The first step in leaf sampling is to collect the leaves from the locations that are being sampled. In Phase II, a single species of leaves, Honeylocust (Gleditsia Tracanthos), was chosen, based on its good sampling results in Phase I and its widespread availability. Due to the unavailability of this species in some locations, a similar species, Black Locust (Robinia Pseudoacacia), was substituted for some samples. The samples were collected in a synoptic pattern of 15 leaves from rural to urban areas in and outside Syracuse (see map and chart) during an extended dry period (9/4/99), as well as two more collected from a suburban area after the first rain, II-16 (9/18/99) and II-17 (9/25/99). A photograph was taken of each of the trees from which each of the leaves were collected. The leaves were wrapped in aluminum foil, which was then placed into a plastic bag and stored in a freezer at approximately -20 degrees Celsius.

The scanning electron microscope (SEM) was chosen for this experiment because of its usefulness in analyzing small particles on surfaces. Preparation of samples for the SEM requires several steps. These steps are as follows.

From each location, one leaf was selected at random from the aluminum foil. A portion of approximately 0.5-1 cm² was cut from each of the leaves using a razor blade. A strip of double-sided adhesive carbon tape was placed on a 5x5 cm aluminum sampling plate (See fig.1). The leaf samples were placed adaxial (top) surface up, due to the results found in Phase I.

For the rinsed samples, three leaves were selected for analysis: II-3, II-4, and II-16. Each leaf was cut into three parts using a razor blade. One of the three parts from each leaf was placed on the sample plate without rinsing. The next part of each leaf was squirted 10 times with distilled water, and named SR (hence the names 3aSR, 4aSR, 16aSR). This was then placed onto the sample plate next to the non rinsed leaves. Finally, the third part of each leaf was squirted 40 times, named LR, and placed on the sample plate.

The aerosol filter samples used for comparison in were samples collected in 1995 from rural (s#9) and an urban (s#8) sites, and in 1997 from a suburban area (ns#27), by Mark Abraham as part of his particulate air pollution research [Abraham, M.E., 1998] (kindly supplied for reanalysis in this project.). These archived samples were already prepared and could be placed immediately onto the sample plate for analysis. These filter samples were selected as their location corresponded to several of the leaf samples collected for this project, despite the differences in the time of collection.

Normally, non-conductive materials, or “wet” samples, such as leaves, are dried and coated with a thin layer of conductive carbon or metal before use in the SEM to prevent a buildup of a charge on the surface of the sample. However, the Aspex PSEM that was used for this experiment has a “variable pressure” feature that allows control of the air pressure inside the sample chamber. With higher pressure, more air molecules are present to conduct the charge, so it doesn’t build up on the sample surface. This enables the study of non-conductive samples. Therefore, instead of being dried and coated, the leaf samples were kept in their natural state to minimize the number of variables and to shorten the preparation procedure.

The sampling plate with the leaf samples was placed into the SEM sample chamber immediately after preparation. Using the variable pressure feature, the chamber was evacuated to a pressure of 0.2 torr. The filament was saturated and the electron beam was accelerated to a voltage of 20 kV, the maximum voltage for this SEM. The maximum voltage was used to provide the best backscattered electron (BE) imaging and x-ray information.

The method used to count and measure the particles on the leaves was automated analysis. In this type of analysis, which is another feature of the Aspex PSEM, the area of analysis and the parameters are set, and the computer program controls the analysis.

The first step to setting up the automated analysis is to set the basis of the analysis. In this step, the run number is set, the samples are named and the run letters (e.g. run 507a) for each sample are set. After this the parameters for the analysis are set. This tells the program the limits of the analysis, such as the maximum number of particles to count, the minimum and maximum sizes for particle analysis, the maximum amount of time, etc. This also sets the magnification for the analysis. The maximum number of particles ranged from 500 to 1500 (up to 2500 and 4000 in the counting-only runs). The particle size limits set for this project were 0.12 µm minimum and 10.63 µm maximum. The maximum time set was 5 hours. The magnification for this project was set at 1000x. Also set up in the parameters is whether the automated run will include the x-ray analysis or whether it will be counting and measuring only. The latter mode allows for much faster run time, but does not provide the elemental composition data of the particles found.

The next step is to set the stage coordinates for each leaf sample. This sets the corners of the area of each leaf to be analyzed, which, in this experiment, was a quadrilateral. This also sets the focus for each coordinate, so that if the sampling surface is not flat, the computer can calculate the slope of the surface and set the focus for each part of the surface.

The final step in setting up the automated analysis is setting the particle detection threshold. This setting ensures that the background surface (in this case, the leaf surface) is not analyzed, by setting the minimum brightness of particles that are analyzed. This is set so that the maximum number of particles possible and the least possible background “noise” (folds in the leaf surface, etc.) will be analyzed (fig.2a).

The automated analysis measures many parameters of the particles, of which two were primarily used in this analysis. First, the computer calculates the average diameter (Dave), which is the average of 16 different diameters of the particle. Next, the SEM does an x-ray analysis of the particle, which determines which elements are present in the particle and places them in a spectrum. This spectrum (see fig.2b), along with the average diameter and the coordinates of the particle, is stored in the memory so that the data can be viewed at a later time. For every tenth particle, a digital image of the particle was stored.

The data collected from the automated analysis were stored on several 100-megabyte Zip disks. This data was opened on a PC using the Zepview program, a program made for viewing data from automated analyses. From there, the data was exported into Microsoft Excel format and from there into Statistica. Excel and Statistica were used to determine median Dave, concentrations, and, chemical classification statistics (if available) of the particles, after which the different leaves and aerosol filter samples were compared. These comparisons were displayed in graphs.

Ten different runs were set up, using five different sample plates. The numbers of these runs were 363, 389, 395, 415, 428, 429, 432, 433, 476, and 507. Runs 363 and 395, 389 and 507, 415, 428, and 432, and 433 and 476 were repeats of the same sample plates. Runs 415, 428, and 432 used rinsed leaves, and run 429 used aerosol filter samples (see chart).