SOP 15. Odor Evaluation and Intensometry

 

Determination of Odor Concentration

 

Odor concentrations will be determined using a dynamic dilution venturi olfactometer (AC’SCENT International, St. Croix Sensory, Inc., St. Paul, MN). The olfactometer uses precise airflow rates of odorous air and clean air to control the dilution of the odorous air. The olfactometer is capable of creating 14 dilution ratios between 8 and 65,500. The forced-choice ascending concentration series method will be employed. Very dilute concentrations are used with the concentration increasing until the odor is perceived by 50% of the odor panelists. This is considered the odor threshold value. The panelists smell three different air streams, two consisting of clean air and one consisting of the odor. The panelists are required to indicate electronically which stream is odorous. They must guess if no odor is perceived. They indicate if they identified or guessed the odorous stream as well. Only a correct response that was perceived and not guessed is considered a “true” response.

 

The strength of an odor sample is measured by determining the number of dilutions required to reach the odor detection threshold (ODT). The dilutions to threshold (DT) is defined as the dilution of an odor sample that cannot be distinguished from odorless air by 50% of the members of an odor panel. Stated another way, DT is the number of dilutions with odor-free air required for an odor to be just detected by 50% of the odor panel or until the least definitely perceptible odor is achieved. Stated still another way (Norén, 1977), DT is given as the dilution factor, expressed as a logarithm, that is needed to dilute the polluted air with fresh air so that the threshold value is reached. Dilutions to threshold (DT) is assessed by diluting samples with a known amount of odor-free air and presenting the dilutions to a panel of human subjects using an olfactometer, a dilution apparatus. Thus, the mixture at the odor detection threshold is a barely detectable odor. The number of times a sample has to be diluted to reach the odor detection threshold is a measure of the odor strength. The DT increases with greater odor strength because more odor-free air is needed to dilute the sample to its ODT.

 

DTs will be measured with a dynamic dilution forced-choice olfactometer (a dilution apparatus). This olfactometer (AC’SCENT International Olfactometer, St. Croix Sensory, Stillwater, MN) meets the olfactometry standards in the United States (ASTM, 1991) and Europe (CEN, 1999). The odor panel consists of eight people that have been screened to determine their odor sensing ability (ASTM, 1986). The olfactometer continuously delivers a precise mixture of sample and dilution air to a panelist through a Teflon-coated presentation mask. The dilution ratio of the mixture is the ratio of total diluted sample flow volume to the odor sample flow volume. For example, a dilution ratio of 10,000 is achieved with 2 cc/min of sample flow and 20 L/min of total flow. Olfactometer airflow rates are calibrated prior to and after each odor evaluation session using a precision airflow calibration device (GILIBRATOR-2ä, Sensidyne, Clearwater, FL).

 

Starting with an extremely high dilution ratio, a step-by-step series of ascending concentrations (step factor = 2) is presented to each panelist. The olfactometer can create 14 dilution ratios, ranging from 23 to 216. A triangular forced-choice test is conducted whereby the panelist sniffs three sequential sample coded gas streams at each dilution step. One gas stream is randomly assigned to have the odor while the other two gas streams are odor-free. The three gas streams are directed one at a time to the mask.

 

The panelist selects which of the three presentations is “different” (even if no difference is perceived) and thus contains the odor (ASTM, 1991). The panelist declares by pressing a button whether the selection is a “guess” (no perceived difference), “detection” (selection is different from the other two), or “recognition” (selection smells like something). Initial samples are so dilute that they cannot be distinguished from odor-free air. Higher and higher odor concentrations (2-fold increases) are presented to the panelist until the sample is correctly detected and recognized.

 

An individual best-estimate DT estimate is calculated by taking the geometric mean of the last nondetectable dilution ratio and the first detectable dilution ratio. The panel DT is calculated as the geometric mean of the individual DTs and is sometimes reported as the log DT. Retrospective screening of each panelist threshold is applied to the panel DT (CEN, 1999) to remove outlying individual results (van Harreveld et al., 1999). All averages of odor concentrations and emission rates are reported as geometric means because they typically exhibit lognormal distributions (Verdoes and Ogink, 1997).

 

The sensitivity of different dynamic olfactometers to reference odorants has varied by a factor of 10 or more in interlaboratory comparison tests. Similar variance can occur within laboratories, thus citation of panel dilutions to thresholds for a reference odorant, e.g. n-butanol, is strongly recommended so that results can be compared and related to current standards (Smith and Dalton, 1999; Watts, 1999). Therefore, a reference odorant, n-butanol gas (certified at a concentration around 60 ppm) is normally included in each odor session and is evaluated like the other samples, and is used to trace and archive panel sensitivity by calculating its odor detection concentration (ODC). The ODC is the concentration of a chemical at its ODT. To improve repeatability, accuracy and transferability of results of olfactometry, the CEN TC264 draft standard (CEN, 1999) will require that the mean ODC for the last 20 samples of n-butanol to be somewhere between 20 and 80 ppb for each panelist and that the odor panel’s ODC for n-butanol be around 40 ppb (van Harreveld et al., 1999).

 

An odor unit (OU) is defined as the amount of odorant(s) in 1.0 m3 of odorous gas at the panel ODT (Thièle, 1986; VDI, 1986). Even though the DT is really a dimensionless number, it is expressed as OU/m3 in order to calculate odor emission rate, thus the panel DT is an abstract measure of odor concentration. The product of odor concentration and volumetric airflow (e.g. from ventilation exhausts of buildings, or across land) gives the rate of odor emission in OU/s. The odor emission rate can be regarded as the total odor load per unit of time leaving a particular process. Odor emission values are used in atmospheric dispersion models to calculate odor nuisance distances.

 

The gross odor emission rate (OU/s) from a livestock building is the product of the ventilation airflow rate (m3/s) and the odor concentration (OU/m3) in the exhaust air. Since incoming air may be odorous, the difference between ventilation inlet and outlet odor concentrations is used to calculate the net odor emission or only the odor generated in the room (Miller et al., 2001; Smith and Dalton, 1999; Watts, 1999). To allow comparison with other research results, odor and gas emission rates are normalized to the number and weight of animals by dividing the total emission rate by the number of animal units (AU=500 kg live weight) (Ni et al., 2000; Oldenburg, 1990; Wathes et al., 1997). Area-specific building emission rates are determined by dividing the total emission rate by the floor area (Groop Koerkamp et al., 1998).

 

When calculating statistical parameters for DTs, one has to take care that the data are processed in such a manner that the normal frequency distribution applies. The frequency distribution for DTs for odorants is log-normal. Therefore, when calculating statistical parameters, the decimal logarithms of the measured DTs shall be used. To obtain a value in non-logarithmic units, the outcome can be re-converted into its antilog.

 

Olfactometer Operation

 

The laboratory manager will operate the olfactometer using the procedure in Exhibit 6-1.

 

1.      Connect sample bag via a line filter (St. Croix #AC1999-35004).

2.      Press [Prime/Purge]

3.      Press [Start]

4.      Enter panelist number, press [Enter]

5.      Enter starting dilution level (1-14), press [Enter]

6.      Have a panelist evaluate the sample

7.      Enter the panelist’s response:

[1] Guess

[2] Detect

[3] Recognize

[4] Allow panelist to correct an error

8.      Repeat Steps 6-7 for each successive dilution level until the panelist finishes

9.      Press [Stop]

10.  Repeat Steps 4-9 until all the panelists finish

11.  Disconnect the sample

12.  Press the [Prime/Purge] button

13.  Repeat process (Steps 1 to 12) until all samples are evaluated.

14.  Calibrate airflows.

15.  Turn off the air system, wait 30 min, and turn off power.

 

The Data Sense files with the data will be saved often to prevent the loss of any data. Hard copies of data will be stored in the laboratory.

 

Panelists

 

Eight people will serve on the evaluation panel during each session. Every attempt will be made to select a heterogeneous panel in terms of gender, socioeconomic class, etc. The panelists must agree to abide by the following rules which mean they must:

 

1.      be free of colds or other physical conditions affecting the sense of smell,

2.      not smoke or use smokeless tobacco,

3.      not chew gum; eat; or drink anything other than water for at least one hour prior to odor panel work,

4.      not eat spicy foods prior to odor panel work,

5.      be "fragrance-free", not use perfume, cologne, aftershave, or hand lotion the day of the odor evaluation session. Also must not use scented shampoos, scented deodorants or wear clothes dried with scented laundry treatments on the day of the odor evaluation session.

6.      not consume alcohol for at least three hours prior to panel work,

7.      drink only water during odor evaluation sessions,

8.      not discuss their odor selections and answers with other panel members,

9.      keep the odor panel work confidential,

10.  not have been fasting,

11.  not be pregnant, and

12.  not be involved in substance abuse.

 

Determination of Odor Intensity

 

Odor intensity is the relative perceived psychological strength of an odor that is above its detection threshold and is independent of the knowledge of the odor concentration (McGinley and McGinley, 2000). For a single chemical odorant, odor intensity increases as a power function of its concentration. Intensity can only be used to describe an odor at suprathreshold concentrations or concentrations above its ODT.

 

Intensity can be assessed using either category or reference scaling. Because category scale numbers do not reference equivalent odorant concentrations and different category scales are used by different researchers, data cannot be compared between studies. Thus, it is preferred to use reference odorant concentrations as a referencing scale to improve reproducibility and to allow direct comparisons between research studies (Harssema, 1991).

 

Intensity using referencing scales is assessed by either dynamic or static scale methods (ASTM, 1999). The dynamic scale method utilizes a special olfactometer that presents a series of specific concentrations of a reference odorant (e.g. n-butanol) in a continuous flow of air to each panelist. The static scale method utilizes a set of bottles with increasing concentrations of a reference odorant in water. We currently use the static scale and an n-butanol reference scale.

 

The scales used with the dynamic and static scale methods are referred to as the Dynamic Odor Intensity Referencing Scale and the Static Odor Intensity Referencing Scale, respectively. The Dynamic Odor Intensity Referencing Scale is based on the ppm of n-butanol in air (BIA) whereas the Static Odor Intensity Referencing Scale is based on the ppm of n-butanol in water (BIW). The observed intensity values, either the scale number or the equivalent butanol concentration, are used along with other data in the interpretation of odor dispersion models.

 

Field odor inspectors, monitors, plant operators and citizens commonly use the static scale referencing method. The static scale method is used with five concentrations of n-butanol in water (table 1) to assure a geometric interval (3X progression) between each value. Word descriptors assigned to these concentrations are: no odor, very faint, faint, moderate, strong, and very strong. After familiarizing themselves with the various dilutions of n-butanol, the odor panel judges the intensity of a sample by objectively matching it to the intensities they sense from the n-butanol dilutions in water (ASTM, 1999). A small glass funnel is used to present the odorous mixture from the sample bag to the panelist while the bag is manually compressed. The intensity is reported in ppm equivalent to n-butanol in water (ASTM, 1999) or BIW.

 

Odor intensity grows as a power function of the stimulus odor (Stevens, 1960) and follows the equation:

 

BIW = kCn

 

where BIW is the odor intensity (equivalent ppm of n-butanol in water), C is the mass concentration of the odorant in ppm, and k and n are constants that are different for every odorant. The odor intensity can be transformed as follows:

 

log BIW = log k + n log C.

 

Table 1. Concentrations of n-butanol in water. BIW = 83.33x100.477 I = 83.33e1.098 I.

Reference scale #, I

n-butanol in water

n-butanol in air

Odor intensity categories

BIW (ppm)

log BIW

BIA (ppm)

Strength

Annoyance

0

0

0.00

0

No odor

Not annoying

1

250

2.40

24

Very faint

Not annoying

2

750

2.88

72

Faint

A little annoying

3

2250

3.35

216

Moderate

Annoying

4

6750

3.83

649

Strong

Very annoying

5

20250

4.31

1948

Very strong

Extremely annoying

 

The category estimation technique is sometimes used to measure intensity (Misselbrook et al., 1993). Panelists give their perception of intensity according to the following scale:

 

0                    No odor

1                    Very faint odor

2                    Faint odor

3                    Distinct odor

4                    Strong odor

5                    Very strong odor

6                    Extremely strong odor

 

Generally, data generated from category scales are not of equal geometric intervals (Cain and Moskowitz, 1974), interpretation of the scales varies between individuals, and panelists tend to distort the intervals during an odor evaluation session (Nicolai et al., 2000). Thus, geometric means obtained from category scales should be used with great caution.

 

Procedure at Purdue

 

Panelists assess intensity of a sample by objectively matching it to one of a series of n-butanol solutions in water contained in 120-mL wide-necked bottles. This procedure follows the reference scale method (ASTM, 1992). Reference scale solutions are prepared prior to each session using AR-grade n-butanol (Mallinckrodt Baker, Paris, Kentucky, USA), and odor-free de-ionized water, with concentrations ranging from 250 to 20,250 ppm in geometric series with a factor of 3.

 

Sample air is manually compressed, forcing the sample to flow through a Teflon PTFE tube to a glass funnel (6 cm face diameter). Panelists are instructed to sniff the unknown air sample from the glass funnel and then to sniff the reference solutions beginning with the weakest end of the scale, and to match the unknown to the scale, ignoring differences in odor quality. Panelists are trained to shake reference solutions gently before sniffing them and to recheck the unknown against the reference scale any number of times. Odor intensity is reported as concentration (ppm) of n-butanol in water (BIW). Intensity values are geometric means of panel members’ ratings.

 

Persistence

 

The relationship between dilutions to threshold and intensity is important and has several applications. It is used to verify odor dispersion models and to evaluate faint odors whose detection threshold is less than the measuring capability of an olfactometer. The Weber-Fechner Law is more appropriate for predicting intensity from concentration when using category scales whereas the Steven’s Law is more appropriate for ratio scaling or reference scaling (Wagenaar, 1975).

 

Sample Evaluation

 

Dilutions to Threshold: The procedure to measure dilutions to threshold is given in Exhibit 6-1.

 

Persistence: Persistence or the slope of the intensity vs. threshold curve will be made possible by intensity rankings by each panelist at one additional dilution above the odor detection.

 

Procedure to Determine Odor Threshold

 

Each panelist will follow these steps when analyzing a sample:

 

  1. Position the air stream selection dial
  2. Wait for the beginning signal of the panel leader. This is shown by a green light.
  3. Place mask over nose, but not touching face (upper lip touches the mask)
  4. Press sample activation button.
  5. Sniff first air stream from the mask.
  6. Rotate selection dial 1/3 turn
  7. Press sample activation button
  8. Sniff second air stream from the mask
  9. Rotate selection dial 1/3 turn
  10. Press sample activation button
  11. Sniff third air stream from the mask.
  12. Review any of the air streams if necessary.
  13. Select the air stream that contains odor by rotating the dial.
  14. Press one of the buttons:

[Guess]

[Detect]

[Recognize]

  1. Repeat 2-7 if panel leader asks you to evaluate the next level.