'It's fundamental': Quantum dot blinking experiment to teach critical thinking
Laszlo Frazer, Heather F. Higginbotham, Toby D. M. Bell, Alison M. Funston
‘‘It’s fundamental’: Quantum dot blinkingexperiment to teach critical thinking
Laszlo Frazer, † Heather F. Higginbotham, ‡ Toby D. M. Bell, ∗ , ‡ and Alison M.Funston ∗ , † † ARC Centre of Excellence in Exciton Science and School of Chemistry, MonashUniversity, Clayton, Australia ‡ School of Chemistry, Monash University, Clayton, Australia
E-mail: [email protected]; [email protected]
Abstract
Analysis of stochastic processes can be used to engender critical thinking. Quantumdots have a reversible, stochastic transition between luminescent and non-luminescentstates. The luminescence intermittency is known as blinking, and is not evident fromensemble measurements. In order to stimulate critical thinking, students design, per-form, and analyze a semiconductor quantum dot blinking laboratory experiment. Thedesign of the experiment and stochastic nature of the data collected require students tomake judgements throughout the course of the single-particle measurement and anal-ysis. Some of the decisions do not have uniquely correct answers, challenging the stu-dents to engage in critical thinking. We propose that students’ self-examined decisionmaking develops a constructivist view of science. The experiment is visually striking,interdisciplinary, and develops higher order thinking. a r X i v : . [ phy s i c s . e d - ph ] J a n eywords Upper-Division Undergraduate, Laboratory Instruction, Problem Solving / Decision Making,Semiconductors, Fluorescence Spectroscopy, Nanotechnology
Introduction
Critical thinking is an important skill for students to acquire.
While the acquisition ofcritical thinking skills is often given as a goal of laboratory instruction, evidence indicatesit is seldom achieved and it has been posited that expository laboratory experiments maynot develop critical thinking if they have a predetermined outcome. We use the quantumrandomness of single-particle measurements to cause students to make judgments aboutmeasurement quality. We describe an advanced experiment that promotes awareness ofexperimental bias in physical science research. The educational activity is designed forstudents at or above the advanced undergraduate level.There is disagreement about the nature of critical thinking.
Different types of skillsmay be included under the ‘critical thinking’ umbrella. Our investigation focuses on theevaluation aspect of critical thinking. We set an evaluation learning objective for studentsto judge the quality of their measurements. As a second objective, we prompt students toanalyze by identifying the bias that arises from their judgments.Evaluation skills may be related to beliefs about the nature of knowledge. Studentsare often unaware of subjective influences on science. In this experiment, students performan experiment that challenges positivist epistemology or epistimic authoritarianism.
When students identify that their judgments cause bias, conceptual conflict occurs betweenpositivist beliefs and experiences supporting constructivist epistemology.
The conflicthas the potential to change students’ views about the nature of scientific reasoning.The educational value of single-particle measurements is that they challenge stu-dents’ belief that the properties of ensembles are identical to the properties of individual2articles.
This type of belief is reinforced by the wide-spread use of ensemble measure-ments in chemical education, such as in optical spectroscopy, rheology, or NMR. However,single-particle measurements can detect rare events that are masked in ensembles. For ex-ample, the discovery of stable isotopes proved that neon atoms are not all identical. The apparatus students use in this experiment is primarily used as a research tool for widefield microscopy. Lower cost adaptations are possible. The experiment and the surroundingclassroom activities introduce students to the diffraction limit and the valuable applicationsof high resolution imaging. Quantum dots are semiconductor nanoparticles. They exhibit broad absorption spectraat energies above their bandgap along with narrow photoluminescence. Quantum dots gen-erally outperform molecular dyes in photostability, and can be investigated in air at roomtemperature. These properties, along with their inherent polydispersity, make them idealfor accessible single particle measurements at the undergraduate level. Here, we focus onquantum dot properties resulting from bottom-up colloidal synthesis. A distinctive propertyof such nanoparticles is that their attributes, particularly their quantum yield, are highlysensitive to modification of the surface due to their high surface to volume ratio.Single quantum dots are well known to blink. Blinking occurs when a photolumi-nescent particle (or molecule) temporarily ceases to emit light. Illuminated quantum dotsgenerally exhibit two-state behaviour, consisting of a brightly luminescent on-state and adark off-state. The ground state is ignored by the two-state theoretical framework for blink-ing. However, there can also be additional gray states caused by trap states and Augerprocesses. These have an intermediate brightness, between the brightness of the on-stateand off-state.The photoluminescence of single quantum dots has properties that can contribute tostudents’ awareness of measurement bias. The lifetime of the blinking off-state is distributedaccording to a power law. The origin of the power law is a current topic of investigation.There is no typical duration for an off-state because the distribution of off-state durations has3o statistical moments. If a quantum dot is not detected by measuring photoluminescence, itis impossible to determine if it is absent, incapable of luminescence, or simply in the off-statefor a long time.The power law probability distribution function cannot be normalized over its entiredomain. There are short off-periods that cannot be resolved and long off-periods whoseends are missed. The presence of gray states can further complicate the experimenters’decisions about the presence of a particle and criteria for detecting an off-state. Even inthe absence of a gray state, the quantum dot may switch between on- and off-states inthe middle of an acquisition, producing a data point that is between on and off. Whenmultiple particles are separately measured, the peak brightness of one may be less thanthe background brightness of another, resulting in a need to discard data or use multiplecriteria to detect blinking. Additionally, photoluminescence observed from a diffractionlimited spot, nominally attributed to a single quantum dot, may in fact be due to two (ormore) nanocrystals separated by a distance less than the diffraction limit.We use a pedagogical design where students make decisions about single quantum dot lu-minescence measurements. Then the students analyze the relationship between decisions andresults. The process is an example of constructivism in action. The action is constructivistbecause what is learned depends on the students’ choice, and not solely on the instructor oron the physics. The experiment is designed for final year undergraduates, masters students,or first year doctoral students in chemistry, (bio)physics, nanotechnology, materials science,or quantum information.
Experimental Methods
A protocol written for a student audience and the survey protocol are included in the Sup-porting Information. 4 aterials
Toluene (HPLC+), poly(methyl methacrylate) (GPC Standard, M W = Ω , Milli-Q) was used for all the procedures.CdSe quantum dots and CdSe/CdS/ZnS core-shell-shell quantum dots were syn-thesized in advance according to literature methods.Glass coverslips were cleaned in advance. The coverslips were soaked in chloroform for30 minutes, rinsed, then sonicated sequentially in acetone, m aqueous NaOH and deion-ized water for 20 minutes each respectively, with extensive rinsing between solvents. Thecoverslips were stored in a clean beaker in deionized water until needed.A number of pre-cleaned coverslips were rinsed with ultrapure water and dried with astream of N in preparation for sample deposition by each student. All glassware includingsample vials, dried coverslips and pasteur pipettes were placed in an UV-ozone cleaner for15 minutes immediately prior to use.Quantum dots of either type were serially diluted in toluene or poly(methylmethacrylate)/toluene solution. Optionally, m hexadecylamine or 1-octadecanethiol wereincluded in the solvent during dilution to change the surface chemistry of the quantum dots.The solution was spin-coated onto pre-cleaned coverslips at 5000 rpm for 60 seconds.Spin coating is used in single particle imaging because it distributes the particles with a lowdensity.In the interests of time, each student was limited to preparing and analyzing one solutionof quantum dots. Dilution in pure toluene relies on the presence of the surfactants/ligandsfrom the synthesis to maintain the colloidal stability of the nanoparticles. This particulardilution and its subsequent deposition (via spin coating) on the coverslip should be carriedout as quickly as possible to minimise dissociation of the ligands upon dilution as this canreduce the quantum dot luminescence. 5 ata collection An inverted microscope (Olympus IX71) set up in a wide-field configuration as illustrated inFig. 1 was used for single particle measurement. Samples were illuminated using a 488 nm,200 mW Toptica iBeam Smart 488-S-HP-10901 G0 solid state laser diode. The laser beamwas expanded to a flat field using a lens system and focussed onto the back of the objective.A 1.49 or 1.4 numerical aperture, × magnification objective was used. A dichroic filterseparated the incident laser excitation and the outgoing luminescence. The luminescence ofmultiple quantum dots was recorded as a video using an Andor iXon Ultra EMCCD in adarkened room. The luminescence and blinking were readily visible to the dark-adjusted eyethrough microscope eyepieces. Luminescence blinking was recorded in 100 second, 10 frameper second videos, allowing about five students to complete the experiment per hour. VideoCameraLens 1 Lens 2 Dichroic Mirror488 nmLaser ObjectiveSample EyepiecesRemovable Mirror
Figure 1: Simplified diagram of the wide-field fluorescence microscope.6 azards
Lasers may cause eye injury and should not be viewed directly. Quantum dots may be toxic. Spin coaters should be guarded and interlocked. Working in the dark is a trip hazard.
Data Analysis
Using a convenient custom-made data analysis package based on a menu-and-dialogue com-puter interface, pixels capturing the luminescence from a presumed single quantum dot aredefined. Background areas are also defined. The analysis software automatically generatesa photon trajectory (brightness as a function of time bin, as measured with video frames)from the summed light intensity of all pixels defined as quantum dot luminescence, withbackground counts from an area of identical size subtracted. Multiple trajectories are con-catenated to increase statistical power easily at the cost of accuracy. A brightness thresholdidentifies off-times in the trajectory, from which the software automatically generates a his-togram of off-times. The software reduces a selected domain of durations in the histogram toa power law exponent by log transformation and Poisson weighted linear regression. Thisexponent describes the temporal distribution of blinks. We do not introduce autocorrela-tion/power spectral density analysis or Bayesian estimation, which are more complexbut do not require a threshold, or change point analysis. Possible Modifications
Commercial quantum dots in organic solvents, with luminescence in the visible region andexchangeable surfactants/ligands can be substituted. Core-shell quantum dots are signifi-cantly easier to measure due to their generally higher quantum yields. Wide field fluorescenceinstruments are widely available because they are used for epifluorescence imaging in bi-ology. Quantum dots absorb light over a wide range of the spectrum. Illumination may becarried out using visible irradiation at any energy higher than the quantum dot band-gap.7any lasers or filtered arc lamps provide suitable illumination. A large numerical apertureobjective is essential to ensure efficient light collection. Any high quantum efficiency CCDvideo camera capable of recording at least ten frames per second could be used as a detector.The surface chemistry of the nanoparticles is manipulated to be either the surfactants/ligandspresent in a standard colloidal synthesis of CdSe core-shell nanoparticles, alkylamine (pri-mary amine) functionalised or alkanethiol functionalised. The latter two conditions areachieved by a straightforward dilution of the nanoparticles in solutions containing an excess(1 mM) of the ligand. The alkyl chain lengths of the ligands are relatively unimportant pro-vided colloidal stability of the nanoparticles is maintained (generally true for C8 or longeralkyl chains), and so substitutions of these ligands with those containing different alkyl chainlengths is possible. Substitution for secondary amines is also possible. Care should be takento ensure the ligand does not introduce impurities which fluoresce. Pedagogical Design
The experiment is organized as a conventional inquiry-based instructional activity, as illus-trated by Fig. 2. However, student decision making is not limited to the initial hypothesis-formation step. Instead, as shown in the green box, opportunities for decision making areinterspersed throughout the task. We give some examples of the most interesting decisionsbelow.
HypothesizeShellLigandPolymer MakeDilution MeasurePositionAccept/Reject AnalyzeSelect ParticleSelect BackgroundThresholdTime Domain CompareDecision/ResultRelationsipStatistical Validity ReportEpistemic View D ec i s i o n s Figure 2: Students perform a series of steps. For each step, we give examples of decisionsstudents could make. Decisions are opportunities for spontaneous or prompted evaluationand analysis. In other words, they are opportunities for critical thinking.Prior to undertaking the laboratory practical, students have attended lectures introducing8hem to the phenomenon of quantum dots. The topics covered include quantum confinement,colloidal stability and the passivation of quantum dots via ligands/surfactants or shelling witha wider bandgap material on the (ensemble) quantum yield for quantum dots. Referencepapers for these effects are provided throughout the lectures. Students are also directedto consult the Supporting Information and its references to inform themselves before theactivity.To initiate inquiry based learning, students form a hypothesis relating the choice ofquantum dot type, inclusion of polymer, and choice or omission of ligand on the power lawexponent. These low cost options allow students to explore the factors that determine theblinking and luminescence quantum yield of quantum dots. During sample preparation,students’ choice of a sufficiently low quantum dot concentration is essential to achieving themeasurement of single quantum dots. Whilst very general guidance regarding the concentra-tions required is given, the concentration tolerance of the stock solution to form a sample onwhich single particle measurements can be carried out is high. The students are required todecide the actual concentration of the stock solution to spin coat following a prelaboratorydiscussion considering the diffraction limit, and processes involved in spin coating. It is notunusual for the students to make a few attempts to prepare a sample with appropriate dis-persion, allowing them to appreciate the diffraction limit and particle density requirementsfor single particle imaging. The measurement portion of the activity is primarily expositorybecause of time and safety constraints. While measuring fluorescence, students select a re-gion of the coverslip to measure and may decide to reject their sample preparation in favorof a new sample preparation design.Students make a series of judgments to reduce a video file to a measurement of the powerlaw exponent for the quantum dot off-state duration. First, they select pixels in the videothat contain a single quantum dot. Here, students are judging the number of particles presentin these pixels, as well as which ‘single’ quantum dots should be selected for measurementin their analysis. Next, students select the pixels that they use to measure the background9rightness. The background brightness can be time- and space-dependent, leading to thepossibility that students may inadvertently select pixels corresponding to a dim quantum dotor a background of different average brightness compared to the quantum dot measurement.They need to decide if the background area is appropriate.The students decide a brightness threshold to set, which is then used to calculate theoff-times (duration continuously below threshold) of the quantum dots. This decision is com-plicated by the concatenation of data from different single quantum dots which may havedifferent blinking and/or background brightnesses. Usually, this decision involves ambiguitybecause of the stochastic nature of blinking. Students select a domain of durations withinthe resulting histogram to exclude invalid data before calculating the power law exponent.The optimal domain is ambiguous because the onset of duration-dependent errors, such asblinks that continue beyond the end of the video, is gradual. Care is taken to explain theproperties of histograms to the students owing to the difficulty students face understand-ing histograms.
Measurement bias inherent to quantum dot properties is relevant tostudents’ decisions. Students may fail to measure a quantum dot which happens to be offfor longer than the experiment. When a student inadvertently selects two quantum dots,an off state will only be measured when both quantum dots are in the off state. Then offstate durations are underestimated. When selecting a threshold, students may misclassifygray states. Since the power law probability distribution function is not normalizable, themeasurements are inevitably biased by the students’ chosen time domain. These types ofbias are example opportunities for students to achieve the evaluation learning goal.Students test their hypotheses using their own together with their classmates’ results.They may decide upon decision-result relationships and statistical validity at this point. Intheir laboratory reports, students are instructed to compare their measurements with theirhypothesis about the samples. Subsequently, they describe how they made judgments aboutthe data.Finally, students were prompted to consider why measurements might vary. Students10an choose to support different epistemic views. These are illustrated in Table 1. Thepedagogical design emphasizes the constructivist aspects of science.Table 1: Perspectives on the epistemology of the experiment. Each example illustrates astudents’ belief about the causes of variance.Epistemic View Illustrative ExampleAuthoritarian The expert’s measurement is more accurate.Positivist The measurements are inherently random.Constructivist The experimenter’s choice of analysis method changed the results.In our implementation, the experiment is targeted at students who are transitioning tobeing researchers. The experience gained serves as an introduction to optical research. Italso informs students about aspects of the interdisciplinary research fields of single-particlespectroscopy, super-resolution microscopy, nanotechnology, quantum information, andexcitonics.We implemented the experiment in an honours course at a research intensive Australianuniversity. An honours course is a fourth year of tertiary education completed after a threeyear bachelor’s degree. It can be used as a prerequisite for enrollment in doctoral study.The students enrolled have an average grade of 70% or better from undergraduate chemistrystudies. Honours students complete a nine month capstone project and coursework. Inour context, enabling progression to doctoral research is an objective of honours education.About ten students perform the experiment per year.
Quantum Dot Blinking Results
Fig. 3 shows an example video frame of single quantum dot luminescence. Students ob-serve less than one layer of luminescent quantum dots. The single quantum dot emissioncan be observed using the eyepieces of the microscope, with both the emission colour andblinking obvious. Fig. 4(a) and (b) are examples of the large variation in blinking behav-ior caused by students’ choice of sample preparation conditions. In (a), the quantum dots11ithout shells are mostly off. In (b), the core-shell-shell quantum dots are mostly on, witha larger peak brightness and much clearer on/off contrast. The photon trajectories obtainedunder each condition is consistent with literature measurements. The noise level is about × photons /
100 ms .Example Fig. 4(c) shows how the histogram changes as a function of the choice of thresh-old. Fig. 4(d) indicates that the power law exponent also depends on the selected threshold.The region where the exponent is insensitive to the threshold, (2 – × photons /
100 ms ,spans an order of magnitude. Students can, and occasionally do, use the interaction be-tween the threshold and power law exponent to judge the quality of the data. Thresholdinsensitivity suggests a replicable measurement of the exponent. A high degree of sensitivitysuggests unsuccessful classification of on and off states. For example, in Fig. 4(c–d), a nega-tive threshold incorrectly classifies background noise as short-lived off states. The exponentsreported by students ranged from 1.1 to 2.1 across 62 experiments, with a mean of 1.5 andstandard deviation of 0.2. Since the random error estimates are typically 0.01, students’choices explain most of the variation in the results. These random error estimates includevariation across both blinks and particles.Figure 3: An example of a student-recorded image of quantum dot luminescence. Theexperiment is visually striking because students can view the blinking of single quantumdots by eye. 12 B r i g h t n e ss ( P h o t o n s /100 m s ) Time (s)(a) Blinking Example -50 0 50 100 150 200 250 300 350 0 10 20 30 40 50 B r i g h t n e ss ( P h o t o n s /100 m s ) Time (s)(b) Blinking Example 1 10 100 0.1 1 10 O ff E v e n t s Duration (s)(c) Off Duration HistogramThreshold(10 Photons/100 ms)-500100 300 0 0.5 1 1.5 2 2.5-50 0 50 100 150 200 250 300 350 E x p o n e n t Threshold (10 Photons/100 ms)(d) Power Law Exponent
Figure 4: Examples of the interaction between student decisions and blinking results. (a,b) Contrasting student-recorded single quantum dot photon trajectories. (a) Quantum dotswithout shells in poly(methyl methacrylate) and hexadecylamine. (b) Core-shell-shell quan-tum dots in poly(methyl methacrylate) with only the surfactants/ligands present from thesynthesis. (c) Authors’ blinking off duration histograms generated from multiple studentmeasurements under condition (b), for various brightness thresholds. (d) Authors’ calcu-lated power law exponent as a function of threshold for the same data set as (c). The errorbars indicate random error estimates. The choice of threshold changes both systematic andrandom error. 13 tudent Decision Making Results
Our investigation of students’ decision making was approved by the Monash UniversityHuman Research Ethics Committee.
Student Laboratory Reports
We performed a retrospective, qualitative analysis of 45 students’ reports to investigate whatstudents thought about the decisions they made. These students received no instructions todescribe their decisions. Fig. 5 illustrates the cognitive processes we inferred from students’descriptions. These congnitive processes contrast with the conventional scientific methodand the experiment workflow. Some students simply reported actions, such as “A thresholdwas set.” In these cases, it is unclear if students are unaware that they made a decision, ifthey chose to hide their decision-making because they believe decision making is unscientific,or if they lack the writing ability to clearly articulate their decisions.Many students reported their decisions using decision words such as “subjective,” “choose,”“judge,” “exclude,” “bias,” or “select.” These words indicate the onset of critical thinking.Students often wrote in the passive voice, leaving it open to our interpretation if the stu-dents believed they made the choice or that some external decider provided it. We chooseto interpret passive voice as indicating that the student made the choice. Passive voicecommonly indicates author actions in formal scientific writing. Students may or may nothave identified criteria for their choices, such as “only the brighter dots or dots with moreinteresting blinking were chosen” or “two separate NPs overlaid in the same selection andwere identifiable [for omission] by three distinct intensity levels.”Some students proceeded to state that their decisions had effects. For example, “. . . thesubjectivity of the applied [brightness] threshold may have introduced errors into the data,”“results could have varied by the wavelength of laser used,” or “errors . . . may be due to thesubjective nature of choosing which spots appear to be due to a single or multiple emitters,14 xpositoryMaterials
ExperimentIdentify DecisionsIdentify CriteriaIdentify EffectsForm Hypothesis
Report knowevaluateevaluateanalyzeanalyzesynthesize
Figure 5: An illustration of students’ thought processes, as inferred from their reporting ofthe decisions they made. 15otential biases when selecting which spots to monitor, subjectivity in the choice of theintensity above which the spots are considered ‘on’ and below which they are considered ‘off’and the subjectivity in choosing a data fit range.” These types of comments indicate thatstudents understand their choices were important because they changed the results.Some students identified the dependent variable, such as, “intensity.” Only rarely didstudents identify an effect without the decision that caused it, such as, “If there were over-lapping quantum dots, then the intensity value may have increased.” In this case, the studenthas not described their choice of dilution or their choice of image pixels as playing a rolein the “overlapping” of quantum dots that were not spatially resolved in the microscope.Identification of effects implies students have transitioned from the “evaluate” portion of theBloom Taxonomy of Learning to the “analyze” portion.As an additional level of complexity, some students formed specific hypotheses aboutthe relationship between a decision and a dependent variable. For example, “When thethreshold was too high, it meant that the nanoparticle[s] were in the ‘on’ state too muchwhich would have resulted in a less steep gradient.” This contrasts with students who didnot have any hypothesis about the direction of change in the dependent variable. Thesestudents have entered the “synthesize” portion of critical thinking, which goes beyond ourlearning objectives.Finally, a few students conducted an experimental test of their hypotheses, leading to abetter experiment or better error estimates. Examples include, “If the sample was too diluteit became more difficult to locate the quantum dots, and therefore the experiment had to berepeated with a more concentrated solution,” or “moving the [brightness] threshold line bya small amount could change the final µ value by up to ± µ refers to the power lawdistribution exponent, which is the main parameter describing blinking.16 tudent Reflections First Survey
To investigate students’ views of the decisions they made, we used a survey. Fifteen studentsvoluntarily participated by completing written free responses to four prompts. The students’reflective comments were collected immediately after they finished analyzing their data withthe provided software.The first prompt probed students’ knowledge of the decisions they made when excludingdata. In response to the question “During the prac, how did you determine which data touse, at each stage of the analysis?” 73% of the participants identified at least one decisionabout excluding data. The remaining students referred to authorities, including writteninstructions or instructors, or indicated that they guessed. One student wrote, “I used somedata that in hindsight I probably shouldn’t have,” indicating that they learned more aboutdata selection as they completed the data analysis.The second prompt investigated students’ decisions when setting the threshold, which isused to convert brightnesses to blink durations. We asked, “What did you consider whensetting the threshold for on and off? How did you feel about the threshold selected?” 53%of students were able to identify at least one criterion they used to decide the threshold.For example, one student wrote “I chose a threshold between on and off brightness...” Thecriteria given varied. The difficulty of identifying these criteria depends on the sample thestudent was investigating. When measuring quantum dot cores, the on-state intensity is notsufficient to allow students to clearly identify a particular brightness level as indicating anon-state.We asked, “What sources of experimental bias could be in the experiment?” to find outif students could identify their choices as sources of bias. 93% of students identified at leastone source of bias. 80% of students identified at least one source of bias that was related totheir decisions. Overall, every student participant indicated that they were thinking about17heir decisions in at least one of three ways: identification of a decision to exclude data,identification of criteria for setting a threshold, or identification of bias related to a decision.The final question probed students’ analytical skill. It was “How could bias change yourresults?” Students had a lot of difficulty with this question. While 60% of students identifieda potential effect of bias, only 20% suggested that bias might change the value of the powerlaw exponent, µ . None of those students successfully identified a relationship between abias and the distribution of off-state durations. Two factors can contribute to this failure:First, students had not compared µ measurements across different sample conditions at thisstage in the activity. Therefore they may not have been prompted to realize µ was a result.Second, students may have inadequate mathematical preparation to understand power laws. Second Survey
At the end of students’ lab reports, a second survey was conducted. The same fifteenstudents participated. Students received a grade for answering the survey, but the graderswere blind to the student’s consent to participate in the research. First, we asked “Howdid you decide which data to use in your report?” At this stage, 93% of students identifiedat least one decision making criterion. None of the students referred to authorities. Thisindicates a pronounced increase in students’ awareness of their own decisions between thefirst and second survey. This may indicate students were learning as they prepared theirreports, or it may indicate greater student engagement with the second survey.To probe students’ views of the relative importance of sample preparation decisions anddata analysis decisions, we asked “How might your measurements be different from thoseof other students?” 80% of students identified analysis choices as the reason for differencesbetween measurements. 20% of students identified sample preparation choices. Most of that20% gave both reasons.To follow up on that question, we asked “Were the differences between samples causedby bias or were they caused by deliberate differences in the way samples were prepared?18ow could you tell?” 53% of students identified bias as the cause of differences. Analysischoices, inadvertent variation, and random errors were included among the examples of biaslisted by students. 47% of students used internal comparisons to identify bias or lack of bias.For example, “the replicates of the same conditions that were analysed by different peoplevaries [sic] drastically” or “samples that were replicated by different people, as the resultsfor those were reasonably close, and noticeably different when compared to other samplespreparations.”Finally, we probed students’ critical thinking about the literature with “How did youdecide if the results of the prac were the same or different from literature reports?” 50% ofstudents identified methodological choices as the reason they believed their results were dif-ferent from the literature. 33% identified correlations between their data and the literature.20% of students used internal comparisons to argue that their results must be different fromthe literature because students’ results were inconsistent. For example: “Some of the dataobtained was quite conflicting which indicates that at least parts of the data are incorrect.”None of the students mentioned using error estimates. These survey responses give furtherevidence of students’ critical thinking and awareness of the importance of methodologicaldecisions.
Focus Groups
After students completed their reports, we conducted two focus groups with three andfour student participants. As a warm-up, we asked students about the nature of laboratoryinstruction. Interestingly, students did not form a consensus about what laboratory instruc-tion is, except that it involves a task. Next, students were prompted to discuss the specificdecisions they made at various stages of the experiment they performed. As in the reportsand written reflections, students identified multiple decisions they made and discussed howthose decisions related to the results. We asked students about their views on decision mak-ing and the scientific process. The participants indicated that they believed decision making19s integral to the process of science:[ Group 1 ] moderator: So how does decision making fit into the scientific process? student f:
It’s fundamental to the science process. You have to decide what youwanna look at and how you’re going to look at and what implications yourmethod has on what you’re trying to achieve as well. student h:
Very integral. My project’s actually on [redacted] so yeah because they,in industry or independent research there’s like there isn’t always gonna be thatkind of, you know, guide the we need the students or like we as students need tohave these decision making things down pat so you can ask your own questionsand you can kind of you know have some form of independence. student g:
Decision making to science as a whole is very integral. So what methodsyou’re going to use, what you’re going to do, and stuff like that. But in termsof the undergraduate labs here at [redacted], it’s very much do this, do that, usethis, use that, then you try and let you choose some things like perhaps choosing[unclear] choosing what materials we should be using but its a lot very structuredyou just get that, and then you just analyze this.20
Group 2 ] moderator: So, is decision making part of the scientific process? student b:
I think so its under the yeah. student d:
I’d say so.[
Laughter ] student b: Everything’s decision making I feel like. student t:
I’d say it is, but we skipped a lot the decision making ourselves becauseit was over and done. So we came kind of later into the into the scientific process.That’s all I want to say. moderator:
So do the decisions scientists make change our understanding of science? student d:
Yes, I guess they can.[
Laughter ] student t: [unclear] decide to study what they study they probably wouldn’t bestudying they would go about it. student d: Someone could decide not to use less favorable data. It’s not good to dothat but if you’re deciding not to do that then you’re missing out on results thatmay be different from what you find in the end. moderator:
O.K., I see some nodding.These discussions indicate that the participants understand the importance of decisionmaking to science. They gave contrasting reasons, including personal autonomy, lack ofbackground information, scientific integrity, and need for a topic, for why decision makingis part of the scientific process. None of the participants took the positivist position thatscientific knowledge is based purely on evidence, exclusive of the choices of the investigator.We find that the participants value decision making, which is a type of critical thinking, inscientific contexts. 21 onclusions
In conclusion, analyzing the stochastic luminescence intermittencies of quantum dots as partof laboratory instruction develops students’ critical thinking skills. Students are required tomake several choices in order to record blinking videos and reduce the videos to a singleparameter. Not all these choices have unambiguously correct answers. Therefore, the ex-periment presents beneficial challenges to high-performing students, while still easily andconsistently producing measurements. The content is interdisciplinary, relating chemistry,quantum mechanics, and statistics.Students identified that they made decisions, used criteria, and had bias in their mea-surements. None of these observations are consistent with positivist epistemology. Theyare also incompatible with the belief in epistemic authority, which asserts that experts arethe source of truth.
We find that, while students hold a range of beliefs about the ori-gin of scientific knowledge, investigating quantum dots provides students with experiencesupporting constructivist epistemology.
Associated Content
The Supporting Information is available on the ACS Publications website at DOI:10.1021/acs.jchemed.XXXXXXX.Student handout with activity protocol, survey protocol (PDF).
Acknowledgements
We gratefully acknowledge the technical assistance of Rosalind Cox, Esther Miriklis, RileyHargreaves, Tich-Lam Nguyen, Paul Mulvaney, and Anum Nisar. We acknowledge TinaOverton for helpful discussions. We also acknowledge support from Michael Grace. A. M.F. and L. F. acknowledge support by the Australian Research Council (ARC) via the ARC22entre of Excellence in Exciton Science (CE170100026). T. B. acknowledges support fromthe Australian Research Council (DP170104477).
References (1) Kogut, L. S. Critical thinking in general chemistry.
Journal of Chemical Education , , 218.(2) Warren, A. R. Quantitative critical thinking: Student activities using Bayesian updat-ing. American Journal of Physics , , 368–380.(3) Holmes, N.; Wieman, C. E.; Bonn, D. Teaching critical thinking. Proceedings of theNational Academy of Sciences , , 11199–11204.(4) Domin, D. S. A review of laboratory instruction styles. Journal of Chemical Education , , 543.(5) Schoffstall, A. M.; Gaddis, B. A. Incorporating guided-inquiry learning into the organicchemistry laboratory. Journal of Chemical Education , , 848.(6) Danczak, S.; Thompson, C.; Overton, T. ‘What does the term Critical Thinking mean toyou?’ A qualitative analysis of chemistry undergraduate, teaching staff and employers’views of critical thinking. Chemistry Education Research and Practice , , 420–434.(7) Aloisi, C.; Callaghan, A. Threats to the validity of the Collegiate Learning Assessment(CLA+) as a measure of critical thinking skills and implications for Learning Gain. Higher Education Pedagogies , , 57–82.(8) Anderson, L. W., Krathwhol, D. R., Eds. A taxonomy for learning, teaching, and as-sessing: a revision of Bloom’s taxonomy of educational objectives ; Longman, 2001.239) Ryan, A. G.; Aikenhead, G. S. Students’ preconceptions about the epistemology ofscience.
Science Education , , 559–580.(10) Raviv, A.; Bar-Tal, D.; Raviv, A.; Biran, B.; Sela, Z. Teachers’ epistemic authority:Perceptions of students and teachers. Social Psychology of Education , , 17–42.(11) Hornikx, J. Epistemic authority of professors and researchers: Differential perceptionsby students from two cultural-educational systems. Social Psychology of Education , , 169–183.(12) Njegovan, B. R.; Vukadinović, M.; Nešić, L. G. Characteristics and types of authority:the attitudes of young people. A case study. Sociológia , , 657–673.(13) Blue, J. Assessing astronomy students’ views about the nature of scientific inquiry. Physical Review Physics Education Research , , 010141.(14) Conley, A. M.; Pintrich, P. R.; Vekiri, I.; Harrison, D. Changes in epistemological beliefsin elementary science students. Contemporary Educational Psychology , , 186–204.(15) Bodner, G. M. Constructivism: A theory of knowledge. Journal of Chemical Education , , 873.(16) Bodner, G.; Klobuchar, M.; Geelan, D. The many forms of constructivism. Journal ofChemical Education , , 1107.(17) Harbron, E. J.; Barbara, P. F. The Poisson Distribution and Single-Molecule Spec-troscopy. An Undergraduate Analytical Laboratory Experiment. Journal of ChemicalEducation , , 211.(18) Zimmermann, J.; van Dorp, A.; Renn, A. Fluorescence microscopy of single molecules. Journal of Chemical Education , , 553.2419) Ito, T. Observation of DNA molecules using fluorescence microscopy and atomic forcemicroscopy. Journal of Chemical Education , , 680.(20) Tversky, A.; Kahneman, D. Belief in the law of small numbers. Psychological Bulletin , , 105.(21) Rabin, M. Inference by believers in the law of small numbers. The Quarterly Journalof Economics , , 775–816.(22) Griffiths, I. W. JJ Thomson–the centenary of his discovery of the electron and of hisinvention of mass spectrometry. Rapid Communications in Mass Spectrometry , , 2–16.(23) Coltharp, C.; Xiao, J. Superresolution microscopy for microbiology. Cellular Microbi-ology , , 1808–1818.(24) Huang, B.; Bates, M.; Zhuang, X. Super-resolution fluorescence microscopy. AnnualReview of Biochemistry , , 993–1016.(25) Schermelleh, L.; Heintzmann, R.; Leonhardt, H. A guide to super-resolution fluores-cence microscopy. The Journal of Cell Biology , , 165–175.(26) Whelan, D. R.; Holm, T.; Sauer, M.; Bell, T. D. Focus on super-resolution imaging withdirect stochastic optical reconstruction microscopy (dSTORM). Australian Journal ofChemistry , , 179–183.(27) Whelan, D. R.; Bell, T. D. Super-resolution single-molecule localization microscopy:tricks of the trade. The Journal of Physical Chemistry Letters , , 374–382.(28) Klimov, V. I. Nanocrystal Quantum Dots , 2nd ed.; CRC Press, 2010.(29) Bullen, C.; Mulvaney, P. The effects of chemisorption on the luminescence of CdSequantum dots.
Langmuir , , 3007–3013.2530) Nirmal, M.; Dabbousi, B. O.; Bawendi, M. G.; Macklin, J.; Trautman, J.; Harris, T.;Brus, L. E. Fluorescence intermittency in single cadmium selenide nanocrystals. Nature , , 802.(31) Lee, S. F.; Osborne, M. A. Brightening, blinking, bluing and bleaching in the life of aquantum dot: friend or foe? ChemPhysChem , , 2174–2191.(32) Yuan, G.; Gómez, D. E.; Kirkwood, N.; Boldt, K.; Mulvaney, P. Two mechanismsdetermine quantum dot blinking. ACS Nano , , 3397–3405.(33) Riel, B. An introduction to self-assembled quantum dots. American Journal of Physics , , 750–757.(34) Shimizu, K. T.; Neuhauser, R. G.; Leatherdale, C. A.; Empedocles, S. A.; Woo, W.;Bawendi, M. G. Blinking statistics in single semiconductor nanocrystal quantum dots. Physical Review B , , 205316.(35) Cichos, F.; von Borczyskowski, C.; Orrit, M. Power-law intermittency of single emitters. Current Opinion in Colloid & Interface Science , , 272–284.(36) Verberk, R.; van Oijen, A. M.; Orrit, M. Simple model for the power-law blinking ofsingle semiconductor nanocrystals. Physical Review B , , 233202.(37) Kuno, M.; Fromm, D. P.; Hamann, H. F.; Gallagher, A.; Nesbitt, D. J. Nonexponential“blinking” kinetics of single CdSe quantum dots: A universal power law behavior. TheJournal of Chemical Physics , , 3117–3120.(38) Kuno, M.; Fromm, D.; Hamann, H.; Gallagher, A.; Nesbitt, D. J. On/off fluorescenceintermittency of single semiconductor quantum dots. The Journal of Chemical Physics , , 1028–1040.(39) Gómez, D. E.; van Embden, J.; Jasieniak, J.; Smith, T. A.; Mulvaney, P. Blinking andsurface chemistry of single CdSe nanocrystals. Small , , 204–208.2640) Efros, A. L.; Nesbitt, D. J. Origin and control of blinking in quantum dots. NatureNanotechnology , , 661.(41) Pelton, M.; Grier, D. G.; Guyot-Sionnest, P. Characterizing quantum-dot blinking usingnoise power spectra. Applied Physics Letters , , 819–821.(42) Rabouw, F. T.; Antolinez, F. V.; Brechbühler, R.; Norris, D. J. Microsecond BlinkingEvents in the Fluorescence of Colloidal Quantum Dots Revealed by Correlation Analysison Preselected Photons. The Journal of Physical Chemistry Letters ,(43) van Embden, J.; Mulvaney, P. Nucleation and growth of CdSe nanocrystals in a binaryligand system.
Langmuir , , 10226–10233.(44) van Embden, J.; Jasieniak, J.; Gómez, D. E.; Mulvaney, P.; Giersig, M. Review of thesynthetic chemistry involved in the production of core/shell semiconductor nanocrys-tals. Australian Journal of Chemistry , , 457–471.(45) Talapin, D. V.; Mekis, I.; Götzinger, S.; Kornowski, A.; Benson, O.; Weller, H.CdSe/CdS/ZnS and CdSe/ZnSe/ZnS Core- Shell- Shell Nanocrystals. The Journal ofPhysical Chemistry B , , 18826–18831.(46) Dabirian, R.; Guerrero, V. H.; Loza, D. Construction of low cost spin and dip coatersfor thin film deposition using open source technology. MOMENTO , 13–25.(47) Sadegh-cheri, M. Design, Fabrication, and Optical Characterization of a Low-Cost andOpen-Source Spin Coater.
Journal of Chemical Education , , 1268–1272.(48) Hardman, R. A toxicologic review of quantum dots: toxicity depends on physicochemi-cal and environmental factors. Environmental Health Perspectives , , 165–172.(49) Goldstein, M. L.; Morris, S. A.; Yen, G. G. Problems with fitting to the power-lawdistribution. The European Physical Journal B-Condensed Matter and Complex Systems , , 255–258. 2750) Verberk, R.; Orrit, M. Photon statistics in the fluorescence of single molecules andnanocrystals: Correlation functions versus distributions of on-and off-times. The Jour-nal of Chemical Physics , , 2214–2222.(51) Geordy, J.; Rogers, L. J.; Rogers, C. M.; Volz, T.; Gilchrist, A. Bayesian estimation ofswitching rates for blinking emitters. New Journal of Physics , , 063001.(52) Watkins, L. P.; Yang, H. Detection of intensity change points in time-resolved single-molecule measurements. The Journal of Physical Chemistry B , , 617–628.(53) Peidle, J.; Stokes, C.; Hart, R.; Franklin, M.; Newburgh, R.; Pahk, J.; Rueckner, W.;Samuel, A. Inexpensive microscopy for introductory laboratory courses. American Jour-nal of Physics , , 931–938.(54) Healey, M. In Reshaping the university: New relationships between research, scholarshipand teaching ; Barnett, R., Ed.; McGraw Hill/Open University Press, 2005; Chapter 5,pp 67–78.(55) Savery, J. R. Overview of problem-based learning: Definitions and distinctions.
Essen-tial readings in problem-based learning: Exploring and extending the legacy of HowardS. Barrows , , 5–15.(56) Cummins, R. H.; Green, W. J.; Elliott, C. Prompted inquiry-based learning in theintroductory chemistry laboratory. Journal of Chemical Education , , 239.(57) Deters, K. M. Student opinions regarding inquiry-based labs. Journal of Chemical Ed-ucation , , 1178.(58) Sanger, M. J. The effects of inquiry-based instruction on elementary teaching majors’chemistry content knowledge. Journal of Chemical Education , , 1035.(59) Cords, N.; Fischer, R.; Euler, M.; Prasad, A. Teaching optics with an intra-curricularkit designed for inquiry-based learning. Physics Education , , 69.2860) Kaplan, J. J.; Gabrosek, J. G.; Curtiss, P.; Malone, C. Investigating student under-standing of histograms. Journal of Statistics Education , .(61) Cooper, L. L.; Shore, F. S. Students’ misconceptions in interpreting center and vari-ability of data represented via histograms and stem-and-leaf plots. Journal of StatisticsEducation , .(62) Meletiou-Mavrotheris, M.; Lee, C. Exploring introductory statistics students’ under-standing of variation in histograms. Proceedings of the 4th Congress of the EuropeanSociety for Research in Mathematics Education, Sant Feliu de Guíxols, Spain. 2005.(63) Meletiou, M.; Lee, C. Student understanding of histograms: A stumbling stone tothe development of intuitions about variation. Proceedings of the Sixth InternationalConference on Teaching Statistics, Cape Town, South Africa. 2002.(64) Lee, C.; Meletiou, M. Some difficulties of learning histograms in introductory statistics.Joint Statistical Meetings-Section on Statistical Education. 2003; pp 2326–2333.(65) Pearson, B. J.; Jackson, D. P. A hands-on introduction to single photons and quantummechanics for undergraduates. American Journal of Physics , , 471–484.(66) Dehlinger, D.; Mitchell, M. Entangled photon apparatus for the undergraduate labora-tory. American Journal of Physics , , 898–902.(67) Galvez, E. J.; Holbrow, C. H.; Pysher, M.; Martin, J.; Courtemanche, N.; Heilig, L.;Spencer, J. Interference with correlated photons: Five quantum mechanics experimentsfor undergraduates. American Journal of Physics , , 127–140.(68) Vaughn, S.; Schumm, J. S.; Sinagub, J. M. Focus group interviews in education andpsychology ; Sage, 1996. 29 raphical TOC Entryraphical TOC Entry