Teaching Instrumental Analysis during the Pandemic: Application of Handheld CO2 Monitors to Explore COVID-19 Transmission Risks.

The COVID-19 pandemic has posed a challenge for maintaining an engaging learning environment while using remote laboratory formats. In this work, we describe a Student Choice Project (SCP) in an undergraduate instrumental analysis course that was adapted for remote learning without sacrificing research-based learning goals. We discuss the implementation and assessment of this SCP, selected student results, and student feedback. Students were provided handheld carbon dioxide monitors and charged with designing and implementing an investigation centered on COVID-19 airborne transmission. The real-time monitors provided experience with a new analytical tool that demanded considerations and analysis not common to other methods discussed in the course. Students were motivated by the ability to design their own projects and by the real-world implications of their findings. They performed well for all assessments, reported a positive experience, and recommended these monitors be added to the typical repertoire of instrumentation for the course.

Introduction

Following the global spread of the coronavirus disease (COVID-19) in early 2020, education transitioned to remote learning to limit infections, forcing adaptations from both instructors and students and sparking discussions about the benefits and limitations of various approaches.[1−5] The common challenge of retaining student interest and engagement in the course material was exacerbated by the practical hurdles presented by the pandemic, particularly for inherently hands-on laboratory work which was transitioned to a remote format. In some cases, laboratories were replaced by recordings[1,2] or virtual simulations,[6−8] while others designed at-home experiments.[9,10] Thoughtful implementation of these nontraditional laboratory experiences can effectively engage students and achieve learning outcomes.[11]

In a typical instrumental analysis course, learning goals center on experimental development and problem solving.[12] Upper division chemistry students at the University of Colorado, Boulder take a two-semester instrumental analysis course, culminating in a nine-week Student Choice Project (SCP) to foster research skills. Similar, student-led, course-based undergraduate research experiences have reported increased student engagement and improved learning outcomes.[13−16] Typically, small groups of two or three students establish a research question and hypothesis related to their interests, design and execute an investigation using available analytical techniques (e.g., flame atomic absorption spectroscopy, gas chromatography tandem mass spectrometry, etc.), and present the results. In Spring 2021, restricted in-person access to the analytical laboratory and instrumentation limited the students’ SCP options. We adapted the SCP to use low-cost monitors, which others have used in undergraduate learning outside the laboratory.[14,17,18]

There is mounting evidence for airborne transmission of COVID-19 which involves inhalation of virus-containing aerosol particles exhaled by infected individuals.[19−22] Direct measurements of virus-laden particles is beyond the capabilities of typical instrumental analysis courses, and human-emitted aerosols are difficult to isolate from others, e.g., dust, pollen, vehicular emissions, and indoor emissions. Instead, it has been shown that carbon dioxide (CO2) can be a useful proxy for exhaled human breath and thus relative exposure risk in indoor environments.[23−25] There are limitations with this approach for students to consider: other sources of CO2 exist, and masks and filters reduce exposure to aerosol[26−28] but not CO2.

Here, we present the use of commercially available, easy-to-use, low-cost, handheld CO2 monitors for estimating COVID-19 exposure. We provide an overview of the project and summarize the course materials used to prepare the students. Results of selected student-designed investigations are provided to demonstrate the creative ways students approached the project. Student assessment and feedback demonstrate the effectiveness of this SCP format, and we make recommendations for improved implementation. This approach is a practical solution to remote-learning and is cost-effective for programs that lack instrumentation funding.

Materials and Methods

Course Materials and Resources

The SCP and timeline were briefly introduced to students at the beginning of the Spring 2021 semester. Starting week four, applicable lectures (3, 50 min; slides in the Supporting Information) covered the following:

COVID-19 transmission;

Aerosol size, lifetimes, and emissions;

CO2 from human breath;

CO2 as a proxy for virus-containing particles;

Ventilation and filtration.

Throughout these lectures, students were asked to consider a few hypothetical situations, e.g., riding the bus or attending lecture in person, and determine which has the highest risk for COVID-19 transmission. This poll, which has no clear answer, probed students’ developing understanding of different variables that impact COVID-19 transmission such as indoor space volume, occupancy, and the rate at which indoor air is exchanged with fresh air (air exchange rate, AER). Additionally, students were assigned a homework problem: a box model of CO2 emissions in a home with ventilation (detailed in the Supporting Information).

The COVID-19 Aerosol Transmission Estimator (ATE) developed by Prof. Jimenez[29] was provided as a resource. This spreadsheet employs current literature and models to estimate airborne transmission of COVID-19 and explore trends.[30] For the SCP, students updated values within the calculator, e.g., occupancy or AER, to match their observations and assess transmission risk. This calculator also estimates average CO2 volume mixing ratios (VMRs), allowing for comparisons to measurements and further assessment of risk.

Ideally, the AER was determined experimentally by measuring the temporal decay of CO2 after the removal of all sources, such as measuring overnight. Fitting the decay to an exponential function yields the AER as the exponential variable. Where infeasible (e.g., in a grocery store), students were encouraged to search the literature for representative values. The ATE uses some such values in example cases.

Groups were provided a Temtop M2000 Second Generation Air Quality Monitor (Elitech; ∼$200 USD each) which measures CO2, formaldehyde, and particulate matter with diameters less than 2.5 μm (PM2.5) and 10 μm (PM10) and records data down to 1 min intervals. Some projects benefited from a second monitor. CO2 is measured with a nondispersive infrared sensor (0–5000 ppmv range with ±50 ppmv, +5% accuracy). Formaldehyde is measured with an electrochemical sensor (0–5 mg m–3 range with ±0.03 mg m–3 or ±10% accuracy, whichever is largest). PM2.5 (0–999 μg m–3 range with ±10 μg m–3 or ±10% accuracy, whichever is largest) and PM10 (0–999 μg m–3 range with ±15 μg m–3 or ±15% accuracy, whichever is largest) are measured with a laser particle sensor. Data are exported as .csv files.

Project Description

Potential SCP ideas were discussed during lecture, and then the 34 students were split into pairs. They were tasked with designing and performing a systematic study of CO2 and COVID-19 exposure in a context of interest while considering the applicable limitations when formulating conclusions. Where applicable, students were encouraged to utilize the monitor’s other measurements. The asynchronicity of this project provided flexibility but also posed challenges due to limited interactions with instructors and reduced student engagement, as noted elsewhere.[8,11,31,32] A synchronous element was employed via required, but flexible, weekly meetings with teaching assistants (TAs) online, which were used to assess progress, answer questions, and provide feedback. The 17 groups were divided evenly among three TAs. Meeting length, typically 30–60 min, and structure depended on groups’ progress and needs. All meetings began with progress updates and ended with planning for the following week. Earlier meetings focused on measurement techniques and problem-solving while later meetings focused on interpretation of results. These meetings also served to monitor student growth over the course of the SCP. Below is a brief timeline of major events and deliverables:

Week 1: brainstorm SCP ideas and find a partner;

Week 2: present and write an SCP proposal;

Weeks 3–8: do the measurements and analysis;

Weeks 8–9: present findings and submit a full report.

Hazards

Students were told to comply with state guidelines and mask mandates. Students were advised to not enter locations where they did not feel safe or did not normally go. Additionally, they were told to observe additional precautions as applicable to their measurement’s locations, e.g., a research lab.

Examples of Student Results and Implications for COVID-19 Exposure

Survey of Locations

Figure provides an overview of locations studied in selected projects. For locations measured by multiple groups, the bars represent the averages. Monitors were placed away from sources and measured for minutes to days, depending on the project. One group cleverly used formaldehyde measurements to determine the AER in a well-ventilated research lab, where enhancements in CO2 were difficult to quantify, following a graduate researcher’s experiment. CO2 VMRs provided relative COVID-19 transmission risks, but most groups used the ATE with their observations and AERs for better estimates. All students discussed limitations relevant to their projects, commonly relating to weather conditions, lack of control over public locations, and air mixing rates.

Figure 1

Summary of students’ results including (a) average CO2 VMRs and (b) measured AERs at various locations, (c) average CO2 VMRs in a restaurant’s outdoor and indoor dining spaces given different occupancies, and (d) measured AERs with different ventilation. In (a) and (c), the horizontal line denotes the approximate atmospheric background CO2 VMR. Error bars denote standard deviations.

Average CO2 VMRs and AERs are summarized in Figure a and b. CO2 VMRs are expected to increase with decreasing AER and increasing occupancy, among other factors. One group showed that CO2 increases with restaurant occupancy and decreases with increased AER, as in the case of outdoor dining (Figure c). They recommended outdoor dining when available and indoor dining only with low occupancy. Another group tested different ventilation methods in an apartment and found a dependence on the time of day, likely due to weather conditions, and that ventilation methods, e.g., opening windows and using air conditioning, are not necessarily additive (Figure d). However, it is unclear if the apartment’s air conditioning system introduced outside air or recirculated indoor air. Below, we present results from three creative student projects.

Highlighted Student Project: Determining Safe Distance

One group investigated safe speaking distances by measuring CO2 concentrations at different distances from one and four people speaking. Their hypotheses stated that CO2 concentrations would decrease at greater distances from the source and increase with the number of people. They recorded CO2 for 2 min intervals at different distances away from one person reading the same passage from a book at a constant speed and volume, later repeated with four people. The students identified an exponential decay in CO2 over distance (Figure ), supporting their first hypothesis. When comparing one and four people, the larger group increased the e-folding distance by a factor of 3.5. The students recommended distancing oneself further from groups based on the number of individuals. The students acknowledged limitations in reproducibility for CO2 emissions. With a single monitor, the different distances were recorded for independent speaking events without knowing if the emissions were consistent across events.

Figure 2

CO2 VMR enhancements above atmospheric background (ΔCO2) at different distances from one (red) and four (blue) people speaking with the corresponding e-folding distances, δ. Error bars denote standard deviations.

Highlighted Student Project: Transmission between Rooms

Another group investigated the exchange of PM2.5 and CO2 between the living room and bedroom of an apartment, separated by a closed door. The students used their exhaled breath as a source of CO2 and burned incense to produce PM2.5. The experiment was conducted with and without the apartment’s filtered ventilation system fan. When the fan was off, CO2 and PM2.5 were high in the living room but only increased at a modest rate in the bedroom (Figure ). In contrast, the fan efficiently transmitted CO2 into the bedroom. Similarly, the fan introduced PM2.5 to the bedroom at a greater rate, despite the ventilation system’s filter. The students noted that PM2.5 transport to the bedroom was somewhat hindered by their filter but could be further reduced by filters with better PM2.5 filtration efficiency. The investigators suggested that COVID-19 transmission may increase with the use of ventilation and insufficient filtration in a household where someone is infected, even if isolated from other occupants. They also discussed limitations in their experimental setup: PM2.5 emissions from incense were not necessarily reproducible and create more PM2.5 than humans. Also, an infected person is more likely to occupy the bedroom than the living room, so transmission from the bedroom to the living room should also be investigated.

Figure 3

Transport of CO2 (blue) and PM2.5 (red) from sources in the living room of an apartment (top) to an adjacent bedroom (bottom) with closed doors. The experiment first used no additional air circulation (dashed traces) and then was repeated with the apartment’s filtered ventilation system fan (solid traces).

Highlighted Student Project: Vehicle Air Exchange Rates

A third group investigated the effects of ventilation on the AER in two cars. They used their breath to build up CO2 in the vehicles, exited, and derived the AERs from the decay. As hypothesized, the AER was greater when the windows were open ∼5 cm compared to being closed. With the windows open, AER ranged from 1.7 to 5.5 h–1, and the students found it correlated with wind speed (Figure ). When comparing two cars, they hypothesized that the AER would be similar for both cars due to similar surface area to volume ratios. They found the AER to be much slower for the larger car, which was newer, suggesting it was better sealed and less leaky. Using the ATE’s “subway” case for parameters such as breathing rates along with their measured AERs, vehicle volumes, and reported disease prevalence in the local community, they estimated how long two people could drive together with an absolute COVID-19 transmission risk under 0.1%, per the ATE’s output. They found that two people could travel for 26 min with closed windows and 155 min with cracked windows (∼5 cm) and the average recorded wind speed. The students acknowledged limitations for predicting safe travel times: differences in the leakiness of cars, the exact wind speed and relative direction, the effect of car movement, and the accuracy of the transmission estimator.

Figure 4

Air exchange rates in a vehicle measured at different outdoor wind speeds. Windows were cracked at ∼5 cm. Error bars denote standard deviations.

Learning Objectives and Assessment

Through this SCP, students were able to design a higher-level, open-ended investigation centered on the pandemic and pertinent to their daily lives. In an end-of-semester survey (detailed in the Supporting Information), students reported improved understanding of COVID-19 transmission (Survey Figure S1), positive perceptions of their abilities to do the SCP, and high interest (Figure ). Additionally, students learned research-based skills including the following:

Figure 5

Student survey results including (a) the extent to which students agreed with statements about their abilities to do the project and (b) student interest at different stages of the project. See Figures S2 and S3 for the full survey questions.

Identification of a research question;

Development of a research method;

Adaptation and problem-solving when faced with unanticipated outcomes;

Communication of findings.

For many students, this SCP was their first opportunity to design their own investigation, allowing for growth in understanding of the research process. These broad learning goals were indirectly assessed during different stages of the SCP using the deliverables mentioned previously (templates and guides in the Supporting Information). Traditional instrumental analysis learning goals, e.g., understanding instrumentation, data analysis, keeping notebooks, etc., are not discussed here as they are well documented in the literature.

The planning phase of the SCP was assessed via an oral presentation and written proposal. Students were expected to develop research-planning skills such as identification of a problem (skill A) and formulation of a reasonable means of investigation (skill B). In the ∼5 min presentation, students identified their research topic, posed a hypothesis, and proposed the measurements and data analysis. The purpose of this informal presentation was to garner feedback from peers and instructors to support a more complete and, if necessary, modified proposal that also required background, motivation, and a timetable. During the transition from oral to written proposals, some groups narrowed their topics or modified their methods to better fit the project time frame or to better characterize a smaller scope of transmission. In doing so, they grew in their abilities to identify experimental concerns and construct a higher quality investigation. Proposals were assessed based on the quality of the research question (skill A), the feasibility of the plan (skill B), and the inclusion of sufficient information. Specifically, students were expected to address relevant considerations for their proposed projects, e.g., how to measure AERs or which statistical tests will apply, and address any concerns raised during the presentations. Students performed well with most deductions stemming from insufficient information regarding measurements and analysis. Common pitfalls included misunderstanding the use of CO2 as a proxy for virus-containing aerosol particles, which was often reflected in their question or method. Additionally, some proposals utilized the monitor as a single-sample method where real-time measurements would have been beneficial. This confusion underlines the need to include real-time measurement techniques in undergraduate education to broaden their experiences and analytical tool belt. Most students agreed that they were able to find an interesting research question and design an appropriate experiment (Figure a).

These SCPs were student-designed and provided opportunities to learn skills related to problem-solving and adaptability (skill C). The students experienced unforeseen challenges: reduced access to locations, control of activities in locations, and unexpected CO2 sources, among others. Instructors for in-person laboratories typically monitor students and immediately address experimental problems, but this remote format required more critical thinking on the part of the students. Students were able to resolve most issues themselves but had access to TAs during weekly check-ins for more significant concerns.

Students’ problem-solving and adaptability skills were assessed indirectly through separate creativity and effort grades. All students performed well, as reflected in student perceptions of their problem-solving capabilities (Survey Figure a). Creativity was assessed on the quality of the research topic (skill A) with the highest grades awarded to unique experiments that exceeded classroom examples. Similarly, creativity was assessed for confronting experimental complications with meaningful solutions (skills B and C). This aspect of creativity reflects students’ growth regarding method development as they learned how to address problems with their initial plans. Several groups gradually adapted their methods during the first few weeks of the project as they recognized limitations. In fewer cases, students adapted to their results, as in the cases of deriving an AER from formaldehyde measurements and noticing the correlation of AER and wind speed. Students’ efforts were primarily assessed based on progress, tracked via the weekly meetings, and implementation of their experiments including the use of the real-time monitor (skill B). All students agree that they learned the value of using real-time measurements compared to the single-sample methods taught in the course (Survey Figure a).

Students reported their results via a ∼10 min presentation and a written report, both of which were assessed for coherent scientific interpretation and communication (skill D). Students were graded on the quality of the slides, verbal presentation, scientific argument, and response to audience questions. Students presented well with mostly minor issues related to figure quality, pacing of the presentation, or missing information. There were few significant issues primarily related to the quality of the scientific argument. Following the presentation, students received instructor feedback to address before submitting the final report. The report was assessed similarly to typical lab reports and the presentation. These reports showed improvements over the presentations as students addressed relevant comments and questions brought forth by peers and instructors. Growth in scientific communication was observed during the TA meetings and between the oral and written reports. Interpretation of complicated real-time data initially required help. For example, one group struggled to link an oscillatory signal to a real-world phenomenon which was later attributed to a ventilation system. Over time, the students more confidently interpreted their results as they better understood additional variables. Commonly, students required help brainstorming which type(s) of figures would best display their results, as they were accustomed to well-defined calibration curves, but gradually demonstrated their own creative ideas.

During the SCP, students demonstrated key research skills including the proposal of scientific work, method design, problem-solving, and scientific communication. Students exhibited similar growth regarding these skills relative to previous students performing the in-lab SCP. Additionally, the inclusion of the real-time monitors provided the students a new analytical technique, prompting a different approach to method planning and data interpretation relative to the course’s other instrumentation and previous years’ SCPs. Overall, students achieved the desired learning outcomes in this well-received project as discussed in the next section.

Student Feedback and Project Improvements

Students reacted positively to the SCP (Survey Figure S4) and found that this SCP format was understandable, had reasonable expectations, and contributed to their understanding of the research process. Most students agreed that this format was a good replacement for the in-lab format when considering the limitations. Over time, student interest improved allowing for a productive and engaging learning experience (Survey Figure b). Further, all students supported the addition of the monitors to the suite of instruments available in future semesters.

Flexibility has been noted elsewhere as key for student success[1,2,4,7,31,32] and is a central aspect of the SCP, even during a business-as-usual year. Students appreciated being empowered to design and perform an experiment relevant to their lives while working according to their own schedules (Survey Free Response A). As noted by one student, “My favorite part of this project was using my knowledge of chemistry to evaluate a real-world problem that is currently affecting millions of people.”

Students were asked to identify their least favorite part of the SCP (Survey Free Response B) and offer suggestions for improvement (Survey Free Response C). While the flexibility of this remote experiment was appreciated by some students, others found the lack of structure challenging. The weekly meetings with TAs provided a synchronous element to alleviate some remote-learning difficulties. Per one student: “Keep up the TA weekly meetings because that did help keep everyone on task...” (Survey Free Response C). Most students agreed these meetings were helpful (Figure S4), and we encourage instructors to utilize such meetings to provide students with necessary guidance. In future semesters, where groups can meet, we encourage in-person elements for remote experiments. Furthermore, we suggest sufficient discussion of the pros and cons of remote vs in-person experiments such that the students may make an informed decision.

The SCP has previously been a group project, and there have always been concerns over unequal partner contributions. Several students experienced imbalanced workloads, possibly exacerbated by being remote, and suggested groups be optional. Potential solutions require accountability via peer assessments and open discussion between instructors and students.

Some students expressed concerns that the SCP was complex and overwhelming due to significant differences from prior course material and analytical techniques. Additionally, students experienced issues with the volume of data. As described by one student, “My least favorite part of the project was the analysis of the data because there were so many things that could possibly [be] analyzed...” (Survey Free Response B). Following student suggestions, SCP lectures should begin earlier to provide more lead time to consider research topics before proposal presentations. We suggest the existing box-model homework problem be modified to include multiple, guiding steps (Supporting Information). Additional problems should be included to encourage correct analysis, e.g., fit an exponential decay to sample data to find the AER or use the ATE spreadsheet.

To improve proposed methods, students should be given the monitors earlier, allowing them to better understand how they work. Instructors may also design simple in-lab experiments or assign activities to measure, e.g., cooking, exercising, vacuuming, etc., and have students share their measurements for group analysis and interpretation. A possible in-lab experiment to investigate the AER is provided in the Supporting Information. Doing so demonstrates the benefits and complexity of real-time measurements to the students, promoting more carefully designed projects and improved understanding of such analytical techniques. Similarly, students should give a short ∼5 min update to the class during the second week. In doing so, presenters receive feedback and troubleshooting help to refine their methods while the audience may receive inspiration for their own projects.

Some students’ knowledge of Excel was insufficient for data manipulation and analysis. For example, simple exponential fits in Excel do not allow for a nonzero asymptote, which means that the background CO2 concentrations, which are not necessarily well-known, must be subtracted. The University of Colorado, Boulder and other universities offer software, e.g., MatLab, which can do these fits, or Excel’s “Solver” add-in could be used.

While this SCP was motivated by COVID-19, low-cost, handheld monitors can be applied to study various aspects of indoor air quality, as shown by D’eon et al.,[14] including disease transmission, volatile chemicals, emission sources, and ventilation. Such monitors are available for various species to address different contexts, and, over time, a department can build up a suite of various monitors. With sufficient modifications of motivations and learning goals, low-cost, handheld monitors could be used in student- or instructor-designed projects for a wide range of audiences.

Conclusions

With in-person laboratory learning hampered by the pandemic, handheld CO2 monitors were used in an instrumental analysis course for student-led investigations into COVID-19 airborne transmission. From course assessments and survey responses, students achieved the targeted research-based learning outcomes. The CO2 monitors provided experience with real-time measurements and analysis. Students left the course with a better understanding of COVID-19 airborne transmission. More generally, students expressed an improved understanding of air quality with intentions to use their knowledge of ventilation and filtration in their own lives. Additionally, this project prompted postsemester discussions of the interactions between science and policy in the context of COVID-19 airborne transmission. Students expressed satisfaction with these open-ended projects due to flexibility, applicability, and the ability to design their own project. These factors all contributed to strong student engagement, bolstering the learning experience. This project can be adapted to a variety of contexts related to air quality for a variety of audiences with different skill sets. Moreover, the versatile, low-cost monitors make this form of project accessible to most programs.