Efficacité comparée de divers procédés physiques de séparation des argiles d'un sable de gisement

A Comparison of Different Physical Processes for the Recovery of Clays from an Unconsolidated Sandstone

¹ ENSG Nancy
² Université Toulouse III

Résumé

Divers procédés physiques (mixage, attrition, vibration ultra-sonique) ont été utilisés pour désagréger et disperser les argiles d'un sable de gisement. Le traitement aux ultra-sons s'est avéré le plus efficace : après 10 min de vibration on récupère environ 1,5 fois plus de particules fines (< 40 µm) qu'après 24 h d'attrition et 2 fois plus qu'après 3 min de mixage. Ce temps de 10 min est suffisant pour récupérer la totalité des argiles libres, pelliculaires ou faiblement cimentées, c'est-à-dire les argiles appelées activescar susceptibles d'être en contact avec les fluides extérieurs. Des temps plus longs entraînent une microdivision des plus gros grains, marquée par un net changement de pente de la courbe granulométrique cumulative des particules récupérées, sans effet sur le pourcentage d'argiles recueillies. L'étude critique des conditions expérimentales montre que seuls la puissance et la fréquence des ultra-sons, le volume traité et le temps d'application agissent de façon sensible sur la dispersion des argiles. Un traitement de 10 min avec une puissance de 210 W, une fréquence de 20 kHz et une concentration en solide de 10% permet de récupérer la quasi-totalité des argiles actives d'un sable de gisement. Ce traitement a été retenu en vue du dosage des phases argileuses dans des séries d'échantillons.

Abstract

The characterizing of clays is an essential operation for understanding many problems encountered in production. It goes via a preliminary phase in which species are separated quickly and effectively without altering their nature. Chemical processes disperse clays by solubilizing their cements, amorphous compounds and poorly crystallized oxyhydroxides. But they attack fragile species and lead to mistaken results. Therefore, we have chosen various physical processes (mixing, attrition, ultrasonic vibration) and compared their efficacy with a reservoir sand. This sand was a quartzose arenite in which 96. 6 % of the grains were between 63 µm and 2 mm. The < 40 µm fine fraction, obtained by wet sieving, represented 2. 44 % of the unprocessed sand, and the clayey fraction (kaolinite, illite, montmorillonite, interstratified illite-montmorillonite) was 1. 20%. Mixing was done with a bladed stirer. 40 g of sand and 100 cm³ of demineralized water were stirred for 3 min at a speed of 3000 rpm. Attrition was done by two devices :(1) For attrition by rotation, 400 g of sand and 0. 5 liter of demineralized water were placed in a porcelain jar rotating at a speed of 37 rpm. Tests were performed with water alone and with the adding of a dispersant (0. 3 % sodium hexametaphosphate) and a surfactant (0. 6 mole/liter of sodium sulfonate paraoctylphenyl). (2) Attrition by a Turbula TC2 mixer, which gave the jar a complex rotation movement including shaking and rhythmic rocking. 500 g of sand were thus processed in 1 literof 3 % sodium hexametaphosphate solution for varying times. Ultrasonic processing was done with a Sonifer 830 device. The probe was plunged into a suspension made up of 440 g of sand and 300 cm³ of demineralized water processed for times varying from 5 to 100 min. The efficacy of the processing was measured by the weight percent of the < 40 µm fraction. This percentage was 2. 08 % on the average for mixing, 2. 60 % for simple rotation and 2. 87% for rotation by the Turbula device. The adding of dispersant or surfactant had no significant effect. Increasing the rotation time had a minimal effect, in that the percentage of particles recovered went from 2. 21 to 2. 60 % when time increased from 6 to 24 h. Ultrasonic processing was by far the most efficacious and fastest. Indeed, after 5min, an average of 3. 60% of < 40 µm particles was recovered, 4. 10 % after 10 min and7. 18 % after 100 min. The SEM examination of the fractions processed showed that only ultrasonic processing was capable of recovering all of the free clays, either pellicular or trapped in the dissolution cavities or cracks of the quartz grains, i. e. the so-called activeclays. It was also found that this recovery is selective. Montmorillonite and illite are set free at the beginning, while kaolinite is released all during the tests. At the beginning, it was disordered kaolinite, but after 15-20 min of processing, it was ordered kaolinite coming from the microdivision of mica crystals with which it had been associated. This microdivision is clearly visible, moreover, on the cumulative grain-size curve of the < 40µm particles recovered after ultrasonic processing. This curve has two parts:(1) The first part has a steep positive slope linked to the recovery of activeclays. (2) The second part has a much gentler slope linked to the recovery of clayey aggregates and mica and quartz grains microdivided by ultrasounds. The tests performed and a bibliographic study of various processes and operating conditions led us to choose ultrasounds as the process for separating clays from a reservoir sand. The best results were obtained with the following experimental conditions : instrument frequence = 200 kHz, instrument power = 210 W, suspension = 40 g of sand in 300 ml of demineralized water, processing time = 10 min. This quick and efficacious processing preserves the mineral species.