Surface area measurement of ground rubber using the B.E.T. surface area analyzer.

By Gerald W. Holland, Benfei Hu and Stephen Holland, Baker Rubber Inc.

Ground rubber materials have found wide applications (refs. 1-5). These applications include putting ground rubber back into tires. using ground rubber as frictional materials in brake linings or pads7 blending ground rubber with plastics or thermoelastomers, and blending or mixing ground rubber with asphalt for pavement. In fact7 ground rubber materials use in the asphalt industry is one of the fastest growing market places today. Recent legislation per ISTEA (Intermodal Surface Transportation Efficiency Act) Section 1038 requires states to use a certain portion of ground rubber in asphalt. It is projected this would require as much as 800 million lbs. of ground rubber by 1997.

As ground rubber plays more important roles in industry activities7 additional knowledge about ground rubber is needed to better apply this material to different applications and to obtain desired properties of new end products containing ground rubber. Publications (refs. 6 and 7) on the particle size distribution characterization using sieve analysis and chemical analysis of ground rubber materials have been contributed by Baker Rubber Inc. Whether ground rubber materials are mixed or compounded with thermoplastics, with virgin rubber or with asphalt, the surface area of the materials may affect their handling, processing characteristics, and influence the end products, properties and applications. Little knowledge about the relationship between the particle size distribution of ground rubber materials and surface area is known. The measurement of the surface area is obviously necessary and interesting. Currently little work has been reported in this field. The general practice of measuring the surface area of ground rubber materials is a relatively new area. One can find an ASTM standard D3037, which is "Standard Test Methods for Carbon Black - Surface Area By Nitrogen Adsorption" (ref. 8). But this method is not designed for ground rubber materials. The measurement of surface area of ground rubber materials has been initiated in Baker Rubber Inc. using the Quantasorb B.E.T. surface area analyzer manufactured by Quantachrome Co. The purpose of this article is to demonstrate the application of the surface area analyzer to ground rubber materials and to show the repeatability of tests. The intention of this work is to enhance knowledge regarding ground rubber materials.

Theoretical background of the B.E.T. method

Several methods and theoretical models which are used to measure the surface area have been developed (refs. 9-11). Because of its ease of use, its definitiveness and its straightforward applicability, the B.E.T. (Brunauer, Emmett and Teller) theory (ref. 10) is almost universally employed in surface area measurement. The principle of the Quantasorb surface area analyzer is based on the B.E.T. theory. which is a commonly used method of determining the specific surface of solids by the physical adsorption of a gas on the surface of a solid. The determination of the monolayer capacity [V.sub.m], which is defined as the quantity of adsorbate required to cover the adsorbent with a monolayer, leads to the determination of specific surface areas in the B.E.T. method. Usually a second layer may be forming before the monolayer is complete, but [V.sub.m] is determined from the isotherm equations irrespective of the influence of multilayers. The background knowledge about B.E.T. theory and the principle using the B.E.T. .surface area analyzer to measure surface areas is well discussed by T. Allen (ref. 12), S. Lowell and Joan E. Shield (ref. 13)

The first theoretical equation relating the quantity of adsorbed gas at the equilibrium pressure of the gas was proposed by Langmuir (ref. 9). In his model, adsorption is limited to a monolayer. The Langmuir equation has wider application to chemical adsorption and the adsorption of solute from solution but with limited applicability to physical adsorption. Undoubtedly the most universally used experimental method for surface area determinations is that of reversible gas adsorption. By measuring the quantity of gas adsorbed at various relative pressures and using the B.E.T. theory, one can determine the surface area of almost any dry solid material. Unlike its predecessor theory, the Langmuir model, the B.E.T. theory takes into account that adsorption leads to the formation of multilayers on most surfaces. The B.E.T. theory requires that a linear plot of adsorption data be made using the following B.E.T. equation.

(1) 1/W [([P.sub.o]/P)-1] = C-1/[W.sub.mC] (P/[P.sub.o]) + 1/[W.sub.mC] The terms in the B.E.T. equation are defined as follows: P is the adsorbate (gas) pressure; [P.sub.o] is the saturated equilibrium vapor pressure of the adsorbate at the temperature at which the measurement is made: P/[P.sub.o] is the relative pressure; W is the weight of gas adsorbed at a particular relative pressure; [W.sub.m] is the weight of gas in a single molecular layer; and C is a constant.

Three concentrations of adsorbate are usually used to determine the B.E.T. plot; i.e., the plot of [[W([P.sub.o]P-1)].sup.-1] versus P/[P.sub.o]. Figure 1 is a typical three-point B.E.T. plot on 40 mesh ambiently ground tread rubber with surface area of 0.104 [m.sup.2]/g. According to the B.E.T. equation, the slope s of the straight line in figure 1 is given by:

(2) S = C-1/[W.sub.mC] and the intercept i is:

(3) i = 1/[W.sub.mC] Therefore, [W.sub.m], the weight required to form a monolayer of adsorbed gas, is given by:

(4) [W.sub.m] = 1/S + i Once the value of [W.sub.m] in equation (ref. 4) has been determined, the surface area, S of the sample is calculated from the following equation:

(5) S = [W.sub.m]NA/M

Where M is the adsorbate molecular weight, and A is the adsorbate cross-sectional area. A is 16.2x[10.sup.-20] [m.sup.2] for nitrogen, and 20.5x[10.sup.-20] [m.sup.2] for krypton (ref. 14). N is Avogadro's number, which is 6.02x[10.sup.26] 1/ kmole.

Usually, a B.E.T. plot is made in the range of relative pressures from about 0.05 to 0.35. The reason why the B.E.T. equation is restricted to giving linear plots roughly in the range of relative pressures 0.05 [less than or equal to] P/[P.sub.O] [less than or equal to] 0.35 is assumptions and restrictions made in the derivation of the B.E.T. equation. These assumptions, restrictions and applications are discussed in all modem textbooks on surface chemistry (refs. 12, 13, 15 and 16). Within this valid range of relative pressures, the internal consistency of the B.E.T. method has been demonstrated by Orr and Dallavalle (ref. 17), and by Young and Crowell (ref. 18). According to their measurements on several solids, the degree of correspondence between the specific surface areas obtained with several adsorbates allow confidence to be placed in the method.

The three-point B.E.T. method can be reduced to only one data point to otter the advantages of simplicity and speed but often with some loss in accuracy. The comparisons between the three-point method and the single-point method are well discussed by Lowell and Shield (ref. 13). For this research, we used the three-point method to obtain a higher accuracy.

Experimental procedure

The apparatus used is a Quantasorb B.E.T. surface area analyzer. A data acquisition board is installed in a PC. The computer was connected to the surface area analyzer to acquire data and perform data reductions. The specific surface area was calculated by Quantasorb software version 1.10 and saved in the computer.

The gases used in this experiment were: (a) ultrapure nitrogen, (b) krypton and helium mixtures with a ratio of 0.00036 mole krypton to 1 mole helium, (c) krypton and helium mixtures with a ratio of 0.00075 mole krypton to 1 mole helium, (d) krypton and helium mixtures with a ratio of 0.001 mole krypton to 1 mole helium. Krypton is an adsorbate. Helium is used as an inert carrier gas of krypton. Pure nitrogen is used as a calibration gas. The reason for applying krypton as an adsorption gas lies in the fact that ground rubber materials have low surface area. With krypton as an adsorbate, the ability to use larger samples of low surface area powders facilitates measuring low surface area because of larger signals generated without the problem of thermal diffusion (ref. 13). Three different concentrations of krypton were used to obtain the three-point B.E.T. plot.

The experimental procedures consist of the following steps:

* Purge system - Before turning the system on, the system is purged by inert gases for five minutes to purge out moisture and oxygen for the purpose of protection of the electronics from possible heat oxidation.

* Turn system on - After purging the system for five minutes, the power is turned on, then the computer and monitor is started, and the Quantasorb computer program is invoked. Subsequently one needs to type into the computer the atmospheric pressure, room temperature ([degrees]C) and the three concentrations of krypton.

* Sample preparation - The rubber sample is dried first at 65[degrees]C in an oven overnight. Then the weighed dried sample is degassed at 65[degrees]C for three hours at the outgassing station.

* Testing - After the sample is well degassed, the adsorption and desorption processes are performed. The integration area under the desorption curve is then obtained to calculate the surface area.

* Calibration - The calibration gas is ultrapure nitrogen gas. The calibration will reduce any error from detector nonlinearity to a negligible factor so long as the signal peak is within 15 to 20% of the calibration peak.

Results and discussion

We have conducted the surface area measurement on two groups of ground rubber products. For each material, five replicate tests were performed to allow calculation of the mean value and standard deviation. The five repeat tests were carried out on different days, and some even by different test personnel. The standard deviation of measurements is a critical parameter needed to evaluate the repeatability of tests, and to judge the reliability of the Quantasorb B.E.T. surface area analyzer and experimental procedures.

Figure 2 shows typical adsorption, desorption and calibration signals yielding one point on a B.E.T. plot. The signals were monitored by computer by measuring the change in the thermal conductivity of the gas mixture. Signal areas are integrated by the built-in integrator. From the shape of adsorption and desorption curves, we can judge whether a sample is well prepared. The adsorption and desorption curves are symmetrical and bell-like for well prepared samples. The adsorption and desorption curves are skewed for samples that are not well prepared.

Two groups of ground rubber material were tested. They were ambiently ground rubber and cryogenically ground rubber, respectively. Five samples from each kind of ground rubber were chosen for testing. These samples are:

* Ambiently ground 20 mesh tread rubber.

* Ambiently ground 30 mesh tread rubber.

* Ambiently ground 40 mesh tread rubber.

* Ambiently ground 60 mesh tread rubber.

* Ambiently ground 80 mesh tread rubber.

* Cryogenically ground 20 mesh tread rubber.

* Cryogenically ground 30 mesh tread rubber.

* Cryogenically ground 40 mesh tread rubber.

* Cryogenically ground 60 mesh tread rubber.

* Cryogenically ground 80 mesh tread rubber.

Sieve analysis

The particle size distributions of these samples are characterized by sieve analysis and presented in tables 1-5. TR represents ambiently ground tread rubber materials, and CGR represents cryogenically ground tread rubber materials.

Table 1 - percent weight retrained on sieves and
pan for 20 mesh tread rubber.
      Sieve  20    40    60     80     100    120    Pan
      (mesh)
TR-20 W.R.   6.6   55.7  21.4   7.7    2.4    2.2    4.0
      (%)
CGR W.R.     0.3   36.1  31.1   15.3   5.4    4.9    6.9
-20   (%)
Table 2 - percent weight retrained on sieves and
pan for 30 mesh rubber
      Sieve  30    40    60     80     100    120    Pan
      (mesh)
TR-30 W.R.  1.3   28.2   37.3   14.0   4.8    4.3    10.1
      (%)
CGR W.R.    0.3   20.0   41.5   18.5   6.6    5.1     8.0
-30   (%)
Table 3 - percent weight retrained on sieves and
pan for 40 mesh tread rubber
      Sieve  40    60    80     100    120    140    Pan
      (mesh)
TR-40 W.R.  6.44   8.6   20.6   7.5    4.9     5.9   6.1
      (%)
CGR W.R.    4.04    6.6  22.6   9.5    6.5     6.0   4.8
-40   (%)
Table 4 - percent weight retrained on sieves and
pan for 60 mesh tread rubber
      Sieve  60     80    100    120    140    170    Pan
      (mesh)
TR-60 W.R.   3.7   33.3   13.8   10.9    11.6   8.21  8.5
      (%)
CGR W.R.     0.6   34.3  25.01   2.4     13.0    6.3  8.4
-60   (%)
Table 5 - percent weight retained on sieves and
pan for 80 mesh tread rubber
      Sieve   80   100      120    140    170   200    Pan
      (mesh)
TR-80 W.R.    4.8   20.5    15.9   17.3  12.2    8.9   20.4
      (%)
CGR W.R.     11.8   34.2    13.2   18.8   9.8    3.8   8.4
-80   (%)

The data of particle size distribution presented here are unique for the samples chosen for testing. Other distributions could be obtained depending on processing parameters. They do not imply which method, i.e., ambient grinding method or cryogenic grinding method, provides finer particle size distribution of the ground rubber materials. Nominal size and its distributions are affected by various factors such as grinding time cycle, choice of screens, just to name a few.

Using the data in tables 1-5, the cumulative percent weight retained versus mesh size can be plotted. Figure 3 is the relationship between cumulative percent weight retained and the sieve size for 20 mesh tread rubber. This plot shows that for these two materials the cryogenically ground rubber is finer than the ambiently ground rubber. Figure 4 represents the relationship between cumulative percent weight retained and the sieve size for 30 mesh tread rubber. This plot shows that the particle size distributions of the two samples are very close. Figure 5 represents the relationship between cumulative percent weight retained and the sieve size for 40 mesh tread rubber. Their particle size distributions are close. Figure 6 represents the relationship between cumulative percent weight retained and the sieve size for 60 mesh. This plot indicates that the ambiently ground rubber is finer than the cryogenically ground rubber for these two samples. Figure 7 represents the relationship between cumulative percent weight retained and the sieve size for 80 mesh tread rubber. This plot indicates that the ambiently ground rubber is finer than the cryogenically ground rubber for the two samples.

Surface area

In table 6 are listed the mean values of specific surface areas and sample standard deviations of the ambiently ground rubber materials tested. In table 7 are listed the mean values of specific surface areas and sample standard deviations of the cryogenically ground rubber materials tested. The small standard deviations in the two tables demonstrate that the five replicate test results for a sample are very close, which means that the repeatability of tests is very good and the tested results are reliable. [TABULAR DATA 6 AND 7 OMITTED]

The authors would like to point out here that in the production of ground rubber products and in the sieve analysis process, in order to eliminate the influence of static charge (ref. 6). a certain percentage of talc is used. The surface area of talc is high. The talc used in our laboratory has a specific surface area of 2.179 m2/g. The ground rubber materials discussed in this article contained 5% or less talc, usually much less. We did not make any adjustment to the specific surface areas of ground rubber materials due to the use of talc.

Figure 8 represents the relationship between specific surface area and mesh size for specified rubber materials by using the data in tables 6 and 7. This plot shows that cryogenically ground rubber materials tend to have lower specific surface area than ambiently ground rubber materials with the same or similar particle size distribution. This is true even for some cryogenically ground rubber with finer particle size distribution as shown in figure 3.

Several factors could have influence on the surface area of ground rubber. These factors include the average particle size or nominal particle size, particle size distribution and the morphology of ground rubber material. The nominal particle size and particle size distribution are unique and determined by a specific grinding process. If the influence of morphology of ground rubber is not considered, ground rubber with smaller nominal size should have a larger surface area. But the actual situation is not so simple. The morphology of ground rubber, i.e., the surface shape of ground rubber material has a considerable contribution to the surface area. This is easily understood by looking at the operation principles of these two rubber grinding methods. In the ambient mill grinding method, rubber is ripped and torn apart by open mills at room temperature, leaving the rubber surface rough and irregular, resulting in larger surface area. In the cryogenic grinding method, rubber is "frozen" at a temperature like minus 200[degrees]F and "broken" by the striking action of a hammer mill with little or no tearing action exerted on rubber. The rubber particle surfaces are somewhat smooth, resulting in smaller surface area. Material with coarser particle size distribution tends to give smaller surface area, and material with rough and irregular surface tends to give larger surface area. The difference of particle size distribution usually can not compensate for the difference of morphology. This explains why ambiently ground rubber with coarser particle size distribution compared to cryogenically ground rubber with finer particle size distribution as shown in figure 3 for 20 mesh tread rubber has greater surface area. The relationship between morphology and surface area, and how the morphology of ground rubber material quantitatively contributes to the surface area, needs more investigations to further enhance our knowledge about ground rubber.

Conclusions

From the above results and discussion, we come to the following conclusions:

* The repeatability of tests using the B.E.T. surface area analyzer is very good.

* The B.E.T. surface area analyzer can be used to determine the specific surface area of ground rubber materials.

* The specific surface area of ambiently ground rubber tends to be greater than that of cryogenically ground rubber with the same or similar particle size distribution.

References

[1.] A. Leighton, "The use and handling of particulate rubber," poster session paper, no. VII, ACS Rubber Division, Toronto, May 1983. [2.] Fred J. Stark Jr., "Rubber compositions and method," U.S. Patent 4,481,335(1984). [3.] F.L. Roberts, P.S. Kandehal, E.R. Brown and R.L. Dunning, "Investigation and evaluation of ground tire rubber in hot mix asphalt," National Center for Asphalt Technology, August 1989, Auburn, AL. [4.] Fernley G. Smith and William B. Klingensmith, "Tires as a consumer and source of recycled materials," paper no. 15, ACS Rubber Division, Washington, October 1990. [5.] Ed Fesus and Sam McGrath, "Particulate rubber as a compounding tool," Rubber & Plastics News, June 8, 1992. [6.] Gerald W. Holland, Benfei Hu, Mark A. Smith; "Particle si.e analysis for granulated rubber," Rubber & Plastics News, October 12, 1992. [7.] Gerald W. Holland, Benfei Hu, Mark A. Smith, "Chemical identification of typical recycled tire rubber," Rubber & Plastics News, December 10, 1993. [8.] ASTM Standard D3037 "Standard test methods for carbon black - surface area by nitrogen adsorption," Annual Book of ASTM Standards, Vol. 09.01, Philadelphia, PA, 1991. [9.] I. Langmuir, J. Amer. Chem. Soc. 40, 1368, (1918). [10.] S. Brunauer, P.H. Emmett and E. Teller, J. Amer. Chem. Soc. 60, 309, (1938). [11.] W.D. Harkins and G. Jura, J. Amer. Chem. Soc. 66, 919, (1944). [12.] T. Allen, "Particle size measurement," 2nd ed. John Wiley & Sons, New York, 1975. [13.] S. Lowell, Joan E. Shield, "Powder surface area and porosity," 2nd ed. Chapman and Hall, London, 1984. [14.] Quantachrome Corporation, Quantasorb Surface Area Analyzer Manual. p. XI-3. [15.] S. Gregg, "The surface chemistry of solids," 2nd ed. Reinhold, New York, 1961. [16.] S. Gregg and K. Sing, "Adsorption, surface area and porosity," 2nd ed. Academic Press, New York, 1982. [17.] C. Orr, and J.M. Dallavalle, "Fine particle measurement," Macmillan, New York, 1959. [18.] M.D. Young and A.D. Crowell, "Physical adsorption of gases," Butterworth, 1962.