The Impact Fragmentation Tendency of Limestone Particles in Calcium Looping Cycles in the Presence of Steam and Sulphur Dioxide

Scala, F., Istituto di Ricerche sulla Combustione, Consiglio Nazionale delle Ricerche
Coppola, A., University of Naples Federico II
Esposito, A., University of Naples Federico II
Montagnaro, F., Università degli Studi di Napoli Federico II
Salatino, P., Università degli Studi di Napoli Federico II
This manuscript focuses on the calcium looping process, a flue gas post-combustion treatment aimed at carbon dioxide capture and its further release in a concentrated stream to be geologically stored or utilised in the chemical industry. Calcium looping is based on the reversible carbonation of Ca-based sorbents, such as those derived from natural (and cheap) limestone. It is a process carried out by means of two reactors (a carbonator and a calciner), typically in a dual interconnected fluidised bed configuration.

Processing sorbent particles in a fluidised environment unavoidably leads to attrition and fragmentation phenomena, with consequent changes in the particle size (and residence time) distribution, which in turn may influence the CO2 capture capacity. Moreover, the possible presence, in the combustion flue gas entering the carbonator, of SO2 and/or steam would have relevant effects on the sorbent reactivity. This and other research groups have widely investigated in the recent past the effect of the presence of SO2/H2O in the carbonator atmosphere, with reference to the effect of these species on the CO2 capture capacity and lately on the attrition tendency of the sorbent particles, while the specific effect of sulphur dioxide and steam on the impact fragmentation tendency of the sorbent in calcium looping cycles has been so far mostly neglected.

Accordingly, the aim of this work was to give a contribution in this direction. Calcium looping experiments have been performed in a lab-scale twin fluidised bed apparatus (consisting of two identical interconnected bubbling fluidised bed reactors), able to reproduce a realistic particle thermal/chemical/mechanical history, using a reference high-Ca limestone. Calcium looping tests consisted of ten calcination/carbonation cycles. Carbonation was carried out at 650°C in an atmosphere containing 15% CO2 to simulate a typical combustion flue gas, while calcination was operated at 940°C at 70% CO2 to simulate oxy-combustion conditions. Six different operating conditions for carbonation were tested to study the effect of SO2 and/or H2O, where steam (when present) was fed at 10%, and SO2 (when present) either at 75 ppm (a typical concentration in a pre-desulphurised combustion flue gas) or 1500 ppm (raw combustion flue gas). The CO2 capture capacity was calculated for each carbonation stage and, at the end of the test, sorbent particles were further analysed for the determination of the degree of Ca sulphation.

After the tests, the exhausted sorbent particles were fed to an ex situ impact test apparatus, which is based on the well-established concept of entraining particles in a gas stream at controlled velocity, and impacting them against a target. The test rig consists of a vertical stainless steel eductor tube equipped with a particle feeding device. After feeding, the particles are accelerated by the air flow in the eductor tube. The particle velocity is controlled by regulating the air flow in the eductor tube, by means of a flowmeter. When the particles exit the eductor tube, they impact on a rigid target plate placed in a collection chamber 50 mm below the bottom end of the tube. The target is made of stainless steel and is inclined by 30° with respect to the horizontal. This inclination was chosen as a trade-off between the need of avoiding interference between impacting and reflected particles, and the need of minimising the departure of results from those obtained with a target perpendicular to the particle trajectory. The device is designed so as to minimise the loss of particles entrained by air in the chamber, enabling easy collection of the sorbent particles after impact for further analysis.

Samples (approximately 1.0 g in the particle size range 0.4–0.6 mm) of raw limestone (for reference), and of exhausted sorbent particles generated in the twin bed apparatus under the six different operating conditions, were used for fragmentation tests in the impact testing apparatus. The tests were carried out in air at ambient temperature with the following particle impact velocities: 10, 17, 24, 31, 38, 45 m s–1. After each test, the samples were retrieved from the collection chamber and weighed. The collected particles were then sieve-analysed to obtain their particle size distribution (with related probability density functions), and to calculate the fraction of fragments (<0.4 mm) cumulatively generated. The analysis of data, as a function of the impact velocity, allowed to discuss the impact fragmentation tendency of each sample in the light of the broad particle breakage mechanisms “chipping”, “splitting” and “disintegration”, as defined in literature by this and other research groups. The behaviour was then critically related to the presence, in the calcium looping cycles generating the samples, of steam and/or SO2, which produced sorbent particles containing CaO, CaCO3 and CaSO4 (materials which have different mechanical properties) to various extents as inferred from the analysis of the CO2/SO2 degree of capture.

The raw sample resulted to be the most resistant among all those tested, as it is rich in hard CaCO3. Samples obtained in presence of SO2 resulted to be harder than in absence of it, as higher SO2 concentrations in the carbonator determined thicker CaSO4-based shells around the particles. This is in particular relevant at low impact velocities, where chipping is the main fragmentation mechanism (while splitting/disintegration are active at higher velocities) and the impact tendency of the particles is governed by the mechanical properties of their outer peripheral layer. The presence of steam in the carbonator, orientating the reactivity of CaO towards CO2 rather than SO2 (when present), on the one hand determines particles less resistant than those obtained in absence of steam and in presence of the sole SO2 together with CO2. On the other hand, as steam indeed favours the carbonation reaction of CaO, we have observed particles with a larger CaCO3 fraction and therefore more resistant than those obtained in conditions where both steam and SO2 were absent. From the quantitative point of view, the fraction of fragments resulted, at 10 m s–1, in the range from 1.5% (raw limestone) to 4.7% (sample exhausted in absence of both steam and SO2). At 45 m s–1, these values increased in the range from 5.4% to 23.2%, depending on the sample.