The Literature on Helium in Fenton Hill Zircons

Every source the main article rests on, with what it actually measured and which side it helps.  Sources are grouped by role, not by author, because the same author often appears on both sides of a question.

Everything below was fetched and read during the research for this article except where marked not obtained.  Those exceptions are listed together at the end, because an argument that depends on a source nobody on this side has read should be flagged, not buried.

The measurements everything rests on

Gentry, Glish & McBay 1982, "Differential helium retention in zircons: implications for nuclear waste containment," Geophysical Research Letters 9(10):1129-1130.  http://www.halos.com/reports/grl-1982-helium-in-zircons.pdf

The source of every retention percentage in the dispute.  Individual zircons from the Fenton Hill core, heated to 1000 °C in a mass spectrometer at Oak Ridge.  Retentions of 0.58, 0.27, 0.17, 0.012 and about 0.001 at 105, 151, 197, 239 and 277 °C.

Three things about it are routinely misdescribed.  It draws no young-earth conclusion — it is a paper about whether granite is a safe place to store nuclear waste, and its conclusion is that it is.  Its Q₀ is derived from lead, not from an assumed age, so the "1.5 billion years" does not enter the denominator.  And Gentry himself flagged the thermal-history problem that would later become the whole dispute: "we do not know precisely how long the zircons have been exposed to the present temperatures."

Which way it cuts: neutral.  It is the data.  Its weakness is that Q₀ was computed once, by a method the paper describes only in prose, and the working no longer exists.

Gentry et al. 1982, "Differential lead retention in zircons," Science 216:296-298.  http://www.halos.com/reports/science-1982-lead-in-zircons.pdf

The companion paper, and much more useful than its citation count suggests.  It contains ion-microprobe uranium and thorium on individual grains at several depths, which is what makes a single Q₀ untenable — the 960 m zircons run to roughly 5,300 ppm uranium, more than twenty-fold variation within one sample.  It also states plainly that the samples came from "drill holes GT-2 and EE-2," which RATE's account does not.

Which way it cuts: against RATE, decisively, and it is Gentry's own data.

Zartman 1979, LA-7923-MS, Los Alamos.  https://www.osti.gov/servlets/purl/6111996

Uranium 328.78 ppm, thorium 169.42 ppm, on zircons from 2,903.8 m — the only direct uranium-thorium measurement on Fenton Hill zircon anyone uses.  Also reports the apatite fission-track ages "effectively reset to 0 m.y." below 1,800 m, and a discordant zircon (lead-206/uranium-238 age 1,159 Ma against lead-207/lead-206 1,465 Ma) showing substantial lead loss.

Which way it cuts: against RATE twice.  Its uranium and thorium give Q₀ = 73.7 ncc/µg, five times RATE's figure.  Its reset apatite says the deep rock is hot now.

Farley's diffusion measurements, reported in RATE's own appendices.  Never independently published.

Kenneth Farley at Caltech, contracted through Zodiac Minerals and Manufacturing so that he did not know the client, the goal, or the sample localities.  Biotite in October 2000 and March 2001; zircon in May 2002 and July 2003.

Which way it cuts: nobody disputes these.  They are the strongest part of the project.  Their weakness is that they were never published in the open literature, and that the lowest temperature reached is 175 °C.

RATE's own publications

The most important document is RATE Vol. II Chapter 2, and it is more candid than the movement's summaries of it.  It contains the admission that the first model "wrongly assumed that the zircons were a negligible impediment," the admission that the low-temperature model points rest on extrapolation, the admission that two earlier runs failed, and the sentence describing how the defect line was fitted to three points derived from the assumed age.  Every serious criticism in the main article can be sourced to Humphreys' own text.

The 2011 argon paper deserves separate mention.  It is the strongest supporting argument RATE has that does not depend on Gentry: a different mineral (microcline feldspar), a different isotope (argon-40), and somebody else's data (Harrison, Morgan & Blackwell 1986), reanalysed to give 5,100 (+3,800/−2,100) years.  It is also a reanalysis of a sample the original authors down-weighted, published only in a creationist journal, and — on the reading the main article reaches — it is measuring the same recent heat pulse the helium is measuring.  Humphreys notes that his own lower bound of about 3,000 years coincides with Harrison et al.'s own figure, which is a point in his favour and also a hint about what is really being dated.

The critics

Gary H. Loechelt 2008, "Fenton Hill Revisited," 37 pp.  https://www.asa3.org/ASA/education/origins/helium-gl3.pdf  With a 2009 response at https://www.asa3.org/ASA/education/origins/helium-gl4.htm

The most substantial technical work by anyone on either side.  Loechelt is a semiconductor device engineer with published diffusion modelling in the Journal of Applied Physics, so the competence is documented.  He builds a full forward model — eigenfunction method of Meesters and Dunai, 10,000 eigenfunctions, 826 graded time steps from 1,440 Ma, alpha ejection handled physically rather than in the denominator — and runs four scenarios.

His five charges, in his own ranking: Q₀ far too small; effective radius should be 20 µm not 30; the zircon/biotite boundary condition is applied to one model and not the other; the thermal history figures were read backwards; and single-domain kinetics are inappropriate.  He calls the first three "somewhat inconsequential" and the last two sufficient to invalidate the conclusion.

Which way it cuts: the strongest critique, and the most honest about its own limits.  Its weakness is parameter count.  He fits a multi-domain kinetic spectrum plus a hand-assembled 1.44-billion-year temperature path to five retention numbers, and he applies a single Q₀ across all depths after writing that doing so "is not very good."  His fast-domain gas fraction is pinned only to within two orders of magnitude by his own statement.  What his paper establishes is an existence claim, not a unique fit, and that is enough.

Kevin R. Henke, current version June 2010.  https://www.talkorigins.org/faqs/helium/zircons.html

About 53,000 words, four revisions since 2005, self-published.  A geochemist, and a former young-earth creationist by his own account.

His real findings: the lithology error (the shallow samples are gneiss, not granodiorite); the independent Q₀ reconstruction giving 41; the extraneous-helium hypothesis with measured concentrations from nearby geothermal wells; the standard-deviation objection; the charge that RATE silently converted Magomedov's 1970 Arrhenius plot from natural to base-10 logarithms; and — the cleanest of all — that modern lead-diffusion measurements predict about 1 % lead loss where RATE claimed 50 %.

Which way it cuts: substantially against RATE, but the essay must be used with care.  It misuses Dunai and Roselieb, a paper that never varied pressure, as its central pressure citation.  Its "nine orders of magnitude" pressure framing is physically empty.  Its headline "90,000 ± 500,000 years" is a reductio dressed as a result.  And its tone — "YEC tabloids," "anti-science agenda" — will cost any article that quotes it uncritically.

Rodney Whitefield 2008, "22,000 °K?  Really?"  https://www.asa3.org/ASA/education/origins/helium-rw.pdf

The most underrated document in the dispute, and the shortest.  A retired IBM physicist.  Two arguments, both physical, both independent of Loechelt's modelling.

First: use Gentry's own depth-varying uranium and thorium to correct the 960 m retention, get 0.063-0.068 instead of 0.58, and therefore a zircon helium age near 100 Ma — which then sits in the correct order relative to the borehole's own apatite fission-track ages of 55-67 Ma and its sphene constraint of no more than 250 °C in 1.3 Ga.

Second: the low-temperature "defect" segment has an activation energy of 13.9 kcal/mol, and RATE's own GT-2 biotite over the same interval has 11.4 — against about 35 for zircon.  Since more than 40 % of GT-2 zircons carry non-zircon overgrowths, and the step-heating run went to 500 °C before dropping to 175 °C, he argues the low-temperature helium is coming back out of biotite inclusions loaded during the initial ramp.

Which way it cuts: the strongest single argument against RATE, because the cross-check is a genuine out-of-sample prediction using only pre-existing data.

Timothy K. Christman 2005.  https://www.oldearth.org/RATE_critique_he-zr.htm

A generalisation test, and structurally different from everyone else's work: take RATE's diffusivities at face value and ask what they predict elsewhere.  They predict that surface zircons should be nearly full of helium after 6,000 years.  Ganges and Indus detrital zircons hold under 1 % of theirs.

Which way it cuts: against RATE, and it needs none of the disputed inputs.  Its limit is that it takes detrital zircon helium ages at face value, which a committed accelerated-decay model will not.

Randy Isaac 2007, PSCF 59(2):143-146, with the 2008 exchange at PSCF 60(1):34-39.

Which way it cuts: weakly, and partly the wrong way.  His institutional charge — that presenting RATE as confirming a young earth contradicts RATE's own admission of unresolved problems — is fair and lands.  His technical objections mostly do not.  The factor-of-a-thousand vacuum claim is unsourced and physically unrecoverable, and he substituted a different claim a year later.  His charge that RATE's steady-state equation is "physically meaningless" is refuted by Loechelt, who re-derived the same relation from the standard formalism.  Isaac attacked the algebra, which is correct; Loechelt attacked the inputs, which are not.

Joe Meert 2002.  One published criticism of the helium work, and it was an error.  Humphreys' rebuttal is correct.  Reported here so that it is not counted as evidence on either side.

The mainstream literature RATE has to be consistent with

Reiners et al. 2004, Geochimica et Cosmochimica Acta 68:1857-1887.  The standard reference for zircon helium kinetics: Ea = 169 ± 3.8 kJ/mol, D₀ = 0.46 cm²/s, closure temperature 170-190 °C.  Also establishes that early-step non-Arrhenius behaviour in step-heating runs "would have a negligible effect on the bulk thermal sensitivity and closure temperature of a zircon crystal" — which is the foundation of Loechelt's multi-domain argument.

Guenthner et al. 2013, American Journal of Science 313:145-198.  https://ajsonline.org/article/62284.pdf  The most important paper in the dispute, and it was published eight years after RATE finished.

Helium diffusivity in zircon depends on accumulated radiation damage, and not monotonically.  Up to about 10¹⁸ alpha decays per gram, damage blocks the fast channels and makes the crystal more retentive.  Above about 2 × 10¹⁸ the damaged zones interconnect and diffusivity "increases by about nine orders of magnitude."

Which way it cuts: both ways, hard.  Against RATE's "uniformitarian model," which applies generic kinetics to crystals that are nothing like generic.  Against RATE's central claim, because the Fenton Hill dose computes to 1.98 × 10¹⁸ α/g — right on the turnover — so enhanced leakiness is exactly what the mainstream model predicts here, and was measured in unrelated crystals first.  And in RATE's favour on one narrow point: Humphreys' fitted activation energy of 143.9 kJ/mol is a near-match for Guenthner's own sample TH62Z at 145.96, which means nothing Humphreys measured is anomalous.

Harrison et al. 2009, Geochimica et Cosmochimica Acta 73:1039-1051.  http://sims.ess.ucla.edu/PDF/Harrison_et_al_2009.pdf  Piston-cylinder experiments at 10 and 20 kbar giving an activation volume of 14 cm³/mol for argon in muscovite.  This is the measurement that settles the vacuum objection, and it settles it in Humphreys' favour: at Fenton Hill pressures the effect is 9 to 28 %, not a factor of a thousand.

Wolfe & Stockli 2010, EPSL 295:69-82.  Zircon helium ages down the KTB deep borehole in Germany against known in-situ temperatures — the one natural test of whether vacuum-derived kinetics work in the ground.  Henke's own summary: they "demonstrated that helium diffusivities determined by measurements in a laboratory vacuum were able to suitably predict helium retention in their zircons."  Not obtained (ScienceDirect 403); reported here through the critic who concedes it, which is the safest available citation.

The site's own thermal record

These are the documents that decide the article, and they were all published before the controversy started.

Laughlin & Eddy 1977, LA-6930-MS.  https://www.osti.gov/servlets/purl/7080705  GT-2 completed December 1974, total depth 2,928-2,932 m, equilibrium bottom-hole temperature 197 °C, Precambrian basement at 730 m under 730 m of tuff, limestone and sandstone.  This is what shows Gentry's three deepest samples cannot be from GT-2.

Brookins et al. 1977, LA-6829-MS.  Apatite fission-track ages of 66.8 Ma at 790 m and 55.1 Ma at 1,130 m — the independent clock that Whitefield's zircon helium age has to be older than, and is.

Laney & Laughlin 1981, GRL 8(5):501-504.  Radiohalos in biotite from the same cores: annealing first observed at 1,850 m (134 °C) and complete at 2,120 m (151 °C).  A second independent thermometer, and an awkward one for RATE, whose radiohalo chapter needs halos to survive below 150 °C.

Harrison, Morgan & Blackwell 1986, JGR 91(B2):1899-1908.  Argon-40/argon-39 on microcline from five depths.  Thermal events at about 1,030 and 870 Ma, plus a very recent one: the shallow disturbance about 10 ka, the deep source under about 40 ka, "much younger than the main caldera event at about 1 Ma."  Not obtained — paywalled with no open copy.

Gentry's own table, reworked.  The temperature gradient between his consecutive samples rises from 63 to 95 °C/km below two kilometres.  A conductive steady state cannot do that.  The rock is being heated from below now, and this is derivable from the 1982 paper without reading anything else.

Sources that could not be obtained

Listed so that the reader knows exactly where this article is relying on someone else's transcription.

• Harrison, Morgan & Blackwell 1986, Figure 9, and Kolstad & McGetchin 1978, Figure 11. Loechelt charges that Humphreys read both backwards, mistaking heating curves for cooling ones. Neither original could be obtained. The article relies on Loechelt's transcription plus a verbatim quotation from Harrison et al. that points the same way, and says so in the text. This is the single open item that could still move the conclusion.
• Wolfe & Stockli 2010 abstract, verbatim.
• Sasada 1989, on the fluid inclusions recording a cool episode at 2,624 m.
• Laney & Laughlin 1981 body text; only the abstract was read.
• George P. Drake, "Helium diffusion in zircons," CRSQ 41(3):256-257 (2004) — the one criticism from inside creationism that reached print. Humphreys marks it "(Friendly)." Not obtained.
• Keith Wanser and Robert Brown, both creationists, both raising objections in December 2003 and January 2004 that exist only as conference remarks in Humphreys' own reference list.
• Roger Wiens' March 2005 radiation-damage objection, which exists as a seminar and an email. Humphreys' published answer — that the effect is "only a factor of two" — is asserted without a calculation in any source located.
• Loechelt's letter in Journal of Creation 24(3) (2010) and Humphreys' reply, which is Humphreys' only substantive engagement with Loechelt's model.

Experiments nobody has done

Worth listing, because in a forty-year dispute these are conspicuous.

1. Measure uranium and helium on the same Fenton Hill zircon. This settles Q₀ outright, and it has never been done.
2. Measure helium diffusion in zircon under confining pressure. Humphreys says plainly, "As far as I know, nobody has measured the effect of pressure on helium diffusion in zircon." Still true.
3. Measure helium-3 against helium-4 in the zircons, which would test Henke's extraneous-helium hypothesis and would break either RATE's model or Loechelt's.
4. Publish the Fenton Hill diffusion data conventionally. Farley never did.
5. Measure the radiation damage state of the actual grains. Given Guenthner's nine-order-of-magnitude swing across the threshold these crystals sit on, this is the measurement that would most change what everyone thinks.