Journal of the Air Pollution Control Association

ISSN: 0002-2470 (Print) (Online) Journal homepage: http://www.tandfonline.com/loi/uawm16

Quantitative Odor Measurement John L. Mills , Robert T. Walsh , Karl D. Luedtke & Lewis K. Smith To cite this article: John L. Mills , Robert T. Walsh , Karl D. Luedtke & Lewis K. Smith (1963) Quantitative Odor Measurement, Journal of the Air Pollution Control Association, 13:10, 467-475, DOI: 10.1080/00022470.1963.10468207 To link to this article: http://dx.doi.org/10.1080/00022470.1963.10468207

Published online: 19 Mar 2012.

Submit your article to this journal

Article views: 634

View related articles

Citing articles: 1 View citing articles

Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=uawm16 Download by: [37.44.207.192]

Date: 27 January 2017, At: 13:19

QUANTITATIVE Odor MEASUREMENT JOHN L. MILLS, Principal Engineer, ROBERT T. WALSH, Senior Engineer, KARL D. LUEDTKE, Intermediate Engineer, and LEWIS K. SMITH, Air Pollution Engineer, Los Angeles County Air Pollution Control District, Los Angeles, California

I he subject of olfactometry has been studied by many persons intermittently over a period of time extending back at least as far at 1882, when Ramsey1 used the human olfactory response in measuring the rate at which an odorous vapor diffused from a saturated cotton pad through a long glass tube. From this relatively crude beginning, the science of odor measurement developed slowly up through the 1930's and more rapidly thereafter up to the 1950's. J. F. Mateson2 provides an excellent review and bibliography of the development of odor measurement devices and procedures up to 1955. The quantitative odor measurement method used by the Los Angeles County Air Pollution Control District grew out of the ASTM method,3 a dilution procedure which was developed by Fox and Gex.4 It must be emphasized that the dilution method is purely quantitative and does not differentiate pleasant from unpleasant odors. The dilution method provides odor concentration data based on a zero-standard of no detectable odors. The quantitative results are expressed in terms of concentration, "odor units" per cubic foot. Although the definition of the odor unit, according to the ASTAI, is "one cubic foot of air at the odor threshold," it would be more accurate and informative to define it as "the quantity of any odorous substance or of any given mixture of odorous substances which, when completely dispersed in one cubic foot of odor-free air, produces a median threshold odor detection response in humans." From this definition, it becomes readily apparent that every cubic foot of any odorous gas is capable of contaminating odor-free air in proportion to the odor concentration.

McCord and Witheridge5 report a number of theories of odor production, several systems of classifying odors, over 50 principles for predicting odor production by different classes of chemicals, and tables of odor thresholds. Discussion of these matters in detail is outside the scope of this paper. Significant conclusions which may logically be drawn from this wealth of information, however, include the following, which are of particular value when considering

Volume 1 3, No. 1 0

Prominent among the situations described in Items 4 and 5 are those producing the odors of putrefying or-

AIR

FLOW

N EVALUATION AREA WORK AREA ACTIVATED CARBON 6OO TO

UNIT

12OO CFM

P AIR INTAKE

BENCH

QQQQQQQQQ CHAIRS

That modern technology has not even today produced a truly precise method

October 1 9 6 3 /

(1) Almost all materials must be in the gaseous or vapor state to be detected by the human olfactory system. (2) Characteristic odors are produced by pure chemical substances and by various common mixtures of such substances. (8) Some odorous chemicals lend themselves to detection by chemical means as well as by the sense of smell. (4) In many cases, vaporous materials can still be detected by their odors at dilutions in air far beyond the limit of detection by any presently known chemical tests. (5) In some instances, odors consist of such variably proportioned mixtures of many substances that the use of individual tests would be completely out of the question from a point of view of time and effort.

Characteristics of Odors

Odor Measurement Problems

* Presented at the 56t.h Annual Meeting of APCA, Sheraton-Cadillac Hotel, June 9-13, 196.3, Detroit, Michigan.

quantitative odor measurements:

for measuring most odors attests to the fact that odors are not easily amenable to quantitative measurement. Certain inherent sources of error which exist are associated with the characteristics of human responses to odors, the intrinsic qualities of the odorous materials and the extremely low concentrations in which odorous compounds usually occur. Odor adsorption on surfaces can introduce significant errors during sampling and handling gases in which the odor concentrations are low.

ODOR-FREE

AIR

PLENUM AIR O CONDITIONER O ISO CFM

Fig.

1.

A typical design for an odor-free room or installation for the determination of odor intensities

by the ASTM Dilution Method (D 1 3 9 1 - 5 7 ) .

467

ganic matter. These odors are also those most likely to cause nuisance complaints. In these cases, the only practical method of measuring odor levels involves reliance on human responses.

dling, and diluting the samples of odor-bearing gases or air. These impediments have all been overcome or reduced to a remarkable degree.

Characteristics of Human Olfactory Responses to Odors

The ASTM procedure comprises a static test method. Although most measurements reported in this paper were made by the static method, a dynamic procedure was also developed. Both methods require the use of selected persons as odor panel members. Both methods also require an air-conditioned, odor-free room to make the odor determinations. Certain modifications to the ASTM procedure were developed to accommodate the method to local problems and to accelerate the testing procedures, at the same time maintaining or improving the reproducibility and reliability of the results. These modifications involve both the testing procedure and the determination of results.

That the human olfactory system is extremely variable in its responses to similar concentrations of different odors is attested to by several known facts. The first is the amount of work that was found necessary to measure minimal identifiable odors (M.I.O.'s) or odor identification thresholds for pure substances. The second characteristic of human olfaction that makes it unreliable is that its responses to odors are not necessarily linear with concentration.6 The third disturbing characteristic of human olfaction is that the M.I.O. for any one material will most likely be different for different persons, and it may vary for the same person from time to time.7 The fact that the human sense of smell is subject to relatively rapid "fatigue" or "odor adaptation" is a fourth factor in discounting the human olfactory system as a continuously reliable indicator of odor concentration.8 This characteristic limits the human subject to relatively short time intervals as a useful odor detector. Basis for Quantitative Measurement

When odor nuisance abatement must be judged, the human olfactory organ, despite its failings, is actually the ultimate arbiter as to whether or not effective odor control has been achieved. The most practical solution to the problem has proved to be odor measurement by dilution, using a panel of human odor detectors. In all such instances, odor measurement produces information of far greater value than that afforded by chemical tests, since the results are more closely related to the problem. In essence the method is simple. A sample of the odor-bearing air or gas is diluted with odor-free air to the threshold of detection by a panel of observers. In practice, the procedure encounters complicating factors. In addition to the inherent difficulties with the human sense of smell, the quantitative determination of odor level is plagued with technical difficulties in securing, hanMEDICAL

SYRINGE

(10

Personnel and Facility Requirements

The Odor Panel

Members of the odor panel should be selected carefully, as is pointed out in the ASTM procedure, which describes a suitable screening test. The selectees should be persons neither the most sensitive to odors nor the most insensitive of those screened. The choice of panel members should be limited to those with the most generally reliable olfactory perception. This choice cannot always be made in a single screening. In some instances persons will be quite average in their responses to threshold concentrations of some types of odors and either "blind" or hypersensitive to others. It has been found that consistent and reproducible results are obtained with a panel consisting of at least eight persons. Although a panel of six persons is adequate at times, eight is preferred, because the probabilities of inconclusive results (with the resultant necessity of rerunning the test) are thereby reduced. If possible, the panel members should be allowed to relax in the quiet and comfortable atmosphere of the odor evaluation room for at least 10 to 15 minutes before the test. This will insure that their olfactory senses are not fatigued or dulled by any extraneous GLASS

ml.)

1

/

_ ] / ^

HYPODERMIC

NEEDLE

(tl

^-""•""•"^

CAPILLARY

TUBE

(1 mm. O.D.)

^

GAUGE)

SERUM STOPPER

STOPPER SAMPLE (250

TUBE ml.)

Fig. 2. This equipment is used to sample wet, odorous gas streams without water condensation. Withdrawing the syringe plunger draws the sample into the tube, which has previously been filled with dry, odor-free air.

468

The Odor Evaluation Room

A typical plan for an odor evaluation room is shown in Fig. 1. A wall extends most of the way across to separate the odor-free evaluation area from the work area, in which the various dilutions of the odor somples are prepared. During warm weather, refrigerated air is introduced into the work area from the outside by means of a window-type refrigerating air conditioner with a fresh-air handling capacity of about 150 cfm. During mild weather or cold weather, any source of the same volume of fresh air at room temperature and moderate humidity will suffice. The fresh air inlet should be so located to preclude the possibility of introducing any extraneous odors into the work area from the outside. No concentrated odors nor odor-bearing materials or objects should be released inside or brought into the evaluation area except the odors to be evaluated. The odor room in general should be kept clean, quiet, and free of unnecessary odorous materials. It should not be equipped with drapes, curtains, rugs, or other odor-adsorptive furnishings; and it should contain the minimum amount of bare furniture. An activated carbon adsorption unit is installed to provide odor-free air in the evaluation area. (A commercially available unit containing about 80 lb of activated coconut shell carbon and equipped with a blower capable of delivering 600, 900, or 1200 cfm of filtered and odor-free air has proved to be satisfactory.) All of the air circulates around the open end of the partition and is recirculated through the carbon unit except for the 150 cfm which is exhausted and replaced with fresh air. Measurement Technique-Static Method

The ASTM method calls for taking the sample in a 100-ml Luer-type hypodermic syringe. Because it is difficult to transport a syringeful of odor-bearing gases any distance without disturbing the position of the plunger, the standard method has been modified by using 250-ml gas sampling tubes for taking samples. Sampling Techniques

1 1 I 1 1 11 1 1 1 r-

odors, of which they may not even be conscious. Panel members should not carry out odor evaluations for longer than 30 minutes at a time; and, if tests must be extended, panel members should rest before resuming such tests.

Representative samples are obtained by choosing the location of sampling according to standard air sampling techniques. Factors which may interfere or cause contamination include the use of odorous material (such as rubber tubing) where it may contact the samJournal of the Air Pollution Control Association

pie, the use of improperly cleaned glassware, condensation of odorous material from the vaporous sample on the inner walls of the sample tube, and gas-phase odor adsorption on the glass itself. The problems of condensation and adsorption of odorous material on the inner walls of the sampling apparatus are the most difficult to overcome or even to evaluate. If the gases being sampled are at ambient temperature, there is usually no visible condensation, but this fact provides no assurance that odor adsorption on the glass surfaces is not taking place. The effects of odor adsorption on the inner surfaces of the sampling probe and the sample tube can be minimized by pumping the squeeze bulb a sufficient number of times to make as close an approach as practicable in the sample tube to the conditions of temperature and humidity existing in the gas stream being sampled, thus allowing time for any surface adsorption to reach equilibrium conditions. When a gas stream to be sampled bears a high moisture content, i.e., when its dew point is significantly above 130°F, so much liquid will condense on the inner surfaces of the sample tube upon cooling that the probability of errors increases materially. Extremely wet gases introduce an additional possibility of error by air leakage into the tube through the valve seats after moisture condensation produces a vacuum in the sample tube. An example of such a situation is the problem of sampling the effluent gases from an animal matter rendering cooker, where the moisture content of the gases is in the neighborhood of 97%. For such situations, a second technique has been devised, in which the sample is diluted in the sample tube with sufficient dry, odor-free air at ambient conditions to prevent condensation. Dry, odor-free air is provided by drawing ambient air into the sample tube after passing through a cartridge containing a layer of activated carbon and a layer of moisture absorbent, such as Drierite. It is preferable that this be done immediately before taking the sample. The sample tube must be flushed with the dry, odor-free air in the same manner used for taking a sample by the first technique. With the sample tube thus prepared, the maximum size of sample which may be taken is 5 ml at ambient temperatures between 60 °F and 85 °F and 10 ml at ambient temperatures above 85 °F, providing dilutions of 50:1 or 25:1, respectively, in the sample tube. The assembly of equipment used for this second sampling technique is diagrammed in Fig. 2. In order to prevent moisture condensation in the probe and to insure an accurate sample volume, a pyrex glass capillary tube of slightly less than 1 mm outside diam is used as a October 1963 / Volume 13, No. 10

STEP 1

STEP 2

STEP 3

Fig. 3. Odor samples are transferred to the sample syringe by mercury displacement. Dilution is accomplished by withdrawing the plunger of the dilution syringe to the 100-ml mark after injecting part of the sample. Additional dilution is then made similarly in the panel member's syringe.

sample probe. The capillary tube is inserted through a new cork stopper, size "000," which is then fitted snugly into the end of the sample tube. The assembly of this apparatus is facilitated by the use of an 18-gauge hypodermic needle as a sleeve through which to slide the capillary tube into the cork. The sample is obtained by placing the free end of the capillary tube into the gas stream and withdrawing 5 ml or 10 ml of air from the sample tube with the syringe, thus drawing 5 ml or 10 ml of the wet gas stream through the capillary tube into the sample tube. Care must be exercised not to permit steam to impinge on the outside of the sample tube nor to warm it above ambient temperature by handling. In using either of these two techniques, the stopcock nearest the squeeze bulb or the 10 ml syringe is closed first. The upstream stopcock is left open, and the end of the sample probe is kept in position in the gas stream until the temperature of the sample tube and its contents is virtually at equilibrium with the ambient air.

These procedures are followed for two reasons: (1) the stopcocks of the sample tubes cannot be relied upon to remain gas-tight against a high vacuum, and (2) for all practical purposes the samples thus taken can be considered to be representative of odor concentrations at standard conditions (14.7 psia and 70 °F). In using the second sampling technique, closing the upstream stopcock with the capillary tube still inserted (shearing it) aids in insuring an accurate dilution in the sample tube. Evaluation of Odor Samples

Mercury displacement is used to remove the odorous gases from the sample tubes. The odorous gases are measured and diluted in 100 ml glass syringes. Figure 3 shows schematically the equipment needed. Although the ASTM prescribes a method by which the tip of the syringe is inserted into one nostril, it has been found that panel members prefer to eject a small stream of the odorous gas in front of the nose and to sniff. It is recommended that each panel member 469

_0

CORK STOPPER

« • JET

EVALUATION AREA



1

i -GALLON GLASS SAMPLE

CONTAINER

Fig. 4. Diagram showing equipment assembly for continuous dilution of an odor sample for evaluation of odor concentration by the dynamic method.

be permitted to choose a method by which he himself feels his results are most accurate and reproducible. Each panel member records his reaction as "no odor" or "detectable odor." He then purges the syringe several times with odor-free air. In instances where 100% positive results are obtained from the panel or where the panel members indicate that the odor perceived was strong, there is a tendency to obtain false positive odor perception on the next sample or the next few samples presented, even though one or more of these samples might consist of pure, odor-free air. It has not been determined whether this phenomenon is a result of receptor fatigue or of odor adsorption on the inner surfaces of the panel member's syringe followed by subsequent desorption, or of both these possible causes. In such circumstances, the safest procedure is to allow the panel members to rest for a few minutes and to provide them with clean syringes. In situations where odor samples contain significant proportions of formaldehyde or other aldehydes, erratic panel responses are likely to occur. Aldehydes deaden the sense of smell very quickly. No entirely satisfactory method for overcoming this effect has been found. Measurement Technique-Dynamic Method

When odor concentrations are to be determined from a single source where, from a practical viewpoint, operations may be considered to be under equilibrium conditions, the static measurement method is quite satisfactory. Where operating conditions are highly variable, however, many more odor samples may be needed to learn the extent of these 470

variations. In these circumstances, the static method presents almost insurmountable problems in time demand, the major part of which is expended in cleaning and preparing the equipment. In order to make a large number of odor tests within the available panel time, the dynamic testing method was developed. The advantages gained by this method are as follows: (1) Except for the sample bottle, only a small part of the glassware is exposed to odor concentrations greater than just slightly above the odor threshold. It has been found that under such conditions, air-washing the equipment after a test with odor-free air will remove residual odor to a very high degree—great enough so that complete disassembly and washing are not needed except at infrequent intervals. {2) The dynamic method is adaptable to rapid changes in presenting dilutions of the odors to the panel. Thus, more evaluations can be made by a panel in a given period of time. (8) A larger volume of odorous gas is available to each panel member in making his observations. Panel members have expressed the opinion that this factor makes them more sure of their judgments. Odor fatigue can be thus reduced, since fewer and more diluted samples will provide adequate response data. Sampling Techniques

Larger sample volumes are required in the dynamic method. Whereas in the static method a 250 ml sample is taken, for the dynamic method it has been found most advantageous to obtain a one-half gallon sample and to dilute it 10:1 with dry, odor-free air at

the instant of sampling. This precludes any water condensation in the sampling container for most sampling conditions. A five-gallon bottle, previously cleaned and filled with dry, odor-free air, is partially evacuated until a manometer registers a pressure in the bottle equal to nine-tenths of the ambient air pressure. A short sample probe leading from the bottle is then inserted into the odorous gas source and a valve opened to collect the sample. Prior to filling the sample bottle with dry, odor-free air, a plastic envelope attached to a glass tube is inserted into the bottle and completely deflated, after which the glass tube is sealed with a stopper. After the sample is taken, it can be forced from the bottle at any desired rate with no further dilution, merely by inflating the plastic envelope. The plastic used is a laminated Mylarpolyethylene sheet, which is formed into a bag or envelope, with the Mylar on the outside, by heat-sealing the polyethylene to itself. Mylar is used because it has been found to be the most nearly completely odor-free and impermeable of the plastics available. Since Mylar cannot be heat-sealed to itself, a laminate is required. Polyethylene, which is easily heat-sealable, is unsuitable for use in contact with odorous gases because it is permeable to gases and, except for a few special varieties, has a slight but definite odor of its own. Equipment and Procedure

A diagram of the equipment needed for dynamic odor evaluation is shown in Fig. 4. The odor-free air blower supplies both dilution air and pressure to force the sample from the bottle. Any small blower should prove quite satisfactory for this purpose, since less than two cubic feet per minute of air is needed for the entire procedure. Ductwork is of glass, with welded, ground ball-andsocket or ground tapered joints throughout the system to prevent any odor contamination that might result if flexible tubing were used. The size of the duct for the diluted sample is 1 in., inside diam, and the individual jets are 4 mm, inside diam. The jets are for the use of the panel members. When a reading is called for, the panel member removes the cork and smells the odor-bearing gas as it comes from the jet, recording his reaction after replacing the cork. This procedure allows for several different dilutions of the same sample to be presented to the panel in a relatively short time. About one complete panel reading per minute can be obtained. The panel members prefer the dynamic method because they feel that the extra volume of gases available for smelling removes any doubt as to whether odors are detectable. The dynamic test method is still in Journal of the Air Pollution Control Association

Table I—Summary of Data from Typical Odor Tests

the development state, but sufficient data have been obtained to confirm its value. A disadvantage is the size and weight of the sample container. Where only a few samples need be measured, the static method may be preferable. Determining the Odor Concentration

Test No.

Dilution No.

014c

A B C A B C D E A B C D E A B C D

214J

The odor responses of the panel are quantified by calculating the percent of the odor panel members who detect any odor at all in every dilution of each sample presented. Table I shows several groups of typical data obtained during evaluations of odor samples by the static method.

023"

034' Odor Response Chart

The sensitivities of human responses to stimuli follow a cumulative normal distribution function. Since the range of odor concentrations encountered is extremely large, a practical approach to determining results by a graphic method led to the devising of a special logarithmic-probability graph paper, on which the odor panel response data tend to determine a straight line. Figure 5 shows the plots of the data from Table I on the special logarithmicprobability paper. Occasionally the negative slope of the line is greater than the examples shown, but only rarely is it flatter when using the static measurement method. The dynamic method appears to produce lines of lesser slopes. The ASTM procedure specifies that odor dilution be varied "until the difference between the greatest dilution at which odor is consistently perceived and the next greatest dilution measured is less than 50% of the greatest dilution at which it is consistently perceived." The ASTM also defines the odor threshold as the greatest dilution at which odor is consistently perceived, even though a single observer be used. By using the plotting method shown in Fig. 5, much of the repetition of given dilutions is avoided. Another advantage gained is that all fractional responses of the panel to any of the dilutions are useful in determining the threshold response to a dilution of an odor sample, unless one or more of the individual responses is actually a false positive or a false negative. Experience will soon show the conditions under which an individual response is suspect, e.g., when a person cannot detect an odor at a given dilution but detects it at a greater dilution. Figure 6 shows some comparative results from tests made on replicate samples of the off-gases from a directfired, rotary fish meal drier by both the static and the dynamic methods. Note that the slope of the line for the dynamic method is flatter. This phenomenon appears to be characteristic of dynamic tests, indicating that the true differences October 1963 / Volume 13, No. 10

Dilution Factor" 50 25 10 5,000 3,000 2,000 1,000 500 1,000,000 100,000 10,000 20,000 50,000 100,000 5,000,000 1,000,000 500,000

Number of Panel Members

Number of Panel Members Detecting Odor

Percent of Panel Members Detecting Odor6

6

1 3 5 0 1 3 4 G 0 0 5 4 1

17 50 83 0 14 43 57 86 0 0 83 67 17 88 13 38 75

G 6 7 7 7 7 7 6 6 G 6 6 8 8 8 8

7 1 3 6

a The dilution factor is the volume of the diluted sample evaluated by the panel members divided by the volume of original undiluted sample contained therein. b Zero and 100% responses are indeterminate. c Discharge gases from a jet-condenser and afterburner control system venting a renderingd cooker; afterburner temperature 12G0°F. Discharge gases from a direct gas-fired, rotary fish meal drier. e Discharge gases from a cooker used to render poultry offal. f Uncondensed discharge gases from a surface condenser venting a blood drier and a rendering cooker.

ODOR

RESPONSE CHART

10

Test No. 034 900,000

o.u/scf

10

10" Test No. 023 25,000

o.u./scf

10

Test No. 214 1,400 £

o.u./sct

io"

10 Test No. 014

, 25 o.u./icf

10 5

10

20

30

40

50

60

70

80

90

95

98

RESPONSES REPORTING POSITIVE OF PANEL Fig. 5. The odor concentration, odor units per standard cubic foot, is equal to the number of dilutions of the odorous gas sample required to obtain a threshold, odor response by 5 0 % of the panel members. PERCENT

471

5PONS»E CHiIVRT

ODO R 1

io 4

:

—J

3,000

o.u./4cf

¥ 3

^

OFF-GASES

—

FROM FISH MEAL DRIER

in

fALUA TED BY THE STAJIC METIHOD

io 2

SAMPLI:

No. i

SAMPL E

No. ;

1 10

1

20

30

40

50

60

70

80

90

95

98

10

1,400 o.u./scf

I, OFF-GASES

FROM

EVALUATED

BY THE DYNAMIC

5

10

PERCENT

20

OF

FISH

30

MEAL

40

PANEL

50

DRIER METHOD

60

REPORTING

70

SAMPLE

No. 3

•

SAMPLE

No. 4

•

80

POSITIVE

90

95

98

RESPONSES

Fig. 6. Comparison of odor response data from the static and dynamic methods shows that in the dynamic method, the slope of the line is less.

between the individual responses of panel members to odor stimuli may be less than the data obtained by the static method imply. The samples represented by Fig. 6 are not exact duplicates, although they were all taken in sequence at the same point and under the same operating conditions, as far as all observations indicate. The difference in results (3000 o.u./scf by the static method and 1400 o.u./scf by the dynamic method) may have been caused by unapparent minor changes in the actual odor concentration, by greater odor adsorption on the surfaces of the glassware in the case of the dynamic method or by both these factors. Since in the dynamic method diluted odor samples are in contact with proportionately far greater areas of glass, it is probable that adsorption is the most logical explanation of the differences in results. Very little is known about the absolute magnitude of adsorption, except that it increases with area. Since quantitative odor measurement 472

finds its greatest value in comparing odor samples taken before and after an odor control device, it is recommended that the same method be used in both cases. The effects of errors can thus be minimized. Reproducibility of Data

According to ASTM Standard Method D 1391-57, results that are reproducible within ±50% may be achieved with a single observer, greater accuracy with more observers. The fact that a panel never consists of a single observer should insure reproducibility well within ±50%. This statement is made with the presumption that a plot of the data taken is a reasonably straight line. If data are badly scattered, tests must be repeated until a straight line can be drawn with reasonable assurance. A test of the accuracy of the static method was made using a panel of 13 members. The odor concentration was determined for a given sample and found to be 400 odor units/cu ft. The response data were then replotted, in one trial

leaving off the data reported by the three least sensitive panel members and in another trial leaving off the data reported by the three most sensitive panel members. The threshold odor response determined in the first trial was found to be 500 odor units/cu ft and in the second trial 300 odor units/cu ft. Thus, a deliberate attempt to disturb the results failed to cause a variation of more than 25% from the median. In other instances, data obtained by presenting the same odor sample to two different balanced panels produced threshold odor responses which agreed well within ±50%. To minimize odor fatigue and to retain reproducibility, a panel should not be required to evaluate a large number of samples in succession. Presentation of the more dilute concentrations of an odor first will also tend to prevent fatigue. Practical Use of Quantitative Data Quantitative odor data have proved to be highly useful in evaluating and eliminating odor nuisances. While the odor concentration appears to be the prime factor, it is not the sole criterion in evaluating an odor nuisance. The length of time over which an odor can be detected is probably of second importance. It has been found, for example, that the odors of roasting coffee are quite pleasant to many persons at first exposure but that the continuous presence of these odors evokes loud complaints. The factor of third rank in importance is the odor emission rate at the point of discharge. The odor emission rate is the product of the odor concentration and the volumetric flowrate and is expressed in odor units per minute. In comparing two sources emitting similar odors at the same concentration, the area or the downwind distance over which people will be adversely affected is proportional to the odor emission rate in each case, since dispersion of a large quantity of odor requires more odor-free air to dilute it below the threshold of detection. , Because wind velocity and turbulence are also involved in odor coverage over any given area, the quantitative figures of odor concentration and odor emission rate can but rarely be used to compare one situation with another. Intermittent changes in concentrations of odors are sometimes more complaintevoking than continuous discharges of odors at a uniform concentration. The factor of fourth importance is the quality of the odor. It has been our experience that regardless of the quality of an odor, if it is strong enough and persists long enough, complaints will result. It is recognized, of course, that odors generally considered objectionable will normally evoke complaints sooner than will odors of other types. Journal of the Air Pollution Control Association

TABLE II—Odor Emissions from Typical Industrial Equipment and Odor Control Devices ODOR

LEVELS

EMISSION TYPE OF

AND ODOR

RATES,

LEVELS

AND

EMISSION

RATES,

CONTROLLED

UNCONTROLLED

EQUIPMENT OR OPERATION

RENDERING COOKER (Inedible Charge) Dry Batch Type

AFTERBURNER ODOR CONTROL DEVICES Vent Gas Odor Concentration Range

Modal Odor Emission Rate

o. u./scf a )

o. u./min b )

Direct-Fired (D-F) Afterburner

5,000

(Mode 50.000) 25,000,000

RENDERING COOKER (Edible Charge) Dry Batch Type Wet Batch Type Continuous Type FISH-MEAL DRIER Rotary Drum Type Direct Gas-Fired «**•», i

Temperature0 & . Efficiency^

10,000 l,000,000 g)

2,500^ 350 h) 650 t o . . 7,000 h > l ) 1,000 to 5,000" (Mode 2,000)

o.u./scf

100 to 150 (Mode 120)

90,000

1200°F. 99+%

Jet Condenser, followed by a D-F Afterburner

20 to 50 (Mode 25)

2,000

1200°F. 99+%

Surface Condenser, followed by a D-F Afterburner

50 to 100 (Mode 75)

6,000

1200°F. 99 + %

120,000h)

VARNISH COOKER Batch Type

10,000 t o . . 200,000 ;

700 to ,. 10,000" (Mode 3,000)

COFFEE ROASTER Batch Type

300 to . 30,000 e)

COFFEE ROASTER Continuous Type

500 to ., 1,500" (Mode 1,000) (Estimated) 1,000")

Jet (or Contact) Condenser

100,000 to 10,000,000f> (Mode 500,000) 2,000 to 20,000 (Mode 10,000

Temperature ' Efficiency

o.u./min.

12,000,000

80°F. Negative^

70,000

80°F. 80%

200 to 1,000 (Mode 400)

10,000,000

70°F. 80%

Chlorinationk) Plus Packed Column Scrubber

30 to 50 (Mode 40)

1,000,000

70°F. 98%

Recirculating Spray (Contaci) Scrubber

20,000")

Not Measured

Recirculating Spray (Contaci) Scrubber

100,000h)

Not Measured

Surface Condenser^

6,000

Not Measured

70,000 h)

Packed Column Type Scrubber

50,000.000

Not Measured

(Mode 25,000) 10,000,000

BREAD BAKING OVEN

Surface Condenser

a)

Odor Emission Rate

Not Measured

10,000 to Direct-Fired 70,000 eJ Afterburner (Mode 50.000) 30,000,0001' (Estimated)

AIR BLOWING OF LINSEED OIL

LITHOGRAPHING OVEN Metal Decorating

Vent Gas Odor Type of Odor Concentration Control Equipment

o. u. /min b)

•

AIR BLOWING OF FISH OILS

15,000,000

3,000.000 h) (Estimated)

3,000,0001)

Direct-Fired Afterburner Recirculating Spray Contact) Scrubber followed by a D-F Afterburner

25 to 75 (Mode 50) (Estimated)

3,000

50, 0001)

Not Measured

1200°F. 99 + %

1200°F. 97.5%

10 to 25 (Mode 20)

10,000

1200°F. 99+%

Direct-Fired Afterburner

100 to 400 (Mode 250)

100,000

1200°F. 99%

Direct-Fired Afterburner

50 to 500 (Mode 200)

1,200,000

1200°F. 95%

45o!j>

Catalytic Afterburner

3,000 h)

Direct-Fired Afterburner

150 to . 15,000 n)

Direct-Fired Afterburner

300 to 1,000 (Mode 350) (Estimated)

2,300,000 15,000,000 l,700,000 h) (Estimated)

1,200,0001)

1000°F. m ) 800°F. 1100°F. 50%

900°F. 65%

Not Measured

TALLOW HYDROLYZER ("Fat Splitter")

Not Measured

Not Measured

PHTHALIC ANHYDRIDE MANUFACTURING UNIT

1,800 to 4. 3.500J' (Mode 2,500)

15,000,000

a) b) c) d) e) f) g) h) i) j) k) 1) m) n) o)

Odor Emission Rate

to 500,000 e)

RENDERING COOKER (Blood Drying) Dry Batch Type

Vent Gas Odor Type of Odor Concentration Control Equipment o.u./scf a )

NON - AFTERBURNER ODOR CONTROL DEVICES

Surface Condenser"* 2,000,000 followed by a 2,000 Direct-Fired 750 Afterburner 150 70 Direct-Fired Afterburner

45 to 120 (Mode 75)

Catalytic Afterburner Catalytic Afterburner

Not Measured

940°F. 1100°F. 1200°F. 13OO°F. 1400°F.

500,000

1200°F. 97%

1,800

11,000,000

745°F. 27%

180

1,100,000

815°F. m ) 93%

Odor units per standard cubic foot (70°F. and 14.7 p. s. 1. a.). Odor units discharged per minute, based on average volumetric discharge rate and modal odor concentration. Temperature of gases after leaving flame-contact zone (afterburners); temperature of vent gases in other cases. Odor control efficiency, on a modal odor concentration basis. Odor concentrations in batch processes vary with materials charged and phase of operation. Surface condensers Increase odor concentrations in the vent gases but reduce total odor emission rates. One-hundredfold increase from beginning to end of cycle, One test only. " Samples collected from several points of odor emissions. In continuous processes, odor concentrations vary with temperatures maintained and materials charged. Chlorine (20 p.p.m.) mixed with drier off-gases, which are then scrubbed. More or less chlorine increases odor concentrations. Estimated from two tests only. Maximum temperature at which this catalytic unit can operate. Outlet odor concentration rises and falls with inlet odor concentration. The surface condenser is an integral part of the hydrolyzing unit. Note that low temperature incineration increases odor concentration above condenser vent level.

October 1963 / Volume 13, No. 10

473

For these reasons the District does not attempt to evaluate odor quality in making measurements. If an odor causes complaints, it may be a nuisance regardless of its quality. The general significance of the threshold response, as found, with reference to the expected response of the populace is, of course, dependent upon how well the panel represents the populace. If the odor panel is large enough and chosen with significant care, the results of the odor tests can be extrapolated with relative accuracy in estimating public reaction. Total emission rates measured for odors have been found to correlate positively with the number of complaints received about odor nuisances. Local Odor Problems in Los Angeles

A great number of the odor complaints received implicated two general areas— one including the meat packing plants and the rendering plants and the other the fish canneries and fish meal reduction plants. The only enforcement measure then available to the District was the nuisance statute.9 When an odor nuisance may originate in any one or more of many pieces of similar equipment located in a relatively restricted area, however, it is always difficult and sometimes impossible to pinpoint the exact source of any particularly bad odor about which a specific complaint may be received. In such circumstances, proving specific equipment to be the source of the nuisance may be extremely time-consuming and is frequently impractical to attempt. In 1959, as a result of a period of particularly unfavorable weather conditions, a series of complaints brought about the enactment of a new odor control ordinance, Rule 64.10 This rule requires that all off-gases from heated reduction of inedible animal matter (rendering, drying, dehydrating, digesting, cooking, or evaporating of animal matter unfit for human consumption) be incinerated at 1200 °F for at least 0.3 sec or be processed in a manner equally or more effective for the purpose of odor control. The standards were chosen because meeting them results in destruction of the odor rather than mere reduction of its effect. Condensing out the moisture from off-gases before incineration is not only permitted but encouraged wherever the moisture content is appreciable. ; The effectiveness of odor control is improved by this method; and, in most cases, the cost of the condenser is soon regained by lower incineration fuel costs. Odor problems from stored raw materials and from poor housekeeping are of an entirely different degree of magnitude, since the odor emission rates and odor concentrations change with 474

weather and wind conditions and time. In any event, these odor emissions are of far lesser significance than those from animal matter reduction equipment and must be considered separately when requiring odor abatement. Accomplishments of Rule 64

In order to make enforcement of Rule 64 equitable, it has been necessary to employ the quantitative odor-measurement techniques. By the use of these techniques, the District has essentially eliminated odor complaints and has reduced odor emissions to a small fraction of their former concentrations and quantities. By making measurements both at the inlets and at the outlets of odor control devices, it has been determined that proper incineration is capable of reducing odor discharges by more than 99.99% in some instances and only rarely by less than 99%. In contrast to these results, the effectiveness of odor control by timehonored water-spray scrubbing techniques *has been found to be quite variable—from less than 50% in some installations to about 90-95% at best. It has also been determined that the combination of carefully metered chlorination followed by intensive water scrubbing can produce from approximately 95% to almost 99% odor reduction in the vent-gases from a rotary fish meal drier. Some of the odor levels measured in vent-gases from rendering cookers were found to be so high that a 99% reduction in intensity still permitted the discharge of highly objectionable odors to the air in intolerable concentrations. In these extreme cases, 99.99% odor removal was needed to result in acceptable odor levels in the off-gases. In most instances, however, a 99% odor-intensity reduction produces an acceptable odor level for venting to the atmosphere. Regardless of the odor concentration at the inlet, it appears that the peak permissible odor level in the vented gases from any kind of odor control device should be about 150 odor units/cu ft, with the average preferably 50 odor units or less/cu ft. The enforcement of Rule 64 over the past 3 years has resulted in an effective clean-up of odor nuisances from the animal matter reduction industries in the packing house and rendering plant area. All rendering equipment has been controlled by the use of afterburners or the venting of the noncondensable odorbearing gases to boiler fireboxes. The Rule 64 control of heated reduction equipment in Los Angeles packing houses and scavenger rendering plants today reduces odorous air contaminants by more than 99.9%. All but one of the rotary fish meal driers are still operating under variance from Rule 64. The gas volume from a fish meal drier (over 20,000 cfm) is so

great and the moisture content is relatively so low (varying from about 16-25%) that condensation is technically infeasible and incineration economically unattractive. Even if the cost of incineration were lower, there is not enough natural gas available in the area at present to make incineration feasible. Oil-fired incineration would cost even more. One fish meal drier has been equipped with a full-scale chlorinator-scrubber. This equipment has been successful in controlling odor intensities in the ventgases to less than 50 odor units/cu ft provided that no high temperature combustion gases or flames come in contact with the fish meal and provided that the temperature of the off-gases from the drier can be kept low enough (below 190 °F while drying cooked tuna scrap and below 215 °F while drying scrap from mackerel and sardines). The rates of chlorination and of scrubbing must be controlled within close limits. Chlorine is added to the offgases at the rate of 15-25 ppm, followed by scrubbing with sea water in a packed tower at a rate of not less than 38 gal/1000 cfm of vent gases. Now that this first full-scale control device for a fish meal drier has been approved, the remaining driers will be equipped with comparable units as soon as construction can be completed. Typical Industrial Odor Emission Data

The range of odor concentrations found in the off-gases from industrial operations is large (from a few hundred up to millions of odor units per cubic foot); and the odor emission rates are quite variable, depending upon local conditions. Table II shows typical odor concentrations and odor emission rates for the off-gases from various types of industrial equipment and odor control devices. These data show ranges of concentrations, since typical emissions may vary from plant to plant, from batch to batch and from time to time for the same batch or the same process. It may be noted that, in the case of rendering operations, water spray or jet (contact) condensers in themselves remove a fairly high percentage of odors. The surface condensers do not remove as much of the odors, since there is less water available to dissolve the odorous substances. (It is assumed that the jet condenser tail-water is never permitted to be discharged at a temperature exceeding 140 °F, to prevent re-emission of odors.) Whenever a condenser is used ahead of an afterburner, odor control is better than with the afterburner alone. The concentration of odors in the offgases from an afterburner following a surface condenser is always higher than the corresponding figure for a system using a jet (contact) condenser and an Journal of the Air Pollution Control Association

afterburner, assuming both afterburners are operated at 1200°F. At higher incineration temperatures, the odor control efficiencies of both types improve slightly and equalize. It has been found that no complaints are received concerning the odors of combustion products from the burning of natural gas. Measurements made of the odor concentration in such combustion gases show an average of 25 odor units per standard cubic foot. A good odor control device will have an off-gas concentration of not more than 25 to 50 odor units per standard cubic foot. Summary

The most satisfactory device available for quantitative odor measurement is the human olfactory system. It is much more sensitive than chemical measurements in most cases where odors are nuisance problems, and results are obtained where chemical tests fail. Threshold response odor measurements produce quantitative results more quickly than other types of tests. The definition of the odor concentration as "the dilution factor at the odor perception threshold for 50% of the odor panel" is believed to be a more reproducible standard than that described in the ASTM method.

The odor unit should be redefined from "one cubic foot of air at the odor threshold" to "the quantity of any odorous substance or of any given mixture of odorous substances which, when completely dispersed in one cubic foot of odor-free air, produces a' median threshold odor detection response in humans." The modifications to the ASTM odor measurement method described herein are believed to improve the speed, reliability, and reproducibility of the method. Careful application of this method and the use of good laboratory techniques should produce results considerably more reproducible than ±50%. The application of the threshold response method of quantitative odor measurement is highly successful in determining the actual effectiveness of odor control measures and devices. Control of odors from the "offensive trades" can be achieved by proper incineration for 0.3 sec at 1200°F, but certain of the objectionable organic chemical industry odors require higher temperatures of incineration before adequate odor control is accomplished. Chlorination followed by intensive scrubbing has been found satisfactory under controlled conditions for the off-

gases from a direct-fired, rotary fish meal drier. Odor nuisances from animal matter reduction equipment used in the "offensive trades" can be prevented by reducing the odor concentrations in the offgases to a point where the average is not more than 25 to 50 odor units/cu ft and the absolute upper limit is 150 odor units/cu ft. REFERENCES

1. W. Ramsey, Nature 26: 187 (1882). 2. J. F. Mateson, "Olfactometry: Its Techniques and Apparatus," J. of the Air Poll. Control Assoc. 5: 167 (November 1955). 3. American Society for Testing Materials, "Standard Method for Measurement of Odor in Atmospheres (Dilution Method)," Designation D 1391-57. 4. E. A. Fox, and V. E. Gex, "Procedure for Measuring Odor Concentration in Air and Gases," J. of the Air Poll. Control Assoc. 7: 60, 61 (May 1957). 5. C. P. McCord, and W. N. Witheridge, ODORS Physiology and Control, McGraw-Hill Book Company, Inc., New York, 32-56 (1949). 6. Ibid., 46. 7. Ibid. 8. Ibid., 19. 9. Health and Safety Code, State of California, Section 24243. 10. Rule 64 of Rules and Regulations of the Air Pollution Control District, County of Los Angeles, California.

Air Pollution Control Association Membership Blank 1. Name (or company name) 2.

Mailing address

3.

Individual members: Company, organization, or agency with whom you are affiliated Company or Sustaining members: Industry affiliation or type of business (steel, petroleum, equipment, etc.),

4.

Class of Membership (please check) Individual • Company Member (Local) • Company Member (National) • Government Agency Member • Organization Member • Sustaining Member •

5.

Check enclosed herewith Send bill for dues

6.

Company or sustaining member's delegated voting representative is: Name Title Address

Dues $ 15.00 $ 50.00 $100.00 $ 50.00 (Minimum) $100.00 $250.00 (Minimum)

• D 7. Signed. Title Date

Here is an opportunity to help your Association get its membership campaign off the ground. Use this application blank for new individual, company or sustaining members. If each APCA member brings in just one new member, we'll be well over the 4000 mark. October 1963 / Volume 13, No. 10

475