L B L L L A L B L C L L L L L L L L L L L L A R A R A R B R A R B R R R R D R A R B R R R R R RAC-134-CTR CM8 Coll. Study 2017 November /103

21 Table 13: Intense smoking regime. WATER mean and standard deviation per laboratory code Lab Code Type Mean (mg/t.p.) SD CV [%] Sample size 1B L B L L L A L B L C L L L L L L L L L L L L A R A R A R B R A R B R R R R D R A R B R R R R R RAC-134-CTR CM8 Coll. Study 2017 November /103

22 Table 14: Intense smoking regime. NICOTINE mean and standard deviation per laboratory code Lab Code Type Mean (mg/t.p.) SD CV [%] Sample size 1B L B L L L A L B L C L L L L L L L L L L L L A R A R A R B R A R B R R R R D R A R B R R R R R RAC-134-CTR CM8 Coll. Study 2017 November /103

23 Table 15: Intense smoking regime. NFDPM mean and standard deviation per laboratory code Lab Code Type Mean (mg/t.p.) SD CV [%] Sample size 1B L B L L L A L B L C L L L L L L L L L L L L A R A R A R B R A R B R R R R D R A R B R R R R R RAC-134-CTR CM8 Coll. Study 2017 November /103

24 Table 16: Intense smoking regime. CO mean and standard deviation per laboratory code Lab Code Type Mean (mg/t.p.) SD CV [%] Sample size 1B L B L L L A L B L C L L L L L L L L L L L L A R A R A R B R A R B R R R R D R A R B R R R R R RAC-134-CTR CM8 Coll. Study 2017 November /103

25 Table 17: Intense smoking regime. PUFF COUNT mean and standard deviation per laboratory code Lab Code Type Mean (puff/t.p.) SD CV [%] Sample size 1B L B L L L A L B L C L L L L L L L L L L L L A R A R A R B R A R B R R R R D R A R B R R R R R RAC-134-CTR CM8 Coll. Study 2017 November /103

26 4. Statistical Analysis 4.1 Exclusion of outliers Numerical outlier technique: Cochran & Grubbs test The statistical evaluation of data for this collaborative study followed the methods provided by ISO For outlier testing, the Grubbs and Cochran methods were used. ISO Tests Cochran s test Grubbs test Consistency Within-laboratory variability: Suitable for detecting whether the highest value in a set of laboratory standard deviations is an outlier or not. Between-laboratory variability: Suitable for detecting whether the highest (or lowest) laboratories averages are outliers or not. The protocol Harmonized statistical procedure defined by IUPAC (International Union of Pure and Applied Chemistry) has been applied as well. It consists of sequential applications of the Cochran and Grubbs tests until no further outliers are detected or until a drop of more than 22.2% in the original number of laboratories would occur (see flowchart in Appendix F). First apply the Cochran outlier test, and if an outlying laboratory is identified, then a single value Grubbs test is performed on the individual values of this outlying laboratory (Individual Grubbs). If no individual value is identified as outlier, the outlying laboratory is removed. When an individual value is identified as an outlier, only this value is removed. Afterwards single-value Grubbs test is applied and outlying laboratories are removed. If no laboratory is identified as outlier, the pair-value test is applied (two values at the same end). Remove any laboratory(ies) flagged by these tests, but stop removal if more than 22.2% would be removed. The test was performed with both linear and rotary smoking machines together, but also with rotary, respectively linear smoking machines separately. The results provided in the report for all smoking machines, respectively with linear or rotary only smoking machines are obtained after exclusion of the outliers for those respective categories. Note: The paired value Grubbs test is an extension of the single value Grubbs' test for pairs of outliers. The use is advised in ISO 5725 to detect pairs of outlier undetectable sequentially by using single value Grubbs' test. The outlying laboratories are provided in the Tables 18 to Repeatability and Reproducibility estimation Repeatability and reproducibility limits are calculated for both smoking regimes, each parameter and for both smoking machine types. The results are summarized in tables 18 to 23, with all machines grouped, respectively with linear or rotary machines only, both with ISO and intense smoking regimes. Outlying laboratories are provided in the same tables. The corresponding graphs named individual values are given in Part 5 of this report. 1 ISO :1994 : Accuracy (trueness and precision) of measurement methods and results Part 2: Basic method for the determination of repeatability and reproducibility of a standard measurement method RAC-134-CTR CM8 Coll. Study 2017 November /103

27 Table 18: r&r estimations for CM8 under ISO smoking regime (all machines) p n Global Mean Sr CVr [%] r SR CVR [%] R C G/G2 Stragglers Parameter NA 32 Conditioned.Weight NA 25 TPM NA 32 Water NA NA Nicotine All NA NFDPM NA NA CO NA NA NA Puff.count Table 19: r&r estimations for CM8 under ISO smoking regime (rotary machines) p n Global Mean Sr CVr [%] r SR CVR [%] R C G/G2 Stragglers Parameter NA NA 4B TPM NA NA 16D Water NA NA 31 Nicotine Rotary NA NA NA NFDPM NA NA NA CO NA NA 5A Puff.count Table 20: r&r estimations for CM8 under ISO smoking regime (linear machines) p n Global Mean Sr CVr [%] r SR CVR [%] R C G/G2 Stragglers Parameter NA NA TPM NA 32 Water NA NA Nicotine Linear NA 32 NFDPM NA NA NA CO NA NA NA Puff.count RAC-134-CTR CM8 Coll. Study 2017 November /103

28 Table 21: r&r estimations for CM8 under Intense smoking regime (all machines) p n Global Mean Sr CVr[%] r SR CVR [%] R C G/G2 Stragglers Parameter NA NA NA TPM NA NA NA Water NA 17 Nicotine All janv NA NA NFDPM NA 20 CO NA NA NA Puff.count Table 22: r&r estimations for CM8 under Intense smoking regime (rotary machines) p n Global Mean Sr CVr [%] r SR CVR [%] R C G/G2 Stragglers Parameter NA NA NA TPM NA NA 35 Water NA NA 7 Nicotine Rotary NA NA NA NFDPM NA NA NA CO NA NA NA Puff.count Table 23: r&r estimations for CM8 under Intense smoking regime (linear machines) p n Global Mean Sr CVr [%] r SR CVR [ %] R C G/G2 Stragglers Parameter NA NA NA TPM NA NA NA Water '10-17 NA 7 Nicotine Linear NA NA NFDPM NA NA CO NA NA NA Puff.count 5. Data Representation The following figures show the raw data for CM8 and for all parameters. The plots indicate mean values with 95% confidence interval (CI) highlighted in green (for valid participants) or in red (for participants detected as outlier) for each smoking machine. Laboratories excluded by outlier statistics are denoted by the name of the test rejecting the laboratory (G for Grubbs, G2 for double Grubbs and C for Cochran). The confidence limits at 95% and 99% for the difference between the reference value (average of all laboratories) and the average of one laboratory are plotted in orange and red respectively. RAC-134-CTR CM8 Coll. Study 2017 November /103

29 5.1 ISO smoking regime Figure 1: CM8 Individual observations of conditioned test piece WEIGHT Outlier value in red Figure 2: CM8 Individual observations of mainstream ISO TPM Outlier value in red, straggler value in orange RAC-134-CTR CM8 Coll. Study 2017 November /103

30 Figure 3: CM8 Individual observations of mainstream ISO WATER Outlier value in red, straggler value in orange Figure 4: CM8 Individual observations of mainstream ISO Nicotine Outlier values in red RAC-134-CTR CM8 Coll. Study 2017 November /103

31 Figure 5: CM8 Individual observations of mainstream ISO NFDPM Outlier values in red, straggler value in orange Figure 6: CM8 Individual observations of mainstream ISO CO Outlier values in red RAC-134-CTR CM8 Coll. Study 2017 November /103

32 Figure 7: CM8 Individual observations of mainstream ISO PUFF Count 5.2 Intense smoking regime Figure 8: CM8 Individual observations of mainstream Intense TPM RAC-134-CTR CM8 Coll. Study 2017 November /103

33 Figure 9: CM8 Individual observations of mainstream Intense WATER Figure 10: CM8 Individual observations of mainstream Intense Nicotine Outlier values in red RAC-134-CTR CM8 Coll. Study 2017 November /103

34 Figure 11: CM8 Individual observations of mainstream Intense NFDPM Outlier value in red Figure 12: CM8 Individual observations of mainstream Intense CO Outlier value in red, straggler value in orange RAC-134-CTR CM8 Coll. Study 2017 November /103

35 Figure 13: CM8 Individual observations of mainstream Intense PUFF Count 6. Comparison 6.1 Comparisons of Linear with Rotary smoking machines for test piece CM8 It should be noted that a statistical evaluation of differences in linear and rotary smoking machine results on the basis of just one product (CM8) could be misleading. Furthermore, a simplistic differentiation between linear and rotary smoking machines could also be misleading as both machine types used in this study comprise different models produced by different manufacturers. Furthermore it has to be noted that each measurement from a rotary machine represents a single result from a single pad onto which the smoke of twenty (or ten in intense smoking regime) test pieces is trapped. Linear smoking machines require 4 pads to collect mainstream TPM from the same number of test pieces and hence each measurement on a linear machine is the average of 4 results. This has to be taken into account when interpreting statistical differences in repeatability between machine classes. Due to the aforementioned reasons a pure statistical comparison is considered to have limited value. Graphical comparison of smoke data separated for linear and rotary smoking machines are available in Appendix F. 6.2 Comparison with previous studies This section compares the results of this study with previous collaborative studies of different test pieces conducted by the Routine Analytical Subgroup under the ISO smoking regime. It has to be noted that the number of smoking machines, smoking machine types and participants varies throughout the studies. Table 24 lists the averages, reproducibility and repeatability limits for the test pieces CM6, CM7 and CM8 under ISO smoking regime. RAC-134-CTR CM8 Coll. Study 2017 November /103

36 Table 24: Averages, repeatability, reproducibility of CORESTA collaborative studies from 2008 to 2017 for different test pieces under ISO smoking conditions ISO smoking regime Average Repeatability (r) Reproducibility (R) Year Test piece CM6 CM6 CM6 CM7 CM7 CM7 CM7 CM8 CM8 CM8 TPM (mg/t.p.) WATER (mg/t.p.) NICOTINE (mg/t.p.) NFDPM (mg/t.p.) CO (mg/t.p.) PUFF (/t.p.) TPM (mg/t.p.) WATER (mg/t.p.) NICOTINE (mg/t.p.) NFDPM (mg/t.p.) CO (mg/t.p.) PUFF (/t.p.) TPM (mg/t.p.) WATER (mg/t.p.) NICOTINE (mg/t.p.) NFDPM (mg/t.p.) CO (mg/t.p.) PUFF (/t.p.) There is no systematic tendency for the mean values of CM 8 between 2014 and 2017: some parameters decrease slightly, but nicotine increases slightly. Repeatability and reproducibility figures are quite stable and comparable to those obtained with the other test pieces, taking into account the slightly higher mean values of TPM and nicotine for CM 8 than for the previous test pieces. RAC-134-CTR CM8 Coll. Study 2017 November /103

37 Table 25 lists the averages, reproducibility and repeatability limits for the test pieces CM7 and CM8 under intense smoking regime. Table 25: Averages, repeatability, reproducibility of CORESTA collaborative studies from 2013 to 2017 for different test pieces under intense smoking conditions Intense smoking regime Average Repeatability (r) Reproducibility (R) Year Test piece CM7 CM8 CM8 TPM (mg/t.p.) WATER (mg/t.p.) NICOTINE (mg/t.p.) NFDPM (mg/t.p.) CO (mg/t.p.) PUFF (/t.p.) TPM (mg/t.p.) WATER (mg/t.p.) NICOTINE (mg/t.p.) NFDPM (mg/t.p.) CO (mg/t.p.) PUFF (/t.p.) TPM (mg/t.p.) WATER (mg/t.p.) NICOTINE (mg/t.p.) NFDPM (mg/t.p.) CO (mg/t.p.) PUFF (/t.p.) The mean values obtained for CM 8 are very stable between 2016 and The values for repeatability and reproducibility are also generally well aligned. Table 26 lists the averages, reproducibility and repeatability limits of conditioned weight Table 26: Average weight data, repeatability and reproducibility for test pieces CM8 Study CM CM CM Weight (mg/t.p.) Repeatability (r) Reproducibility (R) The data obtained in the three collaborative tests for the test piece conditioned weight are quite stable in terms of average, repeatability and reproducibility. RAC-134-CTR CM8 Coll. Study 2017 November /103

38 7. Conclusion In summary the analytical results for the CM8 test piece lead to the conclusion that it can be used as a monitor test piece in smoke analysis. RAC-134-CTR CM8 Coll. Study 2017 November /103

39 APPENDIX A List of Participating Laboratories Altria ASL BAT Brazil - Souza Cruz BAT Germany (GmbH) Borgwaldt KC C.I.T. Monte Paz S.A. China National Tobacco Quality Supervision and Test Center Enthalpy Analytical Richmond Essentra Scientific Services Global Laboratory Service Imperial JOURE Imperial Tobacco Polaska S.A. Imperial Tobacco Production Ukraine (ITPU) INCIENSA-CNRBRO ITC Limited ITG Brands Japan Tobacco Inc. JTI Germany GmbH JTI Oekolab Karelia Tobacco Co. KT&G Landewyck Group sarl. Liggett Group LNE Miaoli Tobacco Industry Corp.,LTD MST/Vectis Papierfabrik Wattens PMI AL CC's PMI R&D PT HM Sampoerna Tbk. R. J. Reynolds Tobacco Co. Reemstma/Imperial Tobacco Seita SWM-LTR Tabacalera del Este S.A. Out of the 35 participating laboratories, 8 labs did not perform the intense smoking regime. RAC-134-CTR CM8 Coll. Study 2017 November /103

40 APPENDIX B Experimental Protocol Routine Analytical Chemistry Sub-Group Collaborative Study ANNUAL STUDY OF THE CORESTA MONITOR TEST PIECE CM EXPERIMENTAL PROTOCOL Linda A. Crumpler, Cerulean, USA 10 January 2017 RAC-134-CTR CM8 Coll. Study 2017 November /103

41 Table of Contents Foreword Introduction Objectives Study Coordinator(s) Normative References Methods Target Variables and Data Reporting Study Time Table Data Submission Explanation of Minimum Statistics to be Applied for a Collaborative Study Bibliography Appendix A Système International (SI) Units and their Symbols RAC-134-CTR CM8 Coll. Study 2017 November /103

42 Foreword This CORESTA protocol describes the elements that the Standards Task Force has identified as important for describing elements of an experimental protocol for a collaborative study or a proficiency study as defined below (Horwitz 862): Collaborative study A collaborative study is an interlaboratory study in which each laboratory uses the defined method of analysis to analyse identical portions of homogeneous materials to assess the performance characteristics obtained for that method of analysis. CORESTA recognizes that collaborative studies require considerable effort and should be conducted only on methods that have received adequate prior testing. Proficiency Study A proficiency study is an interlaboratory study consisting of one or more assaysconducted by a group of laboratories on one or more identical materials, by whatever method is in use in each laboratory, for the purpose of comparing the results of each laboratory with those of other laboratories, with the objective of evaluating or improving performance. Proficiency studies often serve as pre-work for selecting a method for a full collaborative study. RAC-134-CTR CM8 Coll. Study 2017 November /103

43 1. Introduction The CORESTA Routine Analytical Chemistry Sub Group has been given the responsibility to organize the annual testing of the CORESTA Monitor test piece. This year the study will include CM8. 2. Objectives This year s study will have expanded objectives that include smoking the monitor test piece under both ISO and a specified intense smoking regime. The study objectives are: To measure mainstream smoke yields of nicotine-free dry particulate matter (NFDPM), nicotine and carbon monoxide (CO) when smoked under ISO and a specified intense smoking regime To establish the conditioned weight of CM8 prior to smoking To determine intra- and inter-laboratory variability for the measured smoke yields and to continue this comparison over time To obtain statistical information about possible differences in smoke yields from different smoking machines 3. Study Coordinator(s) Linda A. Crumpler Linda.crumpler(at)cerulean.com Phone: Mobile: Normative References The relevant Standards for ISO, Intense smoking and statistical analyses are: Definitions and standard conditions ISO 3308 Conditioning ISO 3402 TPM / NFDPM ISO 4387 Carbon monoxide ISO 8454 Nicotine ISO Water (GC) ISO Statistical Analysis ISO Ventilation zone blocking, puffing parameters, and number of test pieces smoked *Health Canada Method T-115 paragraph 16.1 Tar and Nicotine and CO T-115e4.pdf *Note: T-115 has been removed from the Health Canada website and therefore is embedded here. RAC-134-CTR CM8 Coll. Study 2017 November /103

44 5. Methods 5.1. General Guidance As a general guide, experiments conducted under each puffing regime should be conducted according to ISO standard requirements. Exceptions to this guiding principle include the puffing parameters applied for intense smoking regime, the vent blocking applied, the number of cigarettes smoked per sample and the butt length for smoking. The study (the smoking results) and the daily use of the monitor are based on a fixed butt length of 33 mm. Please note that any operating conditions (a) not specified in the above ISO standards, or regarded as optional, (b) not specified in the intense method documents provided as guidelines and (c) any circumstances that may have influenced test results should be noted in the test report. Puffing parameter and vent blocking specifications for each puffing regime are summarized in Table 1. Note: Although the CORESTA Monitor Test Pieces are unventilated it is advised to follow the Health Canada procedures exactly and tape the filter ventilation zone. Table 1: Smoking Regime Specifications Smoking Regime Puff Volume (ml) Puff Frequency* (s) Puff Duration (s) Vent Blocking (%) Description 1 35 ± ± ± ISO 2 55 ± ± ± Intense 5.2. Test Specimen(s) CM8 test piece Sampling The smoking plans will use 200 test pieces for each rotary smoking machine or 220 test pieces for each linear smoking machine included in the study. A composite of all packs should be prepared for conditioning according to ISO 8243:2006. If a substantial period of time is expected to occur between the times when the test pieces are smoked with the different smoking regimes, then two separate composite samples should be prepared, one for each smoking regime Ventilation Zone Blocking for Intense Smoking The blocking of the filter ventilation zone will be achieved by wrapping a single layer of invisible tape, 19.0 mm wide or equivalent, around the complete circumference of the filter and tipping paper of the test articles. RAC-134-CTR CM8 Coll. Study 2017 November /103

45 5.5. Butt Marking For the purpose of this study the CM8 test pieces will be butt marked to a fixed butt length of 33 mm Conditioning The composite samples shall be conditioned for at least 48 hours but no more than 10 days under the following conditions: Temperature: 22 ± 1 C Relative humidity: 60 ± 3% Temperature and relative humidity of the conditioning enclosure (chamber or room) shall be checked and recorded for every smoking run. Test cigarettes to be smoked for the determination of NFDPM, nicotine and CO should be selected at random from the conditioned cigarettes. Any test pieces with visible defects shall be discarded Test Piece Weight The test articles will be weighed for ISO smoking only since the ventilation zone blocking tape applied for intense smoking may vary between laboratories. The weight of the test piece after conditioning is to be determined by weighing all test pieces to be smoked just before they are placed at the smoking machine. The result is reported as the mean weight of the 20 test pieces for linear or rotary smoking machine at ISO puffing parameters Smoking Where smoking laboratories have more than one type of smoking machine (i.e. linear and rotary variants) they are encouraged to include as many types in the study as possible Smoking Plans The smoking plans will incorporate one product (CM8), shown as M in the smoking plans. These plans are designed to smoke the CM8 in all ports for each puffing regime, spread over a number of days. If possible, each participating laboratory should use only one operator (for each smoking machine type) throughout the course of the study. For ISO smoking, one test result is defined as the mean yield obtained from smoking 20 test pieces in a single smoking machine run, therefore from a rotary machine it is the result from one run smoking 20 test pieces and from a linear machine it is the average of four ports, smoking five test pieces per port. This is repeated on five separate days resulting in 100 test pieces being smoked (Tables 2-3). RAC-134-CTR CM8 Coll. Study 2017 November /103

46 For Intense smoking, the number of test pieces smoked per filter pad is reduced to avoid overloading the pad. As a consequence, the number of smoking runs/ports has to be increased so that the same nominal numbers of test pieces are smoked. Thus, a test result from a rotary machine is the mean of two runs, each of 10 test pieces, and from a linear machine it is the average of eight ports, smoking three test pieces per port (Tables 4-5). Note that the side seam of the test pieces should be positioned at random and not in a fixed position since it has been observed that some laboratory staff habitually place cigarettes in a particular orientation. Ports not occupied by the test pieces are left blank, which may be assigned to any other cigarette that is to be smoked by the laboratory, or the ports may be left empty. Table 2: Smoking Plan for 20-port Linear s for ISO smoking (5 test pieces per pad) Port Number Day M M M M 2 M M M M 3 M M M M 4 M M M M 5 M M M M Table 3: Smoking Plan for 20-port Rotary s for ISO smoking (20 test pieces per pad) Day Run Test Piece 1 1 M 2 2 M 3 3 M 4 4 M 5 5 M RAC-134-CTR CM8 Coll. Study 2017 November /103

47 Table 4: Smoking Plan for 20-port Linear s for Intense smoking (3 test pieces per pad) Da y Ru n M M M M 7 M M M M Port Number 8 M M M M M M M M 10 M M M M 11 M M M M 12 M M M M 13 M M M M 14 M M M M 15 M M M M Table 5: Smoking Plan for 20-port Rotary s for Intense smoking (10 test pieces per pad) Day Run Test Piece 6 M 7 M 8 M 9 M 10 M 11 M 12 M 13 M 14 M 15 M 6. Target Variables and Data Reporting The variables to be reported fall into two categories: those which will form the basis of the statistical evaluation of the data and those that are ancillary to the test. The variables to be reported for statistical analysis are: Mean weight of the Monitor after conditioning (just prior to smoking) Mean puff number per test piece Mean TPM per test piece RAC-134-CTR CM8 Coll. Study 2017 November /103

48 Mean nicotine delivery per test piece Mean water delivery per test piece Mean NFDPM delivery per test piece Mean CO delivery per test piece The ancillary variables are: Date of testtype of smoking machine used Temperature (C ) and relative humidity (RH%) of conditioning enclosure Temperature (C ) and relative humidity (RH%) of smoking environment Atmospheric pressure in the laboratory during smoking Ambient air flow in the smoking machine before smoking (mm/sec) Dimensions and rounding of test results Analytical variables: Weight of Monitor mg per test piece ####.# Puff number number ##.## TPM mg per test piece ##.## Nicotine mg per test piece #.### Water mg per test piece ##.## NFDPM mg per test piece ##.## Carbon Monoxide mg per test piece ##.## Ancillary variables: Date of test DD/MM/YY Conditioning temperature degrees Celsius ##.# Conditioning humidity percent RH ##.# Air velocity mm/s ### Laboratory temperature degrees Celsius ##.# Laboratory humidity percent RH ##.# Laboratory barometric pressure kpa ###.# RAC-134-CTR CM8 Coll. Study 2017 November /103

49 Please note that the rounding of the data to the formats specified above must take place after any calculations that may be involved. All calculations will use the laboratory data as recorded using the maximum number of digits available. All data arising from this study will be made available to the laboratories participating in the study. The distribution of data will be done after collation and statistical evaluation. To facilitate the statistical evaluation, all data must be reported using the attached spreadsheets and must be submitted via . Please use the attached spreadsheets as provided without changing the format. RAC-134-CTR CM8 Coll. Study 2017 November /103

50 7. Study Time Table The smoking should be performed during January and February, Strict adherence to the timeline is very important for the data to be compiled for an initial assessment at the next Sub-Group meeting on April, Data not received by March 1, 2017 will be excluded from the formal statistical analysis. Please review your data carefully before submitting as the statistical analysis is very labor intensive. The statistical analysis of the results will follow as closely as possible the recommendations of ISO Should any questions arise please contact the study coordinator, Linda Crumpler. All results must be sent via to: Linda A. Crumpler 8. Data Submission All results must be sent via to: Linda A. Crumpler linda.crumpler(at)cerulean.com linda.crumpler(at)cerulean.com 9. Explanation of Minimum Statistics to be Applied for a Collaborative Study At a minimum Repeatability and Reproducibility estimation will be applied. RAC-134-CTR CM8 Coll. Study 2017 November /103

51 Bibliography Chem. Pure Appl. QUALITY ASSURANCE SCHEMES THE INTERNATIONAL HARMONIZED PROTOCOL FOR THE PROFICIENCY TESTING OF ANALYTICAL CHEMISTRY LABORATORIES (n.d.): n. pag. Web. Council of Science Editors. 8th ed. N.p.: Council of Science Editors, n.d. Print. Form and Style for ASTM Standards. West Conshohocken, PA: ASTM INTERNATIONAL, Print. Horwitz, William. "Protocol for the Design, Conduct and Interpretation of Collaborative Studies." Pure and Applied Chemistry 60.6 (1988): ShareGate. International Union of Pure and Applied Chemistry, De Gruyter, Jan Web. 30 July ISO 5725 Standard And Reference Laboratories Error.pdf. "ISO 5252." ISO 5725 Standard And Reference Laboratories Error (n.d.): n. pag. Web. "IUPAC 2006 IUPAC, Pure and Applied Chemistry 78, "" N.p., n.d. Web. 25 Aug "Numbers, Units, Mathematicial Expressions, and Statistics." Scientific Style and Format; The CSE Manual for Authors Editors, and Publishers. 8th ed. Chicago: The University of Chicago, Print. RAC-134-CTR CM8 Coll. Study 2017 November /103

52 Appendix A Système International (SI) Units and their Symbols SI base units Base Quantity Name Symbol Length meter m Mass kilogram Kg Time second s Electric current ampere A Temperature kelvin K Amount of substance mole mol Luminous intensity Candela cd SI derived units Derived Quantity Name Symbol Equivalent SI Units Frequency hertz Hz s -1 Force newton N m kg s -2 Pressure pascal Pa N/m 2 Energy joule J N m Power watt W J/s Electric charge coulomb C s A Electric potential volt V W/A Electric resistance ohm Ω V/A Celsius temperature Degrees Celsius C K* *Unit degree Celsius is equal in magnitude to unit kelvin SI prefixes Factor Name Symbol Numerical value tera T giga G mega M kilo k hecto h deka da deci d centi c milli m RAC-134-CTR CM8 Coll. Study 2017 November /103