19 Pages

## Introduction

This page provides resources for your written reports. There are lots of useful topics and links here and you should review them before you start writing your report. As with any form of communication, there's plenty of room for individual preference and style but the material presented here should be considered "best practices" for this course.

You can download a copy of what your completed reports should look like here. Most of the formatting will be handled automatically by ShareLaTeX, a cloud-based LaTeX service that we'll introduce in Week 2. You can also view the Written Report Rubric on the course TritonEd site to see how the reports are evaluated.

## Pre-Lab Memos

Pre-Lab questions are available on each experiment's Wiki page in the Introduction section. Your team should work together to answer the Pre-Lab questions because they contain useful concepts or analyses relevant to the experiments you'll perform in lab. One team member is to summarize your results and work in a Pre-Lab Memo; a sample memo is available here. Please note:

• The Memo is to be turned in before you start each new experiment (you know, on account of it being a Pre-lab memo). All Pre-lab memos should be paper copies and turned in to the Pre-lab Dropbox in EBUII 135.
• The memo author (one team member) is responsible for creating the Pre-lab memo. Please write neatly and clearly so that the TAs and I can quickly evaluate your responses, and include a summary page as demonstrated in the sample memo above.
• The memo author is the same team member that will provide the oral presentation at the conclusion of the experiment (see the Written Report Rubric).

## Tips and Resources

### Outlines

Outlines aren't just for crime scenes anymore; they've recently been identified as quite helpful for medium to large reports, especially when working as a team.  Although you're writing various sections independently it's important to maintain a coherent message and style throughout the report.  During at least one of your Research periods, and ideally during one of the early ones, sit down as a team and create a rough outline of the entire report so that everyone has an idea of where their sections fit into the document as a whole.  Even if you don't have any experimental data yet, you should be able to create a decent outline using just the Pre-Lab questions and Standard Operating Procedures by anticipating the equipment available, data to be collected, and analyses to be completed.

As a team, take a few minutes to review Celia Elliott's tips on The Outline [1] before you get started. They're excellent tips and succinct enough to be highly useful.

### Paragraphs

Remember that boring "topic sentence plus four supporting sentences" structure that you were taught way back when?  It's actually kind of useful!  Topic sentences not only give a sense of structure ot the reader, they're quite useful as a tool to help you outline the various sections of your report.  For example, if you prefer to use a bulleted outline form, use full sentences for the bullets and they'll naturally become, if not exactly, at least similar to, your topic sentences for each paragraph.  Again, see Celia Elliott's tips on Building Good Paragraphs [2] before getting started with your writing. Very useful!

### Tools and Resources

• The Elements of Style. The classic by Strunk and White, you should read this small tome cover-to-cover at least once a year.  You'll not only be a better writer, you'll be a better person (but mostly a better writer).
• How to Write a Scientific Paper. Excellent tips on how to do research but somewhat less than relevant unless you're good at planning your experiments prior to entering the lab.
• Textbooks. Much of the report looks quite similar to a textbook, especially the equations, figures, and tables.  If you're stuck trying to figure out what "looks good" or how something "should be" then crack open any of your textbooks and it's very likely that you'll see an example of whatever it is you're trying to do.
• Journal Articles. In much the same way textbooks can be helpful for layout or presentation, so too can journal articles.  Examples of particularly flashy or eye-catching reports (read: difficult to recreate) can be found in elite journals such as Science or Nature, and many of these articles are available free of charge on the web.  More staid but achievable examples can be found in mid-to-upper tier journals such as Angewandte Chemie, Proceedings of the National Academy of Sciences (PNAS), and the Journal of the American Chemical Society (JACS).

As implied by the last point, your references must be appropriate and websites don't count!  Use these two sites to search the wealth of peer-reviewed literature available to you through the university:

Tutorial videos are coming soon for each of these tools!

Plagiarism is absolutely unacceptable in this course and will result in at minimum a 35 pt penalty to the report and referral to the Academic Integrity Office; severe plagiarism will result in course failure and referral to your college Provost for further action.  You must do your own literature search and write in your own words!

There are five ways to use information as a reference; the first four will get you in a lot of trouble:

• Find a good article and copy one or more sentences without attribution.  This is the most blatant form of plagiarism; do this and you'll at minimum repeat the course next year.  This also includes self-plagiarism, the act of recycling work from other courses for use in this course.
• Find a good article and copy one or more sentences with attribution to a different article, presumably in the hope that you won't get caught.  We have a very special set of skills (and software) to track down plagiarism; we will find it, we will catch it, and we will act accordingly.
• Find a good article and copy one or more sentences with attribution.  This is still considered plagiarism because you haven't used your words; an extreme example would be to copy five sentences from five different sources to create a paragraph.  Sometimes this form is acceptable in the context of a literary or legal work where the exact wording is relevant but it's considered inappropriate in science and engineering (and many other fields, for that matter).
• Find a good article, slightly modify one or more sentences, and provide attribution.  Nope, still plagiarism!  If your sentence is substantially similar to the original work, it'll be flagged and we'll see it.
• Summarize a few paragraphs, or better, an entire article, in one or two of your own sentences and provide an accurate citation.  Ok, now this is the way to go!  Here, you've assimilated the information and extracted only those aspects most relevant to your own report, and provided appropriate attribution.  Perfect!

Update (2 Feb 2017): Using figures without attribution is also plagiarism and will be treated in the same manner as the forms of plagiarism listed above. If you need to use someone else's figure then include a citation to the source after the first cross-reference in the main text. For example: "The schematic in Fig. 2[4] shows...", where reference [4] is the figure source. Keep in mind that web sources, even for figures, are not appropriate in written reports.[3]

## The Report as a Whole

Your report should tell a story in a mildly repetitious manner you may have heard before: tell them what you're going to tell them, tell them, and then tell them what you told them.

Think of the various report sections as a way to funnel the reader's attention from the world at large to the experiment you performed.  Briefly, the Introduction and Background provide context, the Methods and Results describe the experiment, and the Discussion and Conclusion state what's important or what's been learned.  Each of these sections are covered in more detail below.

When you sit down as a team to create an outline, focus first on the report as a whole rather than individual sections.  Discuss the story you want to tell by answering questions such as

• Who uses this technology: industry, academia, or both?
• What competitor or alternative technologies exist?
• What are the advantages and disadvantages of this technology in terms of cost, reliability, production, etc?
• How does your experiment provide useful data or insight into this technology?
• What high-impact message do you want the reader to conclude from your report?

With answers to these questions you'll find it much easier to compose the various sections of the report and your readers will appreciate the cohesiveness of the story.

Keep in mind that a scientific or technical report is not a mystery novel!  Unless the essence of your report is to document something that didn't work, the reader isn't interested in what might be called your "beloved failures," those dead-end experiments that were helpful in guiding you to your conclusions but were ultimately secondary to the conclusion itself.  For example, did you have to restart the computer between every run to get the data acquisition software to work properly?  Did you have to calibrate your sensor?  Did you have to try six times to filter your sample?  Crushing blow to the ego that it may be, know that the reader simply doesn't care about these things and would rather you not bring it up at all.

## Report Sections

Each section of the report has a different goal or purpose.  Read these descriptions carefully so that you know how we're evaluating each section.

For each section, a hypothetical shell-and-tube heat exchanger experiment will be used to illustrate or suggest what kind of content might be appropriate.  Lacking an actual experiment and relevant data, we'll assume that we used a simple single-pass, double-pipe heat exchanger to determine the relative effectiveness of co-current and counter-current flow patterns to heat water with high pressure steam.

### Abstract

A fancy word for "summary," the Abstract gives the reader a quick overview of what's to come and is the first opportunity to "tell them what you're going to tell them."  A good abstract is about four to six sentences long and addresses the following points, succinctly summarized by Celia Elliott at the University of Illinois [4]:

1. Give the context and motivation.
2. Tell what you did in sufficient detail so the audience knows if your work is relevant.
4. Tell the reader what you think the results mean, and their implications for future work.

A consequence of this list is that the Abstract is often the very last paragraph in the entire report to be written because it's only after the Discussion and Conclusions have been finalized that the latter points are known. Be sure to review Ms. Elliott's summary on Writing Effective Abstracts [5] and take note of the five things which must never appear in an abstract! You can also look to JACS or Angewandte Chemie to see decent (or sometimes not) examples of abstracts; Nature, Science, and PNAS often have extended abstracts that are a bit longer than what you need to provide.

Regarding our hypothetical heat exchanger experiment, we might use the following concepts for the Abstract:

1. Describe why heat exchangers are useful in chemical processes (1 sentence).
2. Describe the single-pass, double-pipe heat exchanger experiment (1 sentence).
3. Summarize the key results, perhaps a relationship between the Reynolds and Nusselt numbers (1-2 sentence).
4. Place the results in the context of the literature (whether it agrees or disagrees with previous publications) and the laboratory (suggestions for the next team to do the experiment) (1-2 sentences).

### Introduction

The Introduction is the "widest" part of the report in the sense that it represents your opportunity to supply context for your experiment at the broadest level.  A good Introduction will do the following:

• Provide a description of the field in which this technology is found.  For example, heat exchangers are an integral part of nearly all chemical processes, large or small--they recover or re-use heat from reactors, drive the entire process of distillation, heat the water in your homes, and so forth.
• Provide a description of similar or alternative technologies.  The double-pipe heat exchanger is part of the broader category of shell-and-tube heat exchangers.  In addition to the great variety of shell-and-tube configurations, other configurations such as plate-and-frame or finned exchangers are also common.
• Describe the current state-of-the-art.  Is the technology developing or mature?  What challenges are yet to be overcome, and what are some of the approaches that have been attempted to overcome these challenges?  Obviously heat exchangers are a mature technology; check your textbooks--particularly recent copies of Perry's Chemical Engineers' Handbook--to read up on the field in general.

The most critical aspect of these points is that they be well-documented with reliable citations.  Consequently, you need a minimum of three unique[6] citations in the Introduction, and internet sources are not appropriate.  If you're lost, start with either your textbooks and track down the references therein, or check out the relevant Wiki page for your experiment and use those references as a starting points.  Once you have a good idea of the relevant terminology or a few references to start with, use Google Scholar or the ISI Web of Knowledge to track down appropriate support material for the statements you make in the Introduction.

For example, if we learn that plate heat exchangers are an alternative to shell-and-tube then we need to back up this statement with a reference, the most appropriate of which is a peer-reviewed journal article or textbook. After reading Section 11 in Perry's Handbook, we might summarize the description of plate heat exchangers as follows:

Plate heat exchangers consist of a series of thin, metallic sheets sandwiched together and sealed with elastomeric gaskets by compression; such exchangers find use in the food industry because they are easily dismantled for sterilization or expansion.[7]
Note that about five paragraphs have been summarized in a single sentence; a citation is obviously needed! This is appropriate when the plate heat exchanger is only a segment of a given population, which in this case is the larger population of heat exchangers.

### Background

Here, we begin to narrow the field to the specific topic of interest for the report.  If the Introduction is the "30,000 ft overview" of the field, then the Background is the "5,000 ft view" before the ground-level view is provided in the remainder of the report.  It takes a more detailed look at the specific technology of interest. Here are a few examples of the restricted scope of the Background compared to the Introduction, including our heat exchanger example:

• Experiment: Shell-and-Tube heat exchanger:
• Introduction: Heat exchanger styles and uses in industry.
• Background: Shell-and-Tube heat exchanger design and operational principles.
• Experiment: Particle separation using microfluidic flows
• Introduction: Small particle separation techniques in microfluidic devices (e.g., electrostatic, electrodynamic, magnetic, barrier, flow patterns, etc).
• Background: Design considerations of flow pattern-based particle separation including (but not limited to) types of particles that can be separated, separation principle, and device fabrication.
• Experiment: Numerical simulation of molten carbonate fuel cells
• Introduction: Types of fuel cells (hydrogen, methanol, natural gas, etc), their uses, advantages, disadvantages, and variations.
• Background: Hydrogen fuel cell principles, challenges, and the need for numerical simulation to support and facilitate experimental design.

Note that these are merely suggestions; there are many acceptable ways to decide on the content in the Introduction and Background.  For example, the "Numerical Simulation" example could equally have focused on the numerical simulation aspect rather than the fuel cell:

• Experiment: Numerical simulation of molten carbonate fuel cells
• Introduction: Numerical simulation styles, principles, software packages, applications, advantages, and disadvantages.
• Background: Molten carbonate fuel cell basics and appropriate numerical simulation requirements.

Finally, note that many sentences in the Introduction and Background will require unique citations.  For example, if the Introduction to the "Numerical Simulation" example above presents various types of fuel cells then each type of fuel cell probably requires at least a sentence and citation for each fuel cell type.  This is not as tall an order as it might seem: most topics suitable to the Introduction and Background will also be well-documented in the literature, making it easy to find reliable sources using the ISI Web of Knowledge or Google Scholar. As with the Introduction, the Background requires a minimum of three unique citations (see Footnote 5 for a comment about "unique").

The last paragraph in Background should provide a brief overview of the experiment, the tools you used for data analysis, and the type and reason for the analysis, and the major conclusions--in other words, it should be an instance where you "tell them what you're going to tell them" as a transition paragraph from Background to Theory. One convenient way to phrase this paragraph is to begin with a phrase such as "Here, we show..." or "Here, we describe..." A transition paragraph for our heat exchanger example might look like the following:

Here, we use simplified time-dependent mass and energy balances to characterize the overall heat transfer coefficient of a single-pass, double-pipe heat exchanger which used steam to heat liquid water. We demonstrate good agreement between our data and published correlations of heat transfer coefficients over all experimentally accessible operating conditions. We therefore conclude that our simplified model is accurate to within experimental error and still provides scaling behavior consistent with that predicted by more complicated models in the literature.
Naturally you'll have to work closely with your entire team to ensure the accuracy of this paragraph. Sometimes this paragraph is also more appropriate in the Introduction section but this varies from report to report.

### Theory

With Theory the goal is to present the relevant design and analysis principles to enable the reader to understand and follow the detailed Results and Discussion sections.  This presents the beginning of the "narrow" section of the report in the sense that only Theory relevant to your specific experiment should be presented. Theory is typically where equations first appear in the report, although a few equations may not be awry in the Background section (chemical reactions, for example, or scaling relations).  A good Theory section will have the following characteristics:

• Equations and variables will be appropriately and consistently used throughout the document; the Theory section should define and describe all equations and variables needed for the Results and Discussion section, excluding "support" equations such as propagation of error or linear regression. Obviously the Theory author will need to collaborate closely with the Results and Discussion author to ensure that the appropriate equations are provided, a collaboration greatly facilitated if your team sits down together to create an outline before attempting to write the report itself (hint hint).
• Equations will be integrated into sentences:
Incorrect:
The continuity equation is a statement equivalent to conservation of mass for a constant-density system.$\nabla \cdot \vec{v} = \vec{0}$ Where $$\vec{v}$$ is the velocity vector.
In this example, the equation is not integrated into the sentence because it appears immediately after a period: the equation is stranded without the context provided by a sentence. Also, the "W" in "Where" should not be capitalized. Instead, do it like this:
Correct:
The continuity equation,$\nabla \cdot \vec{v} = \vec{0},$where $$\vec{v}$$ is the velocity vector, is a statement equivalent to conservation of mass for a constant-density system.
Here the equation has been integrated properly into the sentence. See your textbooks for piles upon piles of examples if you're not sure how to use an equation in this way. All equations should be numbered and referred to with a cross-reference such as "Eq. 1" rather than "[...] the equation above." You can use LaTeX's simple \label(<labelname>) and \Cref(<labelname>) functions to accomplish this with minimal effort; see the report example in your ShareLaTeX account for examples.
• Derivations will be simplified to one or two lines, if presented at all. General equations such as the conservation equations most common "engineering" equations (e.g., the Bernoulli Equation) can be used without derivation and without citation. More specific equations, such as those derived for the cooling tower, should be presented as a result of applying various assumptions and approximations to the relevant conservation equations. In these cases, a citation to the full derivation can be provided (if you can find such a citation), or you can supply the derivation in an appendix along with a brief, parenthetical statement such as (see Appendix I for derivation) in the report itself.
• Operating principles of relevant experimental equipment or processes will be provided. In our heat exchanger example these principles might take the form of an energy balance or dimensionless number; in other experiments such as the Liposome Nanoparticle or LPCVD Simulation this might be a more qualitative description of the how the simulation or analysis equipment operates.
• Interpretation of relevant equations will be provided. As an engineer you should understand the meaning of the equations, not merely how to use them to arrive at a number. The continuity equation, for example, does not merely state "the sum of the velocity derivatives is zero when density is constant"; it means that mass is conserved in such a system. An incomplete list of quantities for which you should provide interpretation is as follows:
• Timescales. If a timescale appears in your equations, for example in an energy or mass balance, then you should provide a physical interpretation of that timescale.
• Non-dimensional quantities. Nearly all dimensionless groups can be interpreted as the ratio of two quantities (see Table 11.5-3 in BSL[8], for example) and you should provide such an interpretation.
• Adjustable coefficients. The coefficients in a PID controller, for example, have a clear physical interpretation in terms of the system's time-dependent response.
• Differential equations. Each term in an ODE or PDE represents the influence of something on something else and you should elaborate as needed. Often this or a scaling analysis is the basis for eliminating terms from general conservation equations; you can use similar reasoning to explain why some terms remain after simplification.

### Methods

There's a simple question that should guide you when writing the Methods section: How could someone with similar but not identical equipment repeat your experiment? It's not always immediately obvious what should be included and what should be omitted; you'll have to use your own judgment to decide which features of your experiment are critical to its reproduction and which are incidental. Here are a few examples:

• Critical details (include these):
• Flow patterns: The slurry was pumped from the sedimentation vessel to the reactor.
• Operating conditions: A seven-stage, trayed column was operated at atmospheric pressure.
• Critical equipment or steps: The liposome solution was extruded through a 100 nm filter.
• Measured variables: Salt concentration of batch samples was determined by conductivity probe.
• Incidental details (do not include these):
• Specific identifiers or steps: Valve 6 was used to adjust the flow rate of Stream 3. Even if a reader tries to recreate your experiment it's highly unlikely that she will choose to label her equipment in exactly the same way as yours. Instead, describe what was done rather than how it was accomplished: A ball valve was used to control the feed stream flow rate.
• Equipment-specific settings: The reboiler heater was set to 95 until the column reached steady state after 30 minutes. Unless the reader has exactly the same piece of equipment, an equipment-specific setting like this is useless. Again, describe what was done rather than how it was accomplished: The column was operated at total reflux until steady state was achieved, defined as a constant temperature profile and pressure drop across the column.
• Non-critical equipment or steps: The solution was pipetted into a 50 mL beaker with a 5 mL bulb pipette. Rule of thumb: if it's a piece of equipment you might find or use in a high school or first-year chemistry laboratory then it's probably not needed in your report.

### Results

With Results, we establish a data set on which we will base the forthcoming Discussion and Conclusions. Only rarely do raw measurements appear in the Results section; far more commonly the raw data are used as input to an analysis procedure to determine scaling behavior, investigate published or expected relations, or evaluate the effectiveness of a procedure as it relates to a particular metric. Ultimately, the choice of which data are included and excluded is a choice that your group must make based on what story you're trying to tell with your report.

The Standard Operating Procedures for most experiments are deliberately vague in terms of deliverables to eliminate some of the "recipe" aspect of the experiments that you're probably familiar with from your chemistry or physics labs: a series of steps are performed, a variable is measured, and you're supposed to get "the answer" or you did something wrong. While it's true that there are some experiments in this lab that have an "answer" in the sense that there's a particular outcome that you should achieve (for example, the Liposome Nanoparticles should indeed be nanoparticles), most of the experiments are open-ended: you're provided instructions on how to operate the equipment but the questions and relations you choose to investigate are up to you. Naturally, your team's choices will directly influence which data appear in Results.

The following points illustrate the kind of data which generally do or do not appear in the Results, but--with the exception of the "Never included" category, these should be interpreted as general guidelines rather than specific requirements: your team may choose to include something which is generally omitted if it's central to your story, or to omit something that is generally included if it is irrelevant.

• Often included:
• Scaling relations between non-dimensional groups. For excellent examples, review your fluid mechanics textbooks: nearly all correlations are provided in terms of dimensionless groups such as the Reynolds number (think friction factor, drag coefficient, etc).
• Data needed to determine experimental coefficients. Many experimental quantities of interest--rate coefficients, filter properties, heat or mass transfer coefficients--are derived or approximated by the slope or intercept of a linear plot, and such a plot should be included so that the reader can evaluate the accuracy of your data and fitting procedure. Note that error bars are critical to such plots.
• Graphical procedures. Mostly the McCabe-Thiele procedure, but sometimes graphical interpretations are helpful with other analyses such as the Merkel equation.
• Rarely included:
• Intermediate results. Consider a non-ideal flash calculation: the details of the Antoine equation and whatever non-ideal model you used are not as important as the results of the flash calculation itself (note that both details are important in the sense that the reader needs to be told which assumptions and models were used, but that's a responsibility of the Theory author). Other examples of rarely included "intermediate results" are demonstration of steady-state operation, start-up behavior, and the "beloved failures" mentioned in The Report as a Whole.
• Experimental apparatus. These images have virtually no value beyond "ooo shiny!!" and therefore should not be included. Use a process flow diagram if you need to diagram a complex system such as a distillation column, reverse osmosis unit, or heat exchanger.
• Number lists. In most cases, you're better served by a figure or table as opposed to a list of numbers. Consider the following statement:
The conversion at 25 Pa, 50 Pa, and 100 Pa was 30%, 40%, and 45%.
If the goal was to imply a relationship between pressure and conversion then a plot would have been more effective. If the goal was to present a contrast between a few numbers then a table might be more appropriate. In either case, simply listing data in a paragraph is rarely the most effective form of communications.
• Never included:
• Calibration curves. If you made a calibration curve as part of your experiment, put it in the Appendix (be sure to include the prediction intervals and error bars), never the main report. The reader assumes that if you state that concentration was measured by adsorption spectrophotometry then you were using the equipment properly, which implies an accurate calibration curve; it's not a "result" in and of itself.
• Large lists of (usually raw) data. These data have no place in the main report, and are rarely included even as an Appendix. You must retain all experimental data in case it's ever requested but it doesn't have to be included anywhere in the report.
• Only figures and tables (no supporting text). Sometimes called a "data dump," this is as useless as simply listing the raw data. See the comment below regarding combined Results and Discussion sections if you think you're light on supporting text in the Results section.
• Repeated data. If you have a plot of data then there is no need to list that data explicitly in the text. As noted above, there's rarely a need to list data explicitly in the text in any circumstance.

There are four primary methods of communicating data to the reader in a written report. It's again up to you to decide which are most appropriate for what your particular experiment:

• Figures. (includes plots, charts, and images) This is the preferred format wherever possible. See the Wiki page regarding effective figures for more information.
• Tables. A table is a good way to summarize a few numbers, perhaps half a dozen or less. Large tables are often difficult to interpret and can usually be transformed into a plot.
• Numbers. A list of numbers is better suited to a figure or table, but occasionally a single number--a regression coefficient, for example, or HETP--can be reported directly in the text.
• Text. Text alone is best used for qualitative observations which were not or could not be quantified or communicated by any of the above methods.

### Discussion

The Discussion portion of the report is your opportunity to address the last two points listed in The Report as a Whole, reproduced here because they're just that important:

• How does your experiment provide useful data or insight into this technology?
• What high-impact message do you want the reader to conclude from your report?

By far, the most useful activity you can do to aid in writing the Discussion is to plan your experiment with your team before you go in the lab! Answer the Pre-Lab questions, discuss them, think about your experiment, and start your outline before you go into the lab.

To see how the outline influences the Discussion, let's return to our double-pipe heat exchanger example. We'll assume that as part of the outline process we decided that our high-impact message was going to be an evaluation of our exchanger's overall heat transfer coefficient[9] in comparison to the expected values.[10] This decision directly informs nearly every section of the report:

• Introduction: Information regarding typical heat transfer coefficient values for different exchangers should be included.
• Background: Impact of various design parameters on heat transfer coefficient for double-pipe heat exchangers should be included.
• Theory: The overall heat transfer coefficient should be presented, probably derived, along with its various interpretations. The impact of various design parameters should show up somewhere in the appropriate equations.
• Methods: A description of the experiment, equipment, protocols, and measured variables which facilitate determination of the overall heat transfer coefficient.
• Results: Depending on what specific design parameters were listed in the Background and Theory, figures and text describing the effects of these design parameters will show up here.
• Discussion: Interpretation of results in terms of both this experiment--How statistically reliable are the results? Do the trends support those predicted by theory? Why or why not? Do your coefficients agree with published values? Are the trends similar?

Each question can easily require a paragraph or two to answer; with answers to just two or three important questions the entire Discussion is nearly complete (remember, the report length is capped at 10 pages so you've got about 2-3 for the Discussion). Note that direct answers to these questions are excellent topic sentences for Discussion paragraphs!

Remember that you're supposed to be interpreting the data, not merely presenting and describing it. Consider the following hypothetical statements in our Discussion section about double-pipe heat exchangers:

The overall heat transfer coefficient at 30 kg/s was negative, which isn't possible. The outlet hot water thermocouple was probably giving incorrect results which could produce such a result.
The second sentence in particular is about as weak an explanation as you could possible provide: "probably" and "could" imply just about zero confidence in or effort at an interpretation of the results. Replace "outlet hot water thermocouple" with any other piece of equipment and the validity of the statement (if it can be called that) is unchanged. There are actually many problems with these two sentences associated with its lack of specificity, problems which could be addressed by answering the following questions:

• Why is a negative coefficient problematic? Why isn't it possible?
• Why are the thermocouples suspect, and why the outlet hot water thermocouple specifically? Were any other sources investigated? Why were they dismissed?
• How could an erroneous thermocouple reading produce a negative coefficient? How large a departure from reality is necessary to produce the observed result? Is such a magnitude likely? What other effects would this produce?
• Was there something unique about the measurement for 30 kg/s? Presumably other flow rates were measured and the implication is that they're ok but this one is not; why?

These are the kinds of interpretations and explanations you're supposed to provide in the Discussion section. Obviously, you'll want to discuss with your team before committing anything to paper: the engineering analysis and judgment displayed here (and indeed throughout the report) will reflect on the group score, not just the individual author score.

#### Regarding Combined Results and Discussion Sections

Sometimes it's more effective to combine the Results and Discussion sections into a single section. In this way a particular result can be presented (usually in the form of a figure or table), then immediately discussed and interpreted. This could be the case if your results are fairly straightforward and don't need a great number of supporting or descriptive paragraphs, or if the results and interpretation of one experiment directly and sequentially inform the results and discussion of another experiment. If you'd like to combine the two sections then you're welcome to do so; simply re-label the "Results" header as "Results and Discussion" and remove the "Discussion" header.

### Conclusions

And finally we come to the last part of the report, the section where you "tell the reader what you told them." In about a paragraph or two the Conclusions should provide an accurate summary of the relevant observations noted in the Discussion (trends, metrics, etc.) and, importantly, what your team can conclude from these observations.

For example, if the overall heat transfer coefficient of our double-pipe heat exchanger was outside the expected values, perhaps we would conclude that the thermocouple was indeed faulty and therefore the experiment must be repeated after repairing or replacing the thermocouple. Or perhaps we observed that the effect of flow rate on the overall heat transfer coefficient agreed with the predicted behavior, from which we could conclude that our simplified model was an accurate representation of the equipment.

Lastly, you should include a recommendation for future avenues of research. As offered in the example above, an obvious recommendation is to repair faulty equipment; if you offer this recommendation it must be made obvious in the Discussion section that said equipment is indeed the most likely source of whatever error you seek to remedy. Even if the experiment went swimmingly, there are always sources of error that can be further minimized by better protocols or advanced analyses; having recently completed the experiment and analysis, you're in an excellent position to provide such advice.

## References

1. Elliott, C.M. Writing Made Easier - The Outline. http://people.physics.illinois.edu/Celia/Lectures/Outlines_Spr13.pdf (Accessed Dec 30 2015).
2. Elliott, C.M. Building Good Paragraphs. http://people.physics.illinois.edu/Celia/Lectures/Paragraphs.pdf (Accessed Dec 30 2015).
3. There are two exceptions to the "no web sources" rule: LPCVD written reports can reference COMSOL documents, and you can use web sources for images in oral presentations.
4. Elliott, C.M. Writing Effective Abstracts. http://people.physics.illinois.edu/Celia/Abstracts.pdf (accessed Dec 30, 2015).
5. ibid.
6. Citing three different sections of a book does not count as three unique citations, only as one.
7. Shilling, R. et al. Heat-Transfer Equipment. In Chemical Engineers' Handbook, Perry, R., Chilton, C., Eds., 8th ed.; McGraw Hill: New York, 2008, pp 11.54-11-55.
8. Bird, R.B.; Stewart, W.E.;, Lightfoot, E.N., Transport Phenomena, 2nd ed.; John Wiley and Sons: New York, 2007, p 356.
9. Although not without its usefulness, there are better ways to evaluate heat exchanger performance than the overall heat transfer coefficient, namely by using dimensionless quantities.
10. McCabe, W.L.; Smith, J.C.; Harriott, P., Unit Operations of Chemical Engineering, 7th ed.; McGraw Hill: New York, 2005, p 343.
Community content is available under CC-BY-SA unless otherwise noted.