Do you BET on routine? The reliability of N₂ physisorption for the quantitative assessment of biochar’s surface area

Highlights

• Conventional surface area calculations of biochars based on N₂ can be unreliable.

• N₂ induced adsorption stress artificially enlarges the surface area of certain biochars.

• The Brunauer-Emmett-Teller method enhances errors from N₂ adsorption measurement.

• Pore specific methods enable a reliable adjustment of the surface area calculation.

• A modified calculation method is proposed to ensure comparability between biochars.

Abstract

A large specific surface area is one of the structural characteristics which makes biochar a promising material for novel applications in agriculture and environmental management. However, the high complexity and heterogeneity of biochar’s physical and chemical structure can render routine surface area measurements unreliable. In this study, N₂ and CO₂ characterization of twelve biochars from three feedstocks with production temperatures ranging from 400 °C to 900 °C were used to evaluate materials with varying structural properties. The results indicate that the frequently reported peak in the surface area of biochars around 650 °C is an artefact of N₂ measurements and not confirmed by CO₂ analysis. Contradicting results indicate an influence of the structural rigidity of biochar on N₂ measurements due to pore deformation in certain biochars. Pore non-specific calculation models like the Brunauer-Emmett-Teller method do not allow for adjustments to these changes. Instead, the use of a pore specific model and the exclusion of pores smaller than 1.47 nm was found to achieve more representative results. The proposed calculation was validated on an external dataset to highlight the applicability of the method. Our results provide novel insights for understanding the structural evolution of biochar related to production temperature.

Introduction

In the face of climate change, biochar has received an increasing interest as a novel class of materials suitable for long term carbon sequestration as recently recognized by the IPCC [1]. Biochar, the solid product of biomass pyrolysis, can be produced from a wide variety of feedstocks and by a range of production techniques [2]. The properties of the material are shaped by both the pyrolysis conditions, e.g. the highest treatment temperature (HTT), and feedstock parameters such as the lignocellulosic composition or ash content [3]. Due to these various possibilities of feedstock and production parameter combinations, biochars constitute a versatile class of materials with multifunctional properties [4]. Research on biochar application comprises, among other things, the use as a soil amendment [5], [6], in gas and water filtration [7], in metallurgy [8], or as electrode material for batteries and supercapacitors [9], [10]. These different uses require biochar to possess application-specific properties to be efficient. The concept of engineered biochar aims at producing biochar with predefined characteristics by identifying fitting production parameters and feedstocks [11], [12]. Therefore, knowledge about the interplay between biomass characteristics, conversion mechanisms and resulting biochar properties is the basis for tailoring biochar to certain applications. Temperature series of biochars produced at varying HTTs are conducive to study the effects of specific parameters and enable general conclusions about the underlying mechanisms [11], [13], [14]. However, biochars produced at different HTTs exhibit large differences in their structural form and chemical properties and characterization techniques need to be unbiased towards these changes to enable reliable comparisons [3], [15]. As biochar research is still an evolving field, characterization techniques from established research areas such as soil science or related materials like activated carbon are often used to characterize biochar [16], [17]. Nonetheless, as a complex class of materials, the diverse properties can present specific challenges, which might render established characterization techniques unreliable when applied to biochar [18], [19].

The structural rearrangement of biomass with increasing HTT is one of the determining factors for predicting the final product properties. While molecular models on the structural reorganization have been proposed [20], [21], [22], less is known about derived characteristics such as the evolution of surface area (SSA), pore volume (V_pore), and pore size distribution (PSD), impeding accurate pre-pyrolysis predictions of these parameters as required for engineering biochar. The importance of the SSA stems from the fact that it determines the number of potential interaction sites with the surrounding matrix. At the same time, the pore size distribution allows conclusions about their physical accessibility, with both carrying significant implications for the final performance of the material [23], [24], [25]. Emerging applications for biochar such as gas and water filtration often require large surface areas and pore volumes within specific pore size ranges to be efficient. While narrow micropores are a prerequisite for efficient CO₂ separation [26], larger micropores are required for the use as an electrode material [27], and a mixed pore size range for the filtration of pharmaceuticals from wastewater [28]. Biochar engineering therefore depends on an accurate characterisation of these parameters to get an understanding of pore development mechanisms, which can then be utilised for the design of biochars with pre-defined properties.

Conventional methods for assessing the SSA, V_pore, and PSD are based on low-temperature gas physisorption. Although general trends have been observed, such as a rise and fall of SSA with increasing HTT, these are empirically drawn from a limited set of measurement methods and often lack detailed theoretical validation or cross-examination with other techniques [12], [29]. Especially the evolution of SSA and PSD with HTT and feedstock composition remains inconclusive [12]. While N₂ (at 77 K) is the most common probe molecule for the determination of the SSA, alternative adsorbates such as CO₂ (at 273 K), Ar (at 77 or 87 K) or Kr (at 77 K) might also be used, with specific advantages and disadvantages for each method [30], [31], [32], [33]. However, to date, these are seldomly applied for the characterization of biochar. Irrespectively of the type of probe molecule, the calculation of derived parameters such as the SSA is based on an experimental isotherm. Prior to the isotherm measurement, sample preparation requires a degassing step (e.g., under vacuum or He gas atmosphere) and elevated temperature to remove adsorbed matter. As shown by Sigmund et al. [18], biochar already presents significant complications at the degassing stage as structural changes might occur with the application of standard degassing protocols derived from related materials, i.e. activated carbon carbon black. The measurement of the experimental isotherm itself proves similarly challenging, especially with the commonly used N₂. The complex pore structure and low measurement temperature (77 K) can lead to prolonged equilibrium times of up to several days due to diffusion limitations [34]. Correct parameter settings such as the choice of an appropriate equilibrium time are therefore especially important as non-equilibrium conditions can lead to varying results [32], [33], [34], [35]. Another specific characteristic of N₂ isotherms of biochar is open hysteresis, which is incompatible with standard IUPAC classifications or calculation models [30]. Open hysteresis is often simplified as a result of non-equilibrium conditions and mostly ignored in biochar characterization even though its occurrence might be related to structural characteristics of biochars. However, in studies on carbon capture and storage, structural deformation receives greater interest as it can be associated with the adsorbent’s performance [36], [37], [38]. Pore deformation induced by low-pressure N₂ adsorption is the result of elevated localized pressure during the filling of narrow micropores. This adsorption induced stress causes the expansion or contraction of pores if the adsorbent structure is non-rigid. So far, adsorption induced pore swelling has been observed for fossil coal, activated and synthetic carbons, and metal–organic frameworks [39], [40], [41], [42]. However, as observed by Braida et al. [43], benzene sorption can provoke swelling in biochar, demonstrating a similar non-rigidity of biochar’s carbon structure. Therefore, it is plausible that N₂ adsorption induced swelling might occur as well.

Once the experimental isotherm is determined, a variety of models exist to calculate derived quantitative parameters such as SSA, V_pore or PSD. These models can be further divided into pore-specific and pore non-specific methods, which differ in their ability to differentiate between pore sizes and are well described in the literature [32], [33]. The almost universally used pore non-specific method to calculate the surface area from N₂ isotherms is based on the theory proposed by Brunauer-Emmett-Teller (BET), which is also recommended by the European Biochar Certificate (EBC) [44]. Although widely used, the BET method was initially developed for nonporous materials with a uniform surface and is conceptually weak for heterogeneous materials, i.e. biochar [45]. One of the few parameters relevant for the correct application of the BET theory is the applied p/p₀ range, with the original method suggesting a standard p/p₀ range between 0.05 and 0.35, typically reduced to 0.1 and 0.3 in biochar research [32], [33], [46]. Although this range proves to be appropriate for the original purpose, it is only valid in the absence of microporosity or strongly adsorbing sites. In the presence of micropores, the valid p/p₀ range should be identified by applying additional selection criteria as specified by Rouquerol et al. [47]. To determine if the p/p₀ selection is appropriate, the calculated C parameter, which describes the energy of monolayer adsorption, can then be utilized [35]. As the C parameter has to be of positive value, a negative C parameter explicitly indicates a wrong p/p₀ range selection and in consequence an invalid result. Despite semi-automated BET-assistants in modern instruments which suggest a valid p/p₀ range, the often inappropriate use of the standard range can still be found in the published literature [48]. Because of these shortcomings, in practical terms as well as in the theoretical assumptions of the BET theory, the comparability of SSA_(BET) values for biochar is highly variable. Pore specific techniques based on the Density Functional Theory (DFT) or Grand Canonical Monte Carlo simulations (GCMC) are advanced computational methods overcome the main obstacles of pore non-specific models by simulating the adsorption of fluids on a molecular level [49]. For biochars displaying semi-continuous pore size distributions in the micro and mesopore size range, the DFT approach avoids the erroneous averaging approach shown by pore non-specific methods like BET [32], [33], [35]. Especially promising for the analysis of biochars are recently developed kernels for N₂ adsorption on carbons with complex surface structure such as the Heterogeneous Surface Non-Local DFT (HS-NLDFT, Micromeritics) or Quenched Solid DFT (QSDFT, Quantachrome), which despite differences in the calculation procedures, show generally comparable results [50].

Similar to classical approaches such as the BET method, DFT models will be significantly affected by errors in the isotherm measurement. However, no detailed studies on the effects of physisorption parameters on N₂ isotherms of biochar exist today. Due to a lack of standardization and incomplete reporting of experimental parameters, the influence of measurement conditions on the results is difficult to identify as shown by Bachmann et al. [51]. One way to cross-validate the assessment of N₂ isotherms on microporous carbonaceous materials is the use of CO₂ as a complementary analysis in the overlapping pore size range between 0.7 nm and 1.47 nm [52]. The upper range of pore size determination of approximately 1.47 nm for standard instruments limits the sole use of CO₂, but the overlapping range between 0.7 nm and 1.47 nm can be used to assess the reliability of the N₂ isotherm. Due to the slow diffusion of N₂ at 77 K, especially in narrow micropores of biochar and related inconsistencies, Jaggiello et al. [61] proposed a lower p/p₀ limit for N₂ measurements of 10^-3 to avoid non-equilibrated measurements. The analysis is complemented with CO₂ measurements of the narrow micropore range. While the combination of N₂ and CO₂ has several advantages and must be seen as a benchmark method for biochar characterization, it is seldomly used in literature to date.

Surface area measurements should ultimately provide a reliable assessment of biochar properties and ensure comparability between samples. In this study, we analyzed the temperature series of biochar produced from three different feedstocks by pyrolysis at 400 °C, 500 °C, 700 °C, and 900 °C. Physisorption measurements using N₂ and CO₂ as probe molecules allow a comprehensive analysis of the micro and mesopore range. We calculated the surface area of the samples from N₂ isotherms by the traditional BET method as well as more advanced DFT models to allow a comparison of pore specific and pore non-specific methods. Based on the results of pore size distributions calculated from N₂ and CO₂ measurements in the overlapping pore size range from approximately 0.7 nm to 1.47 nm, we propose an adjusted DFT calculation as a more reliable method for the assessment of biochars. To ensure the applicability of our analysis and to avoid any bias introduced within our experiments, we obtained external data from published literature to test the general applicability of our proposed method.