Point Count Method 2.

02
Mike Meier, December 31, 2003 This program helps with the counting and calculations done when using the systematic point count method for measuring the volume fraction of phases. You still have to use test grid and count points that fall on the selected phase, but by using this program tallying the counts and calculating the volume fraction and related sampling statistics is much easier, and the results can be printed.

Registration
Please register this software. To register simply fill out the registration form and mail or email it to me at the address given below. Your registration will give me a better idea of who is using this program. I will not share your personal information with anyone.

Copyright
This program is copyrighted by Mike Meier, July 1, 1999. All rights reserved.

Distribution
Educational Users: Writing this program provided me with an opportunity to develop my Delphi/Windows programming abilities and to make something I could give to the students in my laboratory courses. I am happy to offer it for free to non-profit educational users, including students, staff and faculty. Commercial Users: Private and commercial users pay $10 for each copy of this program. This program represents quite a bit of midnight engineering and an investment in a number of programming tools and many Windows and Delphi-programming books.

Files
The following files are included with this program: 1. Point Count Method 2.02.exe The program 2. Point Count Method 2.02.pdf Documentation (this document) 3. Registration Form.pdf Registration form

Installation
No special installation procedures are required. Simply copy the program to the directory where you want to run it from and then create a shortcut and place it in your "Start Menu".

Feedback and Contact Info

I am interested in hearing from people, particularly students and instructors, who use this program. Please email me at: mlmeier@sbcglobal.net or mail me at: Mike Meier 515 K Street Davis, CA 95616 USA email: mlmeier@sbcglobal.net

Using Point Count Method

This program is very easy to use. The basic procedure involves simply starting the program, setting up a measurement, and entering the counts. After five counts have been entered the results and accompanying statistics will be displayed. Finally, print the report and quit or start another measurement. Step 1. Start the program When the program starts the main form (figure 1) is displayed. This form contains a grid where the counts are entered and buttons for starting a new measurement, displaying the graph, printing the report, etc. Most of these buttons will be disabled until a measurement has been set up.

Figure 1 Point Count Methods main form.

Step 2. Set up a measurement Click the NEW button to bring up the dialog (figure 2) where you can enter basic measurement parameters.

Step 3. Enter the readings Enter the counts by either typing them into the grid or use the keypad (figure 3). The keypad can greatly aid this process by relieving the error-prone job of manual counting. Simply click the +1 button each time an intercept is encountered and the click NEXT button to transfer the count to the grid and start with the next grid.

Figure 3 The keypad can make entering counts much easier.

Step 4. Note the results After five counts have been entered the mean, standard deviation, grain size and other statistics will be displayed on the main form (figure 4). These will be updated each time a new count is entered.

Figure 4 The results will be displayed once five counts have been entered.

Click the GRAPH button to view a bar chart showing the counts and the normal distribution which is calculated from the mean and standard deviation. This graph was added to the program to help students understand the business of sampling statistics better.

Figure 5 This graph helps illustrate the concepts of the normal distribution, mean, standard deviation and sampling statistics.

Step 5. Print the report Click the PRINT button. The report will include time and date, setup parameters (grid size and confidence level), the results, and a listing of each individual count.

Measuring Volume Fraction

Mathematical Basis
Sampling a Two-Dimensional Structure Using Test Points The probability that one of an infinite number of test points falls on a group of particles (phase) contained in a sample area L2 is (1)

where A is the total area covered by the particles and AA is the area fraction of particles. This relationship is the basis of the point-count method which is used to estimate the volume fraction of particles or second phase in a microstructure. This is possible because the volume fraction is equal to the area fraction.

General Methods
Volume Fraction, Areal Analysis 1. Trace the structure onto paper, weight the paper, cut out the parts containing the phase of interest and weigh these parts. The weight fraction equals the area fraction of phase in this section. The average of a number of area fraction measurements gives one the volume fraction of phase. 2. Etch the specimen to color each phase differently. Use a computer to measure the area of each phase (color). The average of a number area fraction measurements is equal to the volume fractions of phases. Volume Fraction, Lineal Analysis Using lines uniformly distributed but randomly oriented, the average length of the line segments crossing the selected phase is equal to the area fraction of the phase viewed in the planar section. The average of a number area fraction measurements is equal to the volume fractions of phases. Volume Fraction, Point Count The fraction of a number of points which fall on the selected phase is equal to the area fraction of that phase. The average of a number area fraction measurements is equal to the volume fractions of phases.

Recommended Method
Measurement of Volume Fraction of Phases Using the Systematic Point Count Method This method is based on the fact that the average fraction of points which fall on grains of a particular phase is equal to the area fraction of the phase in the polished section and that the average of these area fractions is equal to the volume fraction of the phase. This method entails selecting a suitable array of points, overlaying it on the specimen and counting the points which fall on a the selected phase, selecting a new are and counting again, and again until the desired error is obtained. The procedure is as follows:

Selection of grid C The most efficient method will be that which requires the least effort per observation. C Using a reticule grid and observing the polished sample is more efficient than using a micrograph. C A square array of points is simple, symmetric and easy to obtain. C The optimum number of points in the grid large to minimize the effort of changing views (or micrographs) and minimum to reduce the effort of counting. Specifying the Maximum Allowable Error C Bear in mind that the effort involved varies inversely with the square of the error. C Specify the confidence level and the allowable confidence interval. C Obtain the variance from the confidence interval. Counting C Select an area on the specimen where counts are to be made. This selection should be made at random or per a predefined pattern so as to minimize bias. C Count each point which falls on the selected phase. C Assign a count value of to points which fall on interphase boundaries. C After making a few preliminary counts you can estimate the number of points that must be counted to give the desired error. This can be estimated using the following equation

(2)

where F2(VV) is the variance derived from the specified confidence interval. Tabulate the Results C Compute the mean and the confidence interval. C Divide these by the number of points in the grid and multiply them by 100. This gives the results in terms of the volume percent of phase.

Basic Statistics
Measures of Location
Arithmetic Mean - numerical average. Can be greatly distorted by a few large values. (1)

Geometric Mean - closer to the mode than the arithmetic mean (2)

Median - an equal number of items have lower and higher values. Mode - position of highest frequency. The mode of the function f(x) is x2 if the following condition is satisfied: (3)

Skew - location of the mean with respect to the mode

(4)

Measures of Dispersion
Variance - the most commonly accepted measure of dispersion. For continuous variables it is defined as (5)

while for discrete values it is defined as (6)

Standard Deviation - the square root of equations 5 and 6. It can also be written as (7)

Coefficient of Variation - Provides a means for comparing standard deviations from different measurements

(8)

Estimation of Statistical Error

The difference between the population and sample variances and standard deviations is significant only for small samples. Mean of a Random Sample - the average value of a random sample of size n can itself be considered a random variable having a characteristic distribution. The theoretical value of this parameter coincides with the mean of a population (9)

Standard Deviation of the Mean - is different from the standard deviation

(10)

Standard Error of the Mean - the error associated with sampling a population (11)

Confidence
Confident Level - percentage of values which fall within a specified range (distance from the mean) Confidence Interval - the range of values which are within the specified confidence level. It is expressed as the range about the mean. At a 67% confidence level one would say that the values of x fall within the range :x SE while at a 95% confidence level this becomes :x 2 SE and at a 99% confidence level it becomes :x 2.57 SE. 8

Frequency Functions
Normal Distribution - the type one is probably most familiar with (12)

Figure 1 Plot of a normal distribution that has its mean value at 50 and a standard deviation of 30. The cumulative distribution is shown in red.

Log-Normal Distribution - often observed in grain and particle size distributions (is skewed towards lower values of x) (13)

where Fg is the geometric standard deviation.

Figure 2 This is an example of a log-normal distribution. The median value is 9.06 and the average is 13.5.

10

Measuring Volume Fraction of Phases Exercise

Exercise 1 Measuring the Volume Fraction in a Duplex Microstructure
Figure 1 shows an artificial microstructure (generated using Adobe Photoshop) that represents a duplex microstructure. Make a grid using graph paper and use the systematic point count method to determine the volume fractions of the light and dark phases. Your results should reflect a 95% confidence level and the confidence interval should be less than 10% of the mean.

Figure 1 Artificial duplex microstructure.

11

Point Count Method

Volume Fraction Measurement Using the Point Count Method Version 2.02 SESSION Operator: Demonstration Specimen: None Date: 12/31/2003 Time: 8:48:22 PM Comments: Sample for the documentation SETTINGS Confidence Level: 95.0% Grid Size: 25 RESULTS Samples: 8 Total Count: 73.0 Total Grid Count: 200.0 Mean: 0.3650 Standard Deviation: 0.0786 Standard Error: 0.0278 Relative Error: 7.61% Confidence Interval: 0.0556 Relative Interval: 15.22% DATA Sample 1 2

Count 8.0 7.0

Sample 3 4

Count 9.0 6.0

Sample 5 6

Count 12.0 11.0

Sample 7 8

Count 9.0 11.0