Moving early undergraduate chemistry labs online

The last post discussed an advanced physical chemistry lab, and in this one I want to summarise more concrete plans for how we can move an early undergraduate lab online.

The key thing for us at early undergrad stage is teaching chemical technique, and getting students to think about recording data and drawing conclusions from experiments. An important factor is that at this stage, students probably expect that they will be learning about technique. Coming into university from school, their perception will be that they want to learn about chemical techniques, and lots of them.

A problem I have summarised in previous posts as the “swipey-wipe” effect rears its head here. A typical online lab involves students doing a simulation of sorts, getting information from that simulation, and writing up the report. While this likely has some benefit in terms of data processing, I am not sure how much it teaches about experimental technique, because a lot of simulations don’t reflect the tangible reality of laboratory work.

Anyway with resolution rather than confusion on my mind for this post, I’m going to get straight to the template envisaged. This is heavily influenced by Joi Walker’s poster presented at CLEAR20 last week, as well as Lukas Kroos/Nicole Graulich’s poster on decision-making in experimental procedures – using video of those procedures (both posters available on the CLEAR website). All academic credit goes to the work reported by those researchers, and I have linked to papers by Joi below.

Joi Walker’s work is based on getting students to make an academic argument and reason it with evidence. So instead of getting students to play with a simulation and generate data off the bat, I am proposing that we have a lab template that:

  1. Presents students with some data from a described experiment. There should be sufficient data there for students to make a claim.
  2. Students’ first task is to make a claim, based on the evidence presented. If you zoom in really close on Joi’s poster, you’ll see a whiteboard there on how she structures that (what is the guiding question; what is the claim, what is the argument presented; what is the evidence for that argument).
  3. Once students have made their claim, their next task is to decide what further evidence they need to support the claim. It is at this point where we might unleash a simulation on students. I personally prefer video (or at least photos of the real set-up) and Lukas’ poster from CLEAR shows a really exciting way of making that interactive – I am greedy for more on that work.
  4. We somehow provide students with data based on their required further experiment. Students need to use the videos provided to describe the experimental procedure – how they would do the experiment in reality if they were in the lab. This is a way to get students to meaningfully engage with any procedure videos etc we share.
  5. Students add in the data from the “experiment” with the data they were provided, and review their claim. Their discussion outlines whether their claim was supported or not by the additional evidence.
  6. The writing up and presentation of the work needs some further thought (by me). Joi shared some lovely peer-review work, and Marcy Towns, who spoke at CLEAR, outlined a similar approach based around “Claim-Evidence-Reasoning”, which involved some peer review work (recording of that presentation is online next week).

I think this approach ticks a lot of boxes for me. It removes the “game” aspect of doing a simulation to get some data just for the sake of it, and instead turns it into a meaningful activity, with some in-depth considerations by students of experimental procedure, but also a taster about experimental design, and in making judgements on data. It’s also feasible. (Of course after teaching students online, once we get them back in the laboratory, we will be cram-packing in intensive laboratory competencies.)

Example in practice

To do a run through feasibility, I am thinking of some of our early year experiments. Obviously this more suited to general/physical chemistry labs, but one I had in mind was the typical iodine clock kinetics experiment. We could share some data based on some initial concentrations, which could allow students to either deduce the order or reaction, or require just one more piece of the jigsaw. They could make a claim and seek further data to either confirm the order or get the additional data. It would be easy to integrate existing videos we have about the iodine clock, and easy to auto-generate data based on student requests.

The reporting and assessment side of things needs a bit more thought, but I’m certainly happier about this kind of level of laboratory than I am when I finished writing the advanced lab post! In fact it really appeals as it means we can use this as a process to review our current early year lab offering and improve them to include argumentation for future iterations in person or online.

Check out Joi Walker’s publications on ADI:

Sampson, V., Grooms, J., & Walker, J. P. (2011). Argument‐Driven Inquiry as a way to help students learn how to participate in scientific argumentation and craft written arguments: An exploratory study. Science Education95(2), 217-257.

Walker, J. P., Sampson, V., & Zimmerman, C. O. (2011). Argument-driven inquiry: An introduction to a new instructional model for use in undergraduate chemistry labs. Journal of Chemical Education88(8), 1048-1056.

Moving a (physical) chemistry lab online

The last post discussed some epistemological considerations (roll with it) on moving chemistry labs online, sharing some concerns about trying to teach technique via fancy swipey-wipe interactions (roll with it).This one aims to be a bit more grounded. If we were to move a lab online, what might it look like? I am going to go through my first draft of thinking for moving a physical chemistry lab online below. The headline considerations for me are:

(1) not to create busy work for the student for the sake of it;

(2) some bits of lab work aren’t really that great, so let’s not try to just replicate; and in complement,

(3) when we are all back in the actual lab again sometime in 2027, maybe there are some bits of this online malarkey that would actually be quite useful to keep in place.

In other words, my focus is not on replication, but on looking to build a decent learning experience.

Let’s flash!

My prototype experiment is a pretty standard one on flash photolysis of azobenzenes, built on the experiments described nicely in this J Chem Ed paper. We use this as a basis for an experiment in our third year, and this is the year we focus on teaching experimental design, using our “unfinished recipe” approach. That is to say, students do a recipe lab to get used to the experiment, the kinds of experimental protocols, and data outputs, and then follow this up with a non-trivial investigation based on the kinds of things they learned in the first part. Our “Part 2” is based on the recent paper described here; it asks students to explore pH effects, and if they are really good, to go hunting for a transient.

I like this experiment because it is based on photochemistry – enough said? But also it is a good one for students to pull a lot of what they already know – kinetic analysis, simple techniques (UV/vis in kinetics mode), cis­-trans isomerisation, etc – and apply it to a new analysis (for them). The experimental technique, while reasonably straight-forward, requires a bit of black art at times, and you see students who will get really good traces and those who get not so great ones, just based on some timings when they actually do the experiment. I want students to get things like considering timelines for experimental measurement (my homage to Porter), working with trickier kinetic decay curves, and making reasonable judgements about mechanisms from their data. For students who want to push a little more, hunting for the transient based on the information from the Larsen paper is a good trip to go on.

Moving online

Now I know at some point I should write down all the learning outcomes I want but I am not going to do that because it is just too much of a pain. Instead I want to frame my online lab in terms of technical competence and syntactical knowledge. Just a little more detail on what I mean by those are in the previous post, but it will become apparent below if you can’t be bothered to go back to that.

1. Technical competency

As mentioned I don’t really see much point in going down the road of trying to recreate a simulation of the experiment online. They do exist for flash photolysis, and again I make no judgement on these, but my own purpose here is to take students through an actual experiment and understand how they would complete a task to obtain the data they will be presented with. A cost of moving online necessitates, I believe, an acceptance that we cannot teach experimental technique as well as we could do in person. So what can we do? We do have videos of our experimental techniques, made in our labs. At the very least, we could show these videos?

This seems a bit unsatisfactory, and very passive. I want students to really think through the procedure; live through the considerations that they might have to make in practice. Key ones here are the solution concentration (a desirable absorbance of < 0.2 at max) and the kerfuffle of getting a flash (we use an old camera flash) and recording the subsequent decay. The first is reasonably routine – we could provide students with an absorbance spectrum of the stock and ask them what they wanted to do with it; the second is impossible to role-play – simulation or not. Therefore I am thinking of getting students to critique the procedure as presented in the video. This seems like a sensible way to get them to really think through it. Our procedure could be improved, no doubt, in parts, and a nice task for students would be to talk about how they might do that. This is an example of how I think we might actually improve the status quo, and this critique can feed into some discussion later (for example how different events in this photochemical process might be studied).

2. Syntactical knowledge

The complex learning framework that I use to guide my own thinking on labs aims to describe a lab environment in terms of what students need to draw together when completing laboratory work – the combination of, in this case, knowledge of kinetics and analysis, understanding of photochemistry principles and timelines, knowledge of using the UV/vis spectrometer and application of the combination to address a laboratory-based problem. In simple terms, I do not want to use this lab to teach kinetics, but want students to bring their kinetics knowledge to this environment and apply it alongside their other relevant knowledge and experimental skills to address a laboratory-based problem.

How do we move this online?

I think to answer this I need to write out what the component bits are in each stage of the experiment, so that I can see what it is students are “drawing together”.

In part 1 of the experiment, the focus is really on training about the approach:

  • Running each of the solvents in turn (toluene, cyclohexane, acetone), including one at a range of temperatures and extracting kinetic data;
  • Curve-fitting to obtain rate constants, and in one case Arrhenius parameters;
  • Data presentation and interpretation (in terms of mechanism)

This is all pretty routine, but drawing together in an online version means, along with the technical aspects described above, students will need to be given data to run through the processing, and present some summary findings about what this means in terms of the mechanism (essentially a discussion based on polarity). Do we simulate this data based on some student prompts, or do we just give them some data sets? I don’t know yet, but I would like the former. I’d like them to think also about data quality and what might be affecting it.

In part 2 of the experiment, students have to design a pH study of a water-soluble azobenzene, and if they wish, go transient-huntin’ (unlike the Larsen article, we generate point-by-point spectra). This means they need to plan out things like:

  • Deciding on what pH range to study and how to prepare those concentrations ((un)surprisingly this is a pretty major challenge for students);
  • Repeating the analysis for solutions at different pH
  • Exploring different wavelengths beyond the bleaching peak to see if they can find transient
  • Comparing transient and bleaching kinetics

This second part is much more student-driven, and takes them a lot of time in the lab normally just to even scope it out. So in moving this online, I’d like to replicate that – we really want them to think through experimental design. Online, this might need more guidance and dialogue stages; prompt first thoughts on the overall scope – what will the purpose of a pH study be based on what they know from the first part of the experiment (polarity dependence); how could they go about making a study; what are the technical aspects of that (pHs, concentrations, etc); how might they find transients; what would a comparison of transients mean. I can see this working out quite nicely (in my head!) but wonder about how that dialogue would be managed with ~150 students so that each gets to have their own input into “their” experiment. That consideration (workload of demonstrators aside, albeit hugely significant in terms of what can actually happen) is why I think simulated data would be good.


Another significant issue is that of assessment. We could of course continue with the major report we expect of students from these two-part experiments – they are beneficial in preparing students for later pieces of major writing such as their thesis. A concern is that if we need to engage students a lot more in dialogue on the way, they will already be putting quite a lot of work into this – so we need to capture that in what would previously have been a lab mark. Also, I think I would like to hear from students as well; especially in relation to their understanding of the technique and how it works. Perhaps there is scope for that, through an online presentation, and that this would, in some small way, make up for the lack of completion of the experiment. “Category is: explain flash photolysis with 5 pieces of Lego…” Now that would be an improvement on the status quo!

Oh dear – so much to do… And this is only one experiment!

What is an “online chemistry lab”?


The massive shift to online teaching and learning for us in Edinburgh focussed on lectures and tutorials, as we were comparatively lucky in timing – our semester starts and ends very early in the calendar year, which meant that our students were able to complete the majority of their labs. Knowing what I have lived through over the last weeks, I empathise with educators from the other side of the world who are just beginning their teaching year, and those teaching summer semesters: I realise how lucky I am! I mention this as a prelude as I want to discuss what “online labs” are with the luxury of time on my side, but in no way mean to diminish efforts being made by people under extreme time pressures – everyone is doing what they can to fix their local situation. This post is mostly useless as it raises more questions than it answers.

Why labs at all?

Before thinking about what a lab might look like online, we perhaps need to think about why we have labs at all. The literature is really unhelpful in this regard, as there are an array of lists of learning outcomes in labs, but they mostly come from the position that labs happen in the first place, and therefore list what kind of learning can be extracted from them. The place of labs to learn a whole host of non-lab-specific transferable skills is a case in point.

For me there are two and a half reasons why we have labs.

Firstly, labs are in the curriculum for cultural reasons – you couldn’t be a chemist and not have labs. Since von Liebig and perhaps before, labs are an integral part of chemists’ identity.

The second reason is most important to me – labs must offer something that can’t otherwise be learned. So we can learn how to do titration calculations and plot a Beer Lambert law in lectures – what is the lab offering? More generally, we can know all of this chemistry theory, but how do we enact it. For me labs have a distinct pedagogical offering because we can learn techniques, and we can use our knowledge of those techniques, along with a knowledge of chemistry, and pull it together to do something tangible in practice (or on a computer if you are a computational chemist – I am all-inclusive!). This distinction between “knowing” and “doing” is described in the literature as substantive knowledge and syntactical knowledge. This is why “recipe labs” often get a lot of criticism; they are trying to teach technique (poorly, by osmosis) and don’t really require any intellectual input, so they are a poor use of resource and brain capacity.

The half reason is accreditation – we do labs as we “have to” do labs. To be honest, I’m not really sure a University like mine not being accredited would make a difference; of course we love having it, and it is a nice external “check” on what we are doing. But in many ways accreditation is just a reflection on the cultural aspects of chemistry and a (hugely misunderstood) sense of the importance of labs pedagogically, so if the zeitgeist moved away from doing labs in say Edinburgh, Cambridge, Oxford and Durham, I’m not sure it would be an accreditation requirement for much longer. That is to say: accreditation isn’t a reason to have labs.

Moving online – teaching technique

This consideration is important, I think, when we move our labs online. In terms of teaching technique, I worry that with easy access to computer software we’ll try to replicate somehow the in-lab activities, and there will be digital titrations and online spectroscopy before you can blink. (I’m pointing the finger at myself here and here, both from 2011). But even the simulations I have seen which are good, it takes two clicks and a swipey-wipe to get some water into my beaker.  These kind of actions are not associated with the actions in reality. I don’t click to pour water in reality, I hold a beaker, and pour at just the right rate so that the liquid goes into a narrow neck volumetric flask. There is no big red button on any UV/vis spectrometer I have seen that checks the solution for clarity, places the cuvette in the instrument in the right direction and chooses the right wavelength range. So one has to ask (politely) – what benefit is there from the substantial and clever development work going on to create these activities?

I think teaching technique online is going to be really difficult actually, and so the efforts mentioned above are really admirable. I am learning some more advanced aspects of the Adobe CS suite at the moment, and YouTube is my best friend. I look at videos demonstrating something, and then play those videos bit by bit as I do the action, and then I try the action myself. It takes a few goes, because I am not very good and have a terrible memory. But over time, I learn how to do the technique. But I had to practice with reality (which in this case, is easier as reality is also on a computer). How can we ever mimic this for chemistry technique? And an additional note, how do we even begin to mimic the learned caution and care required, to be a safe and competent chemist. Even the masters of online learning, Open University, used to run residential summer schools, although in changing times, these sadly went online too. This is the first thing I am grappling with. (Virtual reality (VR) is an exciting area of development; I have no sense yet whether VR can replicate all of the requirements of teaching technique, but is probably at least an improvement on two dimensions.)

Moving online – syntactical knowledge

We want to teach our chemists to be chemists, not cooks, and being a chemist means having requisite practical techniques to hand and being to use them in situations of your own design. Teaching experimental design is really tough – to teach and to learn – and we’ve invested a lot of effort into it. It involves decision making – both trivial (what solution concentration should I make?) and complex (what is the scope of this experiment and what will it tell me?). Oddly, free from the concerns about technique, this may be easier to replicate online, although I don’t think as a self-contained package (without a lot of effort and perhaps some artificial intelligence), but more as a kind of lab-tutorial, involving ongoing dialogue and lots and lots of decisions for the students to make. Missing from the piece will be the actual “doing” of the experimental work, which is not trivial. If I prepare a solution concentration that’s too high so that I get a bit of precipitate that will distort the spectrum, what do I do in practice? Did I hear a hiss of gas at some stage during a synthetic procedure that may have implications for the reaction? It is hard to replicate those kind of considerations in a mock version online, where you are really relying on the proto-chemist to bring together chemical and technical knowledge, in what I have framed as a “complex learning environment”.

But I see some hope here, certainly for analytical and physical labs, and much of my own thoughts are focussed on this piece and thinking about how we could move at least some of this syntactical knowledge online. I welcome any thoughts and advice!

Links to source

Managing information from student perspective during COVID contingencies

My colleague Chris Mowat and I co-wrote a blog post for our university’s Teaching Matters blog, which I am linking here as it pulls together a lot of our activities into one post, so might be useful for people in a hurry!

The end of academic year is a busy time for students in normal circumstances, but this year, in addition to spending this time reviewing all of their courses and getting ready for assessment, there is a swell in information relating to managing alternative arrangements due to COVID-19 that our honours students have to process. As a combined effort, we have been working to mediate this information in a way that is clear, coordinated, and student-focussed. We share below five headline strategies that are working well for us, what we are doing next, and reflect on how the last few weeks have been. [Link to Teaching Matters blog post to continue reading]

(Graphic courtesy of @ScienceNerdSACR and used with permission) - see post linked for details
(Graphic courtesy of @ScienceNerdSACR and used with permission) – see post linked for details

Managing the open-book exam process

At Edinburgh, we are mostly moving to replacing our 3 hour exams with open book exams. We had initially intended these to be within 24 hour timeframes, but the University has mandated 48. Otherwise, things are as described in the previous post.

So students will need to access an exam paper from a  specific “start time” and submit their written answers no later than 48 hours than that start time. Easy!

Exam Process Guide

Having looked through the various options, I am going with the following 5 step plan based on Blackboard Assignments (rather than Turnitin), described below in terms of front of house and behind the scenes:

For students:

  1. Students will be able to download the paper in advance from Blackboard (called “Learn” here), but it is password protected.
  2. At the start time, the password is released on Blackboard and by email. We didn’t want 150 students trying to download a PDF at the same time.
  3. Students complete their answers on paper.
  4. Students scan their answers using Adobe Scan app to create a PDF. For us, one lecturer corrects one exam question, so we want these uploaded on a question by question basis.
  5. Students upload their answers.

I’ve made a draft video (subtitled) outlining this process, which we are testing robustly this week! We are then going to release to students and allow them play with a mock set-up.

Behind the scenes, this means:

  • Setting timings so that Blackboard courses release in line with exam timetable, and components (e.g. password, answer submission areas) release at the right time
  • Enabling anonymous marking so that the student number doesn’t appear in the file name
  • Allowing multiple submissions so that when students upload the wrong file (it will happen) they can submit the correct file – this needs careful management post hoc.
  • After time window has closed, each question will be downloaded and shared with examiners. I am not going to ask my colleagues to annotate files in any way; they will simply keep a mark tally per exam number related to marking scheme so that they can refer back to that if there are queries.
  • Marks can then be returned in an Excel sheet by exam number, and these can go into “the system” for exam boards.

EASY! 🙂 What can go wrong? (no, really…?)