The People’s Approach to Laboratory Automation
- Credit: Jack Moreh, freerangestock.com
- Fig. 1: Some open-source hardware recently published in the journal HardwareX. A: Dual syringe pump ; B: 3D printing filament recycler ; C: Autosampler ; D: GPS data logger ; E: Pneumatic setup for microfluidics ; F: Drone .
- Matheus Carvalho is the author of the book “Practical laboratory automation made easy with AutoIt” , and has built three different models of open-source autosamplers [10,6,11], in addition to numerous other small devices for laboratory automation. He is editor for the journal HardwareX, and a member of the Open-source hardware association. His scientific interests involve algal photosynthesis and respiration, and stable isotope ecology. He has worked with stable isotope measurement for almost 15 years.
Automation is a desirable trait in any activity, and science is not an exception. After all, who does not want to be freed from tedious / repetitive / hazardous tasks? Who does not want a workforce that never sleeps?
Laboratory automation: desirable but inaccessible
Scientists, being ingenious by necessity of their profession, have since long ago devised ways to improve their efficiency by means of automation . At the same time, entrepreneurs have seized opportunities in the field of laboratory automation, and provided specialized equipment to scientists that have benefitted, and in some cases even made possible, science in diverse fields.
Despite the undeniable benefits proportionated by commercial companies to automate science, there are some particularities in this market that have precluded a more widespread adoption of automation. The main one is the price. Scientific equipment is usually expensive when compared to mass-produced devices like cars, TVs and cell phones. This is a straightforward consequence of the relatively small demand in the scientific market. There are other factors, but there is a more subtle aspect that is hard to be seen by people from outside: the lack of compatibility between instruments from different manufacturers.
For someone who is not familiar with scientific equipment, it is hard to identify their sub-parts. A person goes into the laboratory, and see a “machine that measures carbon”, for example. This machine may, in some cases, be composed by 3 or 4 different machines, all put to work together as a single unit. Some of these units are top-end, cutting edge devices based on the latest developments in sensor technology. However, others will be simple mechanical devices a little more (or maybe a little less) complex than a bicycle. However, such low-tech devices will be charged premium price because the manufacturer will have monopoly over its supply.
An answer to this problem has been a call for standardization of laboratory equipment. This call has been made several times since the dawn of widespread laboratory automation in the early 1990’s, but it has never materialized.
The truth is that there is no compelling reason for manufacturers to spontaneously adhere to standards, as this would result in new production costs and, consequently, loss of competitiveness.
An alternative to standardization that actually works is the use of scripting . For users, it is easy to learn, because there is no need for knowledge in advanced computing, or even in electronics. Therefore, users can mix and match any devices that suit them. For manufacturers, it is cost-free; they do not need to change anything (or they can change everything, if they wish; there is no top-bottom rule here) in the devices that they sell. Even more importantly: scripting opens the door for the widespread adoption of low-cost, open-source devices for laboratory automation.
Open-source hardware for science – origins
There is a growing movement of scientists that make their own devices, and publish the instructions for others to follow. This movement is largely inspired by the overwhelmingly successful open-source software experience. The philosophy behind these movements is that freedom of information is beneficial for the development of any activity, science not being an exception. Some (including Joshua Pearce, one of the most vocal voices advocating open-source hardware for science ) argue that the patent system, which was devised to foster scientific progress, has in fact worked to hamper it. In its present form, the system is closer to a monopoly than to a free market. By its very nature, open-source repels patents.
Scientists are also drawn to the open-source because it closely resembles scientific publishing: the one who gives more receives more, in terms of citations and recognition. Although this resembles an altruistic philosophy, it is not: once others adopt the designs that a certain research group publishes, the group later can benefit from an improved design of their original device published by another group, and so on. Finally, the open-source hardware movement is a bottom-up movement, emanating from the end users themselves, which is attractive to people who reject top-bottom impositions from “know-it-all-field-leader-visionaries”.
The open-source hardware movement has not come from nothing. It is an emergent phenomenon coming from the widespread availability of some key technologies. One of the most important technologies is open-source microcontrollers, like Arduino and Raspberry Pi. Such microcontrollers have enabled people without formal engineering education to build the most varied devices. Another key technology is 3D printing, which enables rapid prototyping at a very low cost. Finally, the rapid dissemination of information on the internet has also been a fundamental factor to allow the emergence of the open-source hardware movement.
Open-source hardware for science – present
For now, most open-source hardware consists of relatively simple devices. Scientists still need to buy commercial devices for specific and highly sophisticated measurements. This may take time to change, or might never change, in some cases. Still, since the lack of compatibility between instruments has been solved, at least marginal savings are possible . In some cases, the laboratory can get a better deal, maybe an autosampler as a freebie, if they tell the vendors that they will build one themselves.
Often, open-source devices cost between 1 and 10% of commercial alternatives. This is of course very attractive, but together with the low price comes the necessity of learning about how to build the devices, as they don’t come assembled. Thus, it is understandable that many scientists will opt for commercial, tested devices that come ready to use, even if this means paying much more than building a device by themselves. This situation is changing, as well. Some companies are already exploring the niche of pre-assembled open-source devices, offering them for competitive prices. This might become a popular alternative in the future, when more scientists become more proficient with enough technical aspects regarding their scientific equipment, and realize that they have more options than their vendors normally show them.
Building open-source hardware for science is becoming easier by the day. This is true for the two cornerstone technologies, microcontrollers and 3D printing, alike. If a few years ago it was necessary to master the operation of microcontrollers, nowadays modified microcontrollers with freely available and relatively user friendly firmware (the software that is in the microcontroller and interacts with the controlling computer) make the task much easier.
Using a 3D printer has also never been so easy. New models arrive to the market on an almost daily basis, suiting virtually any kind of potential user. Inexperienced users will probably go for the cheapest models that can be bought as a single piece (that is, that do not demand assembly by the user). The filament of choice will probably be PLA, which is excellent for prototyping. And, although the best option is to have a 3D printer at hand so that every new interaction of the prototype can be made in a matter of hours, there are hundreds of 3D printing small business that print for a fee.
Even designing a piece of hardware has never been so easy. That is, if a new design is required at all. There are thousands of designs freely available for download which can either be printed straightway, or be slightly modified to suit a specific purpose. Designing from scratch is not a problem, either: incredibly user-friendly programs exist, some do not even require download. It is not exaggeration to say that it is like child’s play.
It is not necessary to build a full piece of equipment to take advantage of open-source hardware. Small components, like valves or cryo-traps, can be built and integrated to existing equipment to complement their function, or substitute broken parts that normally would cost a disproportionate amount of money. “Resurrecting” abandoned devices is also possible using the basis of the open-hardware technology, even if the devices are not open-source to begin with. Also, depending on what the scientist wants to build, chances are that something similar has already been built, either for another scientific application, or for an entirely different purpose. In such cases, it is possible to adapt and modify the existing device so that it can perform the desired task.
It has been possible to find the building instructions for open-source hardware devices scattered in the scientific literature, in specialized journals for each sub discipline. In 2017, as a sign of the growth of the open-source hardware movement, a peer-reviewed journal dealing specifically with such devices was launched: HardwareX. The interested reader can check the journal website (https://www.journals.elsevier.com/hardwarex) for a long list of devices made by scientists for scientists (Fig. 1).
Open-source hardware for science – future
The future for open-source hardware in science is very promising. It is bringing back scientists to the time when they built their devices themselves. This means that devices will be fully customized to the specific needs of a task. But it means more. It means that scientists will become less dependent of high-priced equipment. While this is nice for developed countries, it is essential for developing countries where research funding is meager. It will also mean more opportunities for independent research. There is a growing concern that predominantly government funded research can lead to biased studies. This is not to say that independent studies will never be biased. The point is that the atomization of research funding should, by principle, result at the worst in research fueled by conflicting biases, which in the end is positive for scientific enquire (of course everyone wants research to be bias-free, but unfortunately the equipment used – human beings – has no warranty in this case).
Also regarding funding, the adoption of open-source hardware can be a way of transferring funds from machines to people. As less money needs to be spent in equipment, more money becomes available to employ scientists. In the present crisis of placement for PhDs, such a change would be very welcome. Ironically, it would mean that automation would create jobs instead of destroying them. But in fact it makes sense: PhDs are – hopefully - educated, creative and driven individuals, which cannot be easily substituted by machines (yet). The menial jobs that open-source laboratory robots can perform for a trivial cost might be the tool by which all the investment in the education of such individuals becomes better used by society.
“Knowledge is power”, they say. This is absolutely true in the case of open-source hardware for science. The more you know, the more you can do. And the beauty is that it cannot be taken away. A fantastic knowledge base is being built and propagated to the whole world. The open-source hardware movement is not something imposed by government, regulatory body, IAEA, IUPAC, etc. It stems from the people, and this is its main strength. It is not far-fetched to believe that in some years open-source hardware will be present in most laboratories in a form or another.
Author: Matheus Carvalho
Prof. Matheus Carvalho
Centre for Coastal Biogeochemistry Research
Southern Cross University
 K. Olsen, The First 110 Years of Laboratory Automation : Technologies, Applications, and the Creative Scientist. Journal of laboratory automation, 17, 469-480 (2012). doi: 10.1177/2211068212455631
 M.C. Carvalho, Practical laboratory automation made easy with AutoIt. Wiley VCH, (2016)
 J.M. Pearce, Open-Source Lab: How to build your own hardware and reduce research costs. Elsevier, Waltham (2014)
 V.E. Garcia, J. Liu, J.L. DeRisi, Low-cost touchscreen driven programmable dual syringe pump for life science applications. HardwareX, 4, e00027 (2018). doi:10.1016/j.ohx.2018.e00027
 A.L. Woern, J.R. McCaslin, A.M. Pringle, J.M. Pearce, RepRapable Recyclebot: Open source 3-D printable extruder for converting plastic to 3-D printing filament. HardwareX, 4, e00026 (2018). doi:10.1016/j.ohx.2018.e00026
 M.C. Carvalho, R.H. Murray, Osmar, the open source microsyringe autosampler. HardwareX, 3, 10-38 (2018). doi:https://doi.org/10.1016/j.ohx.2018.01.001
 P.W. Cain, M.D. Cross, An open-source hardware GPS data logger for wildlife radio-telemetry studies: A case study using Eastern box turtles. HardwareX, 3, 82-90 (2018). doi:10.1016/j.ohx.2018.02.002
 K. Brower, R. Puccinelli, C.J. Markin, T.C. Shimko, S.A. Longwell, B. Cruz, R. Gomez-Sjoberg, P.M. Fordyce, An Open-Source, Programmable Pneumatic Setup for Operation and Automated Control of Single- and Multi-Layer Microfluidic Devices. HardwareX, in press (2017). doi:https://doi.org/10.1016/j.ohx.2017.10.001
 J.-L. Liardon, L. Hostettler, L. Zulliger, K. Kangur, N.S.G. Shaik, D.A. Barry, Lake imaging and monitoring aerial drone. HardwareX, 3, 146-159 (2018). doi:doi.org/10.1016/j.ohx.2017.10.003
 M.C. Carvalho, B.D. Eyre, A low cost, easy to build, portable, and universal autosampler for liquids. Methods in oceanography, 8, 23-32 (2013). doi:http://dx.doi.org/10.1016/j.mio.2014.06.001
 M.C. Carvalho, C.J. Sanders, C. Holloway, Auto-HPGe, an autosampler for gamma-ray spectroscopy using high-purity germanium (HPGe) detectors and heavy shields. HardwareX, 4, e00040 (2018). doi:doi.org/10.1016/j.ohx.2018.e00040