Advertisement

Mining, Metallurgy & Exploration

, Volume 36, Issue 1, pp 117–129 | Cite as

Promising Options for Solving the Dolomite Problem of the Florida Phosphate Resources—a Brief Review

  • Patrick ZhangEmail author
  • Shibo Zheng
  • Wenyi Song
  • Chunhui Feng
  • Brij Moudgil
  • Wending Xiao
  • Dapeng Zhang
Article

Abstract

Separation of dolomite from phosphate is the most challenging problem in phosphate mineral processing. Over 50% of the future phosphate reserve in Florida contains too much dolomite to process using the current industry practice. The current phosphoric acid production practice requires less than 1% MgO in the phosphate concentrate as the feed material. The Florida Industrial and Phosphate (FIPR) Institute has collaborated with worldwide experts in the field to address this issue. As a result, the industry is now offered three feasible options. Option 1 offers three methods for reducing MgO content in the concentrate from the Crago process, including adding a dolomite depressant in the rougher flotation step, dolomite flotation of the cleaner concentrate, and scrubbing the cleaner concentrate in quartz sand. These methods could reduce MgO content in the final concentrate by 20–40%. Option 2 involves crushing and grinding of high-dolomite phosphate pebbles followed by dolomite flotation at slightly acidic pH using a new collector that does not require phosphoric acid as a phosphate depressant, achieving a final concentrate analyzing less than 0.9% MgO at about 87% P2O5 recovery. Option 3 is a gravity separation technique using an innovative separation jig, and its full potential remains to be demonstrated.

Keywords

Phosphate Dolomite Flotation Gravity separation 

1 Introduction

1.1 Magnitude of the Dolomite Problem

As the high-grade phosphate deposits deplete rapidly, the global phosphate industry now has to process phosphate ores with ever increasing contaminants, with carbonaceous materials (MgO and CaO) being the major problems. This issue is becoming pressing for the Florida phosphate industry, as it starts mining the ore body from the Southern Extension. The Florida phosphate deposits in the Southern Extension are divided into two zones: an upper zone and a lower zone. As is shown in Table 1 [1], the lower zone is problematic with the pebble fraction containing over 6% MgO and the flotation feed containing enough MgO to produce an unacceptable concentrate by the current industry practice.
Table 1

Analysis of the phosphate deposit from Southern Extension in Central Florida

Zone

wt%

Pebble

Flotation feed

Waste clay

wt%

P2O5

MgO

wt%

P2O5

MgO

wt%

P2O5

MgO

Upper Zone

33

11

27.8

0.52

69

7

0.12

20

9

1.90

Lower Zone

67

8

17.0

6.19

58

7

0.67

34

2

11.50

Although research on separation of dolomite from phosphate has been extensive for processing the dolomitic Florida phosphate pebbles, no practical solution was developed up to about 1995. In a 1994 Florida Industrial and Phosphate (FIPR) Institute study, several seemingly promising flotation processes were evaluated for separating dolomite from phosphate in a composite pebble sample by the same research engineer following the procedures provided by inventors of each of the processes evaluated. The results showed that among the five processes compared, two processes produce a concentrate containing over 1% MgO, and none of the process achieved overall phosphate recovery of higher than 60% [2].

1.2 Brief Literature Review

In order to extend the phosphate mine life in Florida, the industry has to find a way of processing the high-dolomite deposits. Currently, the we-process phosphoric acid manufacturing requires phosphate rock containing less than 1% MgO content, which is not possible using the standard Crago double float process on either the flotation feed or ground high-dolomite pebble from the Southern Extension. Dolomite contamination is also a worldwide problem. Systematic research in beneficiation of carbonate phosphate ore started in the early 1950s. Since then, research efforts in this area have been extensive. The research efforts have produced many processing flowsheets some of which had been pilot-tested or commercialized. These processes include direct flotation of phosphate with a dolomite depressant, reverse flotation of carbonate at mild acidic pH with a phosphate depressant, acid leaching, calcination, and gravity separation methods.

Of all these processes, reverse flotation of carbonates while simultaneously depressing phosphate has been studied most extensively. There are good reasons for this trend: (1) flotation is, in many cases, the least expensive mineral processing operation, (2) flotation of the minor gangue also has many advantages over floating the major phosphate minerals (reducing capital cost and saving reagents, for example), and (3) plant modifications would be minimal if flotation is used for processing the high-dolomite phosphate ores.

As is well understood, the difficulty in separating phosphate from carbonates is due to their similarity in electrochemical properties. Therefore, surface modification is essential for separating carbonates from phosphates efficiently. Besides the additions of pH modifiers, collector extenders, and frothers, the addition of depressant is the most frequently used method of surface modification. Over the years, the following reagents have shown the potential of depressing phosphate [3, 4, 5, 6, 7, 8, 9]: hydrofluosilicic acid, orthophosphoric acid, phosphoric acid, diphosphonic acid, sulfuric acid, aluminum sulfate and tartaric acid, phosphates, dipotassium hydrogen phosphate, sodium tripolyphosphate (STMP), alizarin red S (ARS), ethoxylated alkyl phenol, sodium hexametaphosphate (SHMP), trisodium polyphosphate (TSPP), and starch.

However, there is no universal phosphate depressant for every kind of phosphate deposit. A recent study conducted at the International Fertilizer Development Center compared several phosphate depressants and found that the phosphoric acid depression process gave the best results [10]. Hydrofluosilicic acid, starch, and dipotassium hydrogen phosphate have been extensively tested on a carbonaceous ore from India, and results indicated that dipotassium hydrogen phosphate was superior to the others in depressing phosphate in the test sample. Cornstarch was found to be effective on several high-dolomite Brazilian phosphate ores [11]. According to Wiegel [12], starch is the best phosphate depressant for Florida ores. It is, therefore, evident that there is no universal phosphate depressant. Mineralogical composition plays a critical role in finding the best depressant for a specific ore. Another important feature of most phosphate depressants is their pH dependence, functioning either as a carbonate depressant or an apatite depressant as pH changes.

The most recent screening of phosphate depressants for separating dolomite from phosphate was conducted by FIPR [13, 14]. Results with 10 different dolomite depressants showed the superior selectivity of STPP, SHMP, and TSPP to all other depressants for the phosphate-silica system. Sodium tripolyphosphate (STPP) also found to be an effective depressant for reverse flotation of dolomite from phosphate under slightly acidic conditions, particularly when sulfonate was used as a dolomite collector.

1.3 Recent FIPR Studies

In 1997, IMC-Agrico Company submitted the proposal “Development of New Beneficiation Technology for Florida Dolomitic Phosphate Resources” to Florida Institute of Phosphate Research. The objective of the project was to develop a processing flowsheet for economic recovery of the phosphate values from the Florida high-dolomite phosphate pebbles. The Chinese Lianyungang Design and Research Institute (CLDRI, currently China Bluestar Lehigh Engineering Co.) was the major subcontractor of the project responsible for laboratory testing of the processing alternatives. FIPR Board of Directors approved the proposal at their October 1997 meeting, and the project [15] was initiated in January of 1998.

Under the project described above, five high-dolomite phosphate pebble samples were evaluated on laboratory scale. The processing flowsheet involved grinding the pebble to achieve complete dolomite liberation, dolomite flotation using a mixture of H3PO4 and H2SO4 as pH modifier and the PA-31 (fatty acid soap) as dolomite collector, and cleaner flotation using an amine. The flowsheet of this process is shown in Fig. 1, and lab results are listed in Table 2.
Fig. 1

Flowsheet for processing high-dolomite pebbles in Florida

Table 2

Summary of lab flotation results on five dolomitic pebble samples [15]

Sample

Grinding size-200 mesh (%)

Feed % P2O5

Feed % MgO

Concentrate analysis (%)

%P2O5 recovery

P2O5

MgO

 

FLA-1

55

24.79

1.85

30.71

0.87

82.2

FLA-2

29

25.94

1.19

30.98

0.76

90.1

FLA-3

70

14.95

9.40

31.40

0.96

60.6

FLA-4

32

23.85

2.04

31.30

0.73

81.2

FLA-5

63

25.38

2.88

31.49

0.91

83.6

Since the above lab testing results were encouraging, with 20% higher recovery than most previously developed processes, FIPR approved funding for pilot testing of the CLDRI fine flotation technology. As Table 3 indicates [16], the pilot testing achieved similar results as the lab testing.
Table 3

Summary of pilot testing results on two dolomitic pebble samples

Sample

Feed % P2O5

Feed % MgO

Concentrate analysis (%)

%P2O5 recovery

P2O5

MgO

FLA-6

22.29

3.54

29.70

1.14

76.5

FLA-7

21.95

2.81

29.02

0.81

81.8

According to a feasibility study conducted by Gruber et al. at Jacobs Engineering, based on the pilot testing results, the CLDRI fine flotation technology is both technically and economically feasible for processing Florida dolomitic pebbles.

Subsequently, FIPR formulated PA-31 using local reagents and raw materials. The new collector was designated as USPA-31. Table 4 [17] shows comparative flotation results of the original Chinese PA-31 with the USPA-31 produced using raw materials from the USA.
Table 4

Performance comparison between the Chinese PA-31 and USPA-31

Dolomite collector

USPA-31

Chinese PA-31

Sample

Sample 1

Sample 2

Sample 1

Sample 2

Feed % P2O5

22.45

23.82

22.45

23.82

Feed % MgO

4.64

3.72

4.64

3.72

Concentrate % P2O5

30.31

30.82

29.94

29.64

Concentrate % MgO

0.93

0.95

0.96

1.00

% P2O5 recovery

77.03

80.12

74.19

79.62

Reagent cost, $/t conc.

1.84

1.45

3.19

2.60

The laboratory tests at FIPR demonstrated that the collectors formulated using all US-made raw materials were as good as PA-31. The collector can be used for effectively rejecting the dolomite impurity from Florida high-dolomitic pebbles. Phosphate concentrates with more than 30% P2O5 and less than 1% MgO can be produced from 3.5 to 4.5% MgO pebbles with 75 to 80% overall recoveries.

In 2009, at the request of Mosaic (the largest producer of phosphate rock in Florida), Bluestar Co. conducted dolomite flotation experiments on five (5) high-dolomite pebble samples from Florida. The process uses PA-64 (a fatty acid soap with special additives) dolomite collector developed by Bluestar and an acid mixture (P2O5:H2SO4 = 2:1). Table 5 summarizes the test results from dolomite flotation only.
Table 5

Test results on five high-dolomite pebble samples

Feed

Concentrate

P2O5%

MgO%

P2O5%

MgO%

P2O5% recovery

23.32

3.92

54.53

0.98

73.65

25.38

3.35

55.47

1.01

88.78

22.88

3.96

54.32

0.94

72.68

23.12

3.98

56.24

0.99

73.83

24.44

3.56

56.07

0.96

79.03

Encouraged by these results, Mosaic had planned to build a dolomite flotation plant at Four Corners. However, this was abruptly suspended due to the concern about the cost for the ultimate treatment of the phosphorus containing process water.

2 Option I: Reducing MgO Content in the Concentrate From the Crago Process

In the 1989 FIPR characterization study of the future phosphate resources in Florida, El-Shall and Bogan [1, 2] conducted numerous flotation experiments on flotation feed from both the upper zone and lower zone of the future phosphate deposits. As can be seen from Table 6, the MgO concentration in the upper zone would not pose a major problem. However, the concentrate in the lower zone would average 1.2% MgO, higher than required by the current phosphoric acid production plants that require less than 1% MgO flotation concentrate.
Table 6

Average chemical analyses (weight percent) of flotation concentrates

Zone

P2O5

CaO

MgO

Fe2O3

Al2O3

Na2O

F

Insol

Upper

31.9

46.6

0.43

1.4

0.85

0.62

3.7

3.6

Lower

28.6

44.8

1.21

1.6

0.70

0.69

3.4

4.8

The flotation feed from the upper zone analyzes 6.3% P2O5 and 0.12% MgO, while the feed from the lower zone averages 7.2% P2O5 and 0.67% MgO.

These results suggest that MgO will be a problem with both the pebble and concentrate from the deeper phosphate reserves. Since the ratio of concentrate to pebble will be higher in the future, reducing MgO content in the concentrate by a small margin would allow blending of a large portion of the high-dolomite pebble from the lower zone.

2.1 Experimental

2.1.1 Test Samples

Two Florida phosphate companies (A and B) made special efforts to drill core samples from their high-dolomite deposits (coded as sample A and sample B to protect the identity of the companies). These cores were washed and sized using the standard lab procedures to produce the flotation feed samples for this project. Tables 7 shows chemical analyses and size distribution of the two flotation feed samples.
Table 7

Sizing and chemical analysis of test samples

Size fraction sample A

Analysis (%)

Distribution (%)

wt%

P2O5

MgO

Insol

P2O5

MgO

+ 0.5 mm

8.62

16.96

1.46

42.42

18.81

28.28

− 0.5 + 0.3 mm

21.00

9.78

0.54

68.26

26.42

25.48

− 0.3 + 0.16 mm

61.35

6.08

0.25

80.33

47.99

34.47

− 0.16 mm

9.03

5.83

0.58

79.65

6.77

11.77

Head

100.00

7.77

0.45

74.47

100.00

100.00

Size fraction Sample B

Analysis (%)

Distribution (%)

wt%

P2O5

MgO

Insol

P2O5

MgO

+ 0.5 mm

7.55

13.75

2.14

47.43

15.87

31.33

− 0.5 + 0.3 mm

23.94

7.48

0.53

74.16

27.38

24.60

− 0.3 + 0.16 mm

58.15

5.24

0.30

82.71

46.59

33.83

− 0.16 mm

10.36

6.41

0.51

78.25

10.15

10.24

Head

100.00

6.54

0.52

77.54

100.00

100.00

2.1.2 Reagents

A 5% solution of sodium silicate (Na2SiO3) was prepared as a flotation modifier. Company B provided all the flotation reagents used in their plant, including fatty acid, amine, fuel oil, and diesel. Dolomite depressants evaluated include S711 (a special depressant developed by Bluestar), lignin, alizarin red, carboxymethylcellulose, soluble starch, sulfonated humic acid, flocculant agent NF, tartaric acid, citric acid, polyacrylamide (PAM), polymer JD, Al2(SO4)3, Hengju polymer #1–#14 (all polymeric flocculants), KClO3, JD-02, NH4Cl, alum, and sodium silicate, totaling 30. The following reagents were tested as phosphate depressants: phosphoric acid, sulfuric acid, tartaric acid, dipotassium hydrogen phosphate, sodium tripolyphosphate, and alizarin red S (ARS), triethyl phosphate, boric acid, and starch.

2.1.3 Flotation Experiment

A Denver D-12 flotation machine was used for all the laboratory flotation tests. The machine has a cell volume of 1.2 l and a conditioner volume of 0.8 l. Tap water was used in all tests. Flotation tests were conducted simulating the current practice of “double float” (“Crago”) process. The flotation feed is conditioned at about 70% solids at the desired pH level for about 5 min, and then, water is added to achieve a flotation pulp density of about 30%, followed by flotation of phosphate to completion. The rougher flotation concentrate is acid washed prior to amine flotation.

2.1.4 Mineralogical Analysis

The mineralogical compositions of sample A was examined using x-ray diffraction (XRD) analysis of powdered samples using the x-ray diffractometer Siemens D5000. The powder samples were scanned at 2θ from 5–50°, with scan speed of − 1.2°/min. The 3D mineral liberation analyses were carried out using high-resolution X-ray microtomography (HRXMT) analysis. Polarizing microscopic analysis was also used to identify mineral components.

2.1.5 Scrubbing Test

Two types of scrubbers—a steel scrubber and a glass scrubber—were used for removing dolomite by scrubbing-desliming. The diameter of the steel balls used was 0.8 mm. The scrubber was made of double stainless steel impellers. This type of scrubber can handle high-solid scrubbing, thus requiring a large sample load of 800 g. When quartz granules (ranging from 1.25 to 2 mm) were used as grinding media, a single-impeller glass scrubber was used. A smaller sample of 200 g was used in this case.

3 Results and Discussion

3.1 Mineral Compositions and Degree of Liberation

The deslimed flotation feeds contain less than 1-mm-size white, black, brown, and red particles. Polarizing microscopic analysis identified major minerals to be quartz, francolite, and dolomite, with minor amounts of feldspar and iron oxide. Detailed mineral composition of the three samples is shown in Table 8.
Table 8

Mineralogical compositions (wt%) of test samples

Sample

Francolite

Dolomite

Quartz

Feldspar

Iron oxide

Others

A

18

3

75

1

1

1

B1

15

3

78

1

1

1

B2

12

2

82

1

1

1

The phosphate (francolite) in the test samples exists in the following five structures:
  1. 1.

    Siliceous rock, consisting of francolite cemented with fine quartz particles. In this structure, cemented particles account for 70–90% of the rock.

     
  2. 2.

    Siliceous pellet rock, composed of cemented phosphate rock with a brown color due to iron contamination, and granules of francolite, quartz, and feldspar.

     
  3. 3.

    Granule rock, mainly oolitic shape francolite granules ranging from 0.2–0.8 mm in size, cemented with quartz crumbs inside, impregnated with some − 0.02-mm carbonate particles, and sometimes coated by carbonates.

     
  4. 4.

    Bulk rock, all francolite with three different colors, yellowish brown, black, or opaque.

     
  5. 5.

    Bio-formation fragments, mainly apatite with bunchy or radial shapes, having some features of microorganism structure such as animal teeth.

     

With the exception of the bio-formation fragments, the above discussed phosphate rock types all contain impregnated quartz particles of about 0.02 mm in size and carbonate particles of − 0.02 mm. This type of carbonate impregnation should have a significant effect on MgO content in the final concentrate.

There are three types of carbonate minerals in the samples, as discussed below:
  1. 1.

    Clayey dolomite, consisting of mainly dolomite in fine aggregates, colorless with some showing brownish yellow or gray color due to iron or carbon contaminations, sometimes associated with small amounts of fine silica or francolite.

     
  2. 2.

    Silica-cemented dolomite, composed of dolomite cemented with fine quartz particles and francolite granules, accounting for more than 70% of the dolomite in the test samples.

     
  3. 3.

    Sandy dolomite, fine dolomite particle aggregates cemented with various fine mineral particles such as quartz and francolite, with quartz being isometric particles ranging from 0.1 to 0.2 mm and francolite being homogeneous spherical particles of around 0.2 mm in size.

     

3.2 Mineral Liberation

The flotation feed samples were screened into different size fractions with each fraction measured for francolite liberation; the results are shown in Table 9.
Table 9

Monomer liberation measurements

Size (mm)

Francolite liberation degree (%)

Sample A

Sample B1

Sample B2

+ 0.5

78

85

83

− 0.5~+ 0.3

83

≥ 90

≥ 90

− 0.3~+ 0.16

≥ 90

≥ 90

≥ 90

− 0.16

≥ 90

≥ 90

≥ 90

These results show that liberation extent in the fine fractions does not vary much, since in actual measurements, when phosphate content in a particle is over 80%, this particle is considered to be francolite. However, fine dolomite particles are impregnated in phosphate, which could be observed under microscope after the phosphate particles are crushed. This type of dolomite is hard to liberate even with fine grinding.

The 3D mineral liberation analyses were carried out using high-resolution X-ray microtomography (HRXMT) data to classify particles in each of the samples into 12 grade classes based on both francolite and dolomite volume percent. These analyses were carried out for samples A and B1. Based on the CT data, four types of minerals (gangue, dolomite, francolite, and high-density gangue) were identified/classified and the results are presented in Table 10. The number of particles analyzed for samples A and B were 4225 and 8010, respectively.
Table 10

Mineralogical analyses by HRXMT

Minerals

Sample A

Sample B1

Volume (%)

No. of particles

Volume (%)

No. of particles

Silicate

81.84

4225

80.99

8010

Dolomite

2.11

2.80

Francolite

15.85

16.11

Others

0.19

0.09

The HRXMT 3D liberation spectra for francolite and dolomite were constructed. The spectra show the amount of mineral component of interest, in each grade class. Twelve grade classes 0%, 5%, 15%, 25%, 35%, 45%, 55%, 65%, 75%, 85%, 95%, and 100% by volume were used. The liberation-limited grade/recovery curves were then constructed from the 3D liberation spectra. Figure 2 shows dolomite liberation curve for sample A (CF feed) [18]. It should be pointed out that sample B has nearly identical dolomite liberation curve as sample A.
Fig. 2

Liberation limited grade/recovery curve for sample A

The liberation-limited grade/recovery curve represents a boundary for separation efficiency. The grade and recovery for any actual separation cannot exceed the limit imposed by this curve. In the best case, the actual grade and recovery would fall on the curve, and under these circumstances, improved separation can only be achieved with further liberation by size reduction. If the grade and recovery for an actual separation falls below the curve, then the separation efficiency is limited by other factors (mineral types, surface composition, slime coating, operating conditions, etc.) in addition to liberation limitations.

These results show that a significant portion of the dolomite in these samples is not liberated, and therefore, extensive separation/removal of dolomite will be difficult without further size reduction.

3.3 Base Line Flotation

Table 11 shows flotation results using the “Double Float” process. Since these results would be used to compare the effectiveness of different reagent systems, five parallel tests were conducted to obtain a reliable average value.
Table 11

Analysis of concentrates from the base line flotation experiment

Test sample

P2O5

MgO

Insol

P2O5 recovery

MgO recovery

Sample A

29.83

1.01

8.08

93.42

49.40

Sample B1

27.83

1.27

9.45

92.98

52.41

3.4 Testing of Dolomite Depressants

In these tests, dolomite depressants were added in the fatty acid conditioning stage after the pH modifier was added. Various dolomite depressants were tested at varying points of addition and dosages.

Among the 30 reagents tested, four of the depressants showed some potential to lower MgO content in the final concentrate. They include carboxymethyl cellulose (CMC), soluble starch, polyacrylamide (PAM), and Hengju #9 (a cationic polymer). Further flotation tests were conducted to optimize the dosage and flotation parameters for these depressants, with the results shown in Table 12.
Table 12

Complete flowsheet testing using different dolomite depressants

Depressant (kg/ton)

P2O5

MgO

Insol

P2O5 recovery

MgO recovery

Control

29.83

1.01

8.08

93.42

49.40

CMC, 1.0

30.63

1.04

6.09

88.27

44.85

Hengju #9, 0.05

29.15

1.01

11.03

94.21

46.99

PAM, 0.5

30.00

0.92

7.18

90.31

40.67

Starch, 1.0

30.48

0.89

6.69

90.42

39.74

3.5 Effect of Different Flotation Scenarios

The test results discussed above indicate that adding dolomite depressant alone in the fatty acid flotation step can lower the MgO content to some extent, but not sufficient enough to allow for blending of a large amount of high-dolomite pebbles. Therefore, other approaches were tested to achieve flotation concentrates of less than 1% MgO content.

Two dolomite collectors were evaluated for removing dolomite from both rougher concentrate and cleaner concentrate without grinding. Results showed that no MgO reduction in the final concentrate was achieved when the rougher concentrate was subjected to dolomite flotation. Flotation of the cleaner concentrate reduced MgO content in the final concentrate from 1.01 to only about 0.99%.

Size analyses (Table 7) of the flotation feeds show that a majority of the dolomite is concentrated in the small amount of + 0.5 mm pebble fraction. For example, the + 0.5-mm fraction accounts for 8.62% of sample A and contains 1.46% MgO, or 28.28%, of the total MgO; the corresponding numbers for sample B1 are 7.55%, 2.14%, and 31.33%. Grinding the flotation feed did not yield encouraging results with the final concentrate analyzing close to 1% MgO at the P2O5 recovery ranging from 76 to 86%. This approach was therefore not recommended. In another approach, the rougher concentrate was ground followed by flotation. This approach achieved a small reduction of MgO in the final concentrate, but phosphate recovery was less than 70%. In the final approach, the concentrate from the double float process was ground, followed by a dolomite flotation step to further reduce MgO content. This method resulted in the lowest MgO content with the flotation concentrate analyzing 30.54% P2O5, 0.78% MgO, and 7.75% Insol at P2O5 recovery of 92.58% and MgO rejection of about 23%.

3.6 Scrubbing Tests

Scrubbing tests were initially conducted on the rougher concentrate. After acid washing, the rougher concentrate was scrubbed and deslimed three times, which showed appreciable MgO reduction in the final concentrate. Dolomite in the slimes accounted for 28.32% of the total dolomite in the feed. Encouraged by the above scrubbing test, more scrubbing experiments were conducted to study the effect of scrubbing media type (steel balls versus quartz granules), scrubbing media addition, percent solids, and scrubbing time. The diameter of steel balls used was 0.8 mm. Scrubbing tests with steel balls on the original feed, rougher concentrate, and final concentrate were conducted. Scrubbing of the final concentrate resulted in the highest MgO reduction and the lowest loss of phosphate.

Scrubbing tests were also carried out with quartz granules on the rougher concentrate and final concentrate. Again, the best results were achieved in scrubbing the final concentrate. Overall, quartz was a better scrubbing media than steel balls in this application. Scrubbing test results are summarized in Table 13.
Table 13

Comparison of different scrubbing scenarios and conditions

Scrubbing conditions

Final product analysis

Product

Media

Time (min)

P2O5

MgO

Insol

P2O5 recovery

MgO recovery

Rougher con.

0

30

29.94

0.94

7.11

86.69

41.14

Rougher con.

25% steel balls

30

31.66

0.88

4.46

78.42

32.62

Rougher con.

40% quartz

20

30.41

0.93

7.20

90.66

40.12

Float feed

25% steel balls

30

31.39

0.86

2.90

78.15

34.07

Final con.

25% steel balls

30

30.91

0.75

4.20

81.64

30.49

Final con.

45% quartz

40

30.81

0.94

6.45

97.85

82.32

Final con.

0

40

30.76

0.95

5.92

91.52

42.23

Final con.

25% quartz

40

30.54

0.92

6.59

90.56

39.96

Final con.

50% quartz

40

30.64

0.86

6.31

89.57

37.41

Final con.

75% quartz

40

30.87

0.87

5.88

88.82

37.08

Final con.

40% quartz

40

30.33

0.87

6.32

90.15

38.30

Final con.

40% steel balls

40

30.71

0.81

6.26

72.41

28.18

Based on the extensive laboratory comparative tests, three approaches were selected as potential dolomite removal methods.

3.7 Addition of Dolomite Depressant

In this process, the flotation feed slurry at about 70% solids is first conditioned with a pH modifier and phosphate collector, as is currently practiced in Florida. The dolomite depressant, a polyacrylamide, is added prior to dilution of the slurry to 30% solids followed by flotation. The dosage of the depressant is about 1 kg/ton of feed. This process could reduce MgO content in the concentrate to about 0.81%.

3.8 Reverse Flotation of Amine Concentrate

In this process, the final concentrate from the Crago process is dewatered to about 60% solids and ground to 45.22% passing 200 mesh. The ground feed is conditioned at 30% solids with sulfuric acid (2.75-kg/ton feed) and the dolomite collector USPA-31 (1-kg/ton feed). In this manner, MgO content in the final product can be reduced to about 0.7%.

3.9 Scrubbing with Quartz Sand Media

The final concentrate from the Crago process is dewatered. The scrubbing media (quartz sand) is then added at a quartz to concentrate ratio of 1:2 by weight. The mixture is adjusted to about 60% solids and scrubbed for 40 min in a specially designed scrubber at a speed of 1500 RPM. After scrubbing, the final product contains 0.81% MgO.

Among the three approaches, the dolomite flotation process gives the lowest MgO content in the final concentrate, but it comes with low recovery and high costs. Unless it is absolutely necessary to achieve a concentrate with 0.7% or less MgO, the dolomite flotation is not recommended. The scrubbing process appears to be more promising, since it reduces the MgO content to less than 1% at an acceptable BPL recovery levels (above 90%).

4 Option II: Dolomite Flotation Without Using Phosphoric Acid

As was pointed above, use of phosphoric acid/sulfuric acid mixture in dolomite flotation increases phosphorus (P) content in the flotation process water. This has prevented the Florida phosphate industry from adopting a flotation process developed by FIPR. In order to overcome this hurdle, FIPR provided funding for developing new dolomite collectors that do not require phosphoric acid. The reagent development work was conducted by the China Bluestar Lehigh Engineering Corporation.

4.1 Experimental

4.1.1 Research Samples

The bench tests were conducted on two pebble samples: one was the high-MgO pebble from Mosaic provided by FIPR, BPL 51.16%, MgO 2.20%, and A.I. 16.75%; another pebble sample was from CF, BPL 46.39%, MgO 4.02%, and A.I. 13.99%.

4.1.2 Grinding Mill

Grinding mill model: Laboratory Model XMG-63 with three rollers and four cylinders, with the following specifications: outer size D × L = 142 × 200(mm); inner size D × L = 136 × 150(mm). Grinding medium used was steel rods with varying diameters (17, 13, and 10 mm) and the same length of 145 mm. The proportions of the grinding media were about 40%, 30%, and 30% of the 17-, 13-, and 10-mm rods, respectively. Pulp concentration in mill was 33.2% solids.

4.1.3 Flotation Cell

A model XFD-63, single-cell flotation machine was used for all lab flotation tests. The cell volume is half a liter with an impeller diameter of 44 mm. During flotation, the impeller speed was kept at 2000 RPM with air flowrate of 0.3 l/min.

4.1.4 Reagents

Industrial wet-process phosphoric acid containing 49.43% P2O5 was provided by Sanhuan Chemical Co., Ltd., China. Analytical reagent grade sulfuric acid (98% H2SO4) was used for pH adjustment and as a phosphate depressant. The acid mixture used in flotation tests was prepared by mixing the phosphoric acid and the sulfuric acid at a P2O5 to H2SO4 ratio of 2:1, as 10% w/w solution. Following dolomite collectors manufactured by Bluestar were used as 2% w/w solution—PA-64, PA-66, PA-67.

4.1.5 Flotation Experiments with Acid Mixture

Firstly, flotation experiments were conducted to establish a baseline for comparison with all the proposed approaches under this project. In these experiments, PA-64 was used as dolomite collector, and the phosphate depressant/pH modifier was a mixture of phosphoric acid and sulfuric acid at a P2O5 to H2SO4 ratio of 2:1.

4.1.6 Tests on Fineness of Grinding

It is essential that the ores should be ground to the required fineness for liberating phosphate from gangue minerals and improving separation efficiency. Results showed that when fineness of flotation feed ranged from 60 to 80% passing 200 mesh, the MgO content of concentrate could be less than 1.0%. At higher grinding fineness, MgO content and BPL recovery both decreased. Hence, fineness of grind with − 200 mesh 70% was selected for most flotation tests.

5 Results and Discussion

5.1 Comparison of the Original Collector (PA-64) with the Newly Developed Collector (PA-66)

Table 14 shows that the new collector is suitable for dolomite flotation using sulfuric acid only, achieving both higher recovery and better product grade.
Table 14

Flotation performance of two collectors at optimized dosage in sulfuric acid media only

Collector

Product

wt%

Analysis (%)

Distribution (%)

BPL

MgO

Insol

BPL

MgO

Insol

PA-64

Con.

76.47

54.80

1.05

18.86

80.40

36.53

85.53

Tail 2

6.89

47.99

3.90

11.45

6.34

12.22

4.68

Tail 1

16.64

41.54

6.77

9.92

13.26

51.25

9.79

PA-66

Con.

76.21

55.85

0.87

18.97

81.22

29.34

86.13

Tail 2

4.89

49.63

3.90

11.40

4.63

8.45

3.32

Tail 1

18.90

39.25

7.44

9.37

14.15

62.21

10.55

5.2 Effect of Reagents Dosages

Figure 3 shows the effect of sulfuric acid addition on BPL recovery and MgO content in the final concentrate, while Fig. 4 indicates the influence of collector dosage.
Fig. 3

MgO content of concentrate and BPL recovery as functions of sulfuric acid dosage in dolomite flotation stage I for Mosaic pebble

Fig. 4

MgO content of concentrate and recovery rate of BPL as functions of PA-66 dosage for Mosaic pebble

5.3 Flotation Plant Water Analysis

Water of concentrate and tailings in CF pebble flotation tests with sulfuric acid and acid mixture were analyzed to investigate the change of main composition. The test results are given in Table 15.
Table 15

Analysis of flotation water of concentrate and tailings from CF pebble

Media

Water samples

pH

Ca2+(mg/L)

Mg2+(mg/L)

SO42‑(mg/L)

PO43−(mg/L)

Sulfuric acid

Concentrate

6.38

499.48

114.34

1158

102

Tailing

5.89

696.85

152.94

2251

135

Acid mixture

Concentrate

5.45

321.44

58.15

548

837.6

Tailing

5.33

398.78

94.30

739

1036

Table 15 shows that when sulfuric acid was used alone, PO43− ion concentration in process water is an order of magnitude less than that in phosphoric acid/sulfuric acid mixture. Therefore, the use of sulfuric acid can dramatically reduce the costs for P-containing wastewater treatment.

5.4 Results with Further Optimized Dolomite Collector, PA-67

Flotation results with the new collector PA-67 are plotted in Fig. 5.
Fig. 5

Relations between MgO content and BPL recovery of final phosphate concentrate using new collector PA-67

This figure clearly shows that PA-67 can achieve higher flotation recovery than any of the previous versions of dolomite collector, and at the same level of MgO content in the final product.

6 Option III: Nonchemical Processing Using Packed Column Jig

The packed column jig (PCJ) as a gravity separation device was patented by Yang [19]. PCJ is a gravity separator and operates like a teetered bed separator [20], but the internal packing and a pulsating water flow allows higher throughput capacity and sharper separation. Packing also limits the turbulence and vorticity that takes place in each cell, thus avoiding short-circuiting of coarser particles. This leads to a successful operation with fine particles extending the size range down to a few microns in diameter. PCJ does not require reagents to operate, has a higher capacity per unit cell volume than its competitors, gives sharp separation between ores and silica or other impurities, and is easily scaled up.

Since its invention, the fundamentals underlying PCJ have been delineated and published [21], and small-scale experiments have been carried out on different minerals with varying degree of success [22, 23]. However, the full potential of PCJ has been demonstrated on both pilot scale and commercial scale, only since Mineral Technologies International started collaboration with Professor Wending Xiao of Wuhan Institute of Technology 3 years ago. During this short 3-year period, under the leadership of Professor Xiao, rapid progresses were made to advance the PCJ technology, as highlighted below:
  1. 1.

    Two continuous lab-scale PCJ testing systems were built for testing numerous phosphate samples as well as metallic ores and tailings.

     
  2. 2.

    A portable PCJ pilot plant was constructed and installed on a trailer. The pilot plant comes with full computer control and has been transported to four mine sites for separating dolomite from phosphate, separation of silica from phosphate, and processing metal tailings.

     
  3. 3.

    One industrial-scale PCJ has been installed for processing of a metallic ore.

     
  4. 4.

    Large-scale pilot testing has been conducted to demonstrate the feasibility of PCJ for upgrading phosphogypsum.

     
  5. 5.

    One industrial plant is under construction for separation of dolomite from phosphate.

     

6.1 The PCJ Gravity Separation Technology and Its Recent Achievements

The working principle of PCJ is schematically illustrated in Fig. 6. Detailed descriptions on the fundamentals, operating mechanisms, and control strategies for the PCJ were reported by Yang et al. [23].
Fig. 6

Schematic illustration of PCJ working principle

In summary, PCJ offers the following unique features: long, nearly unlimited separation zone; small footprint; minimal water use; no chemicals, thus being benign to the environment; high throughput; and wide particle size range without limit on the fine side. Recent semi-industrial-scale testing demonstrated the superior performance of the jig over heavy media separation for processing a high-dolomite phosphate ore in China. The following two examples demonstrated PCJ’s potential for nonchemical beneficiation of phosphates.

Example 1: Removal of Silica and Other Gangue Materials From Phosphate

In this application, a high-silica phosphate ore is separated into four size fractions prior to gravity separation using PCJ. Results in Table 16 show close to 90% recovery for every fraction with an acceptable P2O5 grade. These results have not been achieved for gravity separation of phosphate from silica using any other technology.
Table 16

Results on separating SiO2 from phosphate

Size range, mesh

Feed analysis

Concentrate analysis

wt%

% P2O5

% yield

% P2O5

% P2O5 recovery

− 20 + 40

32.34

23.09

70.71

28.88

88.45

− 40 + 100

30.86

22.69

75.84

26.73

89.34

− 100 + 200

10.13

21.99

65.14

28.67

84.92

− 200

26.67

14.88

100.00

14.88

100.00

Example 2: Removal of Dolomite From a Phosphate Ore in Hubei Province

Recent large-scale testing demonstrated the superb performance of PCJ for removing dolomite from phosphate (Table 17).
Table 17

Results on separating dolomite from phosphate

Feed analysis

Jigging concentrate analysis

% P2O5

% MgO

% P2O5

% MgO

15.72

5.36

31.26

1.11

15.87

5.74

29.26

1.16

15.55

5.36

28.78

1.29

16.33

5.27

27.98

1.29

6.2 The Florida Effort to Solve the Dolomite Effort Without Using Chemicals

FIPR recently funded a major project for both laboratory and pilot-scale evaluation of PCJ for separating dolomite from phosphate in Florida high-dolomite phosphate pebbles. A fully computer-controlled lab testing system has been installed for continuous separation experiments. If the lab testing results look promising, a 1-ton/h pilot plant will be constructed for on-site testing.

Initial quick testing demonstrated the potential to remove dolomite from phosphate pebbles. Testing on the 1 by 2-mm fraction of a high-dolomite pebble sample showed efficient dolomite removal by visual observation.

A practice separation on a ground sample also demonstrated recovery of a final product containing only about 1% MgO, Table 18.
Table 18

Preliminary results for separation of dolomite from apatite using PCJ

Sample ID

% P2O5

% Insol

% MgO

% Al2O3

% Fe2O3

% CaO

Jig T1 high MgO

20.97

17.10

3.33

1.54

1.02

34.90

Jig T1 conc

22.86

24.75

1.06

0.79

0.69

34.26

Jig T1 tails

18.11

13.06

5.45

2.19

1.30

33.68

7 Summary Remarks

Several different approaches for apatite-dolomite separation have shown promise for achieving the goal of less than 1% MgO in the phosphate rock feed to the phosphoric acid plants. Most promising of the various approaches developed, so far, are as follows.
  1. I.

    Scrubbing of the Crago process concentrate: This approach includes production of the phosphate concentrate using the Crago process, as is currently practiced, followed by scrubbing of the concentrate using quartz sand. This approach yielded a final product of less than 1% MgO. The scrubbing process offers the following three major advantages: (1) it does not require any chemical; (2) the quartz sand used as scrubbing media is inexpensive and reusable; and (3) the operating cost is low.

     
  2. II.

    Grinding of the high-MgO pebble fraction with subsequent reverse flotation of dolomite using a newly developed collector, and sulfuric acid as the phosphate depressant/pH modifier. This approach resulted in a concentrate that can be blended with the other fractions to yield an overall product with acceptable levels of phosphate recovery and MgO content.

     

Feasibility studies show that both of the above approaches are promising solutions to the dolomite problem in Florida. The practicality and economic feasibility look promising, but remain to be proven at the commercial scale. Most recently, the packed column jig (PCJ), developed in the mid-1990s—a gravity separation technique—has shown promise in separating dolomite from apatite and may prove to be the ultimate solution to the problem through an on-going research program.

Notes

Funding Information

All the research projects presented in this paper were funded by the Florida Industrial and Phosphate Research Institute, Florida Polytechnic University. All Florida phosphate companies participated in the research projects and provided technical input and in-kind supports.

Compliance with Ethical Standards

Conflict of Interest

The authors declare that they have no conflict of interest.

Glossary

PA-31

The earliest version of a dolomite collector developed for processing Florida high-dolomite phosphate pebbles and manufactured in China. It is a fatty acid soap with a surfactant.

PA-46

The second generation of PA-31, a fatty acid soap with a different source of fatty acid.

PA-66

The third generation of PA-31, a fatty acid soap formulated for dolomite flotation without adding phosphoric acid.

PA-67

An improved version of PA-66, a fatty acid soap formulated for improved selectivity by using higher quality fatty acid and a new surfactant.

USPA-31

A dolomite collector produced using raw materials from the USA based on the formulation of PA-31.

References

  1. 1.
    El-Shall H, Bogan M (1994) Characterization of future Florida phosphate resources. FIPR Publication No. 02-082-105, Bartow, Florida, USAGoogle Scholar
  2. 2.
    El-Shall H, Bogan M (1994) Evaluation of dolomite separation techniques. FIPR Publication No. 02-094-108, Bartow, Florida, USAGoogle Scholar
  3. 3.
    Rule AR (1985) Beneficiation of complex phosphate ores containing carbonate and silica gangue. Proc XVth IMPC 3:380–389Google Scholar
  4. 4.
    Hsieh SS, Lehr JR (1985) Beneficiation of dolomitic Idaho phosphate rock by the TVA diphosphonic depressant process. Miner Metall Process 2:10–13Google Scholar
  5. 5.
    Snow RE (1982) Flotation of phosphate ores containing dolomite. U. S. Patent 4,364,824Google Scholar
  6. 6.
    Gruber GA et al. (2001) A pilot-scale demonstration of the IMC/CLDRI/FIPR flotation process for Florida high-MgO pebble. FIPR Publication #02-133-178, Bartow, Florida, USAGoogle Scholar
  7. 7.
    Smani S, Cases JM, Blazy P (1975) Beneficiation of sedimentary Moroccan phosphate ore, part 3. Selective flotation and recovery. Trans SME-AIME 258:176–180Google Scholar
  8. 8.
    Houot R (1982) Beneficiation of phosphatic ores through flotation: Review of industrial applications and potential developments. Int J Miner Process 9:353–384CrossRefGoogle Scholar
  9. 9.
    Lehr JR, Hsieh SS (1981) Beneficiation of high carbonate phosphate ores. U S Patent 4,287,053Google Scholar
  10. 10.
    Lawendy TAB et al (1993) Flotation of dolomitic and calcareous phosphate ores. In: El-Shall H, Moudgil B, Wiegel R (eds) Beneficiation of phosphates: theory and practice. SME, Littleton, pp 231–243Google Scholar
  11. 11.
    Filho L et al (1993) The influence of corn starch on the separation of apatite from gangue minerals via froth flotation. In: El-Shall H, Moudgil B, Wiegel R (eds) Beneficiation of phosphates: theory and practice. SME, Littleton, pp 147–155Google Scholar
  12. 12.
    Wiegel R (1999) Phosphate rock beneficiation practice in Florida. In: Zhang P, El-Shall H, Wiegel (eds) Beneficiation of phosphates: advances in research and practice. SME, Littleton, pp 271–275Google Scholar
  13. 13.
    Zhang P, Snow RE, Bogan M (2002) A screening study on phosphate depressants for beneficiating Florida phosphate minerals. FIPR Publication 02-101-183, Bartow, Florida, USAGoogle Scholar
  14. 14.
    Zhang P, Zheng SB, Sun WY, Ma XQ, Miller JD (2012) Development of reagent schemes for reducing MgO content in the flotation concentrate for processing Florida’s high-dolomite phosphate deposits. FIPR Publication # 02-179-246, Bartow, Florida, USAGoogle Scholar
  15. 15.
    Gu ZX, Gao ZZ, Hwang C (1999) Development of new technology for beneficiation of Florida dolomitic phosphate resources, FIPR Publication No. 02-129-167, Bartow, Florida USAGoogle Scholar
  16. 16.
    Gruber G, Zheng SB, Hwang C (2001) A pilot scale demonstration of the IMC/CLDRI/FIPR flotation process for Florida high-MgO Pebble. FIPR Publication No. 02-133-178, Bartow, Florida, USAGoogle Scholar
  17. 17.
    Gao ZZ, Zheng SB, Guan C, Hwang C (2003) Optimizing the formulation for dolomite collector PA-31 using raw materials from the United States. FIPR Publication No. 02-150-197, Bartow, Florida, USAGoogle Scholar
  18. 18.
    Miller J, Lin CL, Ahmed I, Wang X, Zhang P (2012) Advanced instrumentation for mineral liberation analysis and use in phosphate industry. In: Zhang P, Miller J, El-Shall H (eds) Beneficiation of phosphates: new thought, new technology, new development. SME, Englewood, pp 167–176Google Scholar
  19. 19.
    Yang DC (1996) Device and process for gravitational separation of solid particles. US Patent No. 5,507,393Google Scholar
  20. 20.
    Drummond R, Nicol S, Swanson A (2002) Teetered bed separators -the Australian experience. J South Afr Inst Min Metall 10:385–392Google Scholar
  21. 21.
    Dai Q (1999) Simulation of packed column jigging. Master’s Thesis, West Virginia University, Morgantown, West VirginiaGoogle Scholar
  22. 22.
    Yang DC, Meloy TP (1997) Gravity separation of minus 500 mesh particles. Presented at SME Annual Meeting. Denver, ColoradoGoogle Scholar
  23. 23.
    Yang DC, Bozzato P, Ferrara G (2003) Iron ore beneficiation with packed column jig. J Miner Mater Charact Eng 2(1):43–51Google Scholar

Copyright information

© Society for Mining, Metallurgy & Exploration Inc. 2019

Authors and Affiliations

  • Patrick Zhang
    • 1
    Email author
  • Shibo Zheng
    • 2
  • Wenyi Song
    • 2
  • Chunhui Feng
    • 2
  • Brij Moudgil
    • 3
  • Wending Xiao
    • 4
  • Dapeng Zhang
    • 4
  1. 1.Florida Industrial and Phosphate Research Institute (FIPR)Florida Polytechnic UniversityBartowUSA
  2. 2.China Bluestar Lehigh Engineering CorporationLianyungangChina
  3. 3.Department of Materials Science and EngineeringUniversity of FloridaGainesvilleUSA
  4. 4.Wuhan Institute of TechnologyWuhanChina

Personalised recommendations