Avoiding Interference from Consolidants in Studying Archaeological Bone

A recent study published in the Journal of Archaeological Science positions near-infrared (NIR) spectroscopy as an effective prescreening method for studying collagen in archaeological bone without destroying the samples.^1,2

In this brief Q&A, Christina Ryder, who is a postdoctoral researcher at Texas A&M University and the lead author of this study, explains why the partial least squares regression (PLSR) and random forest (RF) models were restricted to the 2030–2060 nm spectral region and why NIR prescreening improves on percent nitrogen analysis.

The models were restricted to the 2030–2060 nm spectral region to avoid interference from consolidants. How did you identify this optimal window, and how sensitive is the approach to different conservation treatments commonly found in museum collections?

The 2030–2060 nanometer spectral region corresponds to one of the strongest NIR absorption features associated with collagen, and this is where we see combination bands involving NH structure and vibrations of peptide bonds in the protein backbone. It also avoids major water absorption bands that can obscure the collagen signal. So that's why this is a particularly strong region for collagen related spectral features in archeological bone.

Initially, in our 2019 publication, we presented a proof of concept with PLSR models that incorporated five separate wavebands. However, when I actually began to apply these models to archeological and paleontological collections at museums all over the world, I frequently absorbed atypical absorption features in the spectra, so that many of these signals were associated with consolidants and adhesives applied during conservation. No work has been published on yarn thread to detect consolidate conservation material, which includes consolidants and adhesives.

But there are some studies exploring Fourier transform infrared (FT-IR) as a technique. So, is it shocking that we are picking up on it in the NIR? Because these materials are so common in these legacy museum collections, and they're often poorly documented, they introduce spectral features that can interfere with collagen prediction models. All the consolidant material that I have captured thus far does not overlap with this 2030–2060 nm region, but there are some environmental contaminants that are problematic, one being fat guano, which is a nitrogen-rich material, and another complicating consolidant is animal-derived glues. Specifically, glues derived from type one collagen are problematic because I'm detecting collagen. When an animal glue is applied to a material, I can detect if the animal glue is there or not, but it really disrupts the collagen prediction because it is collagen. What is nice with this is this when I stick to the 2030– 2060 region, I can still predict the collagen yield despite the presence of presence of the consolidants.

So today, what's being used most frequently are two types of consolidants, Paraloid B 72 and Butvar. Mastering the identification of these consolidants is critical because we need to know what the consolidant is to remove the consolidant for downstream archaeometry analysis. It's also critical to identify these consultants because once the consultant is applied, it can take certain analyses off the table. For example, a very common consultant applied in archeology in the 1950s to 1960s was an industrial insulator that they used to consolidate bones, and this is a thermosetting polymer.

So, it's critical to identify what samples have contamination because then we know that maybe we cannot do radiocarbon dating on these samples, or maybe we can only do a certain type of radiocarbon dating, or we have to go through a certain collagen extraction process before we can get a reliable radiocarbon date. The manuscript I'm working on right now is a consolidant identification piece, so it goes through all of the common consolidants, past to present, and publishing models to detect these consultants, which will be really useful in my field for people who are working with samples from the past 50,000 years and in museum collections. So, I'm looking forward to finally publishing these models I've been using for quite a few years now.

Your results indicate that NIR prescreening can outperform traditional %N-based methods for selecting samples for radiocarbon dating. What practical improvements could this bring to sampling strategies and success rates in archaeological dating projects?

First, let’s start with a little background on percent nitrogen as a prescreening tool. What's done is about 5 mgs of whole bone is ground and then thrown into a nitrogen analyzer. One of the main limitations of this is that it's inherently destructive, and although the amount of material required is small, it still requires sampling, destructive sampling, and then transporting that material into a laboratory facility for analysis. For many archeological and paleontological collections, transporting both outside of the institution or outside of the excavation site requires permits and curatorial approval, so destructive sampling often occurs before researchers have a clear indication of whether the collagen is preserved because it's easier to shift a small amount of bone or transport a small amount of bone, rather than a giant Mastodon femur.

NIR prescreening changes the workflow. Because NIR measurements are nondestructive and portable, screening can be performed directly at the excavation site. My device actually can run off battery power. I bring it to sites regularly. It's very useful in museum collections because the samples don't even need to leave the drawer that they are in. I oftentimes put my NIR instrument on a movable table, and I'm just go around with the fiberoptic probe if they're thick enough to remain in the drawer. For thinner samples, I do take them out and analyze them on the spectrum because of that penetration depth.

After you select your samples, you initiate the permit process or the destructive sampling request. It's great for curators because they have some assurance that the destructive sampling of their collections is going to yield positive results for researchers, so it saves us an immense amount of time. It also saves us transport, and it's just easier logistically when you're working at an investigation site. Depending on the country you're working in, the permitting process can be very rigorous, and it can take months and even years to get your samples out of country for analysis. So by prescreening, you're ensuring that you're not irresponsibly destroying a finite resource and that the samples you do put all of this effort into exporting are going to yield results. Percent nitrogen analysis is certainly a useful tool, but it doesn't circumvent these permitting issues.

Currently, we have a paper under review right now that is the first large scale application study. In this large-scale application study, we applied NIR prescreening to over 800 archeological and paleontological specimens, and we found a success rate of approximately 93% in select examples that were suitable for radiocarbon dating and paleo proteomic analysis as well. So, the results demonstrate how nondestructive spectroscopic prescreening can significantly improve sample selection before delivery or transport to the destructive laboratory or this destructive analysis.

References

1. Ryder, C.; Celis, G.; Devièse, T. et al. Refining Near-infrared Spectroscopy for Collagen Quantification: A New Predictive Model for Archaeological Bone. J. Arch. Sci. 2026, 185, 106448. DOI: 10.1016/j.jas.2025.106448
2. Wetzel, W.; Spectroscopy Staff. Collagen Preservation in Archaeological Bone Using NIR Spectroscopy. Spectroscopy. Available at: https://www.spectroscopyonline.com/view/collagen-preservation-in-archaeological-bone-using-nir-spectroscopy (accessed 2026-03-31).