"45 thousand-year-old fossil wood encased in 45 million-year-old basalt": Conflict Revisited

One point lost by Bill Nye in the recent debate with Ken Ham was the repeated assertion that YEC researchers had dated fossil wood ~1,000 times younger than the basalt in which it was encased. If the assertion holds, then radiometric dating methods to which Bill Nye appeals as evidence for an old Earth are potentially flawed. Unfortunately, Bill Nye did not seem familiar with the claim, reported originally in 1993, despite my prophetic counsel. ;)

Where did these samples come from, and why were they sent to labs for radiometric dating?

In 1993, several samples of charred and petrified wood were collected during construction of a mining ventilation shaft in central Queensland. The wood samples were seemingly well preserved inside the lava flow that encased them. Since the layers of basalt overlay early Cenozoic sediments, the expected age of the lava flow might be a few tens of millions of years. Not convinced by conventional ages of this geologic column, however, local 'young-Earth' geologists paid to have the basalt dated by the K-Ar method and the wood fragments dated by the radiocarbon method. Their reasoning was simple and seemingly innocent: if indeed the Earth is old, then fossilized wood from an ancient lava flow should contain no radioactive carbon (14C) today and will not yield a finite age; if the Earth is young as we claim, however, then detectable 14C should be present in the fossilized wood and the K-Ar age of the basalt cannot be taken seriously.

A full report of this study by Dr. Andrew Snelling of Answers in Genesis is available for download from the Institute for Creation Research [Snelling, 2000. Creation Ex Nihilo Technical Journal, Vol. 14, No. 2, p. 99-122]. The 24-page publication is rather verbose, reading more like a drawn-out lab report than a scientific study, but it contains all the details necessary to assess the feasibility of Dr. Snelling's extraordinary claims.

Wait, if the fossil wood and basalt are both older than 6,000 years, doesn't Ken Ham undermine his own position by citing this study?

I've described at length how apologists at Answers in Genesis use the radiocarbon and Potassium-Argon (K-Ar) methods to argue for a young Earth. In short, they criticize the assumptions behind each dating method. For example, 'model' K-Ar ages assume that all argon was expelled from the sample before molten rock cooled, starting the radiometric clock. If any argon was trapped in the mineral structure, however, then the rock will appear far older than it actually is. We have known for half a century that excess argon may be trapped in lava flows (especially underwater flows under high pressure), so much research in geochronology is devoted to improving the models. Today, more sensitive instrumentation and the aid of computer models allow geochronologists to identify 'excess' argon and calculate ages with far more precision than even a decade ago.

Given the demonstrated robustness of the K-Ar and Ar-Ar methods with these improvements, YEC's have more recently attacked the assumption that radioactive decay occurred at a constant and known rate throughout history. Many of them suggest that during/after the Flood, radioactive decay rates increased substantially, giving the false impression that most rocks are millions or billions of years old. Despite their attempts to confirm this through scientific investigation, however, the claim remains ad hoc and unsubstantiated. On the other hand, continued success using the K-Ar system in pinning down geological events, as well as confirmation from independent dating methods, gives us ample reason to believe that the model assumptions are valid and verified.

The radiocarbon method is far more complicated than is commonly presented, so YEC's have been successful in twisting its results to support their own paradigm. Traditionally, radiocarbon dating assumes that the relative concentration of 14C (the rare radioactive isotope, compared to the stable forms 12C and 13C) does not change with time. When this assumption is used, the results are reported as "radiocarbon years before present". It is important to distinguish between "radiocarbon years" and "calendar years", because we know that the relative concentration of 14C does indeed change over time. The production rate of radioactive carbon in the atmosphere depends on the magnetic field strength, for example, of the Earth and our sun. The YEC can argue, therefore, that prior to the flood, the relative abundance of 14C was much lower than today. If this assertion holds, then plant and animal remains from before the flood should yield dates many times older than their 'actual' age (4,500–6,000 years).

To document the relative abundance of atmospheric 14C over time, geologists use records that can be dated independently of the radiocarbon method. For example, tree-rings and varved lake sediments preserve annual cycles, which may be counted to obtain the actual age of each layer. When tree-rings and varved sediments are dated by the radiocarbon method, the 'radiocarbon age' of each annual band is compared to the real age to construct a calibration curve. This curve removes the assumption in conventional radiocarbon dating that 14C was constant over time.

Stalagmites provide another tool for gaging the history of atmospheric 14C, since these banded records are datable by the U-Th disequilibrium method or, in ideal cases, counting annual bands. By radiocarbon dating calcite along the growth axis of the stalagmite and comparing these ages to U-Th dates, geologists can test and refine the calibration curve (e.g. Vogel and Kronfeld, 1997). These highly corroborated, independent lines of evidence thoroughly disprove the YEC claim that atmospheric 14C increases significantly only after a recent, global flood.

In summary, yes: Ken Ham undermines his own position by citing this study, though he does not mean to.

Why did the fossilized wood give an age of 45,000 years, and how is this possible if the lava flow solidified millions of years ago?

Dr. Snelling submitted two samples of wood to two independent laboratories for radiocarbon dating by Accelerator Mass Spectrometry (AMS). He argues that since the lab obtained 'finite ages' from organic-derived carbon, the fossilized wood cannot possible be tens of millions of years old. He reports the results as follows:

   Sample                             Lab              Age (Radiocarbon Years Before 1950)
"Wood in drill core"             Geochron          >35,620 ± ?
                                         ANSTO               44,700 ± 950
"Other wood"                    Geochron           29,544 ± 759
                                         ANSTO               37,800 ± 3,450

You should notice three features immediately from this table that warrant suspicion of Dr. Snelling's interpretation:

First, all of these ages are close to the practical limit of radiocarbon dating. Even today, radiocarbon ages exceeding ~40,000–50,000 years are commonly considered suspect, because so little radioactive carbon (14C) is being measured that it becomes nearly impossible to distinguish it from contamination or background interference (more on this later). In 1993, this caution was even more applicable, which is why many important geological studies have been updated in recent years from fresh samples analyzed with better equipment.

Second, neither set of dates overlaps within uncertainty, as the sample age differs substantially from one lab to the other. Ages for the "other wood" samples differ by more than 8,000 years, meaning that Geochron measured more than twice the concentration of 14C in the same sample, compared to ANSTO's analysis. If the 14C in these wood samples were derived from a pre-Flood atmosphere, as Ken Ham and Dr. Snelling claim, then all samples should yield the same age. At the very least, each laboratory should be reporting the same concentration of 14C in the duplicate samples. But they don't. From this fact alone, we should be very suspicious that the measured 14C derived from the wood itself. More likely, we might conclude that the 14C concentration differed because variable amounts of atmospheric contamination were incorporated during each lab's preparation methods. These radiocarbon dates do not reflect the age of the fossilized wood, therefore, and do not substantiate Dr. Snelling's claims.

Third, two of the analyses (first and last samples) yielded large uncertainties, which can result when too little 14C is present, or when the source of 14C is not consistent. In the latter case, the source of 14C might be a mixture of contamination sources: atmospheric CO2 or organic acids still bonded to the fossil wood, residual atmosphere inside the sample chamber, or the tiniest of leaks in the vacuum lines. As one who deals daily with mass spectrometers, I would suggest that all three are likely candidates.

Why?

Mass spectrometers do not count atoms of 14C directly, but compare electric intensities produced by ionized particles hitting Faraday cups at the end of a vacuum tube. In theory, particles of the same mass should all follow the same path along the magnetically charged vacuum tube, so the Faraday cups can be positioned to catch each isotope. In reality, particles of the same mass do not hit the same spot consistently, but produce more of a Gaussian distribution. Imagine a fire hose spraying onto a wall: most of the water hits in the center of the stream, but some veers off to the side. If the mass spectrometer measures only 14C (and not 13C bonded to hydrogen, for example), which is distributed evenly through the sample, then the center of the peak is easily found by the instrument. Wider peaks result from low or inconsistent signals, and help us understand why such large uncertainties were associated with two of Dr. Snelling's samples.

Dr. Snelling provides several additional clues that would lead any other investigator to find better samples for dating. First, the basalt flow encasing the wood was only ~21–25 meters below the surface, meaning that it was long exposed to surface waters percolating downward into the rock. These surfaces waters contain not only modern atmosphere, but organic acids that bond tightly to the wood. The wood fragments themselves show evidence of being altered by intruding waters, as Dr. Snelling notes (p. 8): "Permineralization was too advanced" to identify taxonomically important features under the electron microscope. The porous and jointed (fractured) basalt was also altered (p. 14):

"The basalt in the drill core does, in fact, come from the zone of weathering...where percolating oxidizing ground water readily alters minerals and rock chemistry by dissolving and removing various elements."

Yet this basalt was shipped to a lab for K-Ar dating, after which Dr. Snelling criticized the inconsistent and apparently old results. But that is another issue...

The fact that Dr. Snelling's wood samples were long exposed to modern atmospheric and plant material means that at least some of the measured 14C derives from contamination. Despite Dr. Snelling's rant that the laboratories made thorough efforts to remove contaminants and "staunchly defended [the ages] as valid", he admits himself that it is impossible to exclude all contamination. Acid washes cannot contact every single surface of the sample, perfect vacuums are not obtainable in nature, and electrical interference is a constant reality. He may continue to suggest that contamination could not have been large enough to explain finite ages of 30–40,000 years, but he cannot prove this. On the other hand, we can disprove his claim through his failure to replicate the results of his analyses from independent labs.

Dr. Snelling's article mentions that "a δ13C (V-PDB) value of −25.69‰ [is] consistent with terrestrial plant organic carbon...ruling out contamination." Is he right?

Of all the claims made by Dr. Snelling's article, this one is the most obviously and demonstrably false. Forgive my roundabout answer, but this point is important.

In AMS radiocarbon analyses, the relative abundance of 13C (reported as a δ13C value) is routinely measured, because model radiocarbon ages assume that the δ13C value is exactly -25‰. This value represents a reasonable average for the isotopic composition of plant material utilizing C3 photosynthesis, like an oak tree. However, if you were to radiocarbon date a piece of ancient corn (a C4 plant, whose average δ13C value is much higher, around -13‰), you would have to consider that the corn begins with slightly more radioactive carbon (14C) than something like an oak tree of the same age. The difference results from the fact that during photosynthesis, plants preferentially incorporate the lighter isotope(s) of carbon, but the preference is stronger among C3 plants than C4 plants. Since 14C is much heavier than 13C or 12C, it is not incorporated at the same rate as the other isotopes, and this rate varies among photosynthesizers.

If you mix carbon sources with differing isotopic compositions, the δ13C changes proportionately. For example, let's mix 5 grams of wood with a δ13C value of -25‰ and 5 grams of calcite with a δ13C value of 0‰. The resulting δ13C of the mixture is -12.5‰, or a weighted average of the two sources. So what if the 14C in Dr. Snelling's fossil wood samples derived entirely from contamination? Wouldn't the contaminant shift the δ13C value, as he claims?

No. The concentration of 14C in the modern atmosphere is ~10^-12%, or less than one part per trillion. To contaminate a 14C-free sample with enough modern organic carbon to yield a radiocarbon age of ~45,000 years, less than 0.4% of the sample's mass needs to be derived from modern material. If that material is bacteria (δ13C = -27‰ to -33‰) or humic acid (δ13C = -27‰), the δ13C value of the sample will not shift noticeably, because it is so close to the composition of the fossilized wood. But even if the material were atmosphere (δ13C = -7‰) or mineral carbon (δ13C = 0‰, give or take), the δ13C value of the sample would not shift to an extent detectable by the mass spectrometer. Here is the math:

(0.996 * -25.0‰) + (0.004 * 0.0‰) = -24.9‰

Keeping in mind that the uncertainty is close to 0.1‰, and the actual δ13C of trees ranges by 5–6‰...

Conclusion

Ken Ham's appeal to young fossil wood within old basalt may have caught Bill Nye off guard, but his claim remains unsubstantiated. The actual radiocarbon ages of this fossil wood were not reproducible by independent labs within analytical uncertainty, suggesting that contamination and/or background interference was responsible for much of the detected radiocarbon. Recent advances in AMS radiocarbon dating have focused on how to account for the fact that contamination is always introduced during sample preparation and how to correct for various kinds of background interference. Regardless, radiocarbon ages close to the practical limit of the method are always treated with some suspicion.

Radiocarbon dating of independently datable materials (lakes, tree rings, and stalagmites) provide highly corroborated calibration curves of radiocarbon activity over the past ~50,000 years. These studies alone disprove the notion that a recent, global Flood severely impacted concentrations of atmospheric 14C. Combining these records with the inherent uncertainties of the radiocarbon method, which were less resolved in 1993, we might conclude that Ken Ham's proposed dating conflicts are misleading at best.