Acid pre-analysis treatment: methods and effects

Researchers interested in the organic carbon fraction of bulk samples typically must remove carbonate prior to analysis. This post focuses on important considerations and limitations associated with carbonate removal.

Elemental analysis (EA) of bulk material will convert almost any carbon into carbon dioxide (CO₂) whether it is from an organic or inorganic (e.g. carbonate) source. The presence of carbonate will inflate the total amount of carbon as well as severely bias the stable carbon isotope value (¹³C). Terrestrial carbonates typically have more ¹³C present relative to ¹²C, making the carbon isotope ratio, expressed as δ¹³C, more positive or "heavier". Organic carbon typically has a reduced amount of ¹³C relative to ¹²C, i.e. a more negative δ¹³C. In any case, the organic carbon and carbonate carbon will most likely have different isotopic compositions and they must be separated or analyzed in such a way so that the presence of one does not affect the measurement of the other.

Researchers typically use an acid to remove the carbonate. The species and strength of acid should be considered. While the dominant reaction is the conversion of carbonate to CO₂ and H₂O, organic carbon species can be impacted to a measurable extent by, for example, washing away soluble organic compounds. Also, typical EA of carbon has historically included the analysis of nitrogen. All nitrogen present in a bulk sample will be converted to N₂ during EA for total and/or isotope abundance. The soluble nitrogen species present in a bulk sample can also be washed away during acid pretreatment and / or can react with the the acid leading to measurable effects. Many researchers have described how acid pretreatment, also referred to as acidification, affects their samples.

The purpose of this post is to survey studies that directly compared methods of carbonate removal in an attempt to document the effect of the method itself. Many many publications exist on different approaches for removal of, or separation of, carbonate carbon from organic carbon. These studies are typically geared towards the immediate sample type of interest. Your samples will very likely have a different composition and this is an important consideration when choosing a method. This is not meant to be an exhaustive review, but rather a taste of how complicated this issue is.

Broad Methodology

Selective roasting - This approach aims to heat samples to a lower temperature (e.g. 500 °C) to drive off organics and then at a higher temperature (e.g. 1000 °C) to quantify carbonate. The difference between the higher temperature treatment and the lower temperature treatment provides an estimate of total organic carbon (TOC). The overlapping loss of some carbonate as well as incomplete removal of organic carbon at the lower temperature make this method problematic (Froelich 1980; Weliky et al. 1983). This method is not recommended and is included here for completeness.

Acid Washing - This method involves a large amount of sample in a reaction vessel or test tube typically with dilute acid, sonication, soaking or otherwise incubating while effervescence continues, followed by rinsing with purified water. Take care to ensure you know the particular carbonate species that is present in your samples (e.g., siderite Vinduskova et al. 2019). This method does well at retaining the insoluble, non-volatile portion of the sample while removing the carbonate portion. However it also guarantees the removal of any soluble organic carbon and nitrogen (CN) or volatilized CN compounds. Researchers have reported using hydrochloric acid (HCl), phosphoric acid (H₃PO₄), and sulfurous acid (H₂SO₃) (reviewed and tested in Brodie et al. 2011a). Both Froelich (1980) and Weliky et al. (1983) demonstrate the presence of dissolved organic carbon in solution after sample reaction with H₃PO₄. Loss of organic carbon in the soluble fraction ranged from 5 to 45% (Froelich 1980). Despite potential losses in the liquid phase, Midwood and Boulton (1998) tested four HCl concentrations and incubation durations on two distinct soils and show that δ¹³C was largely unaffected; TOC and total N, however were still affected to varying degrees. Mazumder et al. (2010) also show disparate effects of acid washing on CN, δ¹³C and δ¹⁵N depending on the sample type.

In-capsule aqueous acidification - In an effort to harness the decarbonating potential of acid solution but trying to limit losses due to repeated washing and rinsing, an in-capsule aqueous acidification reaction is allowed to proceed in a silver sample capsule (tin capsules degrade in the presence of acid; so do silver capsules if the experiment runs too long). The sample material inside the capsules is never rinsed and the liquid phase evaporates away after effervescence is finished. Nieuwenhuize et al. (1997) show in-capsule acidification using 25 % HCl exhibits no detectable loss of organic carbon when compared with an alternative method. However, Lohse et al. (2000) demonstrate loss of nitrogen using in-capsule aqueous acidification when compared with no acid treatment. More recently, Kazanidis et al. (2019) demonstrate biases in conclusions when acid pretreated samples vs untreated samples are considered.

In-capsule acid fumigation - Another attempt at driving away carbonate using acidification is via the vapor phase of HCl. Here samples are loaded into silver capsules and placed in a sealed container with a reservoir of 12N HCl. Gaseous HCl reacts with carbonate to form CO₂ and H₂O but does not have an aqueous phase to dissolve any organic C or N. Hedges and Stern (1984) compare this method to a method similar to the acid washing approach without the decanting. That is, they add 1 N HCl and sample material to a reaction vessel but then allow all of the liquid to evaporate at 50 °C. This method, too, avoids loss of material in the liquid phase. In this way, their acid washing method is more like a large-scale in-capsule aqueous acidification. Komada et al. (2008) recommend in-capsule acid fumigation after observing better accuracy and precision as well as reduced labor compared with in-capsule aqueous acidification. They also caution against long fumigations and recommend 6 hours. However, if your samples contain siderite (iron carbonate) or an otherwise difficult to digest carbonate species, you may have to use in-capsule aqueous acidification with stronger concentrations and correct for any losses or effects of said treatment (Larson et al. 2008). Lorrain et al. (2003) successfully used fumigation to decarbonate particle filters while leaving the TOC, total N, δ¹³C and δ¹⁵N unaltered.

Nitrogen

It is important to recognize that while you are interested in removing carbonate to hedge towards better organic carbon data, acid pre-treatment also affects your nitrogen data (Lohse et al, 2000; Harris et al. 2001; Walthert et al. 2010, Brodie et al, 2011b and c). Mateo et al. (2008) is an impressive compilation of published analyses showing that broad effects of acid washing affect δ¹⁵N of marine invertebrates. Acid washing could be removing nitrogen in the same way as organic carbon but it could also be leaving behind nitrogen present in the acid (Wolman and Miller, 1971). Assuming you can achieve complete combustion during your analyses, the best approach is to obtain your nitrogen data with untreated sample material (Kazanidis et al. 2019).

Recommendations

With hundreds to thousands of papers reporting acid-pretreatment of samples and a fraction of those attempting to document the effect of their treatment on the measurement of interest, you may still end up justified in thinking you have a unique sample composition and thus no real direction on how to proceed. Brodie et al. have a collection of three papers from 2011 that provide a thorough review of the above three acidification methods and the affect on CN amounts and δ¹³C and δ¹⁵N. If you are going to read only a paper or two, start with the Brodie series. In the end, the only recommendation is to accept a compromise. Know your samples and know that treating them with acid will affect your results in more ways that removing the carbonate. The reason your percent carbonate results differ when you compare acid washing to methods using CO₂ yield (e.g., Kiel carbonate device), for example, become apparent when you consider forty years of decarbonating.

References