Sampling of Volcano Mountain lavas was intended to date flows of the most youthful appearance and stratigraphic position to determine the age of the most recent eruption (Table 1; Fig. 1). Particular attention was paid to the flow within the small summit crater as the most likely to be the youngest. For redundancy purposes, two to three samples were collected at each location. Site parameters used to guide sampling included a relatively flat, stable surface exhibiting primary depositional features (e.g., ropy textures) and, if possible, minimal vegetation cover. A gas-powered cutoff saw was used to control sampling by slicing the surface to a uniform depth. A rock hammer and chisel were used to liberate the sample from the prepared area. Approximately 2 kg of sample was collected from each site.

A total of 13 samples were collected on six separate lava flow surfaces. Jackson and Stevens (1992) (see also Jackson and Huscroft, 2023) mapped VM flows and named them based on the direction of the flow (north or south) and its stratigraphic position relative to the other flows based on field observations (0–2, where 2 is assigned to the youngest flows) (Fig. 1). Samples VM-01, 02, and 03 were collected from the stratigraphically highest flow in the cinder cone (flow 2Na). VM-06 is from the fifth south flow (flow 2Sa), VM-08 and 09 are samples from the third south flow (flow 1Sa), and VM-10 and VM-11 are samples from the second north flow (flow 1N). Samples VM-04, 05, 12, and 13 are from the fourth south flow (flow 2Sb, possibly the same age as flow 2Sa, Jackson and Huscroft, 2023) but, along with sample VM-07 (flow 2Sa), yielded too little olivine to make a cosmogenic ³He measurement. See Table 1 for sample details.

At elevations ranging from ∼0.7–1.1 km, the area experiences on average 28.8 cm of snow cover for 5.5 months of the year (Government of Canada, 2007). Vegetation, including lichen and moss, covered a significant surface area of the samples, though it is highly variable, from no coverage at VM-03 to covered in a mat of moss 8 cm thick at VM-08.

All hand samples were scrubbed of vegetation, rinsed in water, dried in air, and jaw-crushed to millimeter-sized grains. The abundance of coarse olivine crystals varied among samples; only samples VM-01, 02, 03, 06, and 08 through 11 had sufficient olivine for cosmogenic ³He measurement. From these, matrix-free, millimeter-size olivine crystals were handpicked from the crushate based on their distinct lack of cleavage and green color.

Helium was extracted from olivines by crushing under vacuum for 2 min and then re-crushed for another 5 min. The two steps were combined into a single He result. Previous work with this same crusher design indicates these crush durations will not cause release of matrix-sited ³He (Blard et al., 2008). Prior to dropping samples into the crusher anvil, 2 min “no sample” crushes were measured and used to blank-correct He measurements. Maximum blank crush levels of ³He and ⁴He were 2000 atoms and 0.009 pmol, respectively.

For some samples, material recovered from the vacuum crushes was powdered in a mortar and pestle under ethanol and sieved to <30 µm, dried, weighed, and wrapped in Sn foil packets for fusion analysis. For other samples analyzed by fusion, the handpicked olivines did not undergo the initial vacuum-crush step, only powdering under ethanol. Ethanol was used to circumvent (Cox et al., 2022) trapping of atmospheric He during powdering (Protin et al., 2016).

Sn-foil-wrapped samples for high temperature vacuum extraction analysis were loaded into the dropper arm of a double-wall furnace and evacuated overnight. The dropper arm was not heated to eliminate any possibility of baking-induced loss of ³He. The furnace was outgassed directly into the vacuum pump for 1 h at 1400 °C prior to analysis. Samples were then dropped into a carbon liner in the furnace and heated to ∼1200 °C for 25 min. Re-extracts were measured directly after each extraction to ensure the complete release of helium. When the re-extract signal exceeded the pre-analysis hot blank, the re-extract was repeated. All sample gas in excess of blank was combined in the concentrations reported here. While these olivines were not fused during these analyses (the temperature was held sub-melting to minimize He blanks), complete He extraction is ensured by the re-extraction steps. For consistency with prior work, we nevertheless refer to these measurements as “powder fusions”. Two olivine samples were fused in a lithium borate flux at ∼1000 °C for 25 min in a molybdenum liner. The flux was used in these fusions to reduce blank He levels by lowering extraction temperature (Farley et al., 2020). Hot blanks at ∼1200 °C for ³He and ⁴He were 22 000 atoms and 0.009 pmol, respectively. Blanks measured on the tin foil are negligible compared to this hot furnace blank. Blank corrections made to sample gas concentrations were mostly <2 % but always <5 %.

About 750 mg of powdered (<30 µm) sample was wrapped in Sn foil packets and loaded into the vacuum furnace as described above. The Sn packets were then dropped into the furnace and He successively released at 800, 1000, and 1400 °C each for 30 min duration. These temperature steps were selected based on previous studies reporting the release of cosmogenic ³He below 1000 °C (Kurz, 1986a; Swindle et al., 2023). A thermocouple and a pyrometer were used to calibrate furnace temperatures as a function of furnace output power, after which heating steps were maintained using constant power output.

Hot blanks were performed on an empty furnace at each of the scheduled temperature steps to use as blank corrections on He quantities at each step. Blank levels were <1.5 % of extracted He for all steps. Re-extracts were measured on the material after step heating to ensure total helium release. When the re-extract signal exceeded the pre-analysis hot blank, the re-extract was repeated. When the re-extract exceeded the hot blank, the excess He was added to the high temperature step.

Helium was purified and measured as previously described (Horton et al., 2019; Swindle et al., 2023). Gas extracted from the furnace was passed through a charcoal U-trap to remove contaminants such as CO₂, H₂O, and Ar. Remaining reactive gases were removed using a hot SAES NP10 getter, and H₂ was removed using a cold SAES NP10 getter. Helium was then cryofocused on charcoal at 14 K and released at 34 K into a Helix SFT mass spectrometer. ³He was measured on a pulse-counting electron multiplier and ⁴He on a Faraday cup. Instrument sensitivity and internal consistency were monitored by running a standard containing ∼3.1 pmol of He (3He/4He ratio = 2.05 RA) throughout our experiments. This standard was created from pure ³He and ⁴He using a capacitance manometer, and the amounts delivered are known to better than ±1 %.

Corrections for snow cover, vegetation cover, and self-shielding were considered for every sample. The snow correction factor was calculated using Eq. (3) in Vermeesch (2007) based on the spallogenic neutron attenuation length, average snow thickness, and snow density. A snow density of 0.3 g cm⁻³ was assumed. Topographic shielding corrections were made for VM-01, 02, and 03 due to their location inside the VM cinder cone (see red outline in Fig. 1). Topographic shielding for the other samples was negligible. Shielding factors are listed in Table 1.

Analytical results for the crushing and fusion experiments are listed in Table 2. The three samples (VM-02, VM-06, and VM-09) analyzed by crushing come from three different flows (2Na, 2Sa, and 1Sa) yet have indistinguishable crush 3He/4He ratios (7.99±0.19, 8.08±0.19, and 8.23±0.2 RA). Crush-release He concentrations in these samples are high for basalt-hosted olivines, ranging from 3.38 to 8.20 pmol g⁻¹.

Six powder fusion analyses and the total helium yield from the three step heats (described below) provide information on the matrix-hosted helium in all four of the investigated lava flows. Across these nine analyses, He concentrations range from 1.74 to 6.30 pmol g⁻¹, similar to the crush results. 3He/4He ratios ranged from 7.9±0.2 to 9.6±0.2 RA. For the three lava flows for which both crushing and fusion analyses are available, fusion 3He/4He ratios ranged from the same within error as the crush result (VM-09 from flow 1Sa), to a maximum of 1.5 RA higher (VM-06 from flow 2Sa).

We used these data to compute cosmogenic ³He concentrations via Eq. (1) by assuming that the crush 3He/4He ratio on a given flow can be applied to all samples obtained from that flow. For flow 1N, for which we did not measure a sample by crushing, we computed cosmogenic ³He concentrations by assuming the mean crush 3He/4He ratio of 8.1±0.2 RA obtained on the other flows. This assumption is supported by the step-heat results described below.

Cosmogenic ³He was confidently detected in five of the nine fusion analyses (Table 3). For flow 2Na we obtained cosmogenic ³He concentrations of 2.75±1.25 Mat g⁻¹ (VM-02), 3.27±1.57 Mat g⁻¹ (VM-01), and 4.05±1.82 Mat g⁻¹ (VM-03). One analysis each of flow 2Sa (VM-06) and 1Sa (VM-08) yielded ³He_c of 2.19±0.36 and 2.27±0.64 Mat g⁻¹, respectively. For the remaining four analyses the 1σ analytical uncertainty on ³He_c exceeds the measured value; i.e., the cosmogenic ³He concentration is indistinguishable from zero.

The high analytical uncertainties on ³He_c – typically more than 1 Mat g⁻¹ – in the VM samples arise from the large corrections required for mantle ³He present in the powder fusion analyses (Eq. 1). For example, as shown in Table 2, the fraction of ³He that is cosmogenic in these analyses ranges from a maximum of 15.8 % (VM-06) to <2 % (VM-11, VM-09). Stated differently, the 3He/4He ratios obtained by fusion are very close to the ratios obtained by crushing, and this similarity at least in part reflects preservation of a high concentration of mantle helium through the powdering step.

Because the crush–fusion method was not especially successful in isolating the cosmogenic component, we attempted step heating on two samples from the stratigraphically youngest flow (2Na, samples VM-01 and VM-03) and one from the stratigraphically oldest flows (Flow 1N, sample VM-10). Results are shown in Table 3 and Fig. 2. In all three samples, the 3He/4He ratios decline monotonically with extraction temperature, while the ⁴He yields increase. For example, in sample VM-03, the 800, 1000, and 1400 °C steps yielded 3He/4He ratios of 13.1, 9.0, and 8.2 RA and ⁴He concentrations of 0.67, 0.89, and 3.1 pmol g⁻¹, respectively. Previous work indicates that mantle He in fluid inclusions is released at extraction temperatures >1000 °C (Kurz, 1986a, b; Swindle et al., 2023). In the case of the VM samples, the 3He/4He ratios extracted in the final steps (7.87±0.22, 8.17±0.23, 7.79±0.22 RA) are indistinguishable from the mean 3He/4He ratio obtained by crushing (8.10±0.20 RA), supporting this conclusion.

3He/4He ratios of He released by crushing and in the highest temperature step of the step heats are within the typical range of mid-ocean ridge basalts (Graham et al., 1992), suggesting He in Volcano Mountain derives from the upper mantle. Regardless of its origin, the presence of this mantle He substantially complicates determination of cosmogenic ³He exposure ages. While mantle ⁴He concentrations of crushed olivines are sometimes, but not often (Kurz et al., 1990; Farley and Neroda, 1998; Fenton and Niedermann, 2014; Heineke et al., 2016), as high as the ∼4.5 pmol g⁻¹ we obtained on the VM samples, the more important complication arises from the fact that this mantle He is not effectively removed by powdering to <30 µm. Our powder fusion analyses indicate that about a third of the mantle component remains in the powdered olivine in every case. Our results suggest a mantle component that survives the powdering step, most likely in ≪30 µm fluid inclusions. Evidence to support this interpretation comes from our observation of trails of circular voids (<10 µm) present in backscatter electron images of VM olivine xenocrysts. We interpret these to be secondary inclusion trails that carry mantle He. Survival of the mantle component when crushing this fine is not typical, but it is similar to results obtained on Twin Sisters peridotites (Swindle et al., 2023). Another possibility is that the ubiquitous kink banding and subgrain boundaries observed in VM (Francis and Ludden, 1990) and Twin Sisters olivines (Swindle et al., 2023) could contribute to mantle He retention.

The step-heat method takes advantage of the different release kinetics of matrix-sited and fluid-inclusion-sited He. Diffusion alone is adequate to release matrix-sited helium; in contrast, for fluid inclusion helium to escape the crystal requires it to first dissolve into the olivine, an energetically disfavored process (Trull and Kurz, 1993). Three temperature steps allowed us to be flexible about the temperature that best separates these two components. The evolution of the 3He/4He ratio with temperature thus can reveal two-end-member mixing between mantle and cosmogenic He (assuming radiogenic helium and nucleogenic helium are both absent).

Compared to the 3He/4He ratios obtained on powder fusions, the 3He/4He ratios of the lowest temperature steps in our step-heat experiments are between 1.5 and 4 RA higher (Tables 2, 3). In contrast, 3He/4He ratios of the 1000 °C steps are very similar to the powder fusion values, and about 0.7 RA higher than the highest temperature steps and the associated crush-released 3He/4He ratios. These data provide support for the interpretation that all cosmogenic He is released prior to the highest temperature step, while mantle He is released in progressively greater amounts as step temperature increases.

To calculate the cosmogenic ³He concentration from the step-heat results for each sample, we assumed the final step represents the mantle component, used Eq. (1) to compute the cosmogenic ³He yield in the two lower temperature steps, and combined them to obtain a total. Using this method, we obtained indistinguishable ³He_c concentrations on the two samples of flow 2Na (3.71±0.41 and 3.38±0.33 Mat g⁻¹) and 3.43±0.63 Mat g⁻¹ on flow 1N. While still uncertain (average ±13 %), these cosmogenic concentrations are significantly better determined than by the crush–fusion method.

It is possible that a more detailed step heat could yield even greater degrees of separation between cosmogenic and mantle helium components but at the potential cost of analytical precision associated with averaging multiple analyses and with blank corrections. Were significant amounts of radiogenic and/or nucleogenic helium present in these samples (unlikely given the eruption age of ∼10 kyr; see below), this method would yield inaccurate results in the same way that the crush–fusion method would.

Exposure ages in Table 4 were computed from ³He_c concentrations and corresponding uncertainties using the CRONUS-Earth online calculator (Balco et al., 2008; Phillips et al., 2016; version 3), where the production rate of ³He_c is calibrated using data from Borchers et al. (2016). Exposure ages were obtained for four of the flows. By far the best determined is the exposure age of the youngest flow, 2Na. The two step-heat analyses (VM-01, 03) yield indistinguishable ages of 12.3±1.9 and 11.1±1.6 ka. These are in good agreement with the three crush–fusion pairs (VM-01, 02, 03), with ages ranging from 9.0±4.2 to 13.3±6.1 ka. Assuming independent uncertainties among these analyses, the weighted means of these five ages is 11.4±2.2 ka (MSWD (mean square weighted distribution) 0.17, 95 %), slightly more uncertain than the step-heat analyses on their own. Four analyses were performed on flow 1N, with the step-heat experiment yielding an age of 14.0±3.0 ka and three crush–fusion pairs yielding ages between 2.3±4.8 and 7.4±9.9 ka. The weighted mean of these analyses is 8.7±4.3 ka (MSWD 2.3, 7.4 %), again more uncertain than the step-heat analysis by itself. The individual crush–fusion analyses of flows 2Sa and 1Sa yielded exposure ages of 9.6±1.9 ka and 10.0±3 ka, respectively.

Exposure ages are indicated on the map in Fig. 1 using the weighted means when available. All four flows yield the same exposure age within error (weighted mean of all analyses: 10.5±1.7 ka, MSWD 0.92, 51 %), and there is no indication that the stratigraphically lower flows (1Sa and 1N, weighted mean 9.2±3.5 ka, MSWD 1.8, 13 %) have older exposure than the higher flows (2Na and 2Sa, weighted mean 11.0±1.9 ka, MSWD 0.28, 93 %).

The simplest interpretation of these results is that all of the studied VM lava flows erupted at 10.5±1.7 ka. Alternatively, honoring the stratigraphy, the flows may have erupted over an age range from about 12.7 ka (1S flows) to about 9.1 ka (2Na and 2Sa).