Understanding agglomeration behavior in green sand.

Green sand analyses at two foundries provide a way to determine the extent of agglomeration, helping to solve defects and evaluate material or equipment changes to your sand.

A mass or ball of silica grains, clay and additives with water forms during mulling and aeration, and molding compresses these agglomerated green sand constituents into a near-maximum bulk density. After shakeout, the mass may or may not fully break up during subsequent processing and can be returned to green sand molding stations. Current technical literature provides little description of this condition in present-day green sand. This article - adapted from a 1999 Casting Congress paper - is aimed at describing the particle and agglomerate size distributions through sieve analysis methods and their relationship to clay and fines in green sand.

The complete report provides a more detailed look at the analyses, providing information such as core sand data, base silica grain distributions, cumulative surface area for sand, and a schematic illustration of particle size ranges that can participate in agglomeration.

Sieve Analysis

Green sand has a silica grain distribution that originates in the core sand and enters the system. The core sand sieve analysis must be known because it appears in all subsequent green sand analyses. If a second silica sand is added to change the base silica sieve [TABULAR DATA FOR TABLE 1 OMITTED] analysis from that of the core sand, then the resulting base silica sieve analysis should be calculated. The term "core sand" is used to designate the base silica sand sieve analysis in the system, whatever its source. Agglomeration begins from this base.

For comparison with the base silica sieve analysis, two other sieve analyses are required: the AFS sieve analysis without clay plus the total percent weight of all AFS clay (washed molding sand - WMS) and the dried agglomerated sieve analysis (unwashed molding sand - UWMS). In addition to core sand, WMS and UWMS sieve analyses, percent methylene blue (MB) clay, percent loss on ignition and percent volatile combustible material are required to follow agglomeration behavior.

Foundry Sand Data

Green sand data from two Wisconsin high-volume iron foundries, Grede Foundries, Inc., Reedsburg, and Neenah Foundry Co., Neenah, were used for this study. Both are vertical-parted molding shops and sand sampling, and testing procedures were done according to each foundry's regular procedures. Screen sets were those normally used for control testing.

Grede used a single base silica core sand, however sand is often transferred from other systems and, on occasion, other silica sand can be added. Also, the samples were randomly chosen from the beginning and end of a 5-year period. One sieve analysis data set came from a supplier's laboratory, providing a long-range look at agglomeration behavior in a particular system with numerous inputs.

Neenah data was randomly selected from a shorter time period - January through February and May through July of 1998. In this foundry, a regular addition of silica sand was used to raise the 140-pan-size particles above that from the core sand dilution.

Upon calculating the silica grain base, the number of particles and their surface area for both foundries and comparing the cumulative weight percentage vs. U.S. sieve number curves, the core sand grain fineness number (GFN) difference was revealed at 56.3 for Grede and 67.2 for Neenah.

Agglomeration Analysis

Figure 1 illustrates the effect of percent AFS clay on position of the termination point for three Grede sands of 14.53, 11.53 and 10.1% AFS. In each case, UWMS agglomerates are present on 12, 20 and 30 sieve sands, which are not present in WMS. The UWMS all must terminate at 100%.

Changing from cumulative weight to cumulative numbers of particles or agglomerates greatly changes the graphs. The WMS and UWMS sieve analyses are converted from retained and cumulative percent by weight to retained and cumulative number and surface area of particles using multiplying factors in Table 1. The particles are considered to be spheres. The multiplying factors are based on using U.S. sieve opening sizes as the diameter of particle, or agglomerates, passing through one sieve and retained on the next. The order of magnitude is the same, but shifted accordingly.

In Fig. 2, comparison is made with Grede sands of 11.5 and 10.1% AFS clay and 9.57 and 8.3% MB clay. Note that particle and agglomerate numbers on sieves decline with decreasing clay percentages. Notice the divergence of the particle number curves beyond 70 sieve.

This fact is emphasized when the parameter WMS-UWMS is plotted against sieve number [ILLUSTRATION FOR FIGURE 3 OMITTED]. WMS minus the UWMS is a measure of the agglomeration behavior of the sand. A large WMS-UWMS value indicates strong agglomerating action. The 12-50 sieve agglomerates shown in Fig. 1 do not appear on Fig. 3 because of scale limitations.

Sands from Neenah show particle numbers similar to those of Grede.

AFS and MB Clay Analysis

The sieve fractions of UWMS were analyzed' for percent MB clay for both Grede and Neenah sands. The WMS contains no MB clay. Figure 4 shows the percent MB clay in each sieve fraction. This MB clay distribution appears similar to the WMS-UWMS distribution in Fig. 3. Because the weight percentage of agglomerates is maximum in the 50-140 sieve range, the greatest percentage of the total percent MB present is on these sieves. As expected, the highest percent MB is reported for the 200, 270 and pan, which was analyzed as a composite sample because of its limited weight. Much of the freshly added clay possibly will appear on these sieves. The high percent MB on the 12, 20 and 30 sieve material conforms to its disappearance, or washes away, when comparing the UWMS and WMS sieve analyses.

The total percent AFS clay is comprised largely of the percent MB clay present together with fine silica and carbon. Excluding the latter, the MB is usually about 80-85% of the settling test AFS clay. The AFS clay composite undergoes agglomeration with the MB clay during final mulling. The agglomerates respond to molding and heating from the casting process.

Surface Area

Surface area conversion factors for sieve analysis are given in Table 1. Cumulative surface areas are shown in Fig. 5 for Grede and Neenah core sands. The Grede sands show higher surface area on a sieve than the Neenah sands. This is expected, since the GFN of Neenah core sand is 67.2 vs. 56.3 for the Grede core sands. The Grede core sand has a total surface area of 8375 sq cm and the Neenah core sand has 10,148 sq cm. These area differences account for the WMS and UWMS particle number differences, as well as the cumulative area curves.

Agglomerates and Particles

Agglomerates result from particles sticking together during sand conditioning and molding, and spring from surface interactions of silica grains, clay and water. Figures 2 and 5 reveal the following order of participation. The core sand provides the beginning surface area [ILLUSTRATION FOR FIGURE 5 OMITTED]. Numbers of UWMS agglomerates vary little from core sand and WMS numbers up to 100 sizes because the major cumulative increase in surface area occurs up to the 100 sieve size ([ILLUSTRATION FOR FIGURE 2 OMITTED] as related to [ILLUSTRATION FOR FIGURE 5 OMITTED]). In other words, the major agglomeration increase defined as WMS-UWMS follows the major surface area increase up to the point of divergence indicated on Fig. 3. Then the 140 to pan sizes react with clay to form fine agglomerates until core sand fines are consumed [ILLUSTRATION FOR FIGURE 5 OMITTED]. The percent MB clay and percent AFS clay also will influence the extent of agglomeration.

Fines Accumulation

Examination of Fig. 2 shows that Grede core sand has less fines on 140 to pan than its WMS fines. But, it also was found that Neenah core sand has more fines than its WMS. This again relates to the difference in sieve analyses and GFN of the two base sands. Another source of the higher WMS sand fines is attrition of silica grains in mulling, shakeout and blasting in sand processing. In addition, thermal bonding of clay, coal ash, coked seacoal and soot particles can build up fines or produce coarse agglomerates. Microscopic examination indicates both sources. Fines generation is an inherent result of both particle attrition and agglomeration. Graphs of percent retained vs. sieve number are the simplest way to observe whether fine or coarse particles are increasing on any sieve.

Need for Standardized Testing 

The dried unwashed sieve analysis is not presented as a standard test in the AFS Mold & Core Test Handbook (1989). Inquiry among sand laboratories reveals differences in testing procedure as follows:

* point of sampling the green sand;

* testing for AFS clay by the 20- or 25-micron method produces different results. The 20-micron method was used in this study;

* assuring that both WMS and UWMS results are from the same sample;

* drying of the sand for UWMS sieve analysis is not standardized. Some air dry for 24 hr; some bake 1 hr at 220-230F (104-110C), and some air dry to less than 20% compactibility;

* the screening process time is not standardized. Some screen for 2.5 min and others for 5-15 min before obtaining retained sieve weights;

* sample size is not standardized. Percent MB clay content of UWMS sieve fractions could be gained more easily from larger samples;

* conversion factors for relating weight to particle and agglomerate numbers and surface area should be standardized. In this article, conversion factors are based on spheres having a diameter equal to the sieve opening preceding the one on which the particles are retained. Table 1 used an average value for successive screens. No correction for shape of grain was made.

If the UWMS sieve analysis is to become useful, the testing and reporting procedure should be standardized.

Research on agglomeration in green sand could be fruitful toward better understanding of sand behavior and performance. Suggested subjects include:

* compactibility and moisture - the examples provided were mulled to about 38-43% compactibility. Change in compactibility and related moisture content would change agglomeration results;

* change in grain fineness and distribution;

* additives or pH adjustment;

* change in clay type from sodium bentonite in this arttice to calcium bentonite or mixtures;

* change in the sand processing system mechanisms;

* coated core sand dilution;

* presence of other aggregates;

* carbon or coke fines.

One purpose of this article is to suggest a neglected aspect of sand control testing and process control measurement.

Findings

Agglomeration of particulates is inherent to processing green sand for molding. Study of the extent of agglomeration requires unwashed dried sieve analysis and its percent MB clay plus the standardized AFS clay and washed sieve analysis. Sieve analysis of the core sand entering the system should be known. From this information, several conclusions are supported, including:

* the base silica sand sieve analysis from cores or new sand additions prevails in agglomerate formation;

* percent MB and AFS clay both influence' the extent of agglomeration. Higher AFS clay from higher MB clay are the major contributors to the extent of agglomeration. Higher percent AFS clay leads to a greater percent change in agglomerates;

* AFS GFN, sieve analysis and its surface area exert a dominant effect on agglomerate formation in green sands;

* the number of particles and agglomerates varies much more than the total surface area of WMS, UWMS and green core sand;

* divergence of WMS, UWMS and core sand number of particles or agglomerates begins at 70, 100 or 140 sieve sizes increasing with AFS GFN;

* fines accumulation on 140 to pan sieves is an ongoing process arising from either grain attrition or stable fine agglomerates.

Sand performance in molding and casting may, in part, be related to variation in the agglomerated condition still existing in the sand. The absence of a database of agglomeration conditions prevents such correlation at this time. In this context, a standardized UWMS test procedure is necessary.

This article was adapted from a much larger paper (99-30) presented at the 1999 AFS Casting Congress. A copy of the complete report is available from the AFS Library at 800/537-4237.