Poster Experimental Design Pittcon 2011-Carta

Download as pdf or txt
Download as pdf or txt
You are on page 1of 1

Poster No.

700-10P

Spreadsheet Does Experimental Design for Chemist


Enrique Arce-Medina ESIQIE, Instituto Politcnico Nacional, Edif. 7,Unid. Prof. A.L.M., Mxico 07738 D.F., Mx., [email protected].

Abstract This poster presents how Design of Experiments (DOE) was introduced to students, throughout the use of a spreadsheet, using the Problem Based Learning approach of instruction. By using the template, which students build by themselves, they understand basic concepts and computational aspects of DOE more easily. Answer from students, using written questionnaires, show that some resist to adapt to the active learning approach. Reasons for these are analyzed.

Students can see how DOE directly provides an equation for expected response and, might indicate the factors significance (all factors that have F-ratio > F-critical are significant.) Figure 1 shows an screen shot of a typical spreadsheet developed by a team of students.
1 2 3 4 5 A B C D E F G H I J K L Design is a 16 run Full Factorial Design, allowing to study 4 factors, each at two levels, and all of their 6 interactions
m , exp levels 2 Factor Actual values n, num of factors 4 N, num of trials 16 r, num of replicants 2

Introduction
We are convinced that learner centered teaching is in our students best interest. It is not easy to select and design the learning task we are asking our students to take on, which require them to adopt new learning roles and responsibilities. Students adaptation to new situations takes time. Many students are not ready nor motivated to learning on their own. When students can figure out how to find and use knowledge in meaningful ways to solve real problems, they become independent learners, and is our job to provide opportunities to do so. Self learning is and ability most students lack, another is learning to work with others. An active learning approach that has rapidly gained acceptance by the educational community is the Problem Based Learning (PBL). The didactic tactics of PBL follow a learner-centered style of learning that encourage students to assume responsibility for their learning experience and to shift from passive to active learning patterns. The PBL approach has been implemented in this course in several ways. For example, in assignments to take home, a set of problems comes first, before any type of instruction related to the problem topics takes place. Students learn while searching for knowledge and methods needed to solve the problems. Then after the students hand in the assignments, a discussion session is facilitated in the classroom. Interactive questioning and answering to solve problems, during these sessions, has many similarities with the Socratic Method. The purpose is to summarize what has been studied through the problems that have been assigned for homework. Usually, after the discussion session, a mini-lecture is given by the instructor using Power Point presentations. And also to clarify any misunderstanding of the concepts related to the problems. Another example of PBL tactics used in this course is the assigning to small groups of 3 or 4 students, problems and exercises that the students have to solve during the class session. The students compete to see which group can arrive at the solution first. For the teams that get the correct answer to problems and exercises in the allowed time, get an awarding bonus point. In this way learning is partially constructed, when students build their own interpretations, through problem solving. Using EXCEL in support of PBL This poster presents how Design of Experiments (DOE) was introduced to students throughout the use of a spreadsheet using the PBL approach of instruction. Design of experiments (DOE) is a powerful tool that can be used to reduce the quantity of experimental work required, to improve the precision in the parameter estimation, and to achieve product improvements. Two-level factorial design is a statistically-based method that involves simultaneous adjustment of experimental factors at only two levels: high and low. Two-factor interaction analysis can be constructed with the aid of a canned statistical software, or better yet, with a spreadsheet. First, students work in pairs to implement DOE calculations in a two-factor example taken from a text book. Secondly, the problem is solved and the theory is revised in the classroom in the date due of the assignment. Then a four-factor DOE problem is solved using EXCEL at the computer center with students working in teams, as a case study.

Level 1Level 2

120 130 C tem perature A -1 1 -1 1 -1 1 -1 1 -1 1 -1 1 -1 1 -1 1 3.909375 1.9546875 1 13.94500947

1.5 3.5 Reactants ratio B -1 -1 1 1 -1 -1 1 1 -1 -1 1 1 -1 -1 1 1 3.396875 1.6984375 1 10.528422

0.98 1.15 catalyst concentr C -1 -1 -1 -1 1 1 1 1 -1 -1 -1 -1 1 1 1 1 2.803125

1hr reaction tim e D -1 -1 -1 -1 -1 -1 -1 -1 1 1 1 1 1 1 1 1

3 Yield response 1 A*B 1 -1 -1 1 1 -1 -1 1 1 -1 -1 1 1 -1 -1 1 A*C 1 -1 1 -1 -1 1 -1 1 1 -1 1 -1 -1 1 -1 1 A*D 1 -1 1 -1 1 -1 1 -1 -1 1 -1 1 -1 1 -1 1 B*C 1 1 -1 -1 -1 -1 1 1 1 1 -1 -1 -1 -1 1 1 0.04063 0.02031 1 0.00151 B*D 1 1 -1 -1 1 1 -1 -1 -1 -1 1 1 -1 -1 1 1 -2.7969 -1.3984 125.16 1 125.16 C*D 1 1 1 1 -1 -1 -1 -1 -1 -1 -1 -1 1 1 1 1 0.74688 0.37344 Grand m ean 8.92516 1 8.92516 68.221875 SSe MSe error-dof Sstrial 87.677 17.5354 5 1171.46 1259.14 y1 49.5 60.1 65.1 67.1 45.8 68.8 65.9 76.9 68.4 69.3 67 73 70.1 81.9 74.9 73.5 Yield response 2 y2 51.5 64.7 67 68.3 49 70.8 68.8 80.2 66.2 66.8 68.9 73.8 71.1 83.2 79.8 75.7 ybar 50.5 62.4 66.05 67.7 47.4 69.8 67.35 78.55 67.3 68.05 67.95 73.4 70.6 82.55 77.35 74.6

6 run# m ean 7 1 8 1 1 9 2 1 10 3 1 11 4 1 12 5 1 13 6 1 14 7 1 15 8 1 16 9 1 17 10 1 18 11 1 19 12 1 20 13 1 21 14 1 22 15 1 23 16 1 24 Contrast 68.221875 25 Effects 26 SS 27 dof 28 MS 29 F-ratio 30 Pr(F) 31 32 F-critical 4.49

4.503125 -1.96563 1.44063 -1.984

1.4015625 2.251563 -0.98281 0.72031 -0.992 1 1 1 1 1 3.593

244.5314063 184.620156 125.72016 324.4502 61.8189 33.2064 63.004 0.02641 244.5314063 184.620156 125.72016 324.4502 61.8189 33.2064 63.004 0.02641 7.1695035 18.50257 3.52538 1.89368 0.013510803 0.02282843 0.0439467 0.007705 0.11924 0.22723 0.1165 0.97055

7.13757 0.50898

0.04426 0.50749 total sum of squares

Figure 1. A screen shot of a typical spreadsheet developed by students of the design of experiment for the case study.

Lessons learned
Student feedback is obtained through students questionnaires, informal interviews and e-mail responses to a direct request for information. Responses get polarized; answer from students with high levels of academic achievements reflected that they are ready to take a more pro-active role in their formation, and appreciate the problem based learning instruction as very important for their education. These students, which constitute the majority, expressed positive opinions about the teaching approach adopted in his course. On the negative side, contrary to the highly motivated students there are few low achieving students, which objected to the introduction of the learner-centered activities. These students are not ready to take responsibility for self-learning. Characteristics of these unmotivated students include: (a) complaining of the extra workload required, they would prefer the professor give them all the theory and information they need; (b) disagreement with the mini-lectures on which the Power Point is used, and (c) wanting to be given a lot of examples and asking for more classroom discussion before and after homework assignments have been completed. Transforming a passive learner into an active one is not easy. Students that are over-dependent on teachers, which are not used to take responsibility for their self-directed learning, resist using the active learning methods.

Case study
The synthesis of hesylresorcinol was used as case study to determine the optimal conditions of reaction temperature, reactants ratio, reaction time, and catalyst level, to maximize yield. Students save a lot of time by letting Excel take care of the calculations of effects and residuals.

Conclusion The DOE is entered in a spreadsheet template in a 24 design matrix where columns represent data for the control It has been found out that Excel is a simple and efficient tool to introduce students to DOE. By using the template, which students build by factors and response at each experiment. Each column contains 8 pluses and 8 minuses to ensure that factors are not themselves, they understand basic concepts and computational aspects of DOE more easily. correlated. The effect of each factor is calculated by simply averaging the responses at the plus level and subtracting There is no doubt, preparing for active learning requires much more work from both instructors and students than traditional teaching. Teachers the average at the minus levels. An ANOVA analysis was implemented in the spreadsheet to pick significant factors. must have to adopt the role of facilitator of the students learning process, while students have to be more responsible of their own learning.

You might also like