Factors governing the formation of porosity in metal loaded cellulose during pyrolysis and the effects of pore structure on reactivity in O₂ and NO

Abstract

A series of carbons have been made from cellulose that has been loaded with varying amounts of calcium, potassium and iron. Small angle X-ray scattering showed that the calcium-loaded chars produced carbons with mass fractal properties whereas potassium loaded chars produce carbons with surface fractal properties in the mesopore region. Calcium loaded chars were more microporous than potassium loaded chars. Iron loaded cellulose produced a Small Angle X-Ray Scattering (SAXS) pattern with two linear regions. The high q-scattering was characteristic of a highly microporous material. Differences in carbon properties were attributed to crosslinking reactions induced by the metals during pyrolysis. The reactivity of the carbons in O₂ and NO was determined and the ratio of the reactivities, R(NO)/R(O₂) noted. It was found that the presence of calcium and potassium both decreased R(NO)/R(O₂), but iron increased R(NO)/R(O₂) for a certain loading range. The latter was attributed to the reduced catalytic activity of iron for O₂ gasification and the pore structure of the iron loaded carbons.

Introduction

The possibility of the removal of NO from flue gases by using fixed beds of charred coal is attractive because it is potentially both cheap and efficient. It also offers a new market for coal products, using coal itself to improve the environmental impact of coal combustion.

One problem in the use of carbons as NO reductants, which has received little attention, is the much higher reactivity of carbon in O₂ compared to NO. Combustion flue gases typically contain 4–5% of O₂. Although a number of studies have shown that carbon reactivity in NO is enhanced by the presence of small amounts of O₂ [1], [2], [3], it is nevertheless desirable to choose a carbon for which the reactivity in NO, R(NO), is as large as possible compared to O₂, R(O₂).

The ideal carbons for development should be easily available and as cheap as possible and there is interest from coal producers to use chars derived from coal. Coals are extremely complex materials and the reactivity of a coal char will depend on a number of factors including the graphitic ordering, the presence and kind of any porosity and also the nature of any catalytically active species in the ash. In order to identify the best possible coals for development, it is first necessary to understand the important factors that determine R(NO), R(O₂) and R(NO)/R(O₂). Consequently, there have been a number of investigations of model systems, either pure carbon [4] or model carbons loaded with known inorganic species such as calcium [5], potassium [6] and iron [7].

Although carbon reactivity in O₂ and NO have been investigated individually, there have been few studies that directly compare them in the same carbon. As mentioned, it is known that carbon reactivity in NO is augmented by the presence of O₂ the factors governing R(NO)/R(O₂) are still largely uninvestigated.

Ruiz and Hall [8] have investigated the effects of porosity in low temperature NO gasification. They found no correlation with surface areas and R(NO) in agreement with earlier workers and also no correlation between R(NO) and R(O₂). The best correlation for R(NO) was with the fractal dimensions of the carbons in the mesoporous region. The most open porous structures, either surfaces with low surface fractal dimension or mass fractals, exhibited the highest reactivity.

The purpose of this paper is to investigate the effects of metal catalysts on gasification in NO and O₂. The metals chosen were calcium, potassium and iron because these gasification catalysts are commonly found in coals. One problem in this kind of study is that the presence of metals during carbonisation of the cellulose may produce different kinds of pore size distribution. Since it is known that the porosity in itself can influence R(NO)/R(O₂) it may not be possible to completely decouple the catalyst and porosity effects. Therefore, before the effectiveness of individual catalysts can be evaluated, it is necessary to understand the pore structure of carbons loaded with metals and how the presence of metals influences the development of porosity during pyrolysis.

The best possible characterisation of porosity is essential and special emphasis is placed on the interpretation of Small Angle X-Ray Scattering (SAXS) data. This is because SAXS supplies two important pieces of evidence. Firstly, it can identify and distinguish the presence of fractal structures. For surface fractals Schmidt [9] and Kjems and Schofield [10] have shown that:

$$\text{I(q)=I}{}_0\text{Γ(D−1)}\text{sin}\text{[}\text{π}\text{(D−1)/2]q}{}^{\mathrm{-(6-D)}}$$

where I₀ is a constant and Γ( ) is the gamma function, D is the surface fractal dimension and q is the scattering wave vector defined by

$$\text{q=}\text{4}\text{π}\text{λ}\text{sin}\text{θ}$$

with λ being the X-ray wavelength and θ being half of the scattering angle.

For mass fractals Sinha et al. [11] and Teixeira [12] have shown that:

where I_om is a constant and D_m is the mass fractal dimension.

Therefore, when scattering plots become linear on log–log plots when either surface or mass fractals are present. If the slope of the graph is less than 3.0, then scattering is from a mass fractal. If the slope of the graph is greater than 3.0, then scattering is from a surface fractal.

Ruiz and Hall [8] have argued that the nature of fractal objects is important in determining R(NO). Also, SAXS is sensitive to both open and closed porosity. A useful parameter derived from small angle scattering is the Porod Invariant, defined by

$$\text{PI}\text{=}\text{∫}\text{0}\text{∞}\text{q}{}^2\text{I(q)}\text{d}\text{q}$$

wherein the PI values are related to the total void fraction of the carbon under investigation. During carbonisation many carbons loaded with metals form structures with significant amounts of porosity that is closed to the external surface [13]. Such porosity may open during gasification, and surface areas of the unactivated chars may not relate to those during gasification. Conventional gas adsorption is different to SAXS in that it is only sensitive to open porosity. Therefore, comparison of SAXS and gas adsorption data is often useful in detecting pore opening. Although it is not possible to integrate over the entire q-range, PI values are nevertheless useful for monitoring how porosity changes from sample to sample [8].