Side stream smoke, which is emitted directly from the source between puffs and not exhaled by the smoker, has been shown to produce greater emissions than mainstream smoke . Based on observations of the participant, the inhalation time of the puff was about 2 s, and the exhalation time was ~2–4 s, making the total puff time about 6 s. We applied the same 3-puff protocol to the Absolute Xtracts vaping pen, which carried out an internal 15-s pre-heat mode prior to the start of puffing and produced little side stream emissions between puffs. An important advantage of the 3-puff protocol in our experiments is that it avoided the extremely high PM2.5 concentrations expected to occur in the 43 m3 room if a marijuana joint had been smoked completely in the room, thus allowing the participant to avoid exposure to unacceptably high concentrations. Based on our interviews with experienced cannabis smokers and information available on the Internet, we concluded that marijuana smoking often differed from tobacco cigarette smoking in several respects. Smoking a marijuana joint often takes place in a group setting, where more than one person smokes, following the rule, “take two puffs and pass it to the left.” We also learned that a marijuana smoker, when smoking alone, often takes just 2 or 3 puffs, then puts the joint out so it can be smoked later in the day. The 3-puff protocol in Fig. 1 includes nearly a minute between puffs for the burning joint to emit side stream smoke, thus producing both mainstream and side stream smoke in a realistic manner. This protocol also has a mathematical advantage for calculating emission source strengths,cannabis vertical farming since the 3-min emission time is much shorter than the residence time of the room, which averaged 115 min for the 60 pre-rolled joints, bongs, glass pipes, vaping pens, and cigarettes.
An objective of this study was to compare emission rates from different sources smoked in the same manner by a human participant. Although a smoking machine may reduce experimental variability, we focused on determining whether the differences between the mean emission rate of the prerolled marijuana joints and the mean emission rates of the other sources, including the tobacco cigarettes, were statistically significant. The statistical methods used in this study are designed to test whether there is a statistically significant difference between the means of two unpaired groups. The unpaired t-test is a parametric test based on estimates of the mean and standard deviation of normally distributed populations from which the samples were drawn. It tests whether the difference between two groups is greater than that caused by random sampling variation. The p value is the probability of being wrong in concluding that there is a true difference between the two groups. The smaller the p value, the greater the probability that the samples are drawn from different populations. We chose the probability p < 0.05 as our criterion for statistical significance. The statistical analyses were performed using Sigma Plot 11 , which employs the Kolmogorov-Smirnov test for a normally distributed population. This program also tests for equal variances. If these conditions are met, it performs the unpaired t-test. If either of these conditions is not met, it informs the user that the data are unsuitable for the unpaired t-test, and it recommends using the non-parametric Mann Whitney Rank Sum Test instead, which performs comparisons based on the ranks of the observations. Table 1 provides summary statistics for the 60 experiments in this study with the various cannabis and tobacco sources, based on the 3-puff protocol. The background concentrations were subtracted from the measured PM2.5 concentrations in Table 1, and the last two columns show the background PM2.5 concentrations were much smaller than the background-corrected maximum PM2.5 concentrations measured in the room for all five sources.
The background-corrected ymax concentrations of PM2.5 observed in the 24 experiments with pre-rolled joints had a mean of 540 μg/m3 and ranged from 143 to 809 μg/m3 . By comparison, the PM2.5 ymax concentrations in the 9 Marlboro tobacco cigarette experiments had a mean of 154 μg/m3 and ranged from 22 to 209 μg/m3 . Each marijuana source produced a larger mean maximum concentration ymax than the tobacco cigarettes. Table 2 shows the summary statistics for the 60 experiments with five different sources. The 24 joints had a mean PM2.5 emission rate of 7.8 mg/min, which was greater than all the other cannabis emission rates and was 3.5 times the mean PM2.5 emission rate of the Marlboro cigarettes of 2.2 mg/min. The mean emission rates of the bong and the glass pipe were 67% and 54% of the joint’s mean emission rate, respectively, and the mean emission rate of the vaping pen was 44% that of the mean emission rate of the joints. The box plots shown in Fig. 3 illustrate the frequency distributions of the PM2.5 emission rates, allowing them to be compared graphically. Only the pre-rolled cannabis joints had enough observations to show the 5th and 95th percentiles of the emission rates , while all the box plots showed the 10th and 90th percentiles . The box boundaries themselves represent the 25th and 75th percentiles, and the bong had the largest spread between these two percentiles. This result was consistent with Table 2, which shows the bong also had the greatest coefficient of variation of 0.71 for the five sources. The mean emission rate in Fig. 3 was higher for the joint than for the cigarette, which also is evident in the emission rate column of Table 2. The median in Fig. 3 showed a pattern similar to that of the mean. Table 3 shows the results of applying standard statistical tests to 10 comparisons of the different methods of smoking marijuana, vaping marijuana, and smoking tobacco cigarettes. In five of the comparisons, the t-test met the requirement that the data were normally distributed but did not meet the requirement of equal variances. In these five cases, Sigma-Plot substituted the non-parametric Mann-Whitney Rank Sum Test for the t-test. With both tests, the criterion for statistical significance was the probability p < 0.05.
The difference between the mean emission rate of the joint and the mean emission rate of the bong was statistically significant , and the differences between the mean emission rate of the joint and the mean emission rates of the glass pipe, vaping pen, and cigarette were highly statistically significant . The probabilities listed above the box plots in Fig. 3 show the statistical significance of the differences between the groups. Although there were n = 24 experiments with joints, there were only n = 9 experiments each with bongs, glass pipes, vaping pens, and Marlboro cigarettes. Comparisons of the bong vs. the glass pipe, the bong vs. vaping, the bong vs. the cigarette, the glass pipe vs. vaping, and the cigarette vs. vaping did not show a statistically significant difference in mean emission rates, which is partly due to the small sample sizes. An exception was the mean emission rate of the glass pipe compared to the mean emission rate of the cigarette, which was statistically significant . In general, groups that did not include the joint were less likely to show a statistically significant difference when compared to groups that included the joint with its high emission rate and larger sample size. The difference between the mean emission rate of the marijuana joints and the mean emission rate of the tobacco cigarettes was highly statistically significant . The largest mean decay rate in Table 2 of 0.690 h− 1 occurred with the grow cannabis in containers vaping pen, while the other four mean decay rates were fairly close together, averaging 0.509 h− 1 . When we compared the differences between the five mean decay rates, we found that only one difference was statistically significant: comparison of the mean decay rate of the 24 marijuana joints with the mean decay rate of the 9 vaping pen experiments. Based on the Mann-Whitney Rank Sum Test, the difference between the mean decay rate of the marijuana joints and the mean decay rate of the vaping pens was highly statistically significant .The measured decay rate φ for the SidePak monitor is the sum of the air exchange rate a and the deposition rate k, as well as the other possible particle losses or gains due to evaporation, condensation, and coagulation. That is, the decay rate φ = a + k + other. If we subtract the observed air exchange rate from the observed decay rate, we are left with a term called the “removal rate” due to aerosol dynamics, which is the sum of the deposition rate k and all the other gain or loss mechanisms, excluding the effect of air exchange. For the 24 cannabis joints, the mean removal rate was 0.085 h− 1 . For the bong, the glass pipe, and the cigarette, the mean removal rates were 0.111 h− 1 , 0.096 h− 1 , and 0.103 h− 1 , respectively. The average removal rate of the four marijuana combustion sources was 0.10 h− 1 , which was smaller than deposition rates listed by Thatcher et al. for a furnished room with a small fan or no fan. In contrast, the mean removal rate of the 9 vaping pen experiments was 0.321 h− 1 , which was the largest removal rate of the five sources and was 3.2 times the average removal rate of the four combustion sources . It is likely that this larger removal rate of the vaping pen was due to volatility of the vaping aerosol and its greater evaporative losses. Evaporation of particles from cannabis vaping is not expected to be as great as evaporation from e-cigarette vaping .
We believe this is an important topic for future research. Since each new marijuana joint included a factory label showing the joint’s percent THC content, we also compared the THC listed for each joint with our measurements of the joint’s PM2.5 source strength. Applying the t-test, we found the relationship between the THC percentage and the source strength was statistically significant . However, this result may occur mainly because the larger joints in our study happen to have higher THC percentages, and their larger size may cause their greater source strength. A more detailed study that controls for the size of the joint would be useful. Our measurements of ultra fine particles > 10 nm used a pair of TSI 3007 condensation particle counters that were collocated with the other monitors in the room during these experiments. The UFP results are summarized in Table S2. Of the five sources, the pre-rolled marijuana joints had the greatest average UFP source strength , while the Marlboro cigarettes had an almost equal UFP source strength . The mean UFP source strengths of the three other methods of consuming marijuana were 1.3 x 1012 particles for bongs, 6.4 x. 1011 particles for glass pipes, and 3.3 x 1011 particles for the vaping pens. Overall, the UFP source strengths of bongs, glass pipes, and vaping pens were smaller than the UFP source strengths of either the pre-rolled marijuana joints or the Marlboro cigarettes. McClure et al. studied 20 heavy users of marijuana, reporting that heavy users smoked an average of 11–12 marijuana cigarettes per day, averaging 13–14 puffs per joint. Since our study compared the PM2.5 emission rates based on 3.0 min of smoking or vaping, we also attempted to estimate the emissions produced by a fully-smoked marijuana joint. We used a precision laboratory scale to measure the weights of the 24 marijuana joints before they were smoked, which ranged from 0.56 to 1.35 g with a mean of 1.024 g . By comparison, the pre-smoking weights of the 9 Marlboro cigarettes ranged from 0.83 to 0.89 g with a mean of 0.863 g . We found that measuring the difference in the weight of a joint before and after it was smoked was challenging, because the water used to put out the joint affected its tightly rolled cannabis leaves, causing the post-smoking weight sometimes to be larger than the original weight. In addition, it was difficult to account for the smoking ashes lost in the water. Therefore, we concluded that comparing the weights before and after smoking a joint would need to use a different method of putting out the joint.