- Email: [email protected]
The influence of phosphate on the properties of clay bricks
Applied Clay Science 10 (1996) 461-475
The i:nfluence of phosphate on the properties of clay bricks V.T.L. Bogahawatta Geomaterials
Unit, School of Engineering,
‘, A.B. Poole
QMW, Universiry of London, London El 4NS, UK
Received 20 March 1995; accepted 3 November 1995
Abstract The fabrication, properties and composition of heated phosphate-bonded clay bricks are reported. The strength of phosphate-bonded clay bodies is shown to be related to the equilibrium pH of the clay mix. The highest strengths were obtained for mixes with equilibrium pH 7 fired at 500°C whllch gave a modulus of rupture 60% higher than the unmodified clay heated under similar conditions but at 800°C firing temperature. The phosphate-bonded products were also of low porosity and improved dimensional stability. An insight into the probable mechanism of strength development of the heated product has been obtained from XRD, SEM and DSC studies. The results show that the properties of the phosphate-bonded clay bodies are critically affected by new mineral phases resulting from the reactions of phosphates with clay which subsequently undergo physico-chemical changes above 550°C.
1. Introdluction
It is economically desirable to reduce the sintering temperature of structural clay products which partially or fully vitrify at the relatively high temperature used in their manufacture. In this respect, the use of bonding materials which react chemically with the matrix particles at low temperatures is attractive for relevant applications. The literature indicates that phosphoric acid and certain phosphates offer good prospects as bonding agents (Kingery, 1950). The bonding property of phosphates with refractory oxides (Bartha et al., 1971) and with silicates (Forest, 1964; Robinson and Segnit, 1967; Tauber et al., 1972) is widely used in the production of refractories of high
’ Present address: NBRO, 99/l Jawatta Road, Colombo 5, Sri Lanka. 0169-1317,/96/$15.00 0 1996 Elsevier Science B.V. All rights reserved SSDI 0169-1317(95)00042-9
462
V.T.L. Bogahawatta, A.B. Poole/Applied
Clay Science 10 (19961461-475
strength, refractory cements and mortars (Kingery, 1952). However, additive systems involving phosphoric binders are little used in fired clay products presumably because of the ease at which strengths are achieved using high temperature and soaking times. In spite of many possible applications, the properties of refractory bonds employing phosphates have not been systematically studied. Kingery (1950), Sudakas and Sinina (1971) and Robinson and McCartney (1964) ascribe the formation of cold-setting and heat-setting bonds in refractory oxide-phosphate systems to thermal dehydration and chemical reactions of phosphates. However, unlike the uniform metal oxide surfaces, clays are characterized by low temperature dehydration with edge and face interactions which make the above mentioned concepts difficult to apply to clay minerals (Secor and Radke, 1985, Bar-Yosef et al., 1988). Reactions between phosphates and soil minerals have received considerable attention from soil scientists, but our understanding of the phosphate-soil reactions remains inadequate. The general consensus is that phosphates react chemically with aluminium containing minerals to form surface compounds as well as insoluble Al-phosphates. However, the details of the mechanisms involved and of the phosphate phases formed are matters of continuing discussion.
2. Experimental
method
The primary objective of this investigation was to develop satisfactory chemically phosphate-bonded clay products with low energy intensive fabrication techniques. Therefore, emphasis was placed upon determining the dependence of strength and other physical and mechanical properties of experimental clay bodies on heating temperature. The experimental studies were designed to determine the heating conditions most favourable for the phosphate/silicate/aluminate reactions in relation to the properties of the finished products, and to explore the possibilities of reducing the temperatures below those normally used in the brick industry.
3. Materials
and methods
3.1. Materials The experiments were limited to a single soil sample (K). This poorly ordered kaolinitic residual sand/silt/clay soil is extensively used by the brick industry in Sri Lanka. It is composed of 30% clay, 25% silt and 45% sand grade materials by weight. It was evenly graded with a maximum particle size of 2 mm. The dominant constituents in the clay grade are kaolinite (46%), quartz (31%) and feldspars (15%). The accessory minerals include vermiculite and hypersthene. The two analar grade phosphatic binders used in the investigation were: 1. Orthophosphoric acid (H,PO,-weight per ml 1.75 g). 2. Diammonium hydrogen phosphate [(NH,),HPO,].
V.T.L. Bogahawatta, A.B. Poole/Applied
Clay Science 10 (1996) 461-475
463
3.2. Method of fabrication The effects of phosphate treatment and the equilibrium pH value of clay mixes prepared under different maturing conditions on the final properties of fired clay products were investigated in relation to the clay-H,PO,-(NH,),HPO, system. For this purpose, seven clay mixes were prepared using different experimental conditions. 1 M :solutions of orthophosphoric acid (pH 0.9) and diammonium hydrogen phosphate (p.H 7.6) were prepared using deionised distilled water and mixed in predetermined proportions to adjust the pH of the bulk solution to desired levels. The levels of pH were selected to give equilibrium pH values ranging from 3 to 7 to clay mixes when treated with a fixed volume of phosphate solution. This volume which imparts proper moulding consistency to a clay mix was predetermined from blank trials. It was found from control tests that 750 g of air-dried clay required 250 ml of 1 M phosphate solution to attain the appropriate moulding consistency. The equivalent composition of 3.16 wt% phosphate ion concentration was maintained throughout the series of clay mixes. A weighed sample of air-dried, disaggregated clay with the required volume of phosphate solution was mixed for 15 minutes. The clay mix was then covered and allowed to mature at room temperature for 20 hours, 40 hours and 90 hours to allow dispersion the additives and to initiate reactions. At the end of each period of maturing, the clay mix was puddled again for 10 min and moulded into a steel mould (160 mm X 30 mm X 15 mm) with a steel plunger in a simple mechanical lever press at 0.5 N/mm2. A uniform moulding moisture content of 20% by weight of dry clay was maintained in all the mixes. The moulding moisture content and equilibrium pH of the clay mix were determined according to BS 1377 (1990). Three batches of specimens were prepared from a clay mix of equilibrium pH 5, with maturing times of 20 hours, 40 hours and 90 hours. A second series consisting of four batches of specimens were prepared from clay mixes of equilibrium pH 3 and 7; with maturing times corresponding to each level of pH of 20 hours and 90 hours. Control specimens were also prepared
Table 1 Composition of experimental on bonding properties Weight dry
mixes prepared to investigate
the effects of equilibrium
pH and time of maturing
Specimen No.
clay (g)
Volume of lM Phos. Soln. (ml)
Mix proportions by wt. H,PO,: (NH,),HP~,
Initial pH of Phos. Soln.
Initial pH of clay mix
Equil. pH of clay mix
Maturing time (h)
K13/1 K13/7 Kl3/13 Kll/l Kl l/7 K17/1 K17/7
750 750 750 720 705 705 705
250 250 250 240 235 235 235
1S:l 1.5:1 1.5:1 14:l 14: 1 1:18 1:18
2.9 2.8 2.8 1.5 1.4 7.5 7.6
6.7 6.7 6.8 6.8 6.8 6.8 6.8
5.1 5.0 4.7 3.0 3.3 7.4 7.4
90 40 20 20 90 90 20
Moulding moisture content relative to dry clay = 20%. Phosphate ion concentration = 3.16 wt% for all specimens.
464
V.T.L. Bogahawatta, A.B. Poole/Applied
Clay Science 10 (1996) 461-475
Table 2 Composition of experimental mixes prepared to investigate the effect of the amount of phosphate the modulus of rupture of heated bars (parts by weight of dry materials) Designation of specimens
Clay (parts by weight)
Phosphate ion concentration (parts by weight)
Average MOR of bars heated at 500°C (N/mm*)
K17 K27 K2.57 K37
100 100 100 100
3.16 6.32 7.90 9.48
2.55 3.57 4.00 4.25
Moulding
(3.06%) (5.94%) (7.32%) (7.66%)
binders on
moisture content 20%.
from an untreated clay mix for comparison purposes. Each batch consisted of at least 12 rectangular bars. The composition of the experimental mixes are presented in Table 1. The specimens were air dried, then oven dried for 24 hours at 105°C. The linear drying shrinkage for each batch of specimens before heating is given in Table 4 as % length. The bars were then heated in a laboratory electric muffle furnace in air at temperatures from 400” to 900°C. The temperature difference between the different firings being typically 100°C. The heating rate was 5.6”C/min with a soaking period of 2 hours at the peak temperature for all firings. Heated specimens were tested for modulus of rupture, bulk density, shrinkage and durability in accordance with appropriate BSI or ASTM methods which were modified where necessary to take account of the specimen size (Table 4). 3.3. Optimisation
of processing
conditions
Having determined the critical equilibrium pH value, time of maturing and minimum heating temperature required to obtain an acceptable product, the second part of the study was designed to ascertain the optimum concentration of phosphate. Four clay mixes were prepared incorporating 1 M, 2 M, 2.5 M and 3 M solutions of orthophosphoric acid and diammonium hydrogen phosphate at the optimum equilibrium pH determined in the first part of the study. The experimental bars prepared as described above, matured for 90 hours and heated at 500°C with 2 hour soaking period. Table 2 presents the composition of the experimental mixes and the results of modulus of rupture of (MOR) heated bars.
4. Results and discussion 4.1. Effect of phosphate
addition on the strength
Mean values for modulus of rupture as a function of heating temperature are shown in Table 3 for the three clay mixes together with the effect of variation of maturing time
V.T.L. Bogahawatta, A.B. Poole/Applied
Clay Science 10 (1996) 461-475
Table 3 Modulus 01’rupture of tired phosphate-bonded
clay bars heated at temperatures
Conditions
Modulus of rupture in N/mm2
at heated temperature
of preparation
465
between 400” and 900°C shown
Eq. pH of Mix
Mat. time (h)
400°C
500°C
600°C
700°C
800°C
900°C
:Kll) fK13)
90 20 40 20
1.54 1.40 1.15 1.92
1.19 1.53 2.08 1.42
0.70 1.37 1.19 1.71
1.24 1.09 1.17 1.24
1.19 1.15 1.18 1.66
1.17 1.14 1.34 1.96
90
0.57
1.32
0.96
0.67
0.72
0.92
90 20 20
2.41 1.98
2.32 2.55
2.10 1.94
2.01 1.86
1.56 1.99 1.60
1.72 1.85 1.73
;iK17) 7 unmodified
clay (K2A)
on strength. An unmodified clay specimen hated at 800°C and 900°C is included for comparis;on. Coefficients of variation lay between 4 and 5%. Marked changes in modulus of rupture with equilibrium pH of the clay mix and firing temperature are shown by the specimens. Highest strengths (2.32-2.55 N/mm’) were associated with clay bodies prepared at an equilibrium pH 7.4 (K17). Specimen bars representing the clay mix of equilibrium pH 5 (K13, maturing time 20 hours) had intermediate strengths whereas an analogous set of bars from the clay mix of equilibrium pH 3 (Kl 1) had the lowest strengths. In general, the modulus of rupture increased sharply when the temperature was raised from 400” to 500°C at which a maximum was attained. Above this temperature, modulus of rupture of both K17 and K13 bodies dropped steeply while that of the Kll body dropped gradually as the temperature was increased to about 600-700°C. An apparent increase in strength above 700°C was discernible in the K13 body. Bodies heated at the optimum temperature of 500°C gave strengths of the order of 60% higher than the normal clay bodies heated at 800°C (1.6 N/mm2 1. Clay mixes within each batch exhibited similar trends irrespective of their extent of maturing. The effect of variation in maturing time of clay mix on the modulus of rupture of fired bars is also illustrated in Table 3. A substantial increase in modulus of rupture with increasing maturing time from 20 hours to 90 hours was evident in the two clay bodies K17 and Kl 1. The gain in strength due to prolonged maturing of the K17 body heated at 500°C was approximately 10%. In contrast, the strength of the clay body prepared at the equilibrium pH 5 (K13), deteriorated gradually on maturing; the strength dropped from 2.08 N/mm2 to 1.32 N/mm’. Since specimen preparation was identical in all cases this anomalous behaviour is difficult to explain but may mark a threshold pH value between acidic phosphate-clay interactions and those occurring in neutral conditions. The influence of the phosphate concentration on the modulus of rupture at a constant temperature (of 500°C) is given in Table 2, which shows a gradual increase in modulus of rupture up to the maximum phosphate addition of 7.66% by weight. However, these strength results are very much better than the heated unmodified clay even at the lowest phosphate ion concentration of 3.06%. The clear pH dependence of strength can be interpreted in terms of the relative
466
V.T.L. Bogahawatta, A.B. Poole/Applied
Clay Science 10 (1996) 461-475
Table 4 Physical properties of phosphate-bonded clay bars heated at temperatures between coeff. = % absorption 24 hour cold water/% absorption 5 hour boiling water) Designation of clay mix
Property
Kll
Bulk density (g/cm’)
Average value of property at temperatures
% Absorption % Asorption
(24 h) (5 h b)
Sat. coefficient % Weight loss % Drying shrinkage % Firing shrinkage
K13
Bulk density (g/cm3)
% Absorption
(24 h)
% Absorption
(5 h b)
Sat. coefficient
% Weight loss
% Drying shrinkage
% Firing shrinkage
K17
Bulk density (g/cm”) % Absorption
(24 h)
% Absorption
(5 h b)
Sat. coefficient Weight loss
400°C and 900°C (Sat.
400°C
500°C 1.45 1.51 20.06 19.51 24.60 24.06 0.81 0.81 4.38 4.43
20 90 20 90 20 90 20 90 20 90 20 90 20 90
h h h h h h h h h h h h h h
mat mat mat mat mat mat mat mat mat mat mat mat mat mat
1.52 1.52 19.22 18.64 23.24 22.56 0.83 0.83 3.15 3.09
20 40 90 20 40 90 20 40 90 20 40 90 20 40 90 20 40 90 20 40 90
h h h h h h h h h h h h h h h h h h h h h
mat mat mat mat mat mat mat mat
1.54 1.52
mat mat mat mat mat mat mat mat mat mat mat mat
20 90 20 90 20 90 20 90 20 90
h h h h h h h h h h
mat mat mat mat mat mat mat mat mat mat
_ _
1.48 19.68 19.28 20.37 23.65 23.03 24.28 0.83 0.84 0.84 3.40 3.16 3.34
0.67 _
1.63 1.61 16.46 16.60 20.82 20.78 0.79 0.80 3.64 3.64
_
600°C
700°C
1.50 1.49 1.50 1.48 20.23 21.24 20.52 21.40 25.03 25.70 25.27 25.95 0.81 0.82 0.81 0.82 5.24 6.67 5.69 6.81 Average = 6.25 Average = 6.45 _ _ 0.67
-
1.46 1.49 1.46 20.85 20.92 21.79 25.43 25.69 26.71 0.82 0.81 0.81 5.15 4.88 5.86
1.46 1.48 1.49 1.45 1.41 1.46 21.87 22.33 21.83 22.12 21.99 22.68 26.53 26.89 26.79 26.93 26.48 27.29 0.82 0.83 0.81 0.82 0.83 0.83 6.70 7.56 6.18 7.16 6.28 7.05 Average = 7.29 Average = 6.77 Average = 6.25 0.67 _ _ 0.67 _
1.62 1.60 17.07 17.48 21.44 21.86 0.80 0.80 4.98 4.98
1.58 1.58 18.44 18.37 23.36 23.24 0.79 0.79 6.87 6.87
1.56 1.53 18.74 19.05 23.53 23.72 0.80 0.80 7.7 1 7.71
shown
800°C
900°C
1.45 1.48 22.40 21.78 26.75 26.29 0.84 0.83 7.33 7.29
1.47 1.45 21.41 21.39 26.32 26.58 0.81 0.82 7.38 7.58
1.33 _
0.67
1.48 1.43 1.46 22.51 22.92 23.10 27.04 27.46 27.45 0.83 0.83 0.84 7.86 7.49 7.57
1.46 1.49 1.43 21.60 21.83 22.79 26.80 26.90 27.72 0.83 0.81 0.82 8.23 7.76 7.93
0.67 _ 0.67
1.33 1.34 0.67
1.56 1.52 19.17 19.30 23.74 23.65 0.81 0.81 7.99 7.99
1.57 1.57 19.10 19.11 24.21 23.89 0.79 0.80 8.37 8.37
V.T.L. Bogahawatta, A.B. Poole/Applied
467
Clay Science 10 (1996) 461-475
Table 4 (continued) Designation of clay mix
Average value of property at temperatures
Property
% Drying shrinkage % Firing shrinkage
20 90 20 90
h h h h
mat mat mat mat
400°C
500°C
600°C
0.68 0.68
Average = 8.65 Average = 8.75 0.68 0.68 0.68 0.68
shown
700°C
800°C
900°C
-
0.68 -
0.68 0.68
abundance of OH groups available for structure formation and the extent of collapse of kaolinite lattices at different levels of pH. The optimum strength achieved at 500°C may be attributed to the favourable structures formed by reactions of phosphates with the clay minerals, since the dehydration of kaolinite occurs in the temperature range 450°C to 525°C. pH appears to have a significant effect on the development of these phosphate-clay reactions Rearrangement of these metastable structures at temperatures above 500°C result in loss of strength. 4.2. The physical properties
of the heated specimens
The bulk density of heated specimens was measured in accordance with ASTM C373-88, water absorption as a percentage of the original weights (24 hours in cold water and 5 hours in boiling water) in accordance with ASTM C67-92a while the percentage weight loss on heating the dried specimen and linear drying shrinkage and the additional linear shrinkage on firing were measured directly. The results expressed as percentages are given in Table 4. These results show that there is a small irregular decrease in bulk density, as heating temperature increases, for specimens from all three mixes. There is also a correlation between modulus of rupture values and bulk density (Fig. 1). The 24 hour water absorption values similarly show small increases with heating temperature for all mixes with a maximum at 800°C. This is interpreted as a gradual increase in porosity with temperature up to 800°C after which sintering, which was confirmed by microscopic examination of the specimens, begins to decrease the porosity. The wet-to-dry shrinkage of bars from the demoulded stage (moisture content 15% by mass) to the oven-dry stage (105°C 24 hours) depended on the composition of the clay mix. Experimental bars made from the three clay mixes Kll, K13 and K17 had wet-to-dry linear shrinkage of 6.3%, 6.8% and 8.7% respectively. However, these phosphate-bonded bodies had remarkably low firing shrinkages when compared with those of conventional clay bodies. The fired clay body K17 had a consistently low dry to fired shrinkage of 0.67% over the range of heating temperature from 400” to 900°C. The low shrinkage of the phosphate-modified heated clay may be due to the quartz type inversions of aluminium phosphate which is believed to occur at lower temperatures than in fsilica (Beck, 1949). The samples heated at 500°C yielded DSC and SEM evidence of the formation of such amorphous products. Differential shrinkage behaviour
468
V.T.L. Bogahawatta, A.B. Poole/Applied
Clay Science 10 ~19%) 461-475
2.5
1.40
t .45
1.50
Bulk
Fig. 1. The relationship determinations.
between modulus
1.55
denslty
II
20 h maturing
??
go h maturing
1.60
1.65
1.70
(glcm3)
of rupture and bulk density. Each point is the mean of at least 3
of the three clay sample types may indicate varying degrees of formation of reaction products in them. An alternative suggestion (R. Bown, pers. commun.) is that the phosphates deflocculate the clay minerals, leading to a closer packing of the drying block and hence greater strength through greater surface contact. The deflocculation effect is greatest at pH 7-8, and hence strength is greatest for specimen (K17). 4.3. The durability of phosphate-bonded
samples
In order to test durability, phosphate-bonded bars prepared from the most promising composition (K17) and heated under the optimum conditions (500°C for 2 hours) were randomly sampled for testing under laboratory conditions. Bars made from a clay mix without any phosphate binder and heated at the same temperature and samples of engineering brick (LBC) satisfying BS specifications were also taken to compare the behaviour of the respective bodies under similar conditions of accelerated weathering. Twelve specimens from each batch were subjected to the test. The procedure used in this study was first to dry the heated specimens at 110°C for 24 hours. They were then immersed in water for 2 hours and after removal placed in an oven overnight at 105-l 10°C. The following morning the samples were cooled in a desiccator and the cycle repeated. A total of 60 cycles of alternate wetting and drying were completed within a period of three months. After the last cycle, specimens were carefully examined for cracks, disintegration and leaching of compounds. The accelerated durability test showed that both the test and control samples did not undergo cracking or disintegration after 60 cycles of test. Weight losses were relatively insignificant. However, there was evidence of leaching of compounds from the test samples with consequent discolouration of external surfaces. In a more severe test of durability a total of 144 samples from the three clay mixes
V.T.L. Bogahawatta, A.B. Poole/Applied
Clay Science 10 (1996) 461-475
469
(Kl 1, K13 and K17) heated at 400”, 500”, 600”, 700”, 800” and 900°C were subjected
to a 5 hour boiling test. The test showed that none of the specimens disintegrated during test indicating the adequacy of the heat treatment and bonding to impart sufficient durability to heated clay. However, water absorption of the specimens increased on boiling. The lowest saturation coefficient is associated with the K17 bars heated at 600°C (Table 4). 4.4. The mineralogy
of the unfired clay
X-ray diffraction methods were used to detect new crystalline phases formed. Heated products were investigated using finely ground randomly orientated samples. Analyses were carried out using a Philips diffractometer with CuK CYradiation and a graphite monochromator. The phosphate treated clays exhibited clear changes in their diffraction patterns with changes in equilibrium pH of the original mix. Two additional reflections appeared in the trace made with phosphate tr:ated clay at equilibrium pH 3 (Kl 1). One reflection of medium intensity at d = 2.71 A probably arose from aluminium phosphate hydrateAlPO, .2H,O (ASTM Index 15311). A second weak reflection at 2.80 A can only be ascribed to a new phosphate hydroxide phase involving calcium ions derived from the Ca-rich clay (ASTM Index 6-0454). A striking feature of the diffractogram obtained with the clay mix cf equilibrium pH 4.7 (K13) was the appearance of a very strong reflection at d = 4.02 Aowith simultaneous reduction in intensity of the characteristic quartz reflection (3.34 A). The relative intensities of the two hypersthene/hoFblende reflections increased considerably shifting the d-spacings to 3.18 and 3.12 A. In addition, ,a series of equally weak diffuse reflections appeared at 2.81,0 2.74, 2.70 and 2.52 A. Positive identification of the additional reflection at 4.02 A is difficult due to limited information available on the SiO,-P,O,-Al,O, system. Robinson and McCartney (1964) quote X-ray data from the work of Jacoby (1957) and Tien and Hummel (1962) relating to the existence ofOa binary compound 3SiO,.2P,O, within the ternary system having a d-spacing of 4.02 A. The binary compound was detected at 500°C the lowest temperature employed by previous workers. They further state that ternary compound formation is unlikely below 50 mol% P,O,. It seems probable that the additional reflection at 4.02 A observed in these experimental clay-phosphate systems arises from a SiO,-P,O, binary compound of this type. No further X-ray data to substantiate this evidence was found, but this observati’on leads to the possibility that binary compounds of the type SiO,-P,O, can exist in natural clay-phosphate systems at room temperature depending on the level at pH at which structural collapse of the clay lattice may occur. The existence of such compounds in temperature ranges lower than that covered by Jacoby has not been reported previously. The trace obtained with the $ay mix K13 (pH 4.7) after maturing for 30 days wa,s free from the reflection at 4.02 A. It was replaced by a very strong reflection at 3.11 A corresponding to a complex Mg, Fe, aluminium silicate hydroxide probably involving phosphate (ASTM 21-149). The quartz reflection remained relatively weak. This shows the metastable nature of the new phases which undergo transformation into hydrated
3.35
3.35
,125 pm
3.35
K13
K13
K13
K17
7.4
4.1
4.1
4.7
3.1
Eq. pH of clay mix
data of unheated
a Effects observed in parentheses,
3.35
Kll
(mm)
Max. aggregate size of clay
of X-ray diffraction
Designation of clay mix
Table 5 Summary
20
20
30 days
20
20
Period of mat. of clay mix
characteristics
Very weak Very strong Weak
Al,(PO,),(OHJs~5H,O New phases
Medium Weak Strong Weak Very strong Weak
Intensity
AlPO, .2H,O Ca-phosphate hydroxide phase 3Si0, 2Pz0, or SiO, P205 New phases (Diminution of quartz peak Na,Mg,Fe,Al silicate hydroxide New phases (Enhancement of quartz peak) New phases Aluminium phosphate hydroxide hydrate
Phases identified a
Diffraction
clay samples treated with phosphate
(,Q
2.94. 2.70. 2.60, 2.55, 2.53
3.00, 2.92, 2.81, 2.70 8.40
2.71 2.80 4.02 2.80, 2.74, 2.70, 2.52 3.11 3.00, 2.94, 2.53, 2.16, 2.01
d-spacing
ASTM 2-75
ASTM 15-311 ASTM 6-0454 (i) Jacoby (ii) Tien et al. ASTM 21-149
Reference
V.T.L. Bogahawatta, A.B. Poole/Applied
Clay Science IO (19%) 461-475
471
products on maturing in aqueous medium. In addition, aDseries of very weak, diffuse reflections appeared at 3.00, 2.94, 2.53, 2.16 and 2.01 A. Increasing disorder of the kaolinite was reflected by the disruption of the characteristic groups of triplets (2.49-2.25 A> and by the simultaneous weakening of the characteristic kaolinite reflections. These modifications of the diffraction pattern may indicate a partial collapse of the kaolinite lattice as a consequence of the phosphate treatment. The diffraction pattern resulting from the clay mix of equilibrium p! 7.4 (K17) exhibits significant changes. A strong reflection had appeared at 8.40 A. This was attributed to a new phase identified as aluminium phosphate hydroxide hydrate (Al,(PO,),(OH),.5H,O) by r$ference to the ASTM Index 2-75. The intensity of the hornblende reflection at 3.12 A was diminished. The quartz reflection remained dominant yhile a series of additional weak reflections appeared at 2.94, 2.70, 2.60, 2.55 and 2.53 A. It may be inferred that the disorder of the kaolinite structure had increased at this pH. It is interesting to note that the evidence of formation of aluminium phosphate hydroxide hydrate was apparent only in the slightly basic clay-phosphate system with the equilibrium pH of 7.4 (K17). This system was also the most promising from the strength standpoint. The system with an equilibrium pH of 4.7 favoured the initial formation of a SiO,-P,O, binary compound while the highly acidic system with equilibrium pH 3 probably contained calcium phosphate hydroxide. The latter had less favourable physical properties. The X-ray data also indicates that lattice breakdown has occurred to a different extent in the clay-phosphate systems investigated. It seems probable that phosphates reacted with the disordered kaolinite crystals, extracting aluminium to form new phosphate compounds. Table 5 is a summary of the phases identified by X-ray diffraction in the phosphate treated [H,PO, + (HH,) 2 HPO,] clay mixes prepared under different conditions. 4.5. Mimralogy of the heated clays X-ray diffraction patterns of phosphate-bonded fired clay specimens showed that no new crystalline compounds were formed up to 400°C. However, in specimens heated at 800°C two crystalline compounds showing weak diffuse reflections at 2.67 and 2.48 A were present. X-ray diffraction results of heated specimens containing different types of phosphatic binders showfd that reactions on heating result in loss of intensity of the characteristic quartz (3.34 A) reflection and in the general weakening of the pattern. These modifications are perhaps associated with the formation of amorphous compounds in the heated bodies. The above effects were most pronounced in the K17 clay mix and were hardly observable in the K13 clay mix. However, extensive structure collapse of kaolinite is evidenced by ill-defined reflections and the appearance of a continuous band in the 35-40” (2 0) range in the diffraction pattern of the Kl 1 clay mix. It was noted that the K17 clay mix gave higher overall values for the modulus of rupture than the Kll clay mix and this suggests that the formation of amorphous compounds probably involving phosphate are more likely to be the cause of the high strengths in phosphate-bonded bodies rather than the development of a collapsed dehydrated structure.
472
V.T.L. Bogahawatta, A.B. Poole/Applied
Fig. 2. Scanning electron micrograph showing clay-phosphate A = amorphous phase, B = clay substrate.
Clay Science 10 (1996) 461-475
reaction phases produced
by heating at 500°C.
In order to investigate the amorphous phases developed in the phosphate treated unheated clay, infrared analyses were carried out on both untreated and treated clay samples using a Unicam 501 100 spectrophotometer using a KBr “in disc” technique over the 400 to 3800 cm-i wave number region. In addition to absorption bands for bound and unbound hydroxyl groups a broad absorption band is present between 100 and 1100 cm- ’ in the spectrum of the phosphate treated sample which contrasts with the sharp absorptions ascribed to microcline, kaolinite and muscovite Si-0 bonds present in the untreated specimen. This is consistent with the appearance of a broad band at 1640 cm-’ corresponding to a hydrated amorphous phase probably involving silica and phosphate. An additional sharp band at 770 cm-’ can be assigned to PO:--0-H vibrations of Na,HPO, .2H,O arising from hydrogen bonding with water (Chapman and Thirlwell, 1964). Thus the spectral differences in the spectrum point to the development of amorphous silicate phase and hydrogen bonding involving phosphates in the treated residual clay. Scanning electron microscopic techniques were also employed to follow changes in microstructure and constitution. A Hitachi S450 electron microscope was used to examine fracture surfaces of the heated clay-phosphate composites, and these were vacuum-coated with a thin film of carbon prior to examination. Scanning electron micrographs, such as illustrated in Fig. 2, revealed several characteristics regarding the physical nature of the phases developed in the clay-phosphate bond on firing.
V.T.L. Bogahawatta, A.B. Poole/Applied
Clay Science 10 (1996) 461-475
Fig. 3. DSC curves for untreated clay (K) (solid line) and phosphate temperature ranges from 20°C to 700°C.
treated clay (K17) (broken
473
line) over
Three distinct types of material can be recognised in phosphate treated specimens heated at 5OO”C, large regions which appear to be amorphous, an intermediate phase of large particle size and fine particles of clay. The large regions are most likely the hydrated aluminium-phosphate phases. Comparitive electron probe microanalyses show the grains of intermediate size to be rich in silica. The physico-chemical changes produced on heating the phosphate treated clays were studied by differential scanning calorimetry. The DSC scans were obtained using a Perkin-Elmer 7 series thermal analysis system at a heating rate of lO”C/min with nitrogen flow rate of 60 cm3/min. Samples of raw and treated clay (- 9 mg) were placed in a flat aluminium pan and heated in nitrogen alongside an empty pan as reference. Typical scans are illustrated in Fig. 3. The untreated clay (K> showed a small endotherm up to about 250°C probably due to the loss of interlayer water of vermiculite followed by the dehydroxylation endotherm in the range 400-500°C with a peak temperature at 475°C. The total areas of the dehydration and dehydroxylation peaks correspond to heats of reaction of 15.5 and 445.4 J/g respectively. The striking feature of the DSC scan of phosphate-treated sample (K17), is the absence of both endotherms in the thermograms. There appears to be a very broad exotherm above 550°C with a peak temperature at 630°C. It is possible that the Iphosphate treatment destroyed the original clay mineral structure of the clay giving rise to compounds which undergo physico-chemical changes above 550°C. It is also likely that this critical change corresponded to the decomposition of hydrated aluminiun-phosphate phases. The loss strength of bars heated at temperatures above 500°C is possibly a direct consequence of such changes in the bonding phases. An alternative suggestion is that unidentified exothermic reactions in the clay-phosphate system mask the endotherms identified for the untreated clay.
474
V.T.L. Bogahawatta, A.B. Poole/Applied
5. The phosphate/
Clay Science 10 (1996) 461-475
clay reaction mechanism
Literature presents a confusing picture of the mechanism involved in clay mineralphosphate reactions. Some compare this to a physical adsorption, others to a chemical reaction. The experimental evidence points to the latter. The model proposed by Hsu and Bates (1964) for reactions of phosphates with aluminium hydroxide and iron oxides may be extended to the experimental system involving kaolinite. In a slightly basic or neutral medium (pH 71, aluminium in the kaolinite lattice is available in the form of A13+. In acidic medium (pH 31, the availability of A13+ from lattice destruction is low and phosphates may form aluminium silicate hydroxide phases by hydroxyl bonding. In moderately acidic medium (pH 5) the formation of a metastable binary phase is initially favoured and by subsequent hydrolysis and polymerisation on ageing stable aluminium silicate hydroxide phases. Thus, precipitation may occur giving rise to aluminium phosphate hydrate when phosphate removes discrete A13+ ions from the lattice. When phosphate breaks only part of the Si-O-Al bonds, aluminium silicate hydroxides may form. Since the OH/Al ratio is different in the various phases, the activity of aluminium in these phases may also be different.
6. Conclusions
The following conclusions may be drawn from this study. (1) The strength of phosphate bonded fired clay products is partly related to the equilibrium pH of the initial clay mix. The equilibrium pH of about 7.4 seems to be the most favourable for the attainment of high strengths. The increase in strength at this level of pH is probably due to the formation of amorphous aluminium phosphate phases involving new and stronger bonds as evidenced by the IRA, DSC and SEM examinations. There is circumstantial XRD evidence to indicate the formation of metastable phosphate-silicate phases. The substantial reduction in porosity and shrinkage is consistent with the development of these new phases. (2) Clay mixtures which gave good strength properties on heating contained 3.16 wt% phosphate ion. The highest strengths were obtained at 500°C where increases as high as 60% were observed compared to the unmodified clay bars fired at 800°C. This strength results from an essentially chemical bonding rather than true vitrification which does not occur at these low temperatures. (3) A mixture of phosphoric acid and diammonium hydrogen phosphate is effective in maintaining the critical equilibrium pH in the clay mixes. (4) These laboratory studies suggest that the normal heating temperature of clay products of this material could be reduced significantly by using a phosphate admixture. Specimens heated at 500°C exhibited low linear shrinkage, passed the limited durability tests as used in Sri Lanka and gave MOR results up to 2.55 N/mm*. A comparison with normal vitrified bricks made of the same raw material is not made in this study.
V.T.L. Bogahawatta, A.B. Poole/Applied
Clay Science 10 (19961461-475
475
References ASTM CCm7-92a, 1992. Method of sampling and testing brick and structural clay tile 04.05. ASTM C373-88, 1988. Test method for water absorption, bulk density, apparent porosity and apparent specific gravity of tired whiteware products 15.02. ASTM Ctl74-88, 1988. Test methods for flexural properties of ceramic whiteware materials 15.02. BSI BS1377 Part 2, 1990. Classification tests, Methods of test for classifying soils and for determining their basic physical properties. Bar-Yosef, B., Kafkafi, V., Rosenberg, R. and Sposito, G., 1988. Phosphorous adsorption by kaolinite and montmorillonite. Soil. Sci. Sot. Am. Proc., 52(6): 1580-1585. Bartha, P., Lehmann, H. and Koltermann, M., 1971. Untersuchungen zur Bindung keramischer Werkstoffe mit H,PO,. Ber. Dtsch. Keram. Gesellsch., 48: 11 l-l 15. Beck, W.R., 1949. Crystallographic inversions of the aluminium orthophosphate polymorphs and their relation to those of silica. J. Am. Ceram. Sot., 32(4): 147-151. Chapman, AC. and Thirlwell, L.E., 1964. Spectra of phosphate compounds: 1. IR spectra of orthophosphates. Spectrochim. Acta., 20: 937-947.; ibid., 27 A: 425-439. Forest, J., 1964. Contribution B la connaissance de ciments lies par l’acid orthophosphorique et a l’ttude de certains phosphate d’aluminium. Rev. Mater. Contr. Trav. Publ., 580: 1-12. Hsu, P.H. and Bates, T.F., 1964. Formation of X-ray amorphous and crystalline aluminium hydroxides. Miner. Mag., 33: 749-768. Jacoby, W.R., 1957. Equilibrium relations in a portion of the system SiO,-P,O,. Publ. No. 22528, 119 pp.; Dissertation Abstr., 17: 2233. Kingery, W.D., 1950. Fundamental study of phosphate bonding in refractories: Part I. J. Am. Ceram. Sot., 33(8): 239-241. Kingery, W.D., 1952. Fundamental study of phosphate bonding in refractories: Part IV. J. Am. Ceram. Sot., 35(31: 61-63. Robinson. P. and McCartney, E.R., 1964. Subsolidus relations in the system SiO, -Al,O,-P,O,. J. Am. Ceram. Sot., 47: 587-592. Robinson. P. and Segnit, E.R., 1967. The influence of phosphoric acid on the forming, constitution and properties of dry pressed steatite bodies. J. Aust. Ceram. Sot., 3(l): 9-17. Secor, R.B. and Radke, C.J., 1985. Spillover of the diffuse double layer on montmorillonite particles. J. Colloid Interface Sci., 103: 237-244. Sudakas, L.G. and Sinina, L.M., 1971. Regulation of the time of binding and hardening of phosphate binders. Inorg. Mater., 7: 537-538. Tauber, E., Pepplinkhouse, H.J., Crook, D.N. and O’Brien, R.J., 1972. Phosphate-bonded zircon refractories. Proc. Aust. Inst. Min. Met., 244: 47-55. Tien, T.Y. and Hummel, F.A., 1962. System SiO,-Pro,. J. Am. Ceram. Sot., 45(9): 422-424.
The i:nfluence of phosphate on the properties of clay bricks V.T.L. Bogahawatta Geomaterials
Unit, School of Engineering,
‘, A.B. Poole
QMW, Universiry of London, London El 4NS, UK
Received 20 March 1995; accepted 3 November 1995
Abstract The fabrication, properties and composition of heated phosphate-bonded clay bricks are reported. The strength of phosphate-bonded clay bodies is shown to be related to the equilibrium pH of the clay mix. The highest strengths were obtained for mixes with equilibrium pH 7 fired at 500°C whllch gave a modulus of rupture 60% higher than the unmodified clay heated under similar conditions but at 800°C firing temperature. The phosphate-bonded products were also of low porosity and improved dimensional stability. An insight into the probable mechanism of strength development of the heated product has been obtained from XRD, SEM and DSC studies. The results show that the properties of the phosphate-bonded clay bodies are critically affected by new mineral phases resulting from the reactions of phosphates with clay which subsequently undergo physico-chemical changes above 550°C.
1. Introdluction
It is economically desirable to reduce the sintering temperature of structural clay products which partially or fully vitrify at the relatively high temperature used in their manufacture. In this respect, the use of bonding materials which react chemically with the matrix particles at low temperatures is attractive for relevant applications. The literature indicates that phosphoric acid and certain phosphates offer good prospects as bonding agents (Kingery, 1950). The bonding property of phosphates with refractory oxides (Bartha et al., 1971) and with silicates (Forest, 1964; Robinson and Segnit, 1967; Tauber et al., 1972) is widely used in the production of refractories of high
’ Present address: NBRO, 99/l Jawatta Road, Colombo 5, Sri Lanka. 0169-1317,/96/$15.00 0 1996 Elsevier Science B.V. All rights reserved SSDI 0169-1317(95)00042-9
462
V.T.L. Bogahawatta, A.B. Poole/Applied
Clay Science 10 (19961461-475
strength, refractory cements and mortars (Kingery, 1952). However, additive systems involving phosphoric binders are little used in fired clay products presumably because of the ease at which strengths are achieved using high temperature and soaking times. In spite of many possible applications, the properties of refractory bonds employing phosphates have not been systematically studied. Kingery (1950), Sudakas and Sinina (1971) and Robinson and McCartney (1964) ascribe the formation of cold-setting and heat-setting bonds in refractory oxide-phosphate systems to thermal dehydration and chemical reactions of phosphates. However, unlike the uniform metal oxide surfaces, clays are characterized by low temperature dehydration with edge and face interactions which make the above mentioned concepts difficult to apply to clay minerals (Secor and Radke, 1985, Bar-Yosef et al., 1988). Reactions between phosphates and soil minerals have received considerable attention from soil scientists, but our understanding of the phosphate-soil reactions remains inadequate. The general consensus is that phosphates react chemically with aluminium containing minerals to form surface compounds as well as insoluble Al-phosphates. However, the details of the mechanisms involved and of the phosphate phases formed are matters of continuing discussion.
2. Experimental
method
The primary objective of this investigation was to develop satisfactory chemically phosphate-bonded clay products with low energy intensive fabrication techniques. Therefore, emphasis was placed upon determining the dependence of strength and other physical and mechanical properties of experimental clay bodies on heating temperature. The experimental studies were designed to determine the heating conditions most favourable for the phosphate/silicate/aluminate reactions in relation to the properties of the finished products, and to explore the possibilities of reducing the temperatures below those normally used in the brick industry.
3. Materials
and methods
3.1. Materials The experiments were limited to a single soil sample (K). This poorly ordered kaolinitic residual sand/silt/clay soil is extensively used by the brick industry in Sri Lanka. It is composed of 30% clay, 25% silt and 45% sand grade materials by weight. It was evenly graded with a maximum particle size of 2 mm. The dominant constituents in the clay grade are kaolinite (46%), quartz (31%) and feldspars (15%). The accessory minerals include vermiculite and hypersthene. The two analar grade phosphatic binders used in the investigation were: 1. Orthophosphoric acid (H,PO,-weight per ml 1.75 g). 2. Diammonium hydrogen phosphate [(NH,),HPO,].
V.T.L. Bogahawatta, A.B. Poole/Applied
Clay Science 10 (1996) 461-475
463
3.2. Method of fabrication The effects of phosphate treatment and the equilibrium pH value of clay mixes prepared under different maturing conditions on the final properties of fired clay products were investigated in relation to the clay-H,PO,-(NH,),HPO, system. For this purpose, seven clay mixes were prepared using different experimental conditions. 1 M :solutions of orthophosphoric acid (pH 0.9) and diammonium hydrogen phosphate (p.H 7.6) were prepared using deionised distilled water and mixed in predetermined proportions to adjust the pH of the bulk solution to desired levels. The levels of pH were selected to give equilibrium pH values ranging from 3 to 7 to clay mixes when treated with a fixed volume of phosphate solution. This volume which imparts proper moulding consistency to a clay mix was predetermined from blank trials. It was found from control tests that 750 g of air-dried clay required 250 ml of 1 M phosphate solution to attain the appropriate moulding consistency. The equivalent composition of 3.16 wt% phosphate ion concentration was maintained throughout the series of clay mixes. A weighed sample of air-dried, disaggregated clay with the required volume of phosphate solution was mixed for 15 minutes. The clay mix was then covered and allowed to mature at room temperature for 20 hours, 40 hours and 90 hours to allow dispersion the additives and to initiate reactions. At the end of each period of maturing, the clay mix was puddled again for 10 min and moulded into a steel mould (160 mm X 30 mm X 15 mm) with a steel plunger in a simple mechanical lever press at 0.5 N/mm2. A uniform moulding moisture content of 20% by weight of dry clay was maintained in all the mixes. The moulding moisture content and equilibrium pH of the clay mix were determined according to BS 1377 (1990). Three batches of specimens were prepared from a clay mix of equilibrium pH 5, with maturing times of 20 hours, 40 hours and 90 hours. A second series consisting of four batches of specimens were prepared from clay mixes of equilibrium pH 3 and 7; with maturing times corresponding to each level of pH of 20 hours and 90 hours. Control specimens were also prepared
Table 1 Composition of experimental on bonding properties Weight dry
mixes prepared to investigate
the effects of equilibrium
pH and time of maturing
Specimen No.
clay (g)
Volume of lM Phos. Soln. (ml)
Mix proportions by wt. H,PO,: (NH,),HP~,
Initial pH of Phos. Soln.
Initial pH of clay mix
Equil. pH of clay mix
Maturing time (h)
K13/1 K13/7 Kl3/13 Kll/l Kl l/7 K17/1 K17/7
750 750 750 720 705 705 705
250 250 250 240 235 235 235
1S:l 1.5:1 1.5:1 14:l 14: 1 1:18 1:18
2.9 2.8 2.8 1.5 1.4 7.5 7.6
6.7 6.7 6.8 6.8 6.8 6.8 6.8
5.1 5.0 4.7 3.0 3.3 7.4 7.4
90 40 20 20 90 90 20
Moulding moisture content relative to dry clay = 20%. Phosphate ion concentration = 3.16 wt% for all specimens.
464
V.T.L. Bogahawatta, A.B. Poole/Applied
Clay Science 10 (1996) 461-475
Table 2 Composition of experimental mixes prepared to investigate the effect of the amount of phosphate the modulus of rupture of heated bars (parts by weight of dry materials) Designation of specimens
Clay (parts by weight)
Phosphate ion concentration (parts by weight)
Average MOR of bars heated at 500°C (N/mm*)
K17 K27 K2.57 K37
100 100 100 100
3.16 6.32 7.90 9.48
2.55 3.57 4.00 4.25
Moulding
(3.06%) (5.94%) (7.32%) (7.66%)
binders on
moisture content 20%.
from an untreated clay mix for comparison purposes. Each batch consisted of at least 12 rectangular bars. The composition of the experimental mixes are presented in Table 1. The specimens were air dried, then oven dried for 24 hours at 105°C. The linear drying shrinkage for each batch of specimens before heating is given in Table 4 as % length. The bars were then heated in a laboratory electric muffle furnace in air at temperatures from 400” to 900°C. The temperature difference between the different firings being typically 100°C. The heating rate was 5.6”C/min with a soaking period of 2 hours at the peak temperature for all firings. Heated specimens were tested for modulus of rupture, bulk density, shrinkage and durability in accordance with appropriate BSI or ASTM methods which were modified where necessary to take account of the specimen size (Table 4). 3.3. Optimisation
of processing
conditions
Having determined the critical equilibrium pH value, time of maturing and minimum heating temperature required to obtain an acceptable product, the second part of the study was designed to ascertain the optimum concentration of phosphate. Four clay mixes were prepared incorporating 1 M, 2 M, 2.5 M and 3 M solutions of orthophosphoric acid and diammonium hydrogen phosphate at the optimum equilibrium pH determined in the first part of the study. The experimental bars prepared as described above, matured for 90 hours and heated at 500°C with 2 hour soaking period. Table 2 presents the composition of the experimental mixes and the results of modulus of rupture of (MOR) heated bars.
4. Results and discussion 4.1. Effect of phosphate
addition on the strength
Mean values for modulus of rupture as a function of heating temperature are shown in Table 3 for the three clay mixes together with the effect of variation of maturing time
V.T.L. Bogahawatta, A.B. Poole/Applied
Clay Science 10 (1996) 461-475
Table 3 Modulus 01’rupture of tired phosphate-bonded
clay bars heated at temperatures
Conditions
Modulus of rupture in N/mm2
at heated temperature
of preparation
465
between 400” and 900°C shown
Eq. pH of Mix
Mat. time (h)
400°C
500°C
600°C
700°C
800°C
900°C
:Kll) fK13)
90 20 40 20
1.54 1.40 1.15 1.92
1.19 1.53 2.08 1.42
0.70 1.37 1.19 1.71
1.24 1.09 1.17 1.24
1.19 1.15 1.18 1.66
1.17 1.14 1.34 1.96
90
0.57
1.32
0.96
0.67
0.72
0.92
90 20 20
2.41 1.98
2.32 2.55
2.10 1.94
2.01 1.86
1.56 1.99 1.60
1.72 1.85 1.73
;iK17) 7 unmodified
clay (K2A)
on strength. An unmodified clay specimen hated at 800°C and 900°C is included for comparis;on. Coefficients of variation lay between 4 and 5%. Marked changes in modulus of rupture with equilibrium pH of the clay mix and firing temperature are shown by the specimens. Highest strengths (2.32-2.55 N/mm’) were associated with clay bodies prepared at an equilibrium pH 7.4 (K17). Specimen bars representing the clay mix of equilibrium pH 5 (K13, maturing time 20 hours) had intermediate strengths whereas an analogous set of bars from the clay mix of equilibrium pH 3 (Kl 1) had the lowest strengths. In general, the modulus of rupture increased sharply when the temperature was raised from 400” to 500°C at which a maximum was attained. Above this temperature, modulus of rupture of both K17 and K13 bodies dropped steeply while that of the Kll body dropped gradually as the temperature was increased to about 600-700°C. An apparent increase in strength above 700°C was discernible in the K13 body. Bodies heated at the optimum temperature of 500°C gave strengths of the order of 60% higher than the normal clay bodies heated at 800°C (1.6 N/mm2 1. Clay mixes within each batch exhibited similar trends irrespective of their extent of maturing. The effect of variation in maturing time of clay mix on the modulus of rupture of fired bars is also illustrated in Table 3. A substantial increase in modulus of rupture with increasing maturing time from 20 hours to 90 hours was evident in the two clay bodies K17 and Kl 1. The gain in strength due to prolonged maturing of the K17 body heated at 500°C was approximately 10%. In contrast, the strength of the clay body prepared at the equilibrium pH 5 (K13), deteriorated gradually on maturing; the strength dropped from 2.08 N/mm2 to 1.32 N/mm’. Since specimen preparation was identical in all cases this anomalous behaviour is difficult to explain but may mark a threshold pH value between acidic phosphate-clay interactions and those occurring in neutral conditions. The influence of the phosphate concentration on the modulus of rupture at a constant temperature (of 500°C) is given in Table 2, which shows a gradual increase in modulus of rupture up to the maximum phosphate addition of 7.66% by weight. However, these strength results are very much better than the heated unmodified clay even at the lowest phosphate ion concentration of 3.06%. The clear pH dependence of strength can be interpreted in terms of the relative
466
V.T.L. Bogahawatta, A.B. Poole/Applied
Clay Science 10 (1996) 461-475
Table 4 Physical properties of phosphate-bonded clay bars heated at temperatures between coeff. = % absorption 24 hour cold water/% absorption 5 hour boiling water) Designation of clay mix
Property
Kll
Bulk density (g/cm’)
Average value of property at temperatures
% Absorption % Asorption
(24 h) (5 h b)
Sat. coefficient % Weight loss % Drying shrinkage % Firing shrinkage
K13
Bulk density (g/cm3)
% Absorption
(24 h)
% Absorption
(5 h b)
Sat. coefficient
% Weight loss
% Drying shrinkage
% Firing shrinkage
K17
Bulk density (g/cm”) % Absorption
(24 h)
% Absorption
(5 h b)
Sat. coefficient Weight loss
400°C and 900°C (Sat.
400°C
500°C 1.45 1.51 20.06 19.51 24.60 24.06 0.81 0.81 4.38 4.43
20 90 20 90 20 90 20 90 20 90 20 90 20 90
h h h h h h h h h h h h h h
mat mat mat mat mat mat mat mat mat mat mat mat mat mat
1.52 1.52 19.22 18.64 23.24 22.56 0.83 0.83 3.15 3.09
20 40 90 20 40 90 20 40 90 20 40 90 20 40 90 20 40 90 20 40 90
h h h h h h h h h h h h h h h h h h h h h
mat mat mat mat mat mat mat mat
1.54 1.52
mat mat mat mat mat mat mat mat mat mat mat mat
20 90 20 90 20 90 20 90 20 90
h h h h h h h h h h
mat mat mat mat mat mat mat mat mat mat
_ _
1.48 19.68 19.28 20.37 23.65 23.03 24.28 0.83 0.84 0.84 3.40 3.16 3.34
0.67 _
1.63 1.61 16.46 16.60 20.82 20.78 0.79 0.80 3.64 3.64
_
600°C
700°C
1.50 1.49 1.50 1.48 20.23 21.24 20.52 21.40 25.03 25.70 25.27 25.95 0.81 0.82 0.81 0.82 5.24 6.67 5.69 6.81 Average = 6.25 Average = 6.45 _ _ 0.67
-
1.46 1.49 1.46 20.85 20.92 21.79 25.43 25.69 26.71 0.82 0.81 0.81 5.15 4.88 5.86
1.46 1.48 1.49 1.45 1.41 1.46 21.87 22.33 21.83 22.12 21.99 22.68 26.53 26.89 26.79 26.93 26.48 27.29 0.82 0.83 0.81 0.82 0.83 0.83 6.70 7.56 6.18 7.16 6.28 7.05 Average = 7.29 Average = 6.77 Average = 6.25 0.67 _ _ 0.67 _
1.62 1.60 17.07 17.48 21.44 21.86 0.80 0.80 4.98 4.98
1.58 1.58 18.44 18.37 23.36 23.24 0.79 0.79 6.87 6.87
1.56 1.53 18.74 19.05 23.53 23.72 0.80 0.80 7.7 1 7.71
shown
800°C
900°C
1.45 1.48 22.40 21.78 26.75 26.29 0.84 0.83 7.33 7.29
1.47 1.45 21.41 21.39 26.32 26.58 0.81 0.82 7.38 7.58
1.33 _
0.67
1.48 1.43 1.46 22.51 22.92 23.10 27.04 27.46 27.45 0.83 0.83 0.84 7.86 7.49 7.57
1.46 1.49 1.43 21.60 21.83 22.79 26.80 26.90 27.72 0.83 0.81 0.82 8.23 7.76 7.93
0.67 _ 0.67
1.33 1.34 0.67
1.56 1.52 19.17 19.30 23.74 23.65 0.81 0.81 7.99 7.99
1.57 1.57 19.10 19.11 24.21 23.89 0.79 0.80 8.37 8.37
V.T.L. Bogahawatta, A.B. Poole/Applied
467
Clay Science 10 (1996) 461-475
Table 4 (continued) Designation of clay mix
Average value of property at temperatures
Property
% Drying shrinkage % Firing shrinkage
20 90 20 90
h h h h
mat mat mat mat
400°C
500°C
600°C
0.68 0.68
Average = 8.65 Average = 8.75 0.68 0.68 0.68 0.68
shown
700°C
800°C
900°C
-
0.68 -
0.68 0.68
abundance of OH groups available for structure formation and the extent of collapse of kaolinite lattices at different levels of pH. The optimum strength achieved at 500°C may be attributed to the favourable structures formed by reactions of phosphates with the clay minerals, since the dehydration of kaolinite occurs in the temperature range 450°C to 525°C. pH appears to have a significant effect on the development of these phosphate-clay reactions Rearrangement of these metastable structures at temperatures above 500°C result in loss of strength. 4.2. The physical properties
of the heated specimens
The bulk density of heated specimens was measured in accordance with ASTM C373-88, water absorption as a percentage of the original weights (24 hours in cold water and 5 hours in boiling water) in accordance with ASTM C67-92a while the percentage weight loss on heating the dried specimen and linear drying shrinkage and the additional linear shrinkage on firing were measured directly. The results expressed as percentages are given in Table 4. These results show that there is a small irregular decrease in bulk density, as heating temperature increases, for specimens from all three mixes. There is also a correlation between modulus of rupture values and bulk density (Fig. 1). The 24 hour water absorption values similarly show small increases with heating temperature for all mixes with a maximum at 800°C. This is interpreted as a gradual increase in porosity with temperature up to 800°C after which sintering, which was confirmed by microscopic examination of the specimens, begins to decrease the porosity. The wet-to-dry shrinkage of bars from the demoulded stage (moisture content 15% by mass) to the oven-dry stage (105°C 24 hours) depended on the composition of the clay mix. Experimental bars made from the three clay mixes Kll, K13 and K17 had wet-to-dry linear shrinkage of 6.3%, 6.8% and 8.7% respectively. However, these phosphate-bonded bodies had remarkably low firing shrinkages when compared with those of conventional clay bodies. The fired clay body K17 had a consistently low dry to fired shrinkage of 0.67% over the range of heating temperature from 400” to 900°C. The low shrinkage of the phosphate-modified heated clay may be due to the quartz type inversions of aluminium phosphate which is believed to occur at lower temperatures than in fsilica (Beck, 1949). The samples heated at 500°C yielded DSC and SEM evidence of the formation of such amorphous products. Differential shrinkage behaviour
468
V.T.L. Bogahawatta, A.B. Poole/Applied
Clay Science 10 ~19%) 461-475
2.5
1.40
t .45
1.50
Bulk
Fig. 1. The relationship determinations.
between modulus
1.55
denslty
II
20 h maturing
??
go h maturing
1.60
1.65
1.70
(glcm3)
of rupture and bulk density. Each point is the mean of at least 3
of the three clay sample types may indicate varying degrees of formation of reaction products in them. An alternative suggestion (R. Bown, pers. commun.) is that the phosphates deflocculate the clay minerals, leading to a closer packing of the drying block and hence greater strength through greater surface contact. The deflocculation effect is greatest at pH 7-8, and hence strength is greatest for specimen (K17). 4.3. The durability of phosphate-bonded
samples
In order to test durability, phosphate-bonded bars prepared from the most promising composition (K17) and heated under the optimum conditions (500°C for 2 hours) were randomly sampled for testing under laboratory conditions. Bars made from a clay mix without any phosphate binder and heated at the same temperature and samples of engineering brick (LBC) satisfying BS specifications were also taken to compare the behaviour of the respective bodies under similar conditions of accelerated weathering. Twelve specimens from each batch were subjected to the test. The procedure used in this study was first to dry the heated specimens at 110°C for 24 hours. They were then immersed in water for 2 hours and after removal placed in an oven overnight at 105-l 10°C. The following morning the samples were cooled in a desiccator and the cycle repeated. A total of 60 cycles of alternate wetting and drying were completed within a period of three months. After the last cycle, specimens were carefully examined for cracks, disintegration and leaching of compounds. The accelerated durability test showed that both the test and control samples did not undergo cracking or disintegration after 60 cycles of test. Weight losses were relatively insignificant. However, there was evidence of leaching of compounds from the test samples with consequent discolouration of external surfaces. In a more severe test of durability a total of 144 samples from the three clay mixes
V.T.L. Bogahawatta, A.B. Poole/Applied
Clay Science 10 (1996) 461-475
469
(Kl 1, K13 and K17) heated at 400”, 500”, 600”, 700”, 800” and 900°C were subjected
to a 5 hour boiling test. The test showed that none of the specimens disintegrated during test indicating the adequacy of the heat treatment and bonding to impart sufficient durability to heated clay. However, water absorption of the specimens increased on boiling. The lowest saturation coefficient is associated with the K17 bars heated at 600°C (Table 4). 4.4. The mineralogy
of the unfired clay
X-ray diffraction methods were used to detect new crystalline phases formed. Heated products were investigated using finely ground randomly orientated samples. Analyses were carried out using a Philips diffractometer with CuK CYradiation and a graphite monochromator. The phosphate treated clays exhibited clear changes in their diffraction patterns with changes in equilibrium pH of the original mix. Two additional reflections appeared in the trace made with phosphate tr:ated clay at equilibrium pH 3 (Kl 1). One reflection of medium intensity at d = 2.71 A probably arose from aluminium phosphate hydrateAlPO, .2H,O (ASTM Index 15311). A second weak reflection at 2.80 A can only be ascribed to a new phosphate hydroxide phase involving calcium ions derived from the Ca-rich clay (ASTM Index 6-0454). A striking feature of the diffractogram obtained with the clay mix cf equilibrium pH 4.7 (K13) was the appearance of a very strong reflection at d = 4.02 Aowith simultaneous reduction in intensity of the characteristic quartz reflection (3.34 A). The relative intensities of the two hypersthene/hoFblende reflections increased considerably shifting the d-spacings to 3.18 and 3.12 A. In addition, ,a series of equally weak diffuse reflections appeared at 2.81,0 2.74, 2.70 and 2.52 A. Positive identification of the additional reflection at 4.02 A is difficult due to limited information available on the SiO,-P,O,-Al,O, system. Robinson and McCartney (1964) quote X-ray data from the work of Jacoby (1957) and Tien and Hummel (1962) relating to the existence ofOa binary compound 3SiO,.2P,O, within the ternary system having a d-spacing of 4.02 A. The binary compound was detected at 500°C the lowest temperature employed by previous workers. They further state that ternary compound formation is unlikely below 50 mol% P,O,. It seems probable that the additional reflection at 4.02 A observed in these experimental clay-phosphate systems arises from a SiO,-P,O, binary compound of this type. No further X-ray data to substantiate this evidence was found, but this observati’on leads to the possibility that binary compounds of the type SiO,-P,O, can exist in natural clay-phosphate systems at room temperature depending on the level at pH at which structural collapse of the clay lattice may occur. The existence of such compounds in temperature ranges lower than that covered by Jacoby has not been reported previously. The trace obtained with the $ay mix K13 (pH 4.7) after maturing for 30 days wa,s free from the reflection at 4.02 A. It was replaced by a very strong reflection at 3.11 A corresponding to a complex Mg, Fe, aluminium silicate hydroxide probably involving phosphate (ASTM 21-149). The quartz reflection remained relatively weak. This shows the metastable nature of the new phases which undergo transformation into hydrated
3.35
3.35
,125 pm
3.35
K13
K13
K13
K17
7.4
4.1
4.1
4.7
3.1
Eq. pH of clay mix
data of unheated
a Effects observed in parentheses,
3.35
Kll
(mm)
Max. aggregate size of clay
of X-ray diffraction
Designation of clay mix
Table 5 Summary
20
20
30 days
20
20
Period of mat. of clay mix
characteristics
Very weak Very strong Weak
Al,(PO,),(OHJs~5H,O New phases
Medium Weak Strong Weak Very strong Weak
Intensity
AlPO, .2H,O Ca-phosphate hydroxide phase 3Si0, 2Pz0, or SiO, P205 New phases (Diminution of quartz peak Na,Mg,Fe,Al silicate hydroxide New phases (Enhancement of quartz peak) New phases Aluminium phosphate hydroxide hydrate
Phases identified a
Diffraction
clay samples treated with phosphate
(,Q
2.94. 2.70. 2.60, 2.55, 2.53
3.00, 2.92, 2.81, 2.70 8.40
2.71 2.80 4.02 2.80, 2.74, 2.70, 2.52 3.11 3.00, 2.94, 2.53, 2.16, 2.01
d-spacing
ASTM 2-75
ASTM 15-311 ASTM 6-0454 (i) Jacoby (ii) Tien et al. ASTM 21-149
Reference
V.T.L. Bogahawatta, A.B. Poole/Applied
Clay Science IO (19%) 461-475
471
products on maturing in aqueous medium. In addition, aDseries of very weak, diffuse reflections appeared at 3.00, 2.94, 2.53, 2.16 and 2.01 A. Increasing disorder of the kaolinite was reflected by the disruption of the characteristic groups of triplets (2.49-2.25 A> and by the simultaneous weakening of the characteristic kaolinite reflections. These modifications of the diffraction pattern may indicate a partial collapse of the kaolinite lattice as a consequence of the phosphate treatment. The diffraction pattern resulting from the clay mix of equilibrium p! 7.4 (K17) exhibits significant changes. A strong reflection had appeared at 8.40 A. This was attributed to a new phase identified as aluminium phosphate hydroxide hydrate (Al,(PO,),(OH),.5H,O) by r$ference to the ASTM Index 2-75. The intensity of the hornblende reflection at 3.12 A was diminished. The quartz reflection remained dominant yhile a series of additional weak reflections appeared at 2.94, 2.70, 2.60, 2.55 and 2.53 A. It may be inferred that the disorder of the kaolinite structure had increased at this pH. It is interesting to note that the evidence of formation of aluminium phosphate hydroxide hydrate was apparent only in the slightly basic clay-phosphate system with the equilibrium pH of 7.4 (K17). This system was also the most promising from the strength standpoint. The system with an equilibrium pH of 4.7 favoured the initial formation of a SiO,-P,O, binary compound while the highly acidic system with equilibrium pH 3 probably contained calcium phosphate hydroxide. The latter had less favourable physical properties. The X-ray data also indicates that lattice breakdown has occurred to a different extent in the clay-phosphate systems investigated. It seems probable that phosphates reacted with the disordered kaolinite crystals, extracting aluminium to form new phosphate compounds. Table 5 is a summary of the phases identified by X-ray diffraction in the phosphate treated [H,PO, + (HH,) 2 HPO,] clay mixes prepared under different conditions. 4.5. Mimralogy of the heated clays X-ray diffraction patterns of phosphate-bonded fired clay specimens showed that no new crystalline compounds were formed up to 400°C. However, in specimens heated at 800°C two crystalline compounds showing weak diffuse reflections at 2.67 and 2.48 A were present. X-ray diffraction results of heated specimens containing different types of phosphatic binders showfd that reactions on heating result in loss of intensity of the characteristic quartz (3.34 A) reflection and in the general weakening of the pattern. These modifications are perhaps associated with the formation of amorphous compounds in the heated bodies. The above effects were most pronounced in the K17 clay mix and were hardly observable in the K13 clay mix. However, extensive structure collapse of kaolinite is evidenced by ill-defined reflections and the appearance of a continuous band in the 35-40” (2 0) range in the diffraction pattern of the Kl 1 clay mix. It was noted that the K17 clay mix gave higher overall values for the modulus of rupture than the Kll clay mix and this suggests that the formation of amorphous compounds probably involving phosphate are more likely to be the cause of the high strengths in phosphate-bonded bodies rather than the development of a collapsed dehydrated structure.
472
V.T.L. Bogahawatta, A.B. Poole/Applied
Fig. 2. Scanning electron micrograph showing clay-phosphate A = amorphous phase, B = clay substrate.
Clay Science 10 (1996) 461-475
reaction phases produced
by heating at 500°C.
In order to investigate the amorphous phases developed in the phosphate treated unheated clay, infrared analyses were carried out on both untreated and treated clay samples using a Unicam 501 100 spectrophotometer using a KBr “in disc” technique over the 400 to 3800 cm-i wave number region. In addition to absorption bands for bound and unbound hydroxyl groups a broad absorption band is present between 100 and 1100 cm- ’ in the spectrum of the phosphate treated sample which contrasts with the sharp absorptions ascribed to microcline, kaolinite and muscovite Si-0 bonds present in the untreated specimen. This is consistent with the appearance of a broad band at 1640 cm-’ corresponding to a hydrated amorphous phase probably involving silica and phosphate. An additional sharp band at 770 cm-’ can be assigned to PO:--0-H vibrations of Na,HPO, .2H,O arising from hydrogen bonding with water (Chapman and Thirlwell, 1964). Thus the spectral differences in the spectrum point to the development of amorphous silicate phase and hydrogen bonding involving phosphates in the treated residual clay. Scanning electron microscopic techniques were also employed to follow changes in microstructure and constitution. A Hitachi S450 electron microscope was used to examine fracture surfaces of the heated clay-phosphate composites, and these were vacuum-coated with a thin film of carbon prior to examination. Scanning electron micrographs, such as illustrated in Fig. 2, revealed several characteristics regarding the physical nature of the phases developed in the clay-phosphate bond on firing.
V.T.L. Bogahawatta, A.B. Poole/Applied
Clay Science 10 (1996) 461-475
Fig. 3. DSC curves for untreated clay (K) (solid line) and phosphate temperature ranges from 20°C to 700°C.
treated clay (K17) (broken
473
line) over
Three distinct types of material can be recognised in phosphate treated specimens heated at 5OO”C, large regions which appear to be amorphous, an intermediate phase of large particle size and fine particles of clay. The large regions are most likely the hydrated aluminium-phosphate phases. Comparitive electron probe microanalyses show the grains of intermediate size to be rich in silica. The physico-chemical changes produced on heating the phosphate treated clays were studied by differential scanning calorimetry. The DSC scans were obtained using a Perkin-Elmer 7 series thermal analysis system at a heating rate of lO”C/min with nitrogen flow rate of 60 cm3/min. Samples of raw and treated clay (- 9 mg) were placed in a flat aluminium pan and heated in nitrogen alongside an empty pan as reference. Typical scans are illustrated in Fig. 3. The untreated clay (K> showed a small endotherm up to about 250°C probably due to the loss of interlayer water of vermiculite followed by the dehydroxylation endotherm in the range 400-500°C with a peak temperature at 475°C. The total areas of the dehydration and dehydroxylation peaks correspond to heats of reaction of 15.5 and 445.4 J/g respectively. The striking feature of the DSC scan of phosphate-treated sample (K17), is the absence of both endotherms in the thermograms. There appears to be a very broad exotherm above 550°C with a peak temperature at 630°C. It is possible that the Iphosphate treatment destroyed the original clay mineral structure of the clay giving rise to compounds which undergo physico-chemical changes above 550°C. It is also likely that this critical change corresponded to the decomposition of hydrated aluminiun-phosphate phases. The loss strength of bars heated at temperatures above 500°C is possibly a direct consequence of such changes in the bonding phases. An alternative suggestion is that unidentified exothermic reactions in the clay-phosphate system mask the endotherms identified for the untreated clay.
474
V.T.L. Bogahawatta, A.B. Poole/Applied
5. The phosphate/
Clay Science 10 (1996) 461-475
clay reaction mechanism
Literature presents a confusing picture of the mechanism involved in clay mineralphosphate reactions. Some compare this to a physical adsorption, others to a chemical reaction. The experimental evidence points to the latter. The model proposed by Hsu and Bates (1964) for reactions of phosphates with aluminium hydroxide and iron oxides may be extended to the experimental system involving kaolinite. In a slightly basic or neutral medium (pH 71, aluminium in the kaolinite lattice is available in the form of A13+. In acidic medium (pH 31, the availability of A13+ from lattice destruction is low and phosphates may form aluminium silicate hydroxide phases by hydroxyl bonding. In moderately acidic medium (pH 5) the formation of a metastable binary phase is initially favoured and by subsequent hydrolysis and polymerisation on ageing stable aluminium silicate hydroxide phases. Thus, precipitation may occur giving rise to aluminium phosphate hydrate when phosphate removes discrete A13+ ions from the lattice. When phosphate breaks only part of the Si-O-Al bonds, aluminium silicate hydroxides may form. Since the OH/Al ratio is different in the various phases, the activity of aluminium in these phases may also be different.
6. Conclusions
The following conclusions may be drawn from this study. (1) The strength of phosphate bonded fired clay products is partly related to the equilibrium pH of the initial clay mix. The equilibrium pH of about 7.4 seems to be the most favourable for the attainment of high strengths. The increase in strength at this level of pH is probably due to the formation of amorphous aluminium phosphate phases involving new and stronger bonds as evidenced by the IRA, DSC and SEM examinations. There is circumstantial XRD evidence to indicate the formation of metastable phosphate-silicate phases. The substantial reduction in porosity and shrinkage is consistent with the development of these new phases. (2) Clay mixtures which gave good strength properties on heating contained 3.16 wt% phosphate ion. The highest strengths were obtained at 500°C where increases as high as 60% were observed compared to the unmodified clay bars fired at 800°C. This strength results from an essentially chemical bonding rather than true vitrification which does not occur at these low temperatures. (3) A mixture of phosphoric acid and diammonium hydrogen phosphate is effective in maintaining the critical equilibrium pH in the clay mixes. (4) These laboratory studies suggest that the normal heating temperature of clay products of this material could be reduced significantly by using a phosphate admixture. Specimens heated at 500°C exhibited low linear shrinkage, passed the limited durability tests as used in Sri Lanka and gave MOR results up to 2.55 N/mm*. A comparison with normal vitrified bricks made of the same raw material is not made in this study.
V.T.L. Bogahawatta, A.B. Poole/Applied
Clay Science 10 (19961461-475
475
References ASTM CCm7-92a, 1992. Method of sampling and testing brick and structural clay tile 04.05. ASTM C373-88, 1988. Test method for water absorption, bulk density, apparent porosity and apparent specific gravity of tired whiteware products 15.02. ASTM Ctl74-88, 1988. Test methods for flexural properties of ceramic whiteware materials 15.02. BSI BS1377 Part 2, 1990. Classification tests, Methods of test for classifying soils and for determining their basic physical properties. Bar-Yosef, B., Kafkafi, V., Rosenberg, R. and Sposito, G., 1988. Phosphorous adsorption by kaolinite and montmorillonite. Soil. Sci. Sot. Am. Proc., 52(6): 1580-1585. Bartha, P., Lehmann, H. and Koltermann, M., 1971. Untersuchungen zur Bindung keramischer Werkstoffe mit H,PO,. Ber. Dtsch. Keram. Gesellsch., 48: 11 l-l 15. Beck, W.R., 1949. Crystallographic inversions of the aluminium orthophosphate polymorphs and their relation to those of silica. J. Am. Ceram. Sot., 32(4): 147-151. Chapman, AC. and Thirlwell, L.E., 1964. Spectra of phosphate compounds: 1. IR spectra of orthophosphates. Spectrochim. Acta., 20: 937-947.; ibid., 27 A: 425-439. Forest, J., 1964. Contribution B la connaissance de ciments lies par l’acid orthophosphorique et a l’ttude de certains phosphate d’aluminium. Rev. Mater. Contr. Trav. Publ., 580: 1-12. Hsu, P.H. and Bates, T.F., 1964. Formation of X-ray amorphous and crystalline aluminium hydroxides. Miner. Mag., 33: 749-768. Jacoby, W.R., 1957. Equilibrium relations in a portion of the system SiO,-P,O,. Publ. No. 22528, 119 pp.; Dissertation Abstr., 17: 2233. Kingery, W.D., 1950. Fundamental study of phosphate bonding in refractories: Part I. J. Am. Ceram. Sot., 33(8): 239-241. Kingery, W.D., 1952. Fundamental study of phosphate bonding in refractories: Part IV. J. Am. Ceram. Sot., 35(31: 61-63. Robinson. P. and McCartney, E.R., 1964. Subsolidus relations in the system SiO, -Al,O,-P,O,. J. Am. Ceram. Sot., 47: 587-592. Robinson. P. and Segnit, E.R., 1967. The influence of phosphoric acid on the forming, constitution and properties of dry pressed steatite bodies. J. Aust. Ceram. Sot., 3(l): 9-17. Secor, R.B. and Radke, C.J., 1985. Spillover of the diffuse double layer on montmorillonite particles. J. Colloid Interface Sci., 103: 237-244. Sudakas, L.G. and Sinina, L.M., 1971. Regulation of the time of binding and hardening of phosphate binders. Inorg. Mater., 7: 537-538. Tauber, E., Pepplinkhouse, H.J., Crook, D.N. and O’Brien, R.J., 1972. Phosphate-bonded zircon refractories. Proc. Aust. Inst. Min. Met., 244: 47-55. Tien, T.Y. and Hummel, F.A., 1962. System SiO,-Pro,. J. Am. Ceram. Sot., 45(9): 422-424.