Virtual Laboratories as a teaching environment

Practical science is currently taught in a labor intensive, hardware dependent fashion. Many of the traditional techniques could be embraced using existing multimedia technologies, thus enabling self-paced, student driven learning. Here we consider the role to be played by virtual laboratories. Virtual laboratories allow students to simulate experiments that may require expensive or dangerous hardware and materials from the safety of their PC. Virtual laboratories come in a number of guises, ranging from screen-shot based applications (J C Waller, N Foster 2000), to virtual reality systems driven by quantum mechanical theory (A Suzuki et Al, 1999). The former aimed at teaching the usage of a specific piece of equipment, the latter being a serious research tool. Recently virtual laboratories have gained press coverage (news.bbc.co.uk, 2001) here they are described as "bring(ing) science to life". In this paper we explore the various aspects of virtual laboratories, and how they can be applied to teaching of undergraduates, Whether they can "bring science to life". Classification of Virtual Laboratories I will classify virtual laboratories into two main categories, depending on how they gain their knowledge. One, has a limited set of facts inserted by the programmer, this is the way that the majority of systems currently work. The second, base their knowledge around a piece of far reaching piece of theory, this allows a far wider range of experiments to be performed. The term Virtual Laboratory is also applied to projects that allow the remote control of laboratory systems (K Keating et Al, 2000) but this is a topic more suited to a discussion of network technologies. Online libraries of knowledge, such as that at Imperial College (H Rzepa, A Tonge, 1998) are sometimes referred to as virtual laboratories, but usually only as a component of a larger tool set. Virtual Reality techniques have been used to create interfaces to existing online knowledge sources, an example of this is http://www.sci.brooklyn.cuny.edu/~marciano a virtual environment that provides hyper-links through interacting with it’s objects. Fact Based Virtual Laboratories Fact based virtual laboratories lend themselves best of supplementing traditional experimental work. In their 2000 paper, Waller and Foster describe a virtual instrument designed as part of a learning experiment. Their goal was to create a system to teach people to use the spectrometer in their lab. The software was then used at part of their undergraduate laboratory course. It is based around a collection of web pages, which contain screen shots from the computer that controls the spectrometer, along with still photos of the key steps of instrument operation. The screen shots are linked using the 'Image Map' technique. This allows different areas of the image to load different pages when clicked. This gives the impression of operating the software. The simulation allows students to take measurements of a number of samples, as if they were operating the real spectrometer. Some students were so convinced by the realism of the simulation that they asked if the software "was somehow interfaced to the real instrumentation" (Waller & Foster 2000). The simulator forms a safe environment, that allows their students to learn to operate the equipment, without the usual fears of damage attached to operating the real spectrometer. It was noted anecdotally by faculty members, that equipment malfunctions, and instances of useless data collection were considerably lower than in previous years. The system they built was specific to that task, and as such is fairly limited in what it can do. D B Armitage, 1999, created a more wide ranging simulator for the same experiment. This provided a simulation whose results were calculated partially on the fly. Waller and Fosters system is limited to only those analysis that they captured. Armitage's software bases it results on data for a number of compounds, and then produces simulated output based on this. This system also allows the user to adjust the variables of the reaction in the same way one would when running a real experiment. It does however lack the realistic interface provided by the screen-shot example, its interface was designed for this application. It's purpose is to help student to understand the procedure of the analysis, not the operation of one specific spectrometer, and so it simplified interface is preferable, highlighting only the more important controls. Oxford University produced a virtual laboratory to complement their first year undergraduate teaching (http://www.chem.ox.ac.uk/vrchemistry/). Their system also falls into the fact-based category, but tries to be a bit less specific, allowing more user control. In their system, a number of reactions have been filmed, and you can call them up by selecting two reactants. After viewing the video clip, the Screen Shots of Armitage's work, (D B Armitage 1999) user is quizzed about the reaction they've just created. This is again limited to the reactions that you can perform, but is closely matched to the experimental work performed in their lab, and is used as a re-enforcement tool. The use of video clips gives the virtual situation a higher degree of realism. Video clips are associated with reality, where as animations are perceived to be fake, although they may be equally accurate. The use of this system benefits the user in that they can repeat the reaction a number of times. Similar repartitions in the lab are often limited for reasons of safety and cost. Equally in a virtual environment potentially harmful reactions can be viewed many times, where as in a real lab situation the reaction may require additional safety procedures which are unavailable in a teaching environment. Derivation based Virtual Labs Derivation based labs allow the user to extend the experiments beyond those envisaged by the programmer. The results derived are based on solution of a mathematical model of the situation. By producing results in this way they extend their usefulness over fact based systems as they can be used to explore new idea in a research situation. Work by Suzuki et Al (1999) proposed a system that allows manipulation of virtual materials at an atomic level. In real world situations Eigler and Schweizer (1990) produced the first documented evidence of this (by spelling their sponsors name using xenon atoms on a nickel surface) It is however a time consuming work, limited to very small scales. Hence the benefits of the virtual material lab are obvious. In the VR environment, atoms can be manipulated in much the same way as macroscopic objects. However when the atoms are positioned, molecular Screen Shot from the Oxford Virtual Laboratory (http://www.chem.ox.ac.uk/vrchemistry/complex/default.html) The result of Eigler and Schweizers work moving atoms. http://www.almaden.ibm.com/vis/st m/atomo.html dynamics theory steps in, and make the atoms behave like real atoms, in terms of interatomic forces and interactions. This view of molecular manipulation was popularized in the film Jurassic Park (Universal Studios, 1993) that shows manipulation of a DNA strand in much the same was as Suzuki describes the manipulation of atoms. Heermann and Fuhrmann (2000) use a similar approach in their teaching of classical mechanics. Their system is based around a differential equation solving engine. Users define objects in their 'experiment' in terms of the physical characteristics (mass, shape, position) and their inter-connections, either chosen from the library, or created using simple java classes. The use of this class based system, allows the user much more freedom to extend the calculations possible. Once the system has been defined the software calculates the behavious of the system, then returns graphical schematic(s) of the system, and results of the calculations either as text or graphical output. Previous software in the same field (examples available from physicsweb.org) limits itself to specific examples, with limited variables. By concentrating on specific systems, previous simulators could use optimized, simplified equations to get results, but the simplification limits their usefulness. It was noted that "the possibility to investigate situations not foreseen by the teacher greatly boosts (student) motivation" (Heermann and Fuhrmann, 2000) hence their work is at a clear advantage. Their calculation system makes no attempt to run in real-time, hence isn't limited to powerful workstation computers in the way that Suzuki's work is. This opens the possibility for students to use the software from their own computers as extension activities, which was found to further boost their motivation. Assessment of teaching "Without assessment, there is no quantitative measure of student performance or effectiveness of teaching" (R Allen, 1998) and here virtual laboratories borrow techniques from Computer Based Teaching (CBT) Packages. By the use of continuous assessment students and staff can be constantly updated with the status of learning, if a topic isn't grasped then it can be repeated, or alternative references provided. The benefits of repetition in the virtual environment have already been mentioned. The Oxford University Virtual Chemistry Laboratory is an example of this, after performing each reaction, users are taken through a few multiple choice questions that prompt thought about the underlying theory, a step The Physics Modeling Environment Heermann and Fuhrmann 2000. that is all to easy to overlook in the teaching laboratory environment where time is often at a premium. Virtual Laboratories have the added benefit that in the case of an error the worst that can happen is a program crash, where as errors in a real laboratory can lead on great expense, both in terms of man power, and repairs (J C Waller, N Foster 2000). This leads to a greater confidence in the student, as they have been able to experiment and test ideas without the worry of breaking apparatus, "The first time they were faced with any instrument, they were more worried about damaging it or making a mistake, than leaning how to use it" (J C Waller, N Foster 2000). It is noted that neither of the derivation-based simulators examined here provide assessment. This is due partly to their wide range of possible experiments, and also to their target being partially research related, as opposed to fact-based simulators that are aimed directly at learning environments. Where these simulators are used in a teaching environment they are used as tools to complete a task, and hence are accompanied by a separate assessment scheme.