Simulating latrine conditions to assess perfume performance against malodour

Abstract To evaluate perfume performance in toilets, we built model toilets in which critical factors such as background malodour, climate, and airflow were controlled. The models were equipped with an odour generator that injected hydrogen sulfide, methyl mercaptan, butyric acid, para‐cresol, and indole, allowing us to accurately and reliably reconstitute toilet malodour headspace. The malodorant concentrations matched the quantitative headspace analysis performed in African and Indian toilets. The toilet malodour headspace performance was validated by chemical and sensory analysis. Olfactory stimuli were presented to participants in different simulated climates to assess the effect of climate on the perception of odours. The sensory data show that increasing temperature and humidity decreased the intensity ratings of odours without altering their quality. Perfume can be delivered in these toilets by forced evaporation to control the headspace concentration, or by delivery systems such as cellulosic pads, liquids, and powders. Our experimental set‐up allowed us to establish dose–response curves to assess the performance of a perfume in reducing toilet malodour and increasing perceived pleasantness.

We initiated our research by identifying molecules that contribute the most to the offending malodour of toilets. Over the course of several field trips we analysed the sludge of toilets in Africa and India, and selected 19 molecules to prepare 40 reconstitutions of toilet malodour. 1 Using 'Sniffin' Sticks', pen-like odour dispensers, we validated the most representative reconstitutions in sensory surveys with local people in India and Africa. 2 We found that a mixture of only four compounds, butyric acid, p-cresol, indole, and dimethyl trisulfide, was necessary to evoke toilet malodour. 2 We used dimethyl trisulfide because it is a liquid at room temperature with an odour that approaches that of hydrogen sulfide and methyl mercaptan, two important components of toilet malodour. 3,4 We developed an analytical method to quantify hydrogen sulfide, methyl mercaptan, butyric acid, p-cresol, indole, and skatole in the headspace of latrines in Africa and India. 5 Knowing the gas-phase concentrations of the major contributors to toilet malodour allowed us to recreate realistic toilet malodour with synthetic compounds by using olfactometers in the laboratory. Consequently, we were able to test perfumes in a stable and precisely defined background of toilet malodour to find optimal perfume formulations and concentrations.
This approach was not enough, however, to evaluate the performance of perfumes to mask toilet malodour. To develop perfumes that are pleasant in the different climates of Africa and India, we need to understand how temperature and humidity modulates the perception of toilet malodour and perfumes. Temperature affects the evaporation rate of odorant molecules and the resulting gas-phase concentrations, a key factor that influences the perception of odours. Consequently, the same liquid mixture of odorant compounds may not have the same smell when presented in climates that differ in temperature. Although barometric pressure and humidity have been shown to influence odour detection thresholds, 6-8 lit-tle is known about the effect of temperature and humidity on perception itself, regardless of the gas-phase concentrations of the odorant molecules. More recently, researchers have reported that the integration of temperature signals is partially located in the olfactory systems of rodents and frogs, 9,10 suggesting that there are interactions between temperature and olfactory systems.
This article describes the development and validation, by chemical and sensory analysis, of an experimental toilet model system that allowed us to evaluate the perception of odours delivered in a controlled manner, in different climates, by varying the temperature and humidity. Inside a climate chamber, we built three model latrines where perfumes can be delivered by delivery systems such as liquids, powders, gels, or cellulosic pads. We also equipped our models with a forced evaporation system inspired by the olfactometer developed by Vuilleumier et al, 11 allowing us to deliver known gas-phase concentrations of odorant molecules. The toilet malodour analysed in urine-diverting toilets (UDTs) in Mukuru (Nairobi) was reconstituted, 5 and we evaluated the performance of a perfume in covering this malodour in four different climates in sensory panels.

| Chemicals
The compounds triethylamine, N-ethylmaleimide (NEM), and methyl octanoate were purchased from Sigma-Aldrich (Buchs, Switzerland), and butyric acid, p-cresol, indole, and L-cysteine were in-house products. The solvents propylene glycol, diethyl ether, methanol, ethyl acetate, and acetone were purchased from Carlo Erba (Val de Reuil, France). For methyl mercaptan and hydrogen sulfide, we used nitrogen mixtures at 15 ppm (v/v) in pressurized cylinders purchased from Carbagas (Carouge, Switzerland). Solid-phase extraction Oasis HLB 1-g cartridges were purchased from Waters (Montreux-Chailly, Switzerland). The perfume was an in-house product with floral tonality, and was a mixture of odorant volatile organic compounds.  Figure 1). The air from the climate chamber entered each latrine via the odour generator placed at a height of 45 cm in the back wall (the odour generator is described in detail below). The air was sucked from the roof of each latrine through a double-sided 82 cm 9 91 cm laminar filter (thick cotton fabric) via a 100-mm diameter aluminium exhaust tube ( Figure 1). The three exhaust tubes were connected to an adjustable fan via a main 100-mm diameter stainless steel pipe. The airflow of each latrine could be adjusted by changing the suction force produced by the fan if necessary, and could be separately adjusted with a damper (SPI 160; Systemair, Skinnskatteberg, Sweden) placed in the exhaust pipe ( Figure 1). A hot-wire anemometer was placed in the main exhaust pipe to measure the main exhaust flow. The airflow inside this pipe was maintained at 51 m 3 /h, distributing 17 m 3 /h in each latrine. The airflow and resulting air changes per hour (roughly 10) were in the range of measurements made in a ventilated improved pit latrine. 12

| Odour generator
To force the evaporation of liquids we modified the lower chamber of the olfactometer, as described by Chappuis et al. 5

| Climate chamber
The climate chamber dimensions were 3. Moreover, we placed a probe (Traceable â hygrometer; VWR International, Radnor, PA, USA) inside the latrine to punctually measure the RH and temperature to ensure that the differences in temperature and RH between the air inside the climate chamber and the air inside the latrines were minimal. We aimed to maintain the temperature difference below 1.5°C and the RH difference below 5%.

| Participants
The participants for the sensory panels were employees from the research centre at Firmenich SA (Geneva, Switzerland).

| Stimuli
The participants were exposed to six odorant mixtures delivered in

| Sensory protocol
The participants were randomly exposed to the odour stimuli in the different climates. As only three latrines were available, the six odours were split into two groups, each containing the malodour alone or the perfume alone, the malodour plus a low dose of perfume, and the malodour plus a high dose of perfume. The first and the last sessions were used as controls to assess the reliability of the panel, and were composed of the malodour alone, the perfume alone, and a mixture of both. The participants entered the climate chamber and directly evaluated the odour of the three latrines by answering a paper questionnaire made with the software FIZZ (Biosystems, Courtenon, France). The questionnaire is available in the Figures S1 and S2. They re-evaluated the odour of each latrine after a 3-min adaptation to the climate. They were asked to rate, on linear scales of 0-10, the pleasantness from 'I don't like' to 'I like', the familiarity from 'not familiar' to 'very familiar', the intensity from 'no odour' to 'very strong', the faecal/toilet character from 'not faecal/toilet' to 'very faecal/toilet', and whether they wanted to enter the latrine from 'not at all' to 'very willingly'.

| Headspace analysis
The Mukuru malodour was released in the model latrines as described above. The climate was set to 25°C at 50% RH. The

| Data analysis
The questionnaires were scanned and the data stored in FIZZ  When the subjects stepped into the climate chamber, they were exposed to the temperature and humidity set up for the experiment; therefore, for the present study, they were asked to make a first evaluation of the odour directly after entering the chamber, and to make a second evaluation after a few minutes of adaptation to the climate. Adaptation had no significant effect on the criteria used to evaluate the odour. The data with and without adaptation were then averaged.
Four descriptors were proposed to the subjects: pleasantness, willingness to enter the toilet, faecal character, and intensity. The panel was reliable, as the results obtained with the panel when we repeated the first and last sessions did not significantly differ (Figure 3). Considering that the panel is reliable and the olfactory stimuli were not delivered in the same order during the first and the last session, we did not find any bias resulting from a potential intrinsic odour of a particular model latrine. No odour was detected when syringe pumps stopped or when the gas was switched off. Therefore, to save panelists time, sniffing blanks were not performed.
The climate significantly affected the intensity, but had no significant effect on the other criteria of pleasantness, familiarity, faecal character, and willingness to enter the latrines. According to the statistical analysis, the increase in temperature significantly decreased the overall intensity (ANOVA, P < 0.0001; Kruskal-Wallis, P < 0.001; Figure 4A). Similarly, but to a lesser extent, an increase in humidity significantly decreased the intensity (ANOVA, P < 0.05; Kruskal-Wallis, P < 0.05), as shown in Figure 4B. The effect of the treatment groups on the willingness-to-enter ratings was much lower than that on the pleasantness ratings.

| DISCUSSION
We were able to validate our gas-phase reconstitution of the Mukuru increased compared with that in normal baric conditions. 6 The authors thought that the decrease in olfactory sensitivity could be explained by the hypobaric condition, which implies gas expansion that consequently reduces odorant concentration. In contrast, they found that humidity increased olfactory sensitivity by decreasing the olfactory detection threshold of single compounds. We showed that humidity slightly decreased the perceived intensity of odours, suggesting a decrease in olfactory sensitivity. However, we tested mix- stably is challenging because they are unstable gases. We overcame this problem by releasing mixtures of both compounds in nitrogen from pressurized cylinders. In addition to these channels for releasing gases, we equipped our model latrines with a forced evaporation system to deliver compounds that are liquid or solid at room temperature. Moreover, the odours were delivered in a continuous flow, ensuring the stable concentration of odours over time. This method greatly facilitated the sensory panels because the first panelist smelled a mixture that was equal to that smelled by the last panelist.
It also allowed us to reconstitute a realistic olfactory background of toilets, and to stably deliver known quantities of a perfume to evaluate its performance to reduce the malodour. We could show that the total concentration of this particular perfume should be above 1.6 lg/L to efficiently reduce our strong background malodour, and to obtain positive pleasantness ratings; however, we showed that Abbreviations: mo, malodour; perf, perfume splitting the olfactory stimuli into two treatment groups created a bias in the data. In fact, the perfume performed less well in group 2 than in group 1, although the gas-phase concentrations of perfume were higher in group 2 than in group 1. Both groups did not have the same olfactory stimuli. Group 1 consisted in the malodour alone, and to two mixtures of the malodour and the perfume. Group 2 consisted in to the perfume alone, and to two mixtures of the malodour and the perfume. Participants may have compared the odour of the three latrines and adjusted their ratings after the evaluation of the three latrines, resulting in lower ratings in group 1 and higher ratings in group 2, when the malodour alone was present. The malodour alone provoked a negative experience in participants, as shown by our data. This response could reduce the time spent analysing and the attention needed for a proper analysis of the malodour, as proposed by Herz, 18 and also by Ferdenzi et al. 19 We found that pleasantness ratings were positively correlated with the concentration of perfume in the presence of the malodour background. This could be explained by the fact that the faecal character was negatively correlated with the concentration of the perfume, and that the perfume alone was pleasant. Toilet malodour is generated by decomposing material, and is generally perceived as disgusting. [20][21][22] The odours of faeces, vomit, or decaying material are associated with potential microbial threats, and they elicit avoidance. 23 Thus, it was not surprising that the willingness to enter the model latrine was positively correlated with pleasantness. This result suggests that a pleasant odour should promote the use of clean and well-maintained toilets as, among other factors, odour plays an important role in the perception of the cleanliness of toilets. 24 We improved the analytical method described by Chappuis et al 5 to quantify hydrogen sulfide, methyl mercaptan, butyric acid, p-cresol, and indole in the gas phase. In the previous method, the air sample was pumped through a wash bottle filled with buffered water containing the derivatization agent NEM. The water trapped the target molecules, and sulphur compounds were derivatized with NEM.
The water was then loaded on Oasis HLB cartridges. In our study, instead of using water as a trap, we directly sampled the air through the Oasis cartridges previously loaded with NEM and triethylamine.
To our knowledge, this is the first time that solid-phase extraction cartridges such as Oasis HLB have been used to directly collect volatile organic compounds and sulphur compounds from the air. These cartridges are mainly used to extract compounds from liquid. [25][26][27][28][29] They contain a solid phase made of a co-polymer of N-vinylpyrrolidone and divinylbenzene, giving them interesting hydrophilic and lipophilic properties. The phase is wettable because of the N-vinylpyrrolidone, and divinylbenzene provides the reversed-phase retention of analytes. 30 Here, we showed that air samples can be pumped through Oasis HLB cartridges at a rate of 1 L/min, and that the sorbent was suitable to trap butyric acid, p-cresol, and indole, as well as being suitable to derivatize sulphur compounds with NEM and trimethylamine. This new method has the advantage of reducing the quantity of laboratory materials needed in the field, and was sensitive enough to quantify the target molecules at concentrations found in our model latrines. Using this analytical method, we demonstrated that the odour generators released the quantity of volatiles that we targeted with relatively high accuracy. In addition to filters distributing air velocities across the model latrines, the level of turbulence inside the odour generators and inside the model latrines was high enough to obtain homogenous gas-phase concentrations of volatile organic compounds. This level of homogeneity is not expected in the field, but it ensures reliable and stable results from laboratory experiments, which are crucial steps before conducting field experiments.
By controlling critical factors such as the background malodour, climate, and airflow, we developed a unique tool that accurately reproduces the toilet environment. This allows us to evaluate the performance of perfumes and various release systems (cellulosic pads, detergents, etc.) in stable and realistic conditions at the analytical and  is not limited to that of toilets. The climate chamber and the odour generator, when equipped with a forced evaporation and gas release system, can produce a great variety of background odours and climates.