Particle Size Optimisation of Sand to Improve Water Treatment Efficiency

Introduction

The availability of safe supplies of drinking water is a worldwide concern owing to the effects of pollution, water-borne disease and increased consumption. ¹ We must continue to develop effective, sustainable, easy to use methods for treating water to allow safe re-use by living organisms.
The natural water cycle uses evaporation and soil filtration as the principal elements for water purification. ²³ As we continue to cover the surface of the Earth with non-permeable materials (buildings and roads) we reduce the effectiveness of the natural water cycle for water filtration.
Sand filtration is a low cost, readily available method for water purification. ⁴ The goal of this project is to define the effect of particle size on the performance of sand as a treatment media for improving the pH and alkalinity of water. The hypothesis of this project is that the particle size composition of sand can be modified to improve its performance as a water treatment media, relative to as-received sand.

Materials and Methods

All purpose sand (Quickrete, inc) was purchased from a garden supply store (Lowes). A series of eight calibrated sieves (Allen-Bradley) were used to separate the all-purpose sand into different particle size fractions based on the sieve sizes, as shown in Figure 1. A triple beam balance was used to weigh all particle size fractions collected from each sieve.

Water treatment columns were prepared by filling 20 cm long, 1.75 cm internal diameter plastic tubes with sand samples to 4 cm height (7 cm³ by volume) and tapping the tubes on a hard surface to remove gaps or holes. Sections of coffee filters were placed in the bottom of the permeation tubes prior to filling with sample to hold the sand in place.
The permeation times of sand samples were measured by adding 200 ml of distilled water to the permeation tubes, collecting every 20 ml of permeate in labeled collection bottles and measuring the time of each 20ml water collection. The permeation time measurement setup is shown in Figure 2.

A pH meter and calcium ion selective electrode (Vernier Scientific) were calibrated using high and low calibration standards (Vernier Scientific). ⁵⁶ The experimental setup used for measuring the pH and Ca²⁺in each of the permeate samples is shown in Figure 3. The calibration standards used for calibrating the electrodes and the actual electrode responses to the standard solutions after running them as samples (before and after running the permeation samples) are shown in Table 1. All electrode responses to the calibration solutions were within 5% of their stated values.

The repeatability of the pH and ion selective electrode measurements was evaluated by performing five repeated measurements on separate samples of tap water. The data from this repeatability test is listed in Table 2. The repeatability relative standard deviation (RSD) for all test methods was equal to or less than 3.3%.

Results

Particle size composition of all-purpose sand

The all-purpose sand used in this study is principally composed of silicate and calcium carbonate particles. The sand particles are composed of a wide range of particle sizes. As shown in Figure 1, calibrated sieves were used to separate the all-purpose sand into eight separate particle size fractions. The weight of each particle size fraction was used to calculate the weight % particle size composition of all-purpose sand, as shown in Figure 4. The data shown in Figure 4 is based on averaged results from five independent particle size separations of all-purpose sand. The individual particle size composition data is listed in Table 3. The relative standard deviations of the particle size composition data ranged from 7.6 to 18.6 %, depending on particle size fraction. This data suggests that some particle size segregation may have occurred in the bag of all-purpose sand during storage and handling.

Impact of sand particle size on water permeation time

The effect of particle size on water permeation time was studied by comparing the water permeation time of all-purpose sand to the permeation times observed for each of the eight sand particle size fractions shown in Figure 4 and Table 3. The average results from two independent permeation tests of all-purpose sand and the eight separate particle size fractions of all-purpose sand are shown in Figure 5. As shown in Figure 5, the <150 micron particle size was the only sample that showed a longer permeation time (slower permeation rate) than all-purpose sand. This observation suggests that removing the <150 micron particle size fraction would significantly reduce the water permeation time (increase the permeation rate) of all-purpose sand.

Samples of all-purpose sand were separated into eight individual particle size fractions as discussed above (see Figure 1). For each sand sample, one particle size was selectively removed and the remaining particle size fractions of sand were recombined. In this manner, eight sand samples were prepared where a different particle size fraction was removed from each sample. 200 ml of deionized water was used to measure the permeation time for each of these samples. Figure 6 and Table 4 show the average permeation time performance for two independent tests of all-purpose sand and sand with one particle size fraction removed. It is interesting to note that none of the particle size modified sand samples exhibited longer permeation times than all-purpose sand. The comparative performance data shown in Figure 6 shows that the selective removal of the <150 micron particle size fraction exhibits the largest reduction in water permeation time (increase in permeation rate). In fact, removing the <150 micron particle size fraction (11.3 wt %) reduced the water permeation time by 55%, from 133.5 to 60.3 seconds.

Impact of sand particle size fraction on water pH

The pH was measured on each 20 ml sample collected from the water permeation test shown in Figure 5 in order to determine if sand particle size has an impact on the efficiency of sand for treating the pH of deionized water. The pH data shown in Figure 7 and Table 5 is the average of two independent permeation tests on each sand sample. As shown in Figure 7, all particle size fractions of sand, including all-purpose sand, raise the pH from an initial value of 6.25 at 20 ml of permeate to a final value of 6.8 after 200 ml of permeate water are collected. Given that the repeatability (1 standard deviation) of the pH test method is + 0.1 pH units, the data shown in Figure 7 indicates that all sand particle size fractions have the same effect on the pH of permeate water (to 95% confidence interval). This observation suggests that increasing permeation rate will increase water treatment efficiency (more water can be treated per unit of time) without reducing the pH treatment capacity of the sand.

The calcium ion (Ca²⁺) concentration in all the permeate samples shown in Figure 5 were also measured using a calcium ion selective electrode in order to determine the effect of sand particle size on the alkalinity of filtered water. The calcium ion concentration data, shown in Figure 8 and Table 6, is highest for the initial 20 ml permeate sample collected from all-purpose sand, as well as all eight of the sand particle size fractions used to filter deionized water. The calcium ion concentration falls off rapidly with increased permeate volumes for all sand particle size fractions tested in this study.

 
It is interesting to note that the initial calcium ion concentrations observed in Figure 8 are correlated to the particle size of the sand. For example, the <150 micron particle size sand showed the highest initial calcium concentration, while the <4760 >2380 micron fraction showed the lowest initial calcium concentration. This relationship between initial dissolved calciumconcentration and sand particle size can be described by a logarithmic equation, as shown in Figure 9, with a correlation coefficient of 0.86.

 
The relationship observed between initial permeate calcium concentration and sand particle size supports several observations, as listed below:
1) The smallest sand particle size fraction releases the most dissolved calcium in the initial 20ml permeate sample.
2) The concentration of dissolved calcium concentration in the initial 20 ml permeate sample is inversely correlated to the logarithm of the sand particle size fraction used the filter the water.
3) Dissolved calcium concentration is greatest for the initial 20 ml permeate sample and falls off rapidly in successive permeate samples from all sand particle sizes.
The relationship between sand particle size, permeate volume and dissolved calcium concentration support a hypothesis where the dissolved calcium comes from dissolution of superfine particles and soluble calcium compounds (i.e CaCO₃) on the surface of the sand particles. In both cases, the soluble calcium would be greatest for the finest particle size fraction of sand, and would decrease with “washing” by increased volumes of deionized water, as shown in Figure 8.

Conclusions

Measurements of permeation time, water pH and dissolved calcium ion were used to evaluate and compare the performance of all-purpose sand, particle size fractions of all-purpose sand and sand with specific particle size fractions selectively removed. The results of this study support a number of observations, as listed below:

• The finest sand particle size releases the highest initial Ca²⁺ concentration (logarithmic relationship)
• After the initial 40-80 ml permeate volume, the impact of particle size on Ca²⁺ concentration is greatly reduced.
• The initial 20-40 ml permeate samples appeared cloudy.
• These observations suggest that ultrafine material is washed from the sand column with initial volumes of permeation water.
• Removing the -150 micron particle size fraction from sand (11.3 wt %) increases the water filtration rate by 55%, while allowing the same overall pH buffering performance.