Fugitive dust control for phosphate fertilizer


Material Information

Fugitive dust control for phosphate fertilizer
Physical Description:
xii, 164 leaves : ill., photos. ; 28 cm.
Rangaraj, Cumbum N., 1955-
Publication Date:


Subjects / Keywords:
Dust control   ( lcsh )
Phosphatic fertilizers   ( lcsh )
Phosphate industry -- Dust control   ( lcsh )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 1988.
Includes bibliographical references.
Statement of Responsibility:
by Cumbum N. Rangaraj.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001119941
notis - AFL6786
oclc - 19976832
sobekcm - AA00004899_00001
System ID:

Full Text







To Ranga for her patience, encouragement and support, to Dhanya for being
the sweet girl she is and to my parents for making this possible.


This research was supported by a grant (Grant Number FIPR 82-01-015)

from the Florida Institute of Phosphate Research (FIPR) and was monitored

by FIPR's Research Director (Chemical Processing), Mr. G. Michael Lloyd,

Jr. I would like to thank them both for their financial support during

my graduate work. I would also like to thank Mr. Floyd Taylor and Mr. E.

Harrison at Agrico Chemical Company and Mr. Harry F. Kannry of National

Wax Company.

I wish to thank the members of my supervisory committee for their

interest and suggestions. I am especially appreciative of Dr. Lundgren

for his encouragement and guidance. His personal interest and confidence

in my abilities have meant a great deal to me.

I would like to thank Mr. Robert W. Vanderpool for making my years

at the University so enjoyable. Finally, I would like to thank Ms. Dona

Ferrell for her invaluable help with the various aspects of preparing

this manuscript.




LIST OF TABLES. . . .. vi

LIST OF FIGURES . . . viii

ABSTRACT. . . . xi




Definition. . . 3
Standards . . 4
Fugitive Dust Emission Sources. . 5
Fugitive Dust Measurement Methods . 7
Dust Suppressants . . 11


Laboratory Tests. . . .. 13
Sample Preparation . . 13
Application of Dust Suppressants . 14
Measurement of Some Fertilizer Properties. .... 15
Emission Factor Measurement. . ... 17
Dust Size Distribution Measurement . 30

Intermediate Scale Field Tests. . ... 33
Apparatus and Operating Procedure. ... 33
Discussion . . 41

Full Scale Field Tests. . . 44
Apparatus and Operating Procedure. . 44
Discussion . . .. 50


Laboratory Tests. . . 52
Effect of Temperature on Test Samples. 52
Effectiveness of the Test Sample Preparation Method. 54
Granule and Dust Characteristics . 58
Product Treatments . 75

Intermediate Scale Field Tests. . 102

Full Scale Field Tests. . .. 102
Dust Suppressant and Coating Technique Evaluation. 102
Further Experiments Pertaining to FSFT Results 124

General Criteria for the Selection of Dust Suppressants 147


APPENDIX. . . . 157

REFERENCES. . . . 160



Table Page

1 Effect of Feed Tube Diameter on the Emission Factor of
GTSP Samples. . . 23

2 Effect of Enclosure Height on the Emission Factor of
Phosphate Rock and White Sand Samples . 25

3 Effect of Pour Time on the Emission Factor of
GTSP Samples . . 28

4 Effect of Heating on the Emission Factor of GTSP Samples. 55

5 Effectiveness of the Test Sample Preparation Method 56

6 Examples of Emission Factors for Various Products 59

7 Variation of Product Quality for GTSP Samples 60

8 Stability of the Moisture Content of Stored GTSP Samples. 63

9 Effect of Sample Size on the Measured Moisture Content of
Untreated GTSP Samples. . . 64

10 Effect of Sample Size on the Measured Moisture Content of
Treated GTSP Samples. . . 65

11 Granule Size Distribution of Samples of Various
Fertilizers . . 66

12 Size Distribution of Samples of Some Non-granular
Materials . . . 67

13 Effect of "drop tests" on Product Size Distribution 70

14 Effect of the Kinematic Viscosity of Oil Blends on the
Dust Release of GTSP Samples. . ... 78

15 Effect of the Kinematic Viscosity of Naphthenic Oils on
the Dust Release of GTSP Samples. . ... 80

16 Effect of the Aniline Point of Paraffinic Oils on the
Dust Release of GTSP Samples. . ... 81


17 Performance of Oil Blends as Dust Suppressants with
GTSP Samples . . 82

18 Performance of Oil Blends as Dust Suppressants with
DAP Samples . . 83

19 Qualitative Characteristics of Waxes. . ... 85

20 Physical Properties of Petrolatum and Slack Waxes 87

21 Performance of Petrolatum Waxes as Dust Suppressants at a
Nominal Application Rate of 1 kg/ton. . 88

22 Performance of Petrolatum and Slack Waxes as Dust
Suppressants at a Nominal Application Rate of 2 kg/ton. 89

23 Performance of Petrolatum Waxes as Dust Suppressants at a
Nominal Application Rate of 4 kg/ton. .. 90

24 Effect of Fertilizer Temperature on the Performance of
Petrolatum Waxes with GTSP Samples Preliminary Tests 93

25 Performance of Wax Emulsions as Dust Suppressants with
GTSP Samples. . . 96

26 Performance of Some Miscellaneous Dust Suppressants 97

27 Dust Concentrations within a GTSP Storage Building. 105

28 Summary of Full Scale Field Test Results with GTSP. 109

29 Emission Concentrations Measured after Transfer Point #1. 112

30 Variability of Dust Emissions from Uncoated GTSP Sampled
at Truck Discharge. . . 114

31 Variability of Dust Emissions from Oil Coated GTSP
Sampled from Belt after Transfer Point #1 . 116

32 Variability of Dust Emissions from Uncoated GTSP Sampled
Simultaneously at Truck Discharge and from Belt after
Transfer Point #1 . . 117

33 Effect of Fertilizer Temperature on the Performance of
Dust Suppressants with GTSP Samples Series #1. ... 129

34 Effect of Fertilizer Temperature on the Performance of
Dust Suppressants Series #2. . ... 133

35 Effect of Temperature on Thin Films of Petrolatum Waxes 141

36 Porosity of Fertilizer Granules . 146


Figure Page

1 Vertical Flow Dust Chamber . 18

2 Photographs of the Vertical Flow Dust Chamber
(a). The Enclosure
(b). The Test Setup . . 19

3 Calibration for the High Volume Air Sampler. 21

4 Effect of Air Flow Rate on the Measured Dust Emission
of GTSP Samples. . . .. 26

5 Effect of Pour Rate on the Measured Emission Factor of
Phosphate Rock Samples . . 29

6 Calibration for Two Configurations of the Vertical Flow
Dust Chamber . . 31

7 Schematic of a Single Stage Impactor . 32

8 Intermediate Scale Field Test Setup
(a). Schematic of the Material Handling System
(b). (i). Cross-section of the Conveyor
(ii). Fertilizer Feed Control Method. 35

9 Photograph of the Front View of the Intermediate Scale
Field Test Setup . . 37

10 Photograph of the Side View of the Intermediate Scale
Field Test Setup . . .. 38

11 Photograph of the Feed Hopper Discharge. ... 39

12 Dust Suppressant Spray System Used for the Intermediate
Scale Field Test Setup . . 40

13 Photographs of the Full Scale Field Test Facility
(a). Truck Unloading Station
(b). Transfer Point #2
(c). Transfer Point #3 . .. 45


14 Details of the Full Scale Field Test Facility
(a). Fertilizer Handling System
(b). Air Sampler Locations. . 47

15 Dust Suppressant Spray Setup for the Full Scale
Field Tests. . . 48

16 Weight Loss due to Heating of GTSP and DAP Samples as a
Function of Time . . 53

17 Deviation of the Emission Factor of Individual Samples
from the Average Emission Factor for that Batch. .. 57

18 Effect of the Moisture Content on the Emission Factor
of GTSP Samples . . 61

19 Hardness of Granules of Various Fertilizers. 69

20 Effect of Handling on the Size Distribution of
Prilled Sulfur . . .. 71

21 Effect of Handling on the Emission Factor of
Various Materials. . . 72

22 Photograph of Crystal Growth on MAP Granules 74

23 Size Distribution of the Dust Emitted by the Handling of
GTSP and DAP Samples . . 76

24 Size Distribution of the Dust Emitted by the Handling of
White Sand and Phosphate Rock. . ... 77

25 Effect of Handling on the Emission Factor for Coated and
Uncoated Samples of GTSP and Prilled Sulfur. 92

26 Effect of Handling on the Mass Fraction of Particles
Larger Than 13.6 Micrometers for Uncoated Fertilizer
Samples . . . 100

27 Relative Particle Release Characteristics of Oil and Wax-
Coated Fertilizers (I Initial, A Aged). 101

28 Performance of Petrolatum Waxes in Intermediate Scale
Field Tests. . . .. 103

29 Nozzle Arrangements at Transfer Point #1 . 108

30 Details of Mixing Technique
(a). Photograph of Mixer for Product on the Belt
(b). Photograph of Mixing Action. . 119

31 Variation of Fertilizer Temperature as Discharged from a
Number of Trucks . . 122


32 Temperature of Fertilizer Samples as a Function of Time
(a). Heat Loss of GTSP Samples Over a Period of Time
(b). Temperature of GTSP Samples Five Hours after
Collection in Five Gallon Buckets. ... 123

33 Effect of Laboratory Mixing Procedure on the Dust
Release of GTSP Samples with an Initial Petrolatum Wax
Distribution of 20 % (Application Rate = 2 kg/ton) 125

34 Effect of the Initial Distribution of Dust Suppressants
on the Dust Release of GTSP Samples (a). NW6364LA and
NW6889 (b). Pet HM (c). P4556 (d). AM303. 127

35 Effect of the Fertilizer Temperature on the Dust Release
of GTSP Samples at Two Application Rates
(a). 3.2 kg/ton (b). 2.0 kg/ton. . .. 136

36 Effect of the Fertilizer Temperature on the Dust Release
of GTSP Samples Coated with NW6889 after Five Hour and
Twenty Four Hour Heating Times . 137

37 Effect of the Fertilizer Temperature on the Dust Release
from Various Fertilizers Coated with Petrolatum Wax
after a Five Hour Heating Time
(a). NW6889 (b). NW6364LA. . .. 138

38 Response to Heating and Cooling for GAGTSP, GAMAP and
FDAP Samples . . 139

39 Photographs of GTSP Granules Showing Evidence of
Petrolatum Wax Absorption (a). NW6889 (b). NW6364LA. 143

40 Photographs of Fertilizer Granule Cross-sections
(a). GTSP (b). MAP . . 144

41 Elemental Spectral Analysis of GTSP Granules Coated with
NW6889 tagged with lead (a). Interior of Heated Granule
(b). Surface of Heated Granule (c). Surface of
Unheated Granule . . 148

42 Elemental Spectral Analysis of MAP Granules Coated with
NW6889 tagged with lead (a). Interior of Heated Granule
(b). Surface of Heated Granule (c). Surface of
Unheated Granule . . .. 149

Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy



Cumbum N. Rangaraj

April 1988

Chairman: Dale A. Lundgren, Ph.D.
Major Department: Environmental Engineering Sciences

A technique for the measurement of fugitive dust emission factors

was characterized and optimum operating parameters were developed. The

technique was based on a vertical flow dust chamber (VFDC) and high

volume air sampler (HVAS) combination.

Extensive tests showed that the technique produced very reproducible

results. The particle penetration characteristics of the VFDC were

determined by using monodisperse test aerosols. An upper penetration

limit of 100 um was determined and the 50 % cut size at 26 liters/sec was

40 um.

Laboratory tests were conducted to evaluate various dust

suppressants with application rates in the 1 kg/ton to 4 kg/ton range.

Oils, in general, were found to be product specific in their performance

and showed a tendency toward decreased performance with age. Petrolatum

waxes were found to be excellent dust suppressants, when used correctly,

with dust suppression effectiveness values better than 90 %.

The laboratory tests were scaled up to intermediate scale field

tests (ISFT) where 5 candidate petrolatum waxes were tested with granular

triple superphosphate (GTSP) at application rates between 1 kg/ton and 4

kg/ton but with fertilizer feed rates of up to 10 tons/hour. Results

obtained were similar to the laboratory test results.

Based on the smaller scale tests, 2 petrolatum waxes with melting

temperatures of about 520C were tested at a GTSP shipping facility where

the nominal process rate was 250 tons/hour. A number of parameters

including nozzle type and location were evaluated, but the best result

obtained was a dust suppression effectiveness of 70 %. It was determined

that the fertilizer temperature varied between about 540C and 770C.

Laboratory work showed that a combination of elevated fertilizer

temperature and time was accompanied by a loss in performance of the

soft, low melting waxes. A higher melting petrolatum wax showed improved

performance in laboratory tests. This loss in performance was correlated

with the porosity of the GTSP granules and the softening point of the

waxes and was shown to be due to absorption of the surface coating into

the granule interior. General criteria for the selection of appropriate

dust suppressants have been identified.


Fugitive dust emissions from granular phosphate fertilizer result

primarily from handling of the fertilizer during the various stages of

manufacture, transfer, storage, shipment and use. Excessive fugitive

dust emissions have a detrimental effect on the sale value of the product

and can be a major nuisance problem.

Fugitive dust emissions from granular phosphate fertilizer can be

caused by a number of factors including:

1. Loss of anti-caking agents due to poor

2. Incorrect granulation and screening of
granular fertilizer.

3. Loss of dust adhered to granule surface and
breakage of crystal growths due to impaction
and attrition.

4. Breakdown and fracture of granules during
material handling operations as at belt
conveyor transfer points or load out areas and
crushing of granules by material handling
equipment such as front end loaders in storage

Fugitive dust can be controlled after generation by conveying the

dust, if technically and economically feasible, to appropriate air

pollution control equipment. The release of fugitive dust can also be

prevented by using dust suppressants. This research is concerned with

the latter approach.


An extensive search of existing literature to determine information

pertaining to dust suppressants and emission factor measurement methods

was conducted. Experimental procedures are described and the various

granule characteristics, including size distribution, hardness and

moisture content, are discussed.

Laboratory tests were performed to study the performance of a range

of dust suppressants and the factors which influence them. Based on the

laboratory tests an intermediate scale field test (ISFT) setup was

designed and assembled so as to evaluate candidate dust suppressants when

used in larger quantities. Results were very comparable with those

observed in laboratory tests.

Two petrolatum waxes, YP2A and NW6364LA, both with melting

temperatures of about 520C were used in full scale field tests (FSFT) at

a GTSP shipping facility. The performance was not found to be as good as

expected from the smaller scale tests. Post field test experiments

conducted in the laboratory showed that a combination of factors

including, fertilizer temperature and porosity, wax melting temperature

and softening point and coating aging time caused absorption of the

surface coating into the granule interior thus leading to a decreased

performance level.

General criteria for the selection of appropriate dust suppressants

have been developed. Requirements for improved performance in field use

are discussed.


Dust emissions from handling granular phosphate fertilizer are a

major problem in the industry. Because of the diffuse nature of the dust

emission, accurate measurement and subsequent control are a major

problem. Background information relating to this problem is discussed in

this chapter.


Industrial emissions are regulated in order to maintain a certain

level of ambient air quality. However, only the ducted industrial

emissions have specific regulations and test methods. Other industrial

process emissions and natural emissions are grouped into a separate

category called fugitive emissions. These fugitive emissions are not

specifically regulated though they might have a significant effect on

ambient air quality. Fertilizer dust is usually considered a nuisance

particulate and when released in a workplace environment the published

Threshold Limit Value (TLV) is 10 mg/m3 (American Conference of

Governmental Industrial Hygienists, 1977).

There are many different definitions of the term "fugitive

emissions." "Fugitive dust" has been defined as particulate emissions

from wind and/or man's activity such as unpaved roads and agricultural

operations and "fugitive emissions" are defined as particulate matter

generated by industrial activities which escape to the atmosphere from

non-ducted sources (Jutze et al., 1977). Industrial process fugitive



particulate emissions can also be defined as particulate matter which

escapes from a defined process flow stream due to leakage, material

handling, inadequate operational control, lack of proper pollution

control technique, transfer and storage. Because these emissions are not

emitted from a stack, they cannot be measured easily by conventional

techniques and their impact on air quality is extremely difficult to



During the initial development of ambient air and industrial

emission standards, fugitive emissions were believed to be minor and

efforts were directed toward control of emissions which could be readily

quantified. With the installation of air pollution control devices on

ducted stationary sources and the discharge of these emissions at

elevations significantly above ground level, the effect of fugitive

emissions on ground level concentrations has become more significant.

The Air Quality Act was passed in 1967 and amended in 1970 and the

new law was referred to as the 1970 Clean Air Act Amendments. The

primary National Ambient Air Quality Standards (NAAQS) for Total

Suspended Particulates (TSP) were

75 ug/m3 annual geometric mean concentration
260 ug/m3 maximum 24 hour concentration not to be
exceeded more than once a year

The corresponding secondary standards were 60 ug/m3 and 150 ug/m3,

respectively, and were described in the Code of Federal Regulations

referred to as 40 CFR 50. The primary standards were aimed at the

protection of public health while the secondary standards defined levels

for the protection of public welfare.


The reference method for the determination of particulate matter

(TSP) was based on the use of a high volume air sampler in an enclosure

of standard dimensions and was also described in 40 CFR 50. Operational

parameters were clearly specified and the upper particle size limit was

stated to be 50 um. A number of studies have been conducted to evaluate

the collection characteristics of the air sampler (Wedding et al., 1977;

Lundgren and Paulus, 1975; Robson and Foster, 1962) and it has generally

been found that particles up to about 60 um were collected.

As of July 31, 1987, EPA promulgated a new standard based on

particulate matter with a carefully defined upper size limit of 10 um. A

new reference method was also proposed. This new standard specifies the

mass concentration of particulate matter less than 10 um (PM-10) and

sampled over a 24-hour period. The idea is to concentrate on that

portion of the total suspended particulate matter that is likely to be

deposited in the thoraic region of the human respiratory tract.

Because PM-10 is only a portion of TSP, the new standard is lower

than the old NAAQS for TSP. The annual average and 24-hour average

primary standards are 50 ug/m3 and 150 ug/m3, respectively. The

corresponding secondary standards are the same as the primary standards.

Depending on the size distribution of the fugitive dust emissions these

lower limits can make the extent of fugitive dust emissions more or less


Fugitive Dust Emission Sources

Fugitive dust emission sources are of both natural and anthropogenic

origin. Early work in the study of fugitive dust emissions was

stimulated by soil erosion problems due to wind. Important anthropogenic

sources, specifically industrial processes, include material transfer and


conveying, loading and unloading, storage piles and unpaved areas and

roads within industrial facilities.

Material transfer is usually accomplished by means of belt, screw or

pneumatic conveyors. A series of conveyors is usually used and the

transfer points are the major sources of dust emissions. Emission rates

for bulk materials are highly variable and often not known (Jutze et al.,

1977). As a result, the effectiveness of control techniques is not

quantitatively determined with any great degree of reliability.

Loading and unloading of bulk material from and to storage are other

sources of dust emissions. Mechanical agitation, dissipation of kinetic

energy on impact and turbulence all lead to generation of dust. Emission

factors vary with product type, moisture content and various process

parameters. Some quantitative data is available but is of questionable

reliability (Jutze et al., 1977).

Large tonnages of bulk materials are often stored in open or

partially enclosed storage piles and storage may be for a short time with

high turnover or for a long time to meet cyclical demand. Storage pile

operations leading to dust emissions include loading onto piles,

vehicular traffic, wind erosion and loadout from piles. The relative

importance of each of these operations depends on factors like storage

pile activity, pile configuration, method of loading and unloading, wind

speed and precipitation. Emission factors (U.S. Environmental Protection

Agency, 1976) and various equations (Jutze et al., 1977; Midwest Research

Institute, 1977; Carnes and Drehmel, 1981) have been developed, but they

are of limited value for general use.

Roads on plant property can be another major source. Vehicular

traffic causes increased mechanical breakdown of material and suspends


particulate matter in the air. The emission factor for roads has been

found to be a function of silt content, vehicle speed and weight and a

number of equations have been developed (Jutze et al., 1977; Midwest

Research Institute, 1977; PEDCO Environmental, Inc., 1976).

Fugitive Dust Measurement Methods

As discussed earlier, reference methods are available to quantify

emissions of particulate matter from ducted sources and so reliable

emission factor data can be developed for such situations. However, no

such single technique exists for the measurement of fugitive dust

emissions. Existing methods can be divided into field scale and

laboratory methods. The field scale methods were aimed at developing

emission factors on the basis of large-scale tests of full scale material

handling operations.

The six most widely used field scale methods are

1. Upwind/Downwind sampling
2. Roof Monitor sampling
3. Quasi-stack sampling
4. Exposure profiling
5. Wind tunnel method
6. Tracer method

Upwind/Downwind sampling (Kolnsberg, 1976) involves the measurement

of particulate matter concentration in the atmosphere upwind and downwind

of the source. Meteorological parameters are also simultaneously

measured. Based on the concentration map obtained and the values of the

meteorological parameters, Gaussian dispersion equations are used to

back-calculate the source emission rate.

Roof monitor sampling (Kenson and Bartlett, 1976) involves sampling

at building openings and has been used with indoor sources. Emission

rates are calculated based on the measured concentration and the exhaust


flow rate through the opening. No meteorological data is needed. Quasi-

stack sampling (Kolnsberg et al., 1976) requires temporarily enclosing

the source and drawing off the emissions through ductwork and measuring

particulate matter concentrations using standard stack sampling methods.

Exposure profiling (Cowherd et al., 1974) is a multi-point sampling

technique where particulate matter concentrations downwind of the source

are isokinetically determined across the plume cross-section. Emission

rate is then calculated by a mass balance approach. In the wind tunnel

method (Cuscino et al., 1983) dust generated by wind blowing over an

exposed surface is measured. A wind tunnel with an open-floored test

section is placed over the surface to be tested and air is drawn at

controlled velocities. Isokinetic samples are collected and used to

calculate dust concentrations. Finally, the tracer method (Hesketh and

Cross, 1983) consists of releasing a tracer at the dust source. Downwind

from the dust source both dust and tracer concentrations are determined

and based on this ratio and the quantity of tracer released the dust

emission rate is determined.

The field scale techniques described above were all developed and

applied to special situations and were often dependant on meteorology.

The techniques are all complicated, time consuming and expensive.

Because of the scale of the tests, the performance of dust suppression

techniques cannot be easily and quickly determined. In addition,

reproducibility is a major problem.

A number of smaller scale techniques for use in the laboratory have

also been developed. A dedusting tower (Hoffmeister, 1979) consisting of

a 8.6 cm diameter glass tube fitted with seven screen stages has been

used. Air is sampled such that air flow is countercurrent to a falling


250 ml sample at a velocity of 0.9 m/sec. Weight loss of the test sample

is used to calculate dust emission factor. Another laboratory scale

technique involves the use of a spouted bed arrangement (Kjohl, 1976)

where 1.2 liters of sample are used in the spouted bed and the dusty air

is sampled through a filter bag. Test conditions are such that particles

up to 200 um are sampled. An analogous technique is one where a

fluidized bed of 400 grams of material, 10 % test sample and 90 % sand,

is used and the dust generated is sampled with a cascade impactor

(Schofield et al., 1979). All these techniques are more representative

of pneumatic type conveying systems. The fluidized bed technique has

been compared with a rotary drum technique and an impact type test

(Higman et al., 1983). The impact type test involves dropping 300 grams

of material into a box and exhausting the box through a cascade impactor

(Wells and Alexander, 1978). All the above tests were more suited to

powders and reproducibilty has been stated to be 15 % to 20 %. The small

sample sizes lead to greater variabilities in dust measurement. In

addition, for moderately dusty materials, the small amount of dust

generated would require more accurate gravimetric analysis. None of the

above techniques really simulate dust generation at transfer points.

A semi-field scale technique where 50 kg of coal was discharged from

a hopper in three minutes through a series of belt conveyors onto a

stockpile (Nakai et al., 1986) is more directly based on an impact type

dust generation process, as at transfer points. Dust concentrations at a

transfer point were measured with an optical device and efforts were made

to correlate emission factors with ambient dust concentrations.

A number of methods based on some means of dropping a test sample in

an enclosed space have been developed. A technique called the powder


spill test column (Cooper and Horowitz, 1986) uses 10 gram samples which

are dropped a distance of 1 m inside a 17 cm diameter column and the air

is exhausted through a 47 mm filter at a flow rate of 52 liters/min. A

particle size limit of 40 um is stated. Another technique used to

evaluate spills and pressurized releases (Sutter et al., 1982; Sutter and

Halverson, 1984) was based on a chamber 2.9 m ih diameter and 3 m high

where small quantities of the sample were discharged and the air was

sampled with high volume air samplers. The ASTM method for determining

an index of dustiness of coal (American Society for Testing Materials,

1975) consists of a 1.5 m tall metal cabinet with a 0.46 m square cross-

section. A minimum of 23 kg of the sample is placed on a tray within the

cabinet and released at a 1.2 m height. After 5 seconds two slides are

inserted 0.6 m below the release point and pulled out 2 minutes and 10

minutes afterwards. The dust settled on the slides is gravimetrically

analyzed and reproducibility of 20 % is claimed. Another technique used

with coal uses a belt conveyor to discharge coal samples into a 0.46 m

diameter chamber of variable height (Cheng, 1973). The chamber is

evacuated with a high volume air sampler and dust is sampled with a

cascade impactor. A variation of the chamber technique called the Totman

dust test device uses a 0.9 m tall chamber of 0.15 m x 0.2 m cross-

section with a chevron type internal material flow arrangement. Because

of this arrangement, unlike other chamber techniques where only one

impact is used, at least 4 impacts occur before the product comes to

rest. The air is sampled in a counter-current manner through a filter

for gravimetric analysis. A review of these and other laboratory

techniques has been published elsewhere (Hammond et al., 1985).


Dust Suppressants

Coating agents have been applied to a very large number of materials

to suit many requirements which include moisture control, prevention of

caking, providing slow release capability and reducing dustiness. The

most commonly used dust suppressant is water. When coal moisture content

was raised from 0.8 % to 1.5 % and mixed briefly in a tumbler, the

emission factor was reduced 70 % (Cheng, 1973) though excessive mixing

created more dust due to breakage. This same effect has been reported

with different kinds of coal (Nakai et al., 1986)and has been reported to

cause agglomeration of coal dust. A number of studies have also

documented the increased adhesive forces between particles and surfaces

with increased relative humidity due to formation of liquid bridges

(Stone, 1930; Van Den Tempel, 1972; Larsen, 1958; Corn, 1961; Ketkar and

Keller, 1975; Corn and Stein, 1965). However, excessive moisture content

with phosphate fertilizer can cause caking problems (Hoffmeister, 1979;

Kjohl, 1976) and decrease granule crushing strength (Kjohl, 1976), thus

leading to increased dustiness due to granule fracture and subsequent

generation of fines.

The most common dust suppressant used in the fertilizer industry is

oil. Oils with high viscosities are suggested to avoid the problem of

absorption into granules and consequent loss of effectiveness

(Hoffmeister, 1979). Oils with high paraffinic content are also

suggested as effective dust suppressants for fertilizer (Frick, 1977).

Extensive work is reported in patent literature on the use of coating

agents to increase granule strength, reduce caking tendency, reduce

dustiness and control moisture content. A list of patents is presented

in the Appendix. Coating agents used have included amines, mineral oils,

surfactants, fillers, acids, waxes and many other materials. These

patents and some others are reviewed elsewhere (Sarbaev and Lavkovskaya,


In the laboratory, dust suppressants have been applied in a rotary

drum where the granules and coating agent are both introduced

(Hoffmeister, 1979). In actual industrial facilities coating agents used

are primarily petroleum oil blends and have been sprayed in screw

conveyors, mixers, on belt conveyors, coolers and material transfer

points and sufficient mixing occurs so as to effectively distribute the

coating agent (Achorn and Balay, 1974).


Extensive experimental work was carried out in order to establish

the nature and extent of the fugitive dust problem associated with

handling phosphate fertilizer. The apparatus and procedures used are

described in this chapter.

Laboratory Tests

Sample Preparation

A supply of uncoated granular phosphate fertilizer was a

prerequisite to any experimental work. Samples of fertilizer were

obtained in quantities of at least 100 kilograms and stored in 19 liter

(5 gallon) plastic buckets with tight fitting lids. The sample buckets

were kept air tight during transfer from the field to the laboratory.

Fertilizer sampling locations were chosen with care and included belt

conveyors, material transfer hoppers and storage piles.

A batch of uncoated fertilizer consisting of about 80 kilograms of

product was poured out of the buckets on to a clean plastic sheet laid

out on the floor. The pile of fertilizer was thoroughly mixed to ensure

that all parts of the pile were homogeneous. Five kilogram test samples

were then made by collecting 8 to 10 scoops of product from various parts

of the pile and stored in polythene bags to provide a stable environment

for the sample. This technique was also used when making test samples of

coated fertilizer.


Five kilograms was chosen as the standard test sample size. This

sample size was considered to be large enough to overcome the possible

variabilities in the fertilizer and more representative of the average

characteristics of the bulk material. This sample size was also the

maximum quantity that could be conveniently handled without spillage

during experiments. In addition, the larger the test sample size the

greater the amount of dust generated and, hence, the greater the accuracy

of gravimetric analysis of the emitted dust.

Application of Dust Suppresants

In actual plant situations the dust suppressant is usually applied

on the fertilizer when it moves past a spray header on a belt conveyor or

at a product transfer point. The dust suppressant is applied as a spray

produced either by a high pressure airless spray system or by a lower

pressure air atomized spray system.

The dust suppressants tested have included vegetable and petroleum

based oils, waxes, petrolatums, emulsions and many other materials. Dust

suppressants which were liquid at ambient temperatures, were dispersed

using an air atomized spray system at a pressure of about 138 kPa (20

psig). A Sears Model 919.156580 portable air compressor was used with a

Sears Model 919.156140 spray nozzle for this purpose. This system was

used because of its similarity to actual industrial practice, ease of use

and availability. This system was designed for use with dust

suppressants which did not require special handling and whose viscosities

at ambient temperature were such that they could be sprayed directly.

However, waxes, which are solid at ambient temperatures, were

sprayed either in the form of water based emulsions or melts. Some

natural waxes were easily emulsified by a process of saponification.


These waxes were tested in both forms, where possible. Emulsification of

petrolatum waxes required a more complicated process using special

emulsifiers and they were, therefore, sprayed only as melts.

The wax emulsions were sprayed without further treatment. The non-

emulsified waxes, on the other hand, were first melted by putting them in

a plastic container immersed in boiling water. Once heated to a

temperature of about 750C the liquid wax was sprayed using an air

atomized nozzle (Spraying System #SU-1) in a siphon arrangement. To

prevent plugging problems due to solidification of wax, the nozzle was

heated to an appropriately elevated temperature by using a heating tape

and variable transformer arrangement.

The test sample to be coated was retained in the storage bag for the

coating operation. The exposed surface layer was first sprayed lightly,

then a new layer was created by mixing the bag contents and this new

layer was sprayed. This process was carried out till the required amount

of dust suppressant was added. The quantity of dust suppressant added

was determined by weighing the test sample before and after application

of the dust suppressant by using a single pan balance with a 20 kg

capacity. Once the coating operation was complete the coated sample was

stored in the polythene bag pending the drop test.

Measurement of Some Fertilizer Properties

Moisture Content. For the purposes of characterization of various

batches of fertilizer, moisture content was determined for at least two

test samples per batch of fertilizer. The technique used was that

recommended by the Association of Florida Phosphate Chemists (Association

of Florida Phosphate Chemists, 1980).


Three 2-gram samples were taken from each test sample to be

evaluated and placed in a vacuum oven (Precision Model #19). The samples

were subjected to a temperature of 500C and a vacuum of 508 mm of mercury

for 2 hours with a stream of dry air being circulated in the oven. The

weight loss of each of the three samples was determined with an

electronic single pan balance (Mettler Model #HK60) and converted to a

"percent moisture content" representation. The average value for the

three samples was calculated and used as a measure of the moisture

content of the test sample.

Size Distribution. Size distribution of the granular fertilizer was

another parameter of interest. A sieving machine (Gilson Model #SS-15

Sieve Tester) with a set of 6 sieves was used. The sieves used were U.S.

Standard 6, 8, 12, 16, 20 and 40 mesh. One-hundred-gram samples were

weighed out using an electronic single pan balance (Sartorius Model

#2355) and then poured into the first sieve. The sieving machine was

operated for 10 minutes. At the end of the sieving cycle the size

fractionated sample was collected in preweighed petri dishes and re-

weighed. The weights of the various size fractions were then used to

calculate the size distribution.

Crushing Strength. Crushing strength of a granule is a measure of

the resistance to fracture. The technique used is also known as the TVA

method (Hoffmeister, 1979). Size fractionated samples were prepared with

the sieving machine as described earlier. For a particular size range a

number of granules were placed on a single pan spring balance with a

weighing range of 0 to 10 kilograms. A load was applied on individual

granules by pressing down on the granules with a steel rod. The scale

reading at the point of granule fracture was noted and the average value


for a number of granules was calculated. This procedure was carried out

for the various size fractions to establish the crushing strength


Emission Factor Measurement

Apparatus and Operating Procedure. Emission factors for coated and

uncoated fertilizer were measured by means of a "drop test" using a

vertical flow dust chamber (VFDC). The VFDC was an enclosure constructed

of 1.3 cm (1/2 inch) thick plywood (Figure 1). The enclosure was 0.6 m

(2 feet) square and 0.9 m (3 feet) high. The top of the enclosure had

two openings: a 0.2 m (8 inch) diameter opening into which a 0.6 m (2

foot) long duct was mounted and a 18 cm (7 inch) by 23 cm (9 inch)

rectangular opening over which a high volume air sampler (General Metal

Works Model #2000) was placed. A baffle separated the two openings in

terms of the air flow characteristics of the enclosure (Figure 2(a)).

The test sample was introduced manually through the 0.2 m diameter feed

tube and fell 1.5 m before striking the floor. Dust was released during

the pouring process and also when the sample struck the floor, due to the

combined action of impaction and attrition. The released dust was picked

up by the high volume air sampler and deposited on a filter for

gravimetric analysis.

About 10 % of the samples in a batch of fertilizer were tested in an

uncoated state to establish an emission factor in units of g/kg for

untreated fertilizer for that particular batch. The remaining samples

were treated with the dust suppressants to be evaluated and then tested.

The test sample was first preweighed to the nearest gram with a 20

kg capacity single pan balance and then transferred from the plastic

storage bag to a pouring bucket. Four 20 cm x 25 cm (8 inch x 10 inch)

Air Inlet

Product Input

Air Outlet (31 liters/sec)


Flow Rate Gauge

02m Dia I
Feed Tube
High Volume
Air Sampler

1.5m I Baffle

(0.6m x 0.6m x 0.9m)

Vertical Flow Dust Chamber.

Figure 1.





r -W


i- -

0 0






glass fiber filters were weighed using a single pan balance (Mettler

Model #H6) equipped with a special attachment for weighing filters. The

VFDC was placed on a plastic sheet spread out on the floor. The first

filter was mounted on the high volume air sampler which was then placed

over the enclosure opening as shown in Figure 2(b). The high volume air

sampler was previously calibrated by using a set of calibration orifices

to develop a correlation between air flow rate and sampler pressure drop

as measured by a magnahelic gage (Figure 3).

The air sampler was turned on and set to operate at a flow rate of

31 liters/sec, unless, specifically stated otherwise. The sampler flow

rate was adjusted with a variable transformer. After 15 seconds the test

sample was steadily poured into the enclosure through the feed tube in a

pouring time of 60 seconds. After an additional 45 seconds of operation

the air sampler was switched off. Thus, the total run time of the

sampler was 2 minutes and this was equivalent to a total of about 10 air

changes in the enclosure, 5 of which were during material transfer.

After the first drop of the test sample the "dirty" filter was

removed from the air sampler and the test sample, now on the plastic

sheet on the floor under the enclosure, was transferred back into the

pouring bucket. The above procedure was repeated 3 more times. The

weight gain of the 4 filters was determined and normalized to given an

emission factor in units of grams of dust per kilogram of test sample.

The average value of the emission factors calculated for the four filters

was determined and represented the average emission factor for the test


Discussion. The VFDC configuration and test procedure described

were established after an extensive evaluation of a number of parameters.

I 40 50 60


Figure 3. Calibration for the High Volume
Air Sampler.




These included baffles, feed tube diameter, enclosure height, air flow

rate and material pour rate.

The baffle in the VFDC was introduced in the basic design to better

define the air flow in the enclosure and to prevent possible "short

circuiting" of the air flows at the enclosure inlet and outlet. Tests

with granular triple superphosphate (GTSP) showed that the presence of

the baffle did have a small, but not negligible, effect on measured

emissions. The principal value of the baffle, however, was that it

permitted a clearer mathematical description of the air flow in the

enclosure. Dust emission factors for test samples were measured using

a "drop test" procedure as described earlier. Four drops were performed

per test sample in a "drop test" as a matter of practice. This was done

in order to obtain an average value for the dust emission factor. A

single drop would usually, but not always, represent a maximum emission

from the test sample and would not be representative of an average

emission resulting from a series of handling of that same test sample.

Four drops would thus permit a more representative estimate of dust

potential of a test sample, especially when comparing different


The effect of various feed tube diameters was also evaluated. As

shown in Table 1 the diameter of the feed tube affects the velocity of

the air at the inlet and so measured dust emission factors were higher

for the smaller diameter feed tube. However, both the 0.15 m and 0.25 m

diameter feed tubes were found to be not quite convenient for regular

operational use. Therefore, a 0.20 m diameter feed tube size was used as

standard. The 0.6 m length was chosen because this would make the

effective height of fertilizer discharge from the pouring bucket, 1.5 m


Effect of Feed Tube Diameter on the Emission Factor of GTSP Samples

Sample Drop Flow Rate Emission Factor Average Feed Tube Diameter
I.D. Number (liters/sec) (g/kg) (g/kg) (m)

1 32 0.0248
A 2 32 0.0234 0.0225 0.15
3 32 0.0211 (gsd=0.0017)
4 32 0.0208

1 33 0.0167
B 2 33 0.0154 0.0162 0.25
3 33 0.0166 (gsd=0.0005)
4 33 0.0162

NOTE: "gsd" is the Geometric Standard Deviation.


from the ground. A height greater than this would have made the process

of fertilizer discharge, which was manual, very inconvenient.

Enclosure heights of 0.9 m and 1.5 m were considered next. The

effective fertilizer discharge height was 1.5 m for both configurations.

The configuration with the 0.9 m enclosure was as shown in Figure 1 while

the configuration with the 1.5 m enclosure height had the feed tube

projecting into the enclosure rather than out of it. Results of tests

conducted with phosphate rock and white sand (Table 2) show that the

measured dust emissions were consistently higher with the 0.9 m

enclosure, probably because of a smaller volume of dead space and a

shorter distance between the point of dust emission and the air sampler.

The difference in measured emission factor was of the order of 10 % and

so this factor did not play a major part in the eventual selection of an

enclosure height. The 0.9 m enclosure was selected as standard because

it was much easier to move around due to its lower weight and smaller


Using the standard inlet size and enclosure height, the effect of

three different air flow rates was evaluated. The air flow rate was

varied by changing the applied voltage to the air sampler. The maximum

possible air flow rate was found to be about 35 liters/sec and, as shown

in Figure 4, operating the sampler at this condition did not result in a

significant increase in measured dust emission. A flow rate of 31

liters/sec was chosen as an optimum value for emission factor measurement

tests. It was observed that a flow rate of 21 liters/sec resulted in

emission factor measurements which were 21 % lower than that at 29

liters/sec and that the air flow rate had a nonlinear effect on measured

dust emission factor. If the air sampler was operated at a flow rate of


Effect of Enclosure Height on the Emission Factor
of Phosphate Rock and White Sand Samples

Product Sample Enclosure Average Overall
Type I.D. Height Emission Factor Average
(m) (g/kg) (g/kg)

Phosphate A 0.90 0.1361 0.1362
Rock B 0.90 0.1363

Phosphate C 1.50 0.1214 0.1206
Rock D 1.50 0.1198

White A 0.90 0.0142 0.0133
Sand B 0.90 0.0123

White C 1.50 0.0118 0.0117
Sand D 1.50 0.0115



o 0.10


0.05 I I
10 20 30

AIR FLOW RATE (liters/sec)

Figure 4. Effect of Air Flow Rate on the
Measured Dust Emission of GTSP


26 liters/sec instead of the optimum 31 liters/sec, the deviation in the

measured dust emission factor would be less than 10 %.

Material pour rate was varied by pouring 5 kilogram test samples of

GTSP in three different pour times, viz., 30, 60 and 90 seconds. As

shown in Table 3, for the pour times evaluated, the variations in

measured dust emission factor as determined by the "drop test" were not

extreme for moderately dusty materials like GTSP. For operational

reasons, the 60-second pour time was found to be most convenient and was

thus established as the standard pour time. Tests conducted with

phosphate rock, a significantly dustier material, showed that pour rate

did have a more significant impact (Figure 5) though the measured

emission factor was that from a single drop. However, the 60-second pour

time is still a valid selection since relative dust emission factor is

the primary parameter of interest.

The cross-sectional area and air flow rate were selected so that the

particle collection characteristics of the VFDC would be similar to that

observed with the high volume air sampler operating in a standard housing

as used for ambient air sampling. The VFDC test procedure simulates the

process of dust generation due to handling of bulk materials as at

transfer points in material conveyors and unloading stations.

Both VFDC configurations were calibrated with monodisperse ammonium

fluorescein aerosols and glass beads. The monodisperse ammonium

fluorescein aerosols were generated with a vibrating orifice aerosol

generator (TSI Model #3050) while the monodisperse glass beads, purchased

commercially, were dispersed from a flask by compressed air. The

fractional penetration of particles of various sizes was determined

gravimetrically for glass beads and fluorimetrically for ammonium


Effect of Pour Time on the Emission Factor
of GTSP Samples

Sample Drop Pour Time Emission Factor Average
I.D. Number (seconds) (g/kg) (g/kg)

1 30 0.0448
A 2 30 0.0383 0.0343
3 30 0.0260 (gsd=0.0089)
4 30 0.0279

1 60 0.0362
B 2 60 0.0536 0.0366
3 60 0.0256 (gsd=0.0122)
4 60 0.0308

1 90 0.0312
C 2 90 0.0491 0.0330
3 90 0.0265 (gsd=0.0111)
4 90 0.0250

NOTE: Enclosure height = 1.50 m (5 feet).
Air flow rate = 25 liters/sec (60 cfm).
"gsd" is the Geometric Standard Deviation.


'T 0.25

o 0.20-

2 0.15

S0.10 -

0.05 I I I I I I I I I I
0 500 1000 1500
POUR RATE (kg/hr)

Figure 5. Effect of Pour Rate on the Measured
Factor of Phosphate Rock Samples.


fluorescein aerosols. With the air sampler operated at 26 liters/sec the

particle penetration characteristics of the two units were found to be

almost identical (Figure 6). The measured 50 % cut point for both

units, when operated in an identical manner, was found to be 40 um.

In summary, the standard VFDC configuration used was like that shown

in Figure 1. Five kilogram test samples were standard as was a 60 second

pour time, a 2 minute air sampling duration and an air flow rate of 31


Dust Size Distribution Measurement

The size distribution of dust emitted due to handling of various

materials was measured using single stage impactors like that shown in

Figure 7. For a given flow rate the 50 % cut size can be changed by

changing the flow area in the impactor or, correspondingly, by using

separate single stage impactors with different nozzle widths. Three

impactors with 50 % cut sizes of 42 um, 25 um and 13.6 um when operated

at 30 liters/sec were used. The calibration of the single stage

impactors has been described elsewhere (Vanderpool, 1983).

The impaction surface was prepared by lining it with aluminum foil

cut to size and then coated with a silicone spray and weighed. The first

impaction surface was mounted in the nozzle section of the 42 um

impactor. Four spacers were placed on the rear side of the impaction

surface and the first pre-weighed filter was laid over it. The high

volume air sampler was then mounted on the impactor and bolted in place.

The impactor-high volume air sampler assembly was then placed over the

enclosure opening. The procedure was then similar to that for the first

drop of the "drop test" for emission factor measurement. For each test a

new sample was used. At the end of the first drop the impaction surface



0 .0


0 0.9 m Enclosure
1.5 m Enclosure

1 10 102

Figure 6. Calibration for Two Configurations of
the Vertical Flow Dust Chamber.



Impactor Nozzle

Impaction Plate



Flow Meter


Figure 7. Schematic of a Single Stage Impactor.


and filter were carefully removed. Prior to the next drop of the test

sample the second impactor nozzle was set up and a similar assembly and

test procedure followed. After the third drop test with the third

impactor nozzle the weight fractions on the impaction surface and filter

were determined and the size distribution calculated.

Intermediate Scale Field Tests

Apparatus and Operating Procedure

From extensive laboratory experiments it was apparent that full

scale field tests to demonstrate the validity of laboratory results would

be much more likely to succeed if intermediate scale tests were first

performed. The intermediate scale field tests were designed to evaluate

possible scale-up problems and to determine the influence of various

operating conditions.

An intermediate scale field test (ISFT) setup was designed to handle

a minimum of about 70 kilograms of fertilizer at a maximum feed rate of

about 10 tons per hour. The setup was composed of two major components

which included the fertilizer handling system and the coating agent spray


The fertilizer handling system consisted of feed and discharge

hoppers and a belt conveyor. The system was made portable by mounting

the conveyor and feed hopper on a modified boat trailer. The boat

trailer was a Harding Model #B-16-7 unit with an overall length of about

5.2 m and a 320 kilogram load capacity. The conveyor and feed hopper

support structure was made of 5 cm x 10 cm (2 inch x 4 inch) pressure

treated wood and was attached to the boat trailer frame with "U" clamps.

The trailer was equipped with a seven foot long tounge which was removed

once the setup was put in place.


A general drawing of the fertilizer handling system is shown in

Figure 8(a). The conveyor selected was a slider bed type conveyor

(Hytrol Model #TT "Thin Trough" conveyor) where the belt runs in a trough

cross-section frame as shown in Figure 8(b)(i). This type of conveyor

had the advantage that the probability of spillage was reduced and the

belt, when in operation, would be relatively smooth running and vibration

free. The conveyor weight was about 200 kilograms and was thus ideally

suited for light duty use as in the present application. The overall bed

length was 4.9 m and the belt was driven by a 3/4 HP motor at a speed of

25 cm/sec through a combination belt and chain drive. The conveyor belt

speed could be changed by changing the sprocket in the chain drive.

The conveyor was mounted on the wooden support structure on the

trailer at an angle of about 15 degrees by using 3 supports of

appropriate height so that the conveyor discharge was about 1.8 m from

the ground. The support heights were adjustable and allowed a variation

of a few degrees in the conveyor inclination if such an adjustment was

desired. The belt tension could also be adjusted by using tensioning

screws provided. The feed hopper was held in place over the belt in a

slotted angle frame so that the relative position of the hopper discharge

with the belt surface was fixed. The hopper was made of 1.9 cm (3/4

inch) plywood and painted so as to resist attack by the fertilizer. It

had an approximate capacity of 255 liters or, equivalently, about 250

kilograms of fertilizer. The downstream end of the hopper discharge was

equipped with an adjustable aluminum slide plate as shown in Figure

8(b)(ii) to allow a measure of control over the product discharge rate

from the hopper. The two sides and the upstream end of the feed hopper

discharge were equipped with rubber skirts to prevent spillage and to

0 03 0.6 1.2

Scale in Meters

255 Liter
Feed Hopper

4.9m Long Slider
Bed Conveyor

-0.5m Bed -
rAm Beldt Jl
3T=010 cm

I (i)

S --

Hopper Wall
1. cm Thick Plywood
/ Bolt with Wing
/Nut and Washer

I Scraper-0.6cm Thick Aluminum
with Sots for Vertical Adjustment

Figure 8.

Intermediate Scale Field Test Setup.
(a) Schematic of the Material Handling
(b) (i). Cross-section of the Conveyor
(ii). Fertilizer Feed Control Method






allow fertilizer flow only in the direction desired. The support

structure overhanging the trailer frame was propped up by concrete blocks

when the system was in use.

The discharge end of the conveyor was semi-enclosed in an enclosure

made of two 0.9 m sections of 0.51 m diameter galvanized pipe. A slot

was cut along the circumference so that the discharge end of the conveyor

was enclosed and the fertilizer discharge was down the axis of the pipe.

The top of the pipe was covered and bottom of the pipe was lower than the

top of the discharge bin (Figure 9). This enclosure helped to protect

the spray droplets and the fertilizer discharge stream from the effects

of wind. In addition to the conveyor supports mounted on the trailer

support structure, a fourth support made of 5 cm x 10 cm pressure treated

wood and slotted angle iron was used to support the overhanging discharge

end of the conveyor where the motor and drive weight was concentrated.

This support was on the ground and was removable (Figure 10).

The fertilizer feed rate could be adjusted by changing the belt

speed or by changing the feed hopper discharge characteristics. The

conveyor was equipped with a single speed motor and so the belt speed

could be varied only by changing the sprocket in the chain drive. It was

much easier, on the other hand, to control the feed hopper discharge

rate. The width of the hopper discharge was about 20 cm and so the bead

laid out on the belt was about 20 cm as shown in Figure 11. However, the

thickness of the bead could be easily varied by adjusting the slide

plate. The feed rate could thus be varied from about 4 tons per hour to

about 10 tons per hour by using the slide plate arrangement.

A line drawing of the coating agent spray system is shown in Figure

12. The spray system was designed to transfer a controlled amount of

Figure 9. Photograph of the Front View of the
Intermediate Scale Field Test Setup.

Figure 10. Photograph of the Side View of the
Intermediate Scale Field Test Setup.


Figure 11. Photograph of the Feed Hopper Discharge.


. .0

> 4J

0 0

c a ci

EI- 4

0 u

S^S S-'--'-Q


coating agent on to the fertilizer granules. The basic spray system

included a portable air compressor (Sears Model #919.156580), two Fitz &

Fitz 1.9 liter pressure containers and 2 nozzles. The nozzles used were

of the pressurized liquid type (Spraying Systems Catalog #1/4TT-730039).

The compressed air supply was divided into two streams, each passing

through a pressure container. The pressure containers were rated at a

peak liquid pressure of about 414 kPa (60 psig). Each pressure container

had two outlets used to provide separate air and liquid flows for air

atomizing nozzles. The air outlet was capped off since the pressurized

liquid nozzles did not need atomizing air. All liquid lines were 9.5 mm

(3/8 inch) diameter copper tubing. The copper tubing and nozzles were

heated by heating tape while the pressure containers were heated by

heating mantles. All heaters were controlled by variable transformers.

Since petrolatum waxes were the primary coating of interest, the pressure

containers were maintained at a temperature high enough to keep the

petrolatum waxes molten and liquid lines were heated to prevent

solidification in the lines. The liquid feed was controlled by adjusting

the regulator pressure on the pressure containers. The nozzles were in

an opposing jet arrangement about 25 cm from each side of the fertilizer

discharge stream and about 15 cm below the discharge end of the conveyor.

At least 3 buckets (about 70 kilograms) of fertilizer were used in

each test. Three 5 kilogram samples of the uncoated fertilizer were

first prepared in the standard manner. The remaining uncoated fertilizer

was then poured into the feed hopper. The line heaters and heating

mantles were all energized and the nozzles were calibrated prior to the

test by timing the consumption of a known amount of hot water. This also

helped to heat the lines and clean them. Hot water was poured into the


pressure containers and the water temperature was maintained by means of

heating mantles. The wax being tested was weighed out into two plastic

bottles which were then placed in a beaker of boiling water till the wax

was completely melted. The bottles were then placed in each of the two

pressure containers. In this manner the wax was not subjected to

excessive local heating, the pressure containers were easily cleaned

after use, successive tests could be conducted more rapidly and cross

contamination was not a problem. After allowing sufficient time for the

nozzles and fluid flow lines to heat up and setting the hopper discharge,

the conveyor was turned on. The wax spray was initiated so as to

coincide with the fertilizer discharge from the conveyor. When all the

fertilizer was used up the wax spray was discontinued by disconnecting

the air supply at the quick disconnect and relieving the line pressure by

using the pressure relief valve on the pressure container. Operating

parameters such as the hopper discharge setting, fertilizer weight,

weight of wax consumed, wax temperature and the wax and fertilizer feed

times were noted. During the test the fertilizer was discharged into the

discharge bin. After the test was complete the fertilizer in the bin was

stirred by using a shovel and then stored in 19 liter buckets. Two to

three test samples of coated fertilizer were then made in the standard

manner for later testing. To verify that the nozzles did not plug during

the test the nozzle calibration for water was rechecked. At this point

the next test if planned, was performed by simply replacing the wax

sample bottles and recharging the feed hopper with a new batch of

uncoated fertilizer.


The development of the ISFT setup and operating procedures was an

evolutionary process. Preliminary tests were conducted without the

discharge enclosure, but excessive fertilizer dust and wax spray blow-off

led to the addition of the discharge enclosure.

Before reaching a decision on the use of the pressurized liquid

nozzles, air atomized nozzles were evaluated. Both internal mix and

external mix nozzles were considered. In using the air atomizing nozzles

the air outlet from the pressure container was connected to the nozzle by

an air hose. In the internal mix type nozzle, compressed air and liquid

are mixed within the nozzle and then ejected from the nozzle. However,

wax solidification due to excessive cooling and losses by overspray due

to extreme atomization were continual problems. Modification of the

setup by regulating the air pressure to the nozzle did not significantly

improve the problem nor did the use of 4 nozzles, each with half the

capacity of the nozzles in the 2 nozzle arrangement. The external mix

nozzles did not have the same wax solidification problem but overspray

losses were still excessive. With the pressurized liquid nozzles, nozzle

plugging due to wax solidification was no longer a problem and overspray

losses were much reduced due to the coarser droplets produced.

The two nozzles were placed 15 cm below the discharge end of the

conveyor, one on each side of the fertilizer discharge. Because of the

close proximity of the nozzle to the underside of the belt, over a period

of time the belt had a tendency to get coated with wax and so a deflector

shield was installed. The nozzles were originally placed 15 cm from the

fertilizer surface but at this distance the spread of the spray was

insufficient to cover the width of the fertilizer discharge. So, the


nozzles were moved back to a distance of 25 cm from the fertilizer


Full Scale Field Tests

Apparatus and Operating Procedure

Full scale field tests were conducted at a fertilizer shipping

facility (Agrico Chemical Co., Pembroke Road, Gibsonton, Florida). This

facility handles granular triple super phosphate (GTSP) and ground

phosphate rock. The GTSP was transported to this facility from the

fertilizer plant by trucks in a travel time of about 1 hour. The

fertilizer handling setup was as shown in Figures 13 and 14(a) with air

samplers placed within the storage building as shown in Figure 14(b).

The nominal fertilizer handling rate was 250 tons/hour.

The coating agent spray system was designed within the facility

constraints to provide a maximum spray rate of about 19 liters/min (5

gpm) at about 414 kPa. Petrolatum waxes were acquired in 208 liter (55

gallon) drum quantities. The spray setup was as shown in Figure 15. The

pump used was a Liquiflo Series 86 Eccentric Impeller pump with a 3/4

H.P., 110V motor. The flowmeter used was an Erdco Series 400 vane-type

flowmeter. Valve 1 was a bypass valve used as flow control, Valve 2 was

a 3-way valve used to switch the flow from recycle mode to spray mode and

Valve 3 was a 1/4 turn valve used to control the supply of compressed

air. All flow lines were 1.9 cm and 1.3 cm black iron pipe and were heat

traced with 220V heating tape and insulated. Four nozzles were aligned

46 cm (18 inch) apart along the axis of the belt conveyor between

transfer point #1 and #2.

The drum of wax was heated by 2 drum heaters (Briskeat Catalog #SRL-

A-DHC-1200) with integral temperature controllers being used to set the



Figure 13. Photographs of the Full Scale Field
Test Facility.
(a) Truck Unloading Station (b) Transfer
Point #2 (c) Transfer Point #3


Figure 13. Continued.


Ulderg id Station
Hopper Z
STruck Unloading
1 I- I


I r
i Transfer I-
iPoint3 1I
I = -I/\'


I Fertilizer
I-- ---

Point 1



Buk rg
(61m x 46m)



Figure 14.

O Sampler Locations

Details of the Full Scale Field Test Facility.
(a) Fertilizer Handling System
(b) Air Sampler Locations


a I
Zoo @5

S< a.






temperature at about 1500C and insulated with fiberglass insulation. The

drum heating process was begun 12 to 24 hours prior to actual use. The

line heaters were then energized and heating was controlled by a variable

transformer. The pump intake was equipped with a suction filter

(Spraying Systems Catalog #HSW) to strain out particulate matter. Four

pressurized liquid type nozzles were cleaned in hot water and mounted in

the spray manifold. Each nozzle was equipped with a 50 mesh strainer.

Valve 2 was first set to position 1 to permit use of the system in

recycle mode, Valve 1 was opened halfway and the pump was then turned on.

The liquid wax was permitted to circulate through the system so that all

the lines and components could be evenly heated. The pump intake was

securely tightened so as to prevent air infiltration which could cause

the liquid wax to foam. A yardstick was taped to the inside of the drum

to permit a secondary measure of liquid consumption.

The oil supply to the existing oil spray system was first shut off.

Once a truck started unloading its load and the fertilizer appeared on

the belt between transfer point #1 and #2, the fertilizer was allowed to

run uncoated for about 45 seconds. The liquid wax was then switched from

the recycle mode to the spray mode by switching Valve 2 to position 2.

Valve 2 was then adjusted to set the flow rate at the required level as

indicated by the flowmeter. It took 30 seconds for material transfer

from transfer point #1 to transfer point #2, 45 seconds for material

transfer from transfer point #2 to transfer point #3 and 30 seconds for

material transfer from transfer point #3 to transfer point #4. A bucket

was half-filled with uncoated fertilizer sampled at transfer point #2 and

then 2 buckets of coated fertilizer were sampled at transfer point #4 a

minute after wax coated product appeared. Once the coated samples were


collected Valve 1 was used to reduce the liquid spray rate by recycling

part of the pumped liquid back to the reservoir and Valve 2 was set to

position 1, to put the spray system in recycle mode. The half filled

bucket of uncoated fertilizer was then completely filled at transfer

point #2. By measuring the fall of the liquid level in the drum and the

time of consumption, the application rate was calculated as a check of

the flowmeter. If more than a few minutes wait was anticipated between

runs compressed air was blown through the nozzles by switching Valve 3 to

the "on" position. Compressed air was provided by a portable air

compressor (Sears Model #919.156580). In this manner, nozzle plugging was

avoided. At the end of a series of tests hot water was circulated

through the spray system in the recycle mode to clean out as much wax

from the lines as possible. No attempt was made to spray water through

the system in the spray mode because of possible caking problems which

could occur in the vicinity of the conveyor.

The coated samples collected were brought back to the laboratory and

5 kg test samples were prepared for further analysis in the standard



The location for conducting the field test was chosen based on a

number of factors, the most important of which was familiarity with the

facility. Power outlets were easily accessible, fertilizer sampling

locations were convenient and the design of the fertilizer handling

system was such that the coating spray system could be situated in a

compact way not too far away from the spray location.

The pump selected was of an eccentric impeller design with a high

density polymer impeller. The maximum pressure and temperature ratings


were 1103 kPa and 900C, respectively. This pump was considered ideal for

the present application because the liquid to be pumped was clean and a

lubricant. No pressure gages were installed because of the possibility

of fouling the internal parts of the gage by solidifying wax. For this

same reason a "sight gage" type vane flowmeter was selected for flowrate

measurement. The deflection of the vane was a measure of flow rate. In

addition, line plugging could be signaled by the "see through" window in

the flowmeter. Pressure relief was provided by a plastic coupling rated

at 828 kPa.

The pump and compressor both had 110 V motors and the power supply

was routed through a 15 amp circuit breaker. As a result, when the

compressed air tank was full the cycling of the compressor caused the

breaker to trip due to the high starting current of the compressor motor.

Thus, in order to operate the pump and compressor simultaneously, a bleed

valve was installed in the compressor outlet so that the compressor would

run continuously without shutting off.

Various combinations of spray location and nozzle size were

evaluated and the results are discussed in a later chapter.


Extensive evaluations were conducted during the course of this

study. Results presented in this chapter are divided into separate

sections: laboratory tests, intermediate scale field tests (ISFT) and

full scale field tests (FSFT). Criteria for the selection of dust

suppressants are also discussed.

Laboratory Tests

Effect of Temperature on Test Samples

The effect of temperature on granular triple superphosphate (GTSP)

and diammonium phosphate (DAP) was studied. Three 30-gram samples of

GTSP, three 20-gram samples of GTSP and three 30-gram samples of DAP were

weighed out in 95 mm diameter aluminum dishes and placed in an oven

(Precision Model #17) at 1050C. Sample weights were measured with a

single pan electronic balance(Sartorius Model #2355). The measured

weight change as a function of time was as shown in Figure 16.

The DAP samples showed a consistent loss in weight with no sign of

equilibration over the time period considered. This loss in weight was

accompanied by a strong smell of ammonia in the vicinity of the oven.

From this observation it was concluded that the DAP 'ranules were

undergoing a process of breakdown and subsequent deammoniation.

The GTSP granules also exhibited a continuous weight loss as a

function of time. However, the rate of weight loss was significantly

reduced after the first 24 hours. This phenomenon was probably due to



O 30 Gram GTSP
0 20 Gram GTSP
S 30 Gram DAP




3 a


5 10 20


Figure 16. Weight Loss due to Heating of GTSP and
DAP Samples as a Function of Time.




accelerated chemical reactions within the granules and subsequent

breakdown by a process called phosphate reversion (Bookey and Raistrick,

1960; Slack, 1968).

In addition, when test samples of GTSP were subjected to elevated

temperatures over a period of time the moisture content of the granules

was significantly reduced. Because of this reduction in moisture content

the measured emission factor (Table 4) was greatly increased.

Effectiveness of Test Sample Preparation Method

Five kilogram test samples were prepared from a given batch of

fertilizer using the technique described earlier. As standard practice

at least two test samples from each batch of product were tested in an

uncoated state. The average emission factor for the batch and the

deviation of the emission factor of each individual sample from the

average emission factor was calculated. This average emission factor

represents the baseline emission level prior to treatment while the

deviation is a measure of relative product homogeneity with regard to

dust emissions.

Specific results for test samples from three batches of fertilizer

are shown in Table 5. A scatter diagram of measured deviation for 178

samples of fertilizer from 89 distinct batches is shown in Figure 17.

From these results it is evident that the sample preparation method and

the measurement method were very effective. The average deviation from

the average emission factor was about 3.5 % and 97 % of the samples had a

deviation of less than 10 % from the average emission factor. Thus, the

calculated average uncoated emission factor for a batch can be considered

to be representative of the whole batch. In addition, since dust

suppression effectiveness is a function of the ratio of coated to


Effect of Heating on the Enission Factor
of GTSP Samples

Sample Sample Moisture Content Enission Factor
I.D. Treatment (%) (g/kg)

R14 None 1.4 0.0166
R3 Heated 0.8 0.0652

AGTSP127 None 0.96 0.0433
AGTSP134 Heated 0.54 0.0628


Effectiveness of the Test Sample Preparation Method

Product Sample Average Emission Average EBission Deviation
Type I.D. Factor of Test Sample Factor of Batch from Average
(g/kg) (g/kg) (%)

AGTSP8 0.0387 -6.5
GTSP AGTSP7 0.0405 0.0414 -2.2
AGTSP6 0.0449 +8.5

GODAP11 0.0918 -6.4
DAP GODAP1 0.1007 0.0981 +2.7
GODAP6 0.1019 +3.9

IGTSP7 0.0229 -8.8
GTSP IGTSP1 0.0247 0.0251 -1.6
IGTSP8 0.0278 +10.8

50 100 150


Figure 17. Deviation of the Emission Factor of
Individual Samples from the Average
Emission Factor for that Batch.




0 0






uncoated emission factor, an accurate value of uncoated emission factor

improves the quality of the calculated effectiveness.

Because of the reproducibility of the emission factor measurements,

this technique was used to screen materials from different sources (Table

6). Dust emission factors in the 0.005 g/kg to 0.1 g/kg range were

measured for various products from many sources. This technique was also

used to monitor the variation of product quality, as shown in Table 7.

For a particular source the measured dust emission factor varied between

0.03 g/kg and 0.08 g/kg over a period of time.

Granule and Dust Characteristics

Moisture, both surface and chemically bound, is present in the

fertilizer granules and, as discussed earlier, sustained high

temperatures lead to moisture loss and severe increases in dust

emissions. This suggests that increased moisture content should have the

opposite effect. The validity of the above observation was borne out by

the results shown in Figure 18. Four samples of GTSP from batch A and

three samples of GTSP from batch B were used. For batch A, sample #1 was

dried, sample #2 was left untreated and sample #3 and sample #4 were

sprayed with known amounts of water to raise their moisture content. For

batch B, sample #1 was left untreated and sample #2 and sample #3 were

sprayed with known amounts of water to raise their moisture content. The

results of drop tests clearly show that just a 20 % increase in moisture

content resulted in significant decreases in dust emission factor and it

appeared that a moisture content of about 1.5 % for GTSP samples could be

very beneficial as far as dust emission reduction was concerned.


Examples of Emission Factors for Various Products

Product Average Bnission Factor
Type (g/kg)

AGTSP 0.0506
GADAP 0.0093
IGTSP 0.0096
IDAP 0.0082
GGTSP 0.0158
GODAP 0.0981
FDAP 0.0309
Phosphate Rock 0.1362
White Sand 0.0133
Sulfur 0.0877

different manufacturers.
different manufacturers.

samples from three

DAP samples from four


Variation of Product Quality for GTSP Samples

Batch aAverage Emission Factor Overall Average
I.D. (g/kg) (g/kg)

A 0.0506
B 0.0720
C 0.0347 0.0457
D 0.0331 (gsd = 0.0143)
E 0.0405
F 0.0435

a. The 6 batches represent product acquired from
the same manufacturer on different occasions.
b. "gsd" is the geometric standard deviation.






Batch A

Batch B

Figure 18. Effect of the Moisture Content on the
Emission Factor of GTSP Samples.




1.7% 1.1% 1.4%
n i F]

-I I-' -


Since moisture plays such an important role in determining product

dustiness, a test was conducted to establish if there was any variation

in measured moisture content as a function of time of storage. Three 2-

gram samples were taken on 3 successive days from a 5-kg test sample of

GTSP and moisture content was measured in the manner described earlier.

Results in Table 8 show that there was no significant change in moisture

content over the time period considered and, correspondingly, the dust

emission factor can be considered to be unaffected by storage, at least

in the short term.

The standard test procedure for the measurement of moisture content

specifies 2-gram test samples. But, a series of tests with larger sample

sizes were carried out to determine if sample size was a significant

factor in the measurement. Results of tests with untreated fertilizer

(Table 9) show that the measured moisture content was quite insensitive

to sample size when the fertilizer was not sprayed with water after

manufacture. However, if in an effort to increase moisture content,

water was externally sprayed on the 5-kg test sample, the smaller 2-gram

sample results in erroneous and scattered results (Table 10). On the

other hand, 10-gram samples resulted in a significantly better

determination of measured moisture.

Size distributions of DAP, GTSP and MAP samples from different

manufacturers were determined by sieving 100-gram samples for 10 minutes

in a Gilson Model #SS-15 Sieve Tester. Calcined phosphate rock and fine

grain white sand were also sieved as a comparative measure. Results of

these sieving tests (Table 11 and Table 12) show that the granular

product was generally in the 2.0 mm to 2.5 mm range and the size

distribution -was fairly narrow. The various size fractions were tested


Stability of the Moisture Content
of Stored GTSP Samples

Day Sample Moisture Average
Number Content bMisture Content
(%) (%)

1 2

2 2

3 2








Effect of Sample Size on the Measured
Moisture Content of Untreated GTSP Samples

Sample Sample Moisture Average
I.D. Size Content Moisture Content
(g) (%) (%)

1 20 1.10 1.13
2 20 1.17

3 10 1.06 1.07
4 10 1.07

5 2 1.13 1.16
6 2 1.20


Effect of Sample Size on the Measured Moisture Content
of Treated GTSP Samples

Sample Serial Sample Measured Average Treatment
I.D. Number Size Moisture Content Moisture Content
(g) (%) (%)

AGTSP109 1 2 0.93 1.01 None
2 2 1.10

AGTSP108 1 2 0.94 1.11 Water
2 2 1.28

AGTSP101 1 2 1.26 1.20 Water
2 2 1.14

AGTSP109 1 10 0.96
2 10 1.04 0.99 None
3 10 0.98

AGTSP108 1 10 1.08
2 10 1.14 1.11 Water
3 10 1.12

AGTSP101 1 10 1.44
2 10 1.42 1.43 Water
3 10 1.42

NOTE: Expected Moisture Content:

AGTSP108 1.2 %
AGTSP101 1.5 %

0s 000

ON ** 3
O- 00

C u

oo m'- m
o > M 0

aom C


* 0
mm C- o





* c






C, v




a, .


;. C
* *

4r 11: ir
* *
ON 00


Size Distribution of Samples
of Some Non-granular Materials

Granule Phosphate White
Size Bock Sand
(un) (wt. % <) (wt. <)

850 94.2 100.0
425 89.6 99.8
212 46.5 72.8
106 2.6 1.3
75 0.44 0.03
53 0.12 -

MMDa(un) 250 190
GSDb 1.60 1.32

a. MMD is the Mass Median Diameter.
b. GSD is the Geometric Standard Deviation.


for granule hardness or crushing strength by the TVA method described

earlier. Results show that the measured crushing strength increased with

increasing granule size as has been observed elsewhere (Jager and Hegner,

1985). For the samples tested, MAP granules were stronger than DAP

granules, which were, in turn, stronger than GTSP granules (Figure 19).

Since product dustiness was determined by drop tests, experiments

were conducted to determine if granule fracture, a possible mode of dust

generation, was measureable. One-hundred-gram samples were extracted

from 5-kg test samples before and after a complete "drop test" and sieved

in the standard manner. The difference in measured size distribution for

DAP and GTSP samples was not significant and could have been due to

sampling variabilities (Table 13). This same behavior was exhibited by

phosphate rock and white sand. However, though the dry sieving technique

used was not sensitive enough to determine if granule fracture occurred,

the results do show that it was not significant.

Similar tests were conducted with prilled sulfur, a brittle

material, and the results are presented in Figure 20. With increased

handling the size distribution exhibited a distinct shift toward the

smaller particle sizes with a corresponding increase in the fraction of

small particles. Examination of the samples also verified that

significant granule fracture occurred. A similar process has been found

to occur with coal, char particles and detergents (Arastoopour and Chen,

1983; Goodwin and Ramos, 1987; Knight and Bridgewater, 1985).

Further study of the drop-wise change in dust emission factor

(Figure 21) indicated the significant difference in response to handling

between sulfur and the other products. The dust release process is a

function of the fracture tendency of materials. Dust release from sulfur


01 I I
10-1 1 10


o l I I I j
10'1 1 10


4 1


10-1 1 10



2 '
0 1 1
10-1 1


4 -


10"1 1



Figure 19. Hardness of Granules of Various Fertilizers.


Effect of "drop tests"
on Product Size Distribution

(a). Granular Materials

Granule IDAP14 IGTSP6
Size Before After Before After
(mn) wt. % < wt. % < wt. % < wt. % <

3.35 97.4 95.8 97.1 97.6
2.36 74.4 71.3 48.3 51.3
1.70 18.8 18.4 6.2 8.4
1.18 0.33 0.39 0.30 0.36
0.85 0.02 0.02 0.01
0.425 0.01 0.01

MMDa(mm) 2.02 2.10 2.22 2.30
GSDb 1.24 1.24 1.23 1.22

(b) Non-granular Materials

Particle White Sand Phosphate Rock
Size Before After Before After
(um) wt. % < wt. % < wt. % < wt. % <

850 100 100 93.1 94.3
425 99.8 99.8 85.2 87.8
212 73.1 76.3 45.9 49.3
106 1.2 1.3 2.6 3.2
75 0.03 0.02 0.42 0.60
53 0.09 0.15

MMD a(m) 190 190 255 245
GSD 1.32 1.32 1.61 1.61

a. MMD is the Mass Median Diameter.
b. GSD is the Geometric Standard Deviation.

U 99.8 0 No Drops
S99.5 A 4 Drops A
99.0 10 Drops 0
< 98.0

95.0 -
w0 90.0
S80.0 A
70.0- v
2 60.0 -
- 50.0
S40.0 -
) 30.0 -
w 20.0 -
0. 0
U 10.0 o
5.0 -
2.0- A
0 0.5-
0.01 1.0 10.0

Figure 20. Effect of Handling on the Size Distribution
of Prilled Sulfur.

0 Uncoated
A Uncoated
- 0 Uncoated

Phosphate Rock
White Sand


. 0.3


0. 0.2


SL 0.1


C 3

$ + I I


Figure 21. Effect of Handling on the Emission Factor
of Various Materials.

* 0

. .. + 4 1 1


was due to significant breakage of prills while with the other materials

fracture was not a significant source of dust. The dust was probably due

to fines in the sample, breakage of crystal growths on the granule

surface (Figure 22) and release of dust bound to granule surfaces by

physical forces. The existence of crystal growths has also been

documented elsewhere (Hoffmeister, 1979; Kjohl, 1976; Jager and Hegner,


The dust size distribution for various products was measured using

the technique described earlier. Products used included GTSP, DAP,

phosphate rock and white sand. Tests were initially conducted with the

Andersen and University of Washington Mark III multi-stage impactors.

These tests were not successful because the optimum operating

characteristics of the multi-stage impactors were not compatible with the

drop test apparatus and operating conditions. The above multi-stage

impactors were designed to measure particle size distributions in the

approximate 0.4 um to 15 um size range with a sample flow rate of about 7

liters/min to 21 liters/min. To ensure that a representative sample was

collected, isokinetic sample conditions had to be maintained by an

appropriate selection of nozzle diameter and sample flow rate and this

required the use of a highly flared short nozzle. By replacing the wood

panel of one side of the VFDC with a plexiglass sheet it was possible to

visualize the flow pattern of smoke injected into the VFDC. This

evaluation revealed that the flow field had characteristics which

prevented accurate sampling with the multi-stage impactor setup used.

Tests were later conducted with a set of 3 single stage impactors

used in the manner described earlier. Use of the single stage impactors

did not interfere with standard VFDC operation and the operating

Figure 22. Photograph of Crystal Growth on MAP


conditions were exactly the same as that of the VFDC. The three single

stage impactors were used at a flow rate of 29.7 liters/sec (63 scfm) and

the corresponding 50 % cut points were 42 uma, 25 uma and 13.6 uma,

respectively. The measured mass median diameter (MMD) and geometric

standard deviation (GSD) for GTSP, DAP, phosphate rock and white sand

were 12 uma and 2.2, 17.5 uma and 1.7, 25.5 uma and 1.8 and 7.4 uma and

3.2, respectively (Figures 23 and 24). The aerosols from the fertilizer

samples were mostly larger than 10 um though, with GTSP, a significant

mass fraction was less than 10 um, and with white sand a major fraction

was less than 10 um. The 10 um size is important because of recent

regulations regarding particle emissions in the less than 10 um size

range and their possible health effects.

Product Treatments

Three principal types of fertilizer were used in the evaluation of

proposed dust suppression agents. These were granular triple

superphosphate (GTSP), diammonium phosphate (DAP) and monoammonium

phosphate (MAP). Dust suppression agents used included oils, waxes,

emulsions and other miscellaneous materials.

Oils. The kinematic viscosities of various oil blends in actual

industrial use, were measured using Cannon-Fenske type glass capillary

viscometers according to procedures described in ASTM method D445-82.

These oils were then applied in the standard manner to GTSP samples. The

coated samples were drop tested immediately and again after an aging

period. In general, the test results (Table 14) reveal that for oils

with kinematic viscosities in the 50 to 250 centistokes range the

performance was poor. In addition, as the viscosity decreased the

performance decreased. Tests were also conducted with naphthenic oils


Z u




95 -

80 -

20 -


2 4 6 810

40 60 80


Figure 23. Size Distribution of the Dust Emitted
by the Handling of GTSP and DAP Samples.



I I I I l I I I l t i






2 4 6 810


20 40 60 80


Figure 24.

Size Distribution of the Dust Emitted
by Handling of White Sand and Phosphate
Rock. (*This point is off the line due
to impactor stage overloading).


80 -
80 -


0 White Sand
V Phosphate Rock




Effect of the Kinematic Viscosity of Oil Blends on
the Dust Release of GTSP Samples

Dusta Sample Kinematicb Initialc Nnrmalized d Final Final
Suppressant I.D. Viscosity Dust Release Dust Release Age Dust Release
(cst) (%) (%) (days) (%)

DCA305 AN2 58 83.6 -
DCA Bell IGTSP7 198 23.5 -
AM302EEF AN4 204 9.4 23.7 11 61.8
AM303 B8-1 232 12.6 15.3 17 27.7

a. Application Rate = 3 kg/ton.
b. At 20C.
c. Initial Dust Release is that determined soon after
application of dust suppressant.
d. Normalized Dust Release is that determined after
an aging period of three days.


with kinematic viscosities of 105, 410 and 755 SUS, respectively.

Results (Table 15) again show a definite decrease in dust release with

increased kinematic viscosity, but with aging the performance was again

severely degraded as manifested by the increased dust release values.

It has been stated in literature (Frick, 1977) that the dust

suppression effectiveness of oils improves with increasing paraffinic

content. The aniline point represents the relative paraffinic content of

oils and is a commonly used measure. Paraffinic oils with aniline points

in the 1020C to 1210C range were acquired from 2 manufacturers and

sprayed on GTSP samples. Drop test results (Table 16) do, indeed, show

that increased aniline points lead to decreased dust release, but the

performance was still average.

A number of other oils, including petroleum and vegetable oil

blends, were evaluated. The results show that most of the oils tested

with GTSP (Table 17) exhibited increased dust release with increasing age

though some oils retained their effectiveness to a greater extent. Most

of the oils tested on DAP (Table 18), on the other hand, showed low

initial dust release levels and smaller increases in dust release with


In summary, of the oil blends tested only some had low initial dust

release values (better than 10 %) and even fewer had low final dust

release values when used with GTSP. With DAP all the oils tested had low

dust release values and exhibited small increases in dust release with

age. This product specific behavior was probably caused by differences

in the interactions at the substrate oil interface leading to migration

of the oil from the granule surface to the granule interior at different

rates. Differences in granule porosity and oil viscosity and the


Effect of the Kinematic Viscosity of Naphthenic Oils
on the Dust Release of GTSP Samples

Dusta Sample Kinematicb Initial c Normalized d Final Final
Suppressant I.D. Viscosity Dust Release Dust Release Age Dust Release
(SUS) (%) (%) (days) (%)

S100 AGTSP137 105 23.2 39.0 8 65.3
S400 AGTSP138 410 11.3 23.0 8 42.5
S750 AGTSP136 755 6.4 17.8 8 36.9

a. Application rate = 2 kg/ton.
b. At 38C.
c. Initial Dust Release is that determined soon after application
of dust suppressant.
d. Normalized Dust Release is that determined after an aging period of
three days.


Effect of the Aniline Point of Paraffinic Oils
on the Dust Release of GTSP Samples

Dust a Sample Aniline Initial b
Suppressant I.D. Point Dust Release
( C) (%)

SP110 AGCN1 101.7 31.8
TUFLO6016 AGTSP50 107.2 32.2
SP120 AGCN2 107.8 20.3
TUFL06026 AGTSP51 113.3 18.0
SP130 AGTSP110 115.6 14.3
TUFLO6056 AGTSP53 121.1 11.5

a. Application Rate = 3.2 kg/ton.
b. Initial Dust Release is that determined soon
after application of dust suppressant.

*. .
o (nLc Ln7 o -

~~~ ~Cj


S<--- I


4) ->





Ln Nj
I I om II

CV %00 D U 0o r o0 Ln

C 'm mm c -m mmc m=r

D o

c. en

Ln -


cm (M m
I I I o


S00 -- n m [(-
CO -m L T M I




w "I





QC7 m Cm LM- U

M C ) kn I.Ln -*

CM M~ m \i nr (A

j NL LzJ
\J C C
c~ (J(.


CM (oj 0

o L.

0 0
c 5

) o

a. a

0 a
-. ,-4

0 *4)

a ro

0 0

4) (
H *V w

a L



.-I U)1 ..
0 l-

a~a 0

-a o I co


corresponding differences in performance suggest that the above

explanation is quite plausible. This aspect is considered again later in

this chapter.

Waxes. Waxes evaluated included natural waxes such as paraffin wax,

microcrystalline wax, candellila wax, carnauba wax and montan wax and

many petrolatum and related waxes. Results of a preliminary qualitative

evaluation are shown in Table 19 and further details on the use and

properties of natural waxes are described elsewhere (Bennett, 1975). The

waxes melt with varying degrees of difficulty. Paraffin,

microcrystalline and candellila waxes formed coatings or films which were

either flaky or powdery in nature and, for this reason, were not expected

to be effective dust suppressants when used as melts. Montan wax did not

melt easily and when it eventually did so, it was "tarry" and did not

spray properly. Carnauba wax, though it melted easily, formed a "grainy"

melt and thus an intermittent, uneven spray was produced. Petrolatum

waxes, on the other hand, melted easily, sprayed easily and formed good,

ductile films that adhered well to substrate materials.

Based on the qualitative evaluations, it was expected that the

natural waxes would give poor results. The melting points of paraffin

wax and candellila wax were 550C and 700C, respectively. Tests were

conducted at an application rate of 2 kg/ton and, as expected, the

performance was very poor. In fact, candellila wax had such poor

adhesive qualities that the coated emission factor was much greater than

the uncoated emission factor (measured dust release = 613 %) thus

suggesting that the coating itself was shedding and contributing to the

overall emission. For paraffin wax the measured dust release was 72 %.

Petrolatum and related waxes were the only materials, among those


Qualitative Characteristics of Waxes


Received in prilled form.
Melts easily.
Sprays easily.
Forms hard, flaky films.





Received as a hard block.
Melts with some difficulty.
Hard to spray.
Forms hard, flaky film.

Received as a hard block.
Melts easily.
Sprays easily.
Forms loose, powdery film.
Significant shrinkage of film on cooling.

Received as fine beads.
Melts with difficulty to a tarry product.
Could not be sprayed
Significant shrinkage of film on cooling.

1. Received as flakes
2. Melts easily.
3. Sprays intermittently
texture of melt.
4. Significant shrinkage

due to grainy

of film on cooling.


1. Received as "pastes" with various
oil contents.
2. Melts and sprays easily.
3. Forms smooth, strong film.




considered, that appeared to have good spray qualities and, thus, the

potential for superior performance. A total of 11 waxes from 3 different

manufacturers were evaluated. Of these Light Plasticrude and NW7098 were

slack waxes while all the others were various grades of petrolatum waxes.

These waxes were classified as having low, medium or high oil content

based on the approximate oil content values provided by the manufacturers

and some of their properties are summarized in Table 20.

Since the waxes were sprayed as melts the ease of melting was an

important consideration. In general, the higher the oil content, the

easier it is to melt a wax. All the petrolatums, except NW6889, melted

and sprayed easily. NW6889 had the lowest oil content and the highest

melting temperature and was a little more difficult to handle. However,

with proper selection of spraying conditions, NW6889 was also sprayed

without undue difficulty.

The effectiveness of the dust suppressant ultimately depends on the

application rate. As a result, 3 different application rates, nominally

1 kg/ton, 2 kg/ton and 4 kg/ton, were used and the results are shown in

Tables 21, 22 and 23, respectively. The most important factors in

judging coating performance are the initial and final dust releases.

Since not all the petrolatum waxes were tested after the same aging

period a normalized dust release was calculated for an averaging period

of 3 days to permit direct comparison of results from different

petrolatum waxes. A loss or decay rate was also calculated and used as

an indicator of the rapidity with which the performance changes. Both

the above parameters were calculated assuming linear variation of dust

release with age. The variation could well be non-linear, but as a first

step the linear assumption provides a quick method of comparing different



O.. 0



o %



0 o%
r># p






m Im mmM I mi S i



-N ( o '.- o N C- m NC -m c-
C- LAn Ln Mn %0 n n Uin Ln L An

_r %0
t- ck 8
(M '.c!

LA0 f
A 11
(\j IT

C 4
o .-l


A -





O -.


C u


0 .


co %D


in cn
ow a.

Cr -I

C- r- C- C- -C- t- t-

CV- a\ -C t- 0-
It ll i a I m 9
00 0000

Ln Cn--nC 0 M O cm-
t- t- E- ? S
I'd 8Q C l m &













am m m i mi
12 2 .2, R


cLm\ 'o -r %0

C-- c

T %u0 ir m '- W04
0 M Ln r- =-
v 1; *

m (NJ

Mm~n mmo *


NO 0 O00 0

co % = 0

5 aoSC





. Q
a M


} 0" r-i
CIx = 4






Cu o

c t

4* 40 Cu

%# 4)
C')0 L

S30 3 0.

' cu Cu

0. Vo N