Methodology

Tephra samples used in this study include two control samples and nine unknown samples. One control sample (n=1) was collected from the Q-I tephra proximally located at the base of Quilotoa volcano and one control samples (n=1) from pumice associated with the Cayambe volcano San Marcos event . All control samples were identified and collected by volcanologists Patricia Mothes and Minard Hall. Unknown tephra samples were collected from the archaeological sites of Cochasquí (n=2), La Vega (n=1), and Zuleta (n=6) during various archaeological investigations carried out between 2019 and 2022. Mothes and Hall tentatively identified two of the unknown samples from Zuleta as Quilotoa AD 1280 and one unknown sample from Zuleta as Cayambe, possibly from the circa AD 1485 San Marcos event based on their mineralogical constituents. The remaining samples were tentatively identified by archaeologists as Quilotoa based on their stratigraphic relationship to archaeological features, their color, and texture.

Sample Preparation:

Sample preparation and instrumentation for this study were based on that outlined in Westgate and Pearce (2017).

Samples were step sieved at 150 μm, 100 μm, and 50 μm with the largest fractions retained as the active samples. Samples were placed in an ultrasonic bath for 10 minutes to remove fragile pumiceous glass. (8 hours).

The larger fraction was mounted in epoxy and polished using 3μm aluminum oxide polishing paper to expose the grains in preparation for laser ablation (48 hours).

Element maps were made for some of the mounts (n=6) using a JEOL 6490 Low-Vacuum Scanning Electron Microscope to aid in the correct identification of volcanic glass (16 hours).

Mounts were scanned using an optical scanner before being loaded into the laser tray (1 hour).

The optical scans were used in Adobe Illustrator to overlay element maps and target specific grains for laser ablation and to aid in relocating these spots for later microprobe analysis. A total of 141 spots were initially selected from the 12 mounts (8 hours).

The spots identified in the scans were selected as targets for spot analysis with the laser (10 hours initially with 2 hours for follow-up spots).

After acquisition was completed on the initial round of 141 spots, an additional 32 spots were selected for a follow-up analysis to test differences between volcanic glass and quartz (1 hour). One of the greatest challenges encountered during this in the course of this study was correctly distinguishing volcanic glass from quartz (Figures 1 and 2). The Quartz has sufficient Si content to appear on the element maps and at first glance has the same consistency as the glass. Once the quartz spots had been identified, it was determined that the trace element concentrations of the quartz spots could be distinguished from the glass by the absence of rare earth elements (Figure 3).

Figure 1: Volcanic glass in a sample mounts during spot selection.         Figure 2: Quartz fragment in a sample mount during spot selection.

Figure 4: Trace elements within two quartz spots (middle frame) bracketed by volcanic glass spots as they appear in the acquisition window during instrument operation.

Analysis:

After setting up the method and tuning the instrument (3 hours total), the instrument runtime took approximately 6 hours for the initial ablation and 1 hour for the follow-up ablation.

The results were processed on iolite4 (6 hours initially, 3 hours during the follow-up) and the data was analyzed in excel (approximately 9 hours so far, ongoing).

Laser ablation was conducted using a 193 nm Excimer laser and analyzed with an Agilent 7500ce quadrupole ICP-MS. Larger shards were analyzed using a 90μm beam and smaller shards using a 50μm beam. All spots were pre-ablated prior to acquisition. Four test spots were run on representative unknowns and one test spot was run on epoxy during the instrument tuning. Laser energy was pulsed at 50% (2 J/cm²) at 10 Hz using an acquisition time of 60 seconds with a sampling period of 0.322 seconds and 20 second dwell times. Analytical conditions were selected after tuning the instrument using NIST 612 glass standard. NIST 612 was used as a the primary external standard and for external calibration in conjunction with ²⁹Si as an internal standard. NIST 610 was used as a secondary external standard. Both external standards were analyzed at the start of the analysis, every hour during the analysis, and after the analysis to account for drift. The 28 analytes chosen for this experiment include ⁷Li, ²⁷Al, ²⁹Si, ⁴³Ca, ⁴⁴Ca, ⁵⁵Mn, ⁶⁰Ni, ⁶³Cu, ⁸⁵Rb, ⁸⁹Y, ⁹⁰Zr, ⁹³Nb, ¹³⁷Ba, ¹³⁹La, ¹⁴⁰Ce, ¹⁴¹Pr, ¹⁴⁶Nd, ¹⁴⁷Sm, ¹⁵³Eu, ¹⁵⁷Gd, ¹⁵⁹Tb, ¹⁶³Dy, ¹⁶⁵Ho, ¹⁶⁶Er, ¹⁶⁹Tm, ¹⁷²Yb, ²⁰⁸Pb, and ²³²Th. Expected weight percent of SiO₂ was collected from existing literature for the internal calibration of the Quilotoa control (65.15% from Mothes and Hall 2008) and the Cayambe control (62.85% from Samaniego 1998) samples. All other unknowns were given an estimated SiO₂ weight percent of 68.5% corresponding with distal Quilotoa tephra (Hall and Mothes 2008).

Instrumentation:

Recovery: