Radiocarbon Dating

Why radiocarbon dating matters

• Timeline of the late Pleistocene and Holocene. Radiocarbon is the dominant chronometer for the last ~50,000 years of human prehistory and history — covering the entirety of behavioral modernity, the peopling of Eurasia, the Last Glacial Maximum, the Neolithic transition, and the rise of complex societies. No other technique has comparable precision (~ ±30 years for 5,000-year-old samples) and reach across this window.
• Direct dating of organic material. Charcoal, wood, bone collagen, plant remains, hair, leather, parchment, papyrus, and shell can all be dated. This is unique among radiometric methods — most others (K-Ar, U-Pb) date inorganic minerals only. Direct dating means the result tells you when this thing died, not when the surrounding rock formed.
• Carolyn Brown's bomb-pulse forensics. 1950s-1960s atmospheric nuclear tests doubled atmospheric 14C (the 'bomb peak') in 1963; values have since decayed exponentially. Tissues incorporate the bomb-peak 14C while alive, so AMS dating of human teeth and DNA can reconstruct year of birth to ~ ±2 years for individuals born 1955-2010 — used in unidentified-remains forensics, art-forgery detection (post-WWII paint vs pre-war paint), and biology of cell turnover.
• Continuous calibration backbone. The IntCal20 calibration curve runs from 0 to 55,000 cal BP, anchored by absolute tree-ring counts back to 14,200 BP and U-Th-dated speleothem and varved-lake records beyond. It has revised the 'true' age of many key archaeological events; e.g. the Iron Age Ipuwer / Hekla 3 eruption shifted by decades after IntCal13 to IntCal20.
• Climate proxy applications. Atmospheric Δ¹⁴C variations themselves record solar activity (Schwabe, Gleissberg, de Vries solar cycles), geomagnetic field strength, and ocean carbon cycling. The Suess effect (post-1850 atmospheric 14C decline from fossil fuels) is a textbook anthropogenic isotope tracer.
• Forensic dating of art and antiquities. The 14C bomb pulse is critical to authenticating modern fakes; pre-1955 'old paintings' must contain pre-bomb canvas. Fakes using post-1955 canvas show elevated 14C and are easily distinguished. AMS dating of the Vinland Map (1973), the Shroud of Turin (1988), and Dead Sea Scroll fragments (1990s) are landmark applications.
• Extends below 50 mg of carbon. Modern micro-AMS techniques can date samples as small as 25 micrograms of carbon — single seeds, individual amino acids from collagen, single particles from atmospheric aerosol. This enables applications like compound-specific radiocarbon (e.g. dating of marine animal lipids while excluding contamination).

Common misconceptions

• The half-life is exactly 5,568 years. The 'Libby half-life' of 5,568 ± 30 years was Libby's 1949 measurement; the true half-life (the 'Cambridge half-life') is 5,730 ± 40 years (1962). All raw radiocarbon ages are reported in 'conventional radiocarbon years BP' using Libby's value to maintain consistency with the older literature; calibration curves convert to calendar years and absorb the ~3% half-life correction internally.
• 14C/12C is constant. It is not. Atmospheric 14C/12C has varied by ~10% over the past 50,000 years due to cosmic-ray flux changes, geomagnetic field changes, and ocean carbon cycling. Calibration is mandatory; raw 'BP' ages and true calendar 'cal BC/AD' ages can differ by hundreds of years in the past 5,000 years.
• Radiocarbon dates rocks. No. Rocks contain no 14C unless contaminated. Radiocarbon dates organic material that exchanged carbon with the atmosphere or oceans (charcoal, bone collagen, shell carbonate). Dating rocks needs U-Pb, 40K-Ar, or 87Rb-Sr — different isotope systems with much longer half-lives.
• Marine and terrestrial samples have the same calibration. They do not. Marine samples are about 400 years 'too old' due to the marine reservoir effect: ocean water is depleted in 14C because deep water has been out of contact with the atmosphere for ~1000 years. Marine20 (and regional ΔR adjustments) corrects this; failing to apply it gives systematic overestimates.
• Old wood = old site. A site dated by burnt wood may be hundreds of years younger than the wood itself if the wood was already old when burnt ('old wood effect') — common in arid regions where dry wood persists. Best practice: short-lived material (twigs, seeds, bone) or multiple dates from different contexts.
• Contamination is rare. Quite the opposite — radiocarbon contamination is the dominant source of error. Modern carbon (humic acids in soil, conservator adhesives, percolating groundwater) only needs to make up 1% of a 30,000-year-old sample to make it date to ~24,000 years instead. Pretreatment (acid-base-acid for charcoal, ultrafiltered collagen for bone, alpha-cellulose for wood) is essential.

Process: from cosmic neutron to calendar date

The technique has three physical stages and one analytic stage. (1) Production: galactic cosmic rays (mostly high-energy protons) collide with atmospheric nuclei to spall off neutrons; about 1% of these neutrons are captured by ¹⁴N: n + ¹⁴N → ¹⁴C + p. The annual global production rate is roughly 7-8 kg of ¹⁴C; with a steady-state atmospheric inventory of ~75 t, this means atmospheric 14C/12C reaches a quasi-equilibrium of ~1.2 × 10⁻¹². The newborn ¹⁴C oxidizes to ¹⁴CO₂ within hours and mixes globally on a timescale of years. (2) Incorporation: photosynthesizing plants take up atmospheric ¹⁴CO₂ at the same ratio (with small δ¹³C-fractionation corrections); animals eat plants; humans eat both; oceans dissolve CO₂ at a slight isotopic offset. While alive, every organism's tissue ¹⁴C/¹²C tracks the atmosphere within ~50 years. (3) Decay: at death, carbon exchange stops; ¹⁴C decays back to ¹⁴N via β⁻ emission with a half-life t₁/₂ = 5,730 ± 40 years. The activity of modern carbon is ~14 disintegrations per minute per gram (dpm/g), or ~0.226 Bq/g.

(4) Measurement and calibration: A 1 mg carbon sample is pretreated to remove contamination (acid-base-acid for charcoal; ultrafiltered hydroxyproline for bone collagen; alpha-cellulose for wood), combusted to CO₂, reduced with H₂ over Fe to elemental carbon (graphite), and pressed into an AMS target. The AMS instrument ionizes the graphite to C⁻, accelerates to MeV energies through a tandem accelerator, strips electrons in a gas-stripper, mass-selects ¹⁴C from ¹³C and ¹²C, and detects ¹⁴C atoms one by one in a gas ionization chamber or solid-state detector. Run time per sample ~30-60 min; precision ~ ±20-40 years for samples from the past few thousand years.

The raw radiocarbon age in conventional years BP (where 'present' is 1950 CE by convention) is computed as t = (5568/ln 2) × ln(F_modern/F_sample), where F_sample is the measured ¹⁴C/¹²C ratio normalized to oxalic acid standard. This raw age is then matched against the IntCal20 calibration curve (Northern Hemisphere terrestrial) using software such as OxCal or CALIB. The output is a probability density over calendar years (BC/AD). Reported dates show the 1σ (68%) and 2σ (95%) confidence intervals — the latter for archaeological publications. Plateaus in the calibration curve (notably 800-400 BC, the 'Hallstatt plateau') give 14C ages with 400-year ambiguity even with perfect measurement, an unavoidable limitation of the technique.

14C vs 40K-Ar vs 87Rb-Sr vs U-Pb dating ranges and accuracies

Method	Parent → Daughter	Half-life	Useful range	Material dated	Best precision
Radiocarbon (14C)	14C → 14N (β⁻)	5,730 yr	~300 to 60,000 yr BP	Charcoal, wood, bone, shell, fabric	±20-40 yr (recent)
U-series (²³⁰Th/²³⁴U)	234U → 230Th (alpha)	245,000 yr	~1,000 to 500,000 yr	Speleothems, corals, carbonate	±0.1% on speleothems
K-Ar (40K-40Ar)	40K → 40Ar (EC)	1.25 Gyr	~100,000 yr to 4.5 Gyr	Volcanic rocks, K-feldspars, micas	±0.5-2%
40Ar/39Ar (improved K-Ar)	40K → 40Ar via 39Ar tracer	1.25 Gyr	~10,000 yr to 4.5 Gyr	Sanidine, biotite, hornblende	±0.1% on Quaternary
Rb-Sr	87Rb → 87Sr (β⁻)	49.6 Gyr	~10 Myr to 4.5 Gyr	Whole-rock, micas (Rb-rich)	±1-2%
U-Pb (zircon)	238U → 206Pb, 235U → 207Pb	4.47 Gyr / 0.704 Gyr	~1 Myr to 4.5 Gyr	Zircon, monazite, baddeleyite	±0.1% on Precambrian
Luminescence (OSL/TL)	Trapped electrons, optical/heat reset	n/a	~100 to 500,000 yr	Quartz, feldspar in sediment	±5-10%
Cosmogenic (10Be)	Spallation production rate	1.36 Myr	~1,000 to 5 Myr	Quartz exposure surfaces	±5-15%

Applications and case studies

• Ötzi the Iceman (5,300 ± 30 years BP). Multiple AMS dates on Ötzi's tissue, hair, and grass-cape fibers from the 1991-discovered Tyrolean glacier mummy returned a calibrated calendar age of 3370-3100 cal BCE — placing him in the late Neolithic / early Copper Age. Stable-isotope and bomb-pulse techniques on the same sample later established his last meal and Alpine birthplace.
• Dead Sea Scrolls (~150 BCE-70 CE). AMS dating of parchment and linen scroll wrappings carried out at the University of Arizona NSF AMS facility (1991, 1995) returned calibrated dates of ~150 BCE to 70 CE, consistent with paleographic estimates and the Roman destruction of Qumran in 68-70 CE. These dates anchored the chronology of Second Temple Judaism.
• Shroud of Turin (1260-1390 CE). Three independent AMS labs (Arizona, Oxford, Zurich) measured Shroud linen samples in 1988 and converged on a calibrated 95%-confidence date of 1260-1390 CE — the medieval period (Damon et al., Nature 337:611). The result is widely accepted by the radiocarbon community despite ongoing dispute from the religious authenticity perspective.
• Lascaux cave paintings (~17,000 BP). Charcoal pigment from the Lascaux Hall of the Bulls dates to 17,000-15,000 cal BP, placing the paintings in the Magdalenian. The very young end of Lascaux painting overlaps with the Solutrean-Magdalenian transition.
• Bomb-pulse forensics on human cells. Atmospheric ¹⁴C from 1950s-60s nuclear tests doubled and then decayed; humans alive in this era incorporated bomb-pulse 14C into their tissue. AMS analysis of single cell-types — heart muscle, neurons, fat cells — measures their year of synthesis to ±2-3 years (Spalding, Bhardwaj, et al., Cell 2005). This established that adult human cardiomyocytes do turn over (~1%/year) and that adult-born neurons exist in the hippocampus.