Aerolis's students in the lab.

Welcome to the Notebook! This page will serve as a source of in-depth information on the biological and technical aspects of the Aerolis project. It will cover everything from Air Pollution to the work packages (Biofunction, Filament and Sensor), 3D Design and additional information on Lab Safety ending with the Downloads section.

Air Pollution

It is known that air pollution has harmful effects on human health. According to the world health organization (WHO), 7 million premature deaths are due to atmospheric pollution, which equals to one eighth of the world annual deaths. The cost of those premature deaths in the WHO countries is $ 1.431 trillion per year.

Air pollution is characterized by the presence of gas and particles in the outside air with harmful impact on human health and/or on the environment. Those pollutants come from natural phenomena (volcanic eruption, organic matter decomposition, forest fire) as well as human activity (industry, transport, farming and residential heating).

Particulate Matter
  • What is PM?

    PM or particulate matter is a combination of liquid droplets and solid particles that can be found in the air. Some are visible with the naked eye e.g. smoke, dust, etc. Others cannot be seen without visual aids such as a microscope. They can be divided into three fractions (US EPA n.d.; Alkalaj & Thorsteinsson 2014):

    • PM10: consists out of inhalable particles with aerodynamic diameters that range around 10 micrometres or smaller.
    • PM2.5: consists out of fine inhalable particles with aerodynamic diameters that fluctuate around 2.5 micrometres or smaller.
    • PM1: consists out of ultra-fine inhalable particles with aerodynamic diameters smaller than 1 micrometre. (this definition may vary, ultra-fine is sometimes called PM0.1)

    These particles come in all shapes, forms and sizes and can consist out of a plethora of different chemicals. So, this indicates that they can come from many different sources either direct or indirect. Examples of such sources are, but are not limited to: car exhaust gases, wild fires, ship emissions, mining, melting glaciers, volcanic eruptions, agriculture, road transport, construction sites, pollutants from industries and power plants… It should also be clear that while residing in the atmosphere, the chemical and physical characteristics may change as they could interact with other particles and could undergo further complex chemical reactions.

  • PM and Health Issues

    Multiple studies indicate that close to major traffic roads there are more people with respiratory symptoms and reduced lung capacity (Alkalaj & Thorsteinsson 2014). Ambient air pollution can also have adverse effects on pregnancy, such as an increase in post neonatal mortality (WHO Regional Office for Europe n.d.) as foetuses are more vulnerable to environmental effects because of their higher cell division rate (Alkalaj & Thorsteinsson 2014).

    The main intake path is via inhalation, where the particles smaller than 10 micrometres in aerodynamic diameter are especially harmful because they can penetrate deep into the lungs and from there potentially enter into the bloodstream. Children, elderly, and people with asthma, lung or heart diseases are the most vulnerable groups and have the biggest chance to be affected by PM exposure.

    Multiple scientific studies have linked particle pollution exposure to a variety of problems, especially connected to the lungs and heart, including: premature death in people with a heart or lung condition or disease, irregular heartbeat, nonlethal heart attacks, asthma or the aggravation of it, decreased long function, increased respiratory symptoms for example difficulty with breathing, coughing or irritation of the airways. And the most dangerous particles would be the Polycyclic Aromatic Hydrocarbons (PAHs) because of their ability to cause damage to DNA, which make them carcinogenic (Alkalaj & Thorsteinsson 2014).

    The particles can be transported over very long distances, mainly by the wind, hereafter they settle in the water and ground. Possible consequences are: depleting nutrients in the soil, changing the biodiversity of ecosystems, acidification of lakes, altering the nutrient balance in rivers and coastal waters, reduction of visibility (haze caused mainly by PM2.5), contributing to acid rainfall, damaging sensitive ecosystems e.g. forests, crops, etc.

  • Removal of PM

    Vegetation is known for its carbon sequestration, but they can remove particulate matter from the air as well e.g. (PM10), sulphur dioxide (S02), nitrogen dioxide (N02), ozone (03) (Alkalaj & Thorsteinsson 2014). Due to vegetation that creates a big and rough leaf surface more eddies are created and turbulent mixing in the neighbourhood of the vegetation will occur. A mixture from mature woodland is apparently three times more efficient at captivating particulate matter than grasslands (Alkalaj & Thorsteinsson 2014). Table 1 shows that, from all vegetation, trees seem to be the best at capturing particulate matter, which is confirmed by Alkalaj & Thorsteinsson (2014).

    It can be concluded that increased stickiness of the surface increases the capture of coarser particles, while the surface roughness and hair density seemed to increase the capturing capacity for finer particles (Kerckhoffs 2014).

    Interestingly, in Table 2, it can be seen that pine species were able to accumulate the most pollutants, with the highest concentrations around the stomatal regions, even though their leaves have no hairs or rough surfaces. The explanation is that long and slim needles collide more easily and more frequent with the pollutants (Kerckhoffs 2014). From the broadleaved species, it can be concluded that White beam was the best, because of the rough and hairy leaf surfaces (Kerckhoffs 2014). Trees with ridged and hairy leaves had the highest deposition velocities. So it is clear that surface-related structures are the most important factors in particle deposition. Also, Birch and Beech species have high deposition velocities (see Table 2), this is in accordance with earlier observations as these trees have hairy and ridged surfaces.

    Table 1 - Removal rates for all pollutants (Kerckhoffs 2014).

    Table 2 - Deposition velocities and removal rates per tree species (Kerckhoffs 2014).

Volatile Organic Compounds
  • What are VOCs?

    Chemical compounds are generally considered as volatile if their vapor consists mainly of molecules in the gas phase, and if sorption of molecules on solid particles in the atmosphere is negligible (Van Langenhove 2013). So, VOCs are a large group of organic molecules that find themselves mainly in the gas phase under atmospheric conditions. Volatility is a physical property, therefore molecules with different chemical structures and reactivity can belong to this group. The two biggest sources of VOCs in the air are exhaust gasses and the industrial use of solvents, but also a lot of household products can emit these VOCs. Pheromones and a lot of other natural chemical compounds emitted from organisms are also considered VOCs, but the focus of this paper will be about VOCs emitted by human activity.

  • VOCs and Health Issues

    The risk for human health of VOCs depends mostly on which compounds are present in the vapor and most effects are due to long term exposure to the compound. They can cause respiratory problems, irritation of the eyes and throat, fatigue and headaches. Exposure to certain concentrations of VOCs like formaldehyde is also associated with an increase in occurrences of asthma among children (Vigez 2017). Also, many VOCs are classified as known or possible carcinogens and toxicants (EPA 2012). Some possible carcinogenic compounds are listed in Table 3 (Chin et al. 2014).

    Table 3 - Unit risks and chronic inhalation reference concentration (RfC) for selected VOCs.

    The most common ways for exposure to VOCs are via drinking water, food and inhalation. However, most of the compounds that are stated to be dangerous are found in the air. Only a few can be found in drinking water and food also appeared to be a minor route of exposure (Wallace et al. 1984). Thus, the most important way to prevent exposure to VOCs is by lowering the concentration in the air or even completely remove them.

  • Removal of VOCs

    There exist multiple methods to remove VOCs out of the air. Figure 1 shows the classification of several VOC control techniques.

    The methods based on recovery of VOCs aren’t usable for Aerolis. Condensation is most efficient for VOCs with boiling points above 100°F (38°C) at relatively high concentrations above 5,000 ppm. Low-boiling VOCs can require extensive cooling or pressurization, which sharply increases operating costs (Khan & Ghoshal 2000). Absorption requires a liquid solvent that, in the case of this project, needs to be replaced from time to time. Adsorption and membrane based recovery processes have the same need of replacement that increases the operational cost and causes a source of waste.

    VOCs can be removed by oxidation, but thermal oxidation systems combust them at temperatures of 1.300 – 1.800°F equivalent to 700 – 1000°C (Khan & Ghoshal 2000). This is too energy expensive to use in an air purification unit and then there is also the need to use heat resistant materials. These temperatures could be reduced to 700 – 900°F or 400 – 500°C using catalytic oxidation (Khan & Ghoshal 2000) but it is too expensive to use in the unit. The most suitable method for the unit would be biofiltration. Microorganisms can convert organic pollutants to water under aerobic conditions. Using biofiltration there is no need for replacement of costly materials like a solvent or an adsorbent, no high energy cost and no waste production. A lot of research has been conducted about the biodegradability of VOCs by microorganisms and especially about the degradation of toluene by bacteria. Because it would be quite impossible to design a purification unit that removes all compounds from the air, toluene is used as a reference for the whole group of VOCs. Thus, the biodegradability of toluene is an important factor when selecting the bacteria for the biofilter in the purification unit.

    Figure 1 - Classification of VOC control techniques (Khan & Ghoshal 2000).

VOC-degrading Microorganisms
  • The Microbiome of a Biofilter

    In order to create a biological air purifying device capable of efficiently degrading VOCs, it is necessary for the biofilter to contain microorganisms with a metabolism capable of utilizing the airborne organic compounds. It is however important to realize that biofilters are inherently open systems, meaning that contamination with microorganisms present in the air is inevitable and keeping the biofilter single species is impossible. Since the degradation of the organic compounds occurs in many steps, different organisms could specialize in different parts of the process (Devinny et al. 1998), meaning that this contamination could actually increase the microbiome’s ability to fulfil its biofiltration job. Thus, the microbiome’s composition would fluctuate greatly, with some species thriving and other failing, before finally roughly stabilizing when the climax community is reached (Devinny et al. 1998).

  • Inoculating a Community of Microorganisms

    Often the biofilter is inoculated with a community of microorganisms, which will then evolve to a microbiome more capable of biodegrading VOCs during the process of biofiltration. A higher biodiversity leads to a more effective and robust biofilter, since the system contains more pathways to metabolize the substrates and metabolites, as well as being more resistant to fluctuation due to functional redundancy (Estrada et al. 2013).

    The usage of activated sludge from wastewater treatment plant for the inoculation of a biofilter has been widely employed. This activated sludge contains a great variety of rugged organisms which have been exposed to the typical wastes of civilization (Devinny et al. 1998). Only a few species will be capable of degrading the substrates, though the rest of the microbial community will form a better environment for growth and will contain several species which will help in biofilm formation (Z. Cheng et al. 2016). This activated sludge can be inoculated directly (for example by Estrada et al., 2013), or be first exposed to the contaminant for a duration for acclimatisation, which could result in a lower start-up time (for example employed by Z. Cheng et al., 2016).

    Another option is to use a medium which already contains a natural microbial community, such as compost. As with the activated sludge, compost contains a variety of microorganisms capable of degrading many components, and the environmental conditions will favour species capable of degrading the contaminant, with a gradual shift towards the climax community as a result (Devinny et al. 1998). The added benefit of a cheap, safe, widely available and easily accessible medium, as well as the lack of the need for a specific inoculum, makes compost a prime candidate for usage in the Aerolis structure.

  • Selection of Specialized Microorganisms

    Even though the species composition of the biofilter will change over time, the inoculation of a biofilter with specific, specialized species can still serve a purpose. It could decrease the acclimatisaton period, allowing the biofilter to reach its maximum elimination capacity (EC) sooner, though the inoculation would not change this maximum EC (Leson & Smith 1997). The biofilter could also be inoculated with species that would not naturally develop in the microbiome, such as transgenic organisms or organisms that cannot be found in the air of the biofilters feed, which could possibly increase the maximum EC.

    One of the first choices that need to be made when selecting species for inoculation of the biofilter is whether it is desired to employ bacteria of fungi. There are species known to be capable of using VOCs as carbon source in both these domains, which makes the question arise whether it is preferred to utilize bacteria, fungi or a combination of both as inoculum.

    Generally, bacteria have a faster substrate uptake and faster growth when compared to fungi (Dorado et al. 2008), but lose this advantage once the conditions get unfavourable, such as a low moisture content, a low pH and increasing nutrient scarcity (Estrada et al. 2013). It has also been suggested that the growth of aerial mycelium may increase the mass transfer of hydrophobic VOCs in fungi (Estrada et al. 2013). However, this mycelium growth will clog the biofilter, which could lead to a low flow rate at high VOC concentration, although this can be partially mitigated by adding mites so they can graze on the fungal mycelium (Cheng et al., 2015). But, since the proposed design will be working at low VOC concentrations, the latter disadvantage is unlikely to pose a serious problem.

    It has been suggested that using a combination of bacteria and fungi would act as a better biofilter as they compensate for each other’s weaknesses. Y. Cheng et al. (2016) confirmed this, concluding that mixed biofilters reached a superior EC when compared to fungal or bacterial biofilters. The combined biofilter also showed better resilience against starvation, which is a great quality in a low maintenance biofilter (Y. Cheng et al. 2016).

    As mentioned earlier, there are a great amount of microorganisms known to have the capacity of VOC degradation and mineralisation. Some bacterial and fungal species able to degrade BTEX (benzene, toluene, ethylbenzene and xylene) are listed in Table 4 and Table 5, respectively. It is also important to consider the pathogenicity of the species, as these could pose a huge health hazard in an open system such as a biofilter. It is import to stress that these species need to be able to compete in the ecological environment of the biofilter, as the capacity of the newly introduced organism to thrive depends on their ability to occupy an ecological niche in the biofilter. We took a look at the BTEX degradation capacities for some bacteria and fungi and by combining information from different scientific articles, we were able to compose the following tables (Jorio & Heitz 1999; Delhoménie 2005; Aranda et al. 2010).

    Table 4 - Some bacterial species capable of BTEX degradation.

    Table 5 - Some fungal species capable of BTEX degradation.

    Microorganisms that can degrade multiple compounds are particularly interesting with respect to the goals of this design. Only 9 species qualify for this; 2 Bacteria and 7 Fungi. The following paragraphs provide a short description of these microorganisms.

    Pseudomonas pseudoalcaligenes and Pseudomonas putida are Gram-negative, aerobic Gammaproteobacteria. The Pseudomonas genus is widespread through nature because of its diverse metabolism. P. putida, in specific, is a saprotrophic soil bacterium. It’s degradation abilities have already been used in bioremediation (e.g. oil spills) (Phi Doan et al. 2016). The main advantage, aside from its degradation abilities, is that P. putida is a very safe bacterium compared to some other Pseudomonas species (de Castro et al. 2010). Therefore, it is exceptionally interesting. P. pseudoalcaligenes is a less effective BTEX degrader and it also belongs to the Pseudomonas aeruginosa group and has been reported as a rare opportunistic nosocomial human pathogen (Hage et al. 2013).

    Cladosporium resinae and Cladosporium sphaerospermum are common moulds from the Ascomycota. Cladosporium is mainly found on living and dead plant material. They are not often pathogenic to humans, but have been reported to cause infections (Sosa et al. 2012). The spores are significant allergens that can affect people with respiratory diseases (e.g. asthma) (Tham et al. 2017). A major disadvantage is that Cladosporium has been associated with odour problems due to the fact it produces VOCs on its own (Micheluz et al. 2016).

    Exophiala lecanii-corni is an ascomycete. It has been implicated in causing cutaneous phaeohyphomycosis (tinea nigra) (Liou et al. 2002).

    Paecilomyces variotii is a widespread mold belonging to the Ascomycota. It is often found on substrates such as soil, wood and food products. P. variottii is an uncommon pathogen for humans, although it can act as an opportunistic pathogen in immunocompromised individuals (Steiner et al. 2013).

    Phanerochaete chrysosporium is a saprophytic crust fungus from the Basidiomycota. It is the model white rot fungus because it has the ability to degrade lignin, while leaving the white cellulose nearly untouched. Because of this, P. chrysosporium is pathogenic to dying plants, but it is not a known pathogen of humans or animals. Infected plants can be recognized by the white patches of cellulose due to the disappearance of lignin. Because of P. chrysoporium’s specialized degradation abilities, extensive research has been conducted in order to enhance the bioremediation of a diverse range of pollutants. As a result, P. chrysosporium is the first member of the Basidiomycota to have its complete genome sequenced (Ulmer et al. 1983; Fulekar et al. 2013; Martinez et al. 2004).

    Trametes versicolor is an edible saprotrophic mushroom from the Basidiomycota. One of its metabolites, polysaccharide-K, is used in conjunction with chemotherapy to improve the treatment of various cancers (Fisher & Yang 2002).

    Trichosporon beigelei is a yeast that belongs to the Basidiomycota. It is an opportunistic pathogen that is associated with trichosporonosis in immunocompromised individuals and with white piedra (tinea blanca) (Colombo et al. 2011; Walzman & Leeming 1989).

    Safety is an important criterion. Not all of the 9 species above are safe to use outside of specialized laboratories. Some of them are pathogens or they produce allergens. This means that only Pseudomonas putida, Phanerochaete chrysosporium and Trametes versicolor are left to be used in this design. Since the research on T. versicolor is still rudimental and there has already been conducted a lot of research on P. putida and P. chrysosporium, this project might focus on the implementation of the latter two. Additionally, both organisms can degrade the whole BTEX range.

  • Molecular Mechanisms of VOC Degradation.

    In this part, the molecular mechanisms of VOC degradation by Pseudomonas putida and Phanerochaete chrysosporium will be discussed, in particular the degradation of toluene. BTEX biodegradation in Pseudomonas putida relies on the formation of Krebs cycle intermediates. The cell invests energy and oxygen in exchange for these energy carrying molecules, which are essential for growth. When toluene enters the cell, 2 different pathways can be activated: the TOL pathway and the tod pathway.

    Firstly, TOL plasmid is an umbrella term for a collection of toluene degrading plasmids, the archetype being pWW0. pWW0 is a self-transmissible plasmid that belongs to incompatibility group P-9. It is a very large plasmid, about 117kb in size. Approximately 40kb is needed for the catabolic pathway and the regulatory genes, the remainder is used for conjugation and replication among other things (Burlage et al. 1989).

    As can be seen in Figure 2, the TOL pathway has 3 components: the upper, meta and ortho pathway. The upper pathway is responsible for the initial degradation of toluene to benzoate. The resulting benzoate can go through either the meta-cleavage pathway or the orthocleavage pathway. These latter pathways can convert benzoate to Krebs cycle intermediates via a catechol intermediate. In case of the meta-cleavage pathway, acetaldehyde and pyruvate will be formed. In the case of the ortho-cleavage pathway, succinate and acetyl-CoA will be formed (Burlage et al. 1989). The big difference between the meta-cleavage pathway and the ortho-cleavage pathway is that the ortho-cleavage pathway is regulated by the chromosome, while the meta-cleavage pathway is regulated by the plasmid itself. This means that strains with an inactive meta-pathway can still degrade benzoate, but they will use the ortho-cleavage pathway and thus they will form succinate and acetyl-CoA (Panke et al. 1998).

    Figure 2 - Biodegradation of toluene by P. putida through alternative meta or ortho lower pathways (Panke et al. 1998).

    Secondly, there exists a second pathway: the tod pathway. When toluene is sensed by the TodS/TodT two-component regulatory system, this activates the tod operon (Koh et al. 2016). Most notable in this operon are the todA, todB and todC genes. The gene products of these three structural genes form the toluene dioxygenase enzyme complex. This enzyme is responsible for the oxidation of toluene to cis-toluene dihydrodiol. After this, cis-toluene dihydrodiol dehydrogenase (expressed by todD) converts cis-toluene dihydrodiol to 3- methylcatechol (Zylstra et al. 1988). 3-methylcatechol is then further transformed by metacleavage (Zylstra et al. 1988) to pyruvate and acetyl-CoA by a series of enzymes that are not discussed here (Cho et al. 2000). Much like the TOL’s ortho-cleavage pathway, the tod pathway is regulated by chromosomal housekeeping genes.

    Figure 3 – Toluene oxidation (Zylstra et al. 1988).

    Figure 4 - Tod pathway (Zylstra et al. 1988).

    BTEX degradation by P.chrysosporium is not yet well understood. It is reported that P.chrysosporium mineralizes BTEX compounds to CO2 under nonligninolytic conditions. Under these conditions, P.chrysosporium will not produce extracellular lignin peroxidases or manganese-dependent peroxidases (used for lignin degradation). So that means that these enzymes are not involved in the process (Yadav & Reddy 1993)



Fusion Protein

To attach our microorganisms to Aerolis's surface, biotin has been integrated in the filament. Hence, our microorganisms must display streptavidin (mSA2 protein) on the outer membrane surface for effective binding to the 3D-shape. In addition, to enhance condensation, it should also express INP. For this purpose, two constructs were made:

  1. A custom made fusion protein of streptavidin with the membrane binding region of INP (INP_NC-mSA2).

  2. The wild type InaZ encoding INP (INP).


In this approach, E. coli will bind to the biotin on the surface of the 3D printed structure using the membrane-bound streptavidin. The WT InaZ (containing the membrane binding domain) can then perform its normal nucleation function.

Several control constructs were used to test the function of the different parts:

  1. A custom made fusion protein of GFP with the membrane binding region of INP (INP_NC-mGFPuv2). This allows us to visually assess whether the membrane binding region of InaZ functions properly.

  2. A fusion protein of the membrane binding domain of InaZ with GFP and streptavidin (INP_NC-mGFPuv2-mSA2). With this construct we can confirm that the membrane binding domain doesn’t influence the streptavidin binding to biotin.


Two aspects of a truncated INP fusion protein can be analyzed for normal functionality.

  1. Protein folding: to visualize correct folding of the protein we also fused, instead of the mSA2 protein, the mGFPuv2 protein to INP_NC.
  2. Ice Nucleating capability:
    • a. Cells were grown overnight in LB medium at 37°C.
    • b. Deionized water was super-cooled to approximately -10°C in a cooled 50% glycerol water bath.
    • c. The capability of ice nucleation was tested in two ways:
      • i. In the same tube, first 50 µL of lysate containing mGFPuv2 was added and subsequently 50 µL of one of the lysates with an INP(NC)-fusion protein.
      • ii. To prove the ice nucleation reaction is not just caused by increasing the volume from 50 to 100 µL with any lysate, 50 µL of the lysate containing mGFPuv2 was also added multiple times to the same tube.

The fusion with mGFPuv2 exhibits fluorescence and both the lysates with INP_NC-mSA2 and mGFPuv2 were active as a nucleus for ice formation.


Figure 5 - Left tube: INP_NC-mSA2; Right tube: INP_NC-mGFPuv2.

  1. DeLeon-Rodriguez, N. et al. Microbiome of the upper troposphere: Species composition and prevalence, effects of tropical storms, and atmospheric implications. Proc. Natl. Acad. Sci. 110, 2575–2580 (2013).
  2. Pratt, K. A. et al. In situ detection of biological particles in cloud ice-crystals. Nat. Geosci. 2, 398–401 (2009).
  3. Lundheim, R. Physiological and ecological significance of biological ice nucleators. Philos. Trans. R. Soc. B Biol. Sci. 357, 937–943 (2002).
  4. von Stetten, D., Noirclerc-Savoye, M., Goedhart, J., Gadella, T. W. J. & Royant, A. Structure of a fluorescent protein from Aequorea victoria bearing the obligate-monomer mutation A206K. Acta Crystallogr. Sect. F. Struct. Biol. Cryst. Commun. 68, 878–82 (2012).
  5. Ito, Y., Suzuki, M. & Husimi, Y. A novel mutant of green fluorescent protein with enhanced sensitivity for microanalysis at 488 nm excitation. Biochem. Biophys. Res. Commun. 264, 556–60 (1999).


Two different approaches were tested to biologically publicize the biotin on the PLA filament:

Next, we used fluorescent protein visualization to demonstrate that microorganisms displaying streptavidin (mSA2) are attached to a PLA surface by taking advantage of the strong biotin-streptavidin binding.


In this first approach, we tried to impregnate the biotin in the PLA. Therefore, we looked for a chemical condition in which a vast amount of PLA and biotin could be dissolved without changing or destroying their integrity. This condition was obtained by heating biotin-saturated dimethylformamide (DMF) to 130°C and dissolving PLA in it.

experimental setup

Figure 6 - Experimental setup for the impregnation of biotin.

After thoroughly mixing the dissolved PLA and biotin, the PLA (with biotin) can be crashed out of solution by adding the solution to an excess of ethanol, saturated with biotin at room-temperature.

In this way we obtained biotin that is completely captured by PLA in a nicely dispersed fashion. Moreover, it’s most likely that the biotin will remain in position and interwoven in the PLA fibers, even in circumstances of fog, dew, water, ...

Next, the obtained crashed out and dried PLA can be supplied to an extruder that will generate a filament which can subsequently be used as feed for a 3D printer.

Figure 7 - The set-up for the extruder.

Figure 8 - The final generated filament.


In this second approach, we made some kind of biotin-lacquer which can be painted on the filament and even on already printed structures. This lacquer was created by over-saturating a solvent with PLA and biotin, that classically dissolves PLA. After a simple trial and error essay, dichloromethane was selected as the optimal solvent, due to its fast working method of action and easy preparation.


Figure 9 - By simply submerging a PLA fragment in this solution, a biotin-PLA coating will be applied to the surface of the filament.

Proof of Concept

To prove the microorganisms do indeed stick to the filament, one option is to make fusion proteins containing both streptavidin and a fluorescent protein (INP_NC-mGFPuv2-mSA2).


This was tested in 2 different setups. On one hand the biotin impregnated filament was used and on the other hand glass slides coated with PLA and biotin:

In the following picture (A), the filaments labeled with ‘10’ are coated with the mSA2-mGFUuv2 protein and filaments labeled “11” with mGFPuv2. Under UV light, there was no green color visible for either protein coatings. In the picture B, the same lysates are applied to PLA + biotin coated glass slides. Here we clearly see under UV light that the mSA2-mGFPuv2 (left slide) sticks to the surface while the mGFPuv2 does not (right slide). This tells us that making fusion proteins which include a fluorescent protein can work in some cases to prove the presence of a protein of interest, but clearly the method was not sensitive enough in case of our filament and creating fusion proteins to prove the binding of a protein can be very labor intensive.


Figure 10 - Comparison between mSA2-mGFUuv2-coating and mGFPuv2-coating on filaments and glass slides.

  1. Ben‐Shabat, S., Kumar, N., & Domb, A. J. (2006). PEG‐PLA Block Copolymer as Potential Drug Carrier: Preparation and Characterization. Macromolecular bioscience, 6(12), 1019-1025.
  2. Li, D., Frey, M. W., Vynias, D., & Baeumner, A. J. (2007). Availability of biotin incorporated in electrospun PLA fibers for streptavidin binding. Polymer, 48(21), 6340-6347.
  3. Salem, A. K., Cannizzaro, S. M., Davies, M. C., Tendler, S. J. B., Roberts, C. J., Williams, P. M., & Shakesheff, K. M. (2001). Synthesis and characterisation of a degradable poly (lactic acid)-poly (ethylene glycol) copolymer with biotinylated end groups. Biomacromolecules, 2(2), 575-580.
  4. Weiss, B., Schneider, M., Muys, L., Taetz, S., Neumann, D., Schaefer, U. F., & Lehr, C. M. (2007). Coupling of biotin-(poly (ethylene glycol)) amine to poly (D, L-lactide-co-glycolide) nanoparticles for versatile surface modification. Bioconjugate chemistry, 18(4), 1087-1094.


As mentioned before, the scope of the sensor consists of two parts: the quantitative measurement and report of the air quality and the visualization of air quality by means of an LED that changes colour. The selected sensor system can run on batteries and the goal is to achieve this by implementing a source of renewable energy into the design, for example by using a small-scale solar panel as an external power supply.


This sensor system will be implemented by using Arduino, an inexpensive open-source electronics platform based on easy-to-use hardware and software. Arduino boards are able to read inputs (sensors, buttons, knobs, etc.) and turn them into an output (motors, LED’s, etc.). This is done by sending a set of instructions to the microcontroller on the board. The instructions are written in the Arduino programming language, a dialect of features from the C and C++ programming languages.

A joint effort by Seeed Studio and Arduino has resulted in the development of many Arduino derived Seeeduino boards. For example, the Seeeduino Lotus board used in this project is an ATMEGA328 microcontroller developed by Seeed Studio that is very similar to the Arduino UNO board. Another development by Seeedstudio is the Grove System. This is a modular, standardized connector prototyping system that is ideally suited for connecting sensors, actuators, displays and communication modules. A wide range of sensors is also offered by Seeedstudio. For the purpose of this project, the Grove Air Quality Sensor v1.3 is used.

The Grove Air Quality Sensor is responsive to a wide scope of harmful gases, as carbon monoxide, alcohol, acetone, thinner, formaldehyde and so on. Due to the measuring mechanism, this sensor can't output specific data to describe target gases' concentrations quantitatively. But it's still competent enough to be used in applications that require only qualitative results.

Hardware Setup

The Seeeduino Lotus is equipped with on-board Grove connectors that will be used to plug in the Grove Air Quality Sensor. Additionally, a common cathode LED was installed on a breadboard. It was connected to the ground (GND) pin and to three of the PWM1 pins of the Seeeduino Lotus. This was done by connecting the leads to the pins with jumper wires through 220 Ω resistors.

The next step was the installation of the Arduino Integrated Development Environment (IDE) on a personal computer. The Seeeduino Lotus Driver was then installed to achieve compatibility between the Seeeduino Lotus board and the Arduino IDE. Once these tasks were completed, the Seeeduino AVR Boards were added to the Arduino IDE Board Manager and the Seeeduino Lotus was selected and connected to serial port COM3.

arduino board setup

Figure 11 - Schematic hardware setup for an LED-equipped Arduino board.

seeeduino board setup

Figure 12 - Real-life hardware setup with LED and gas sensor functionality.

Software Setup

In order to test the hardware setup, an open-source AirQuality_Sensor Library demo version was uploaded to the Seeeduino Board. After successful testing of the Grove Air Quality Sensor, the aforementioned library was stripped down to its roots and modified to contain RGB LED functionality. This resulted in three files:

  1. The Project.ino sketch: this is the main file, it contains all of the instructions to make the sensor do its job. It functions by creating the “sensor” object in the “Lotus” class. Further on it declares parameters such as the pin locations of the sensor and the LED’s, the time between measurements and also the calibration parameters. After declaration, the sketch goes into the setup part to initialize the sensor and the serial connection. Then it enters the loop part, this part consists of continuously measuring and visualizing the air quality data.
  2. The Lotus.h header file: this file defines the “Lotus” class. It contains all variables used by the functions in the Lotus.cpp file and all information on their accessibility. Public variables can be accessed from the sketch, private variables can only be accessed from within the class itself. A header file lists everything that’s inside the class.
  3. The Lotus.cpp source file: this file contains all of the functions defined in the “Lotus” class. The functions are called from the Project.ino sketch.

The Arduino.h header file is by default included in the library. This header gives the other library files access to the standard Arduino functions (e.g. pinMode, AnalogRead, AnalogWrite, delay, map). What this entire program does, is displaying text and values on the serial monitor and at the same time, translating these values to colours on an RGB LED.

All software developments can be followed on the Aerolis GitHub repository, where we are currently working to extend the Seeeduino's functionality to include datalogging via an Ethernet Shield. All of this is done with the final goal of making data from a network of different sensor devices available via an Android or iOS application.


For the calibration of the sensor, a number of locations were sampled. The output voltage of the sensor was measured on every location. From these data, it was possible to define which voltages could be considered as clean air and which voltages should be considered as polluted air. The sampling results can be consulted in Table 6 and Table 7.

Table 6 - Outdoor sampling results.
Outdoor Locations Mapped voltages (from 0-5 volts to 0-1023) AQI (from 0 to 100)
Forest in Buggenhout, Belgium 32 97
Fresh air in Lebbeke, Belgium 34 93
Fresh air in Gent, Belgium 37 88
Outdoor trash 42 80
Pure exhaust gases 500 to 600 -683 to -850
Ventilation shaft at Vrijdagmarkt (Gent, BE) N/A 70
Urban air at Sint-Baafsplein (Gent, BE) N/A 80

Table 7 - Indoor sampling results.
Indoor Locations Mapped voltages (from 0-5 volts to 0-1023) AQI (from 0 to 100)
Ventilated bedroom 47 72
Unventilated bedroom 76 23
Ventilated kitchen 55 58
Kitchen (while cooking) 104 -23
Refrigerator 72 30
Trash bin 63 45
Bathroom 43 78
Unused toilet 44 77
Used toilet 93 -5
Storage room 45 75

After the sampling, it was decided that the lowest pollution environments had a mapped voltage of 30 and that the environments with the highest pollution had a mapped voltage of 90. Since it is desired to have the sensor calibrated to values between 0 for the highest pollution and 100 for the lowest pollution, the following formula was used to convert the obtained sensor voltage to a more meaningful air quality index:

AQI formula

Where AQI is the air quality index, V is the mapped voltage measured by the sensor, L is the mapped voltage of the reference lowest pollution and U is the mapped voltage of the reference highest pollution.

For the RGB LED calibration, some hardware issues had to be taken into account. Differences in the resistors resulted in a different brightness for each colour at full output voltage. This had to be compensated by programming a lower output voltage for colours with high brightness.

3D Design

The 3D modeling of Aerolis was accomplished with 3D computer graphics software toolsets SideFX Houdini and Blender. This kind of software is capable of creating designs in a procedural way and ultimately delivering a high-quality STL-file of the model, which is required for 3D printing it.

Sample 3D modeling workflow in Houdini.

The prototypes, cross-sections and details were 3D printed on an Ultimaker 2+, a high-end consumer-grade printer that makes use of the fused filament fabrication (FFF) technology. For the final model, we have collaborated with Materialise, a specialized 3D printing firm who are able to provide us with a high-quality and well-finished 3D print, due to their expertise in selective laser sintering (SLS).

Lab Safety

Our experimental work is done in a GMO class 2 laboratory at the Department Biochemical and Microbial Technology at Ghent University.

The Faculty of Bioscience Engineering at Ghent University, where our lab is situated, has its own facultary Biosafety Committee (only Dutch version of the website available). This committee makes sure the researchers are aware of the International, European and Belgian legal framework concerning biosafety. All activities of this committee take place in close consultation with, amongst others, the facultary Environment, Hygiene and Safety Committee, the facultary Ethics Committee and with some Rectoral Services (only accessible from within the UGent intranet).

In addition we have committed ourselves to both the national legal and ethical requirements compliance and the international and EU regulation compliance:

Every person who works in one of the labs of Ghent University must know and act according to certain regulations. This means that each team member signed the Laboratory and workplace regulation in which he/she declares to have taken note of these regulations, to have received a specimen of these regulations and to follow the guidelines/obligations which arise from these regulations. This also includes always wearing a lab coat, safety glasses and gloves when necessary. Guidelines more specific for working with GMOs can be found in the Biosafety in the laboratory document provided by the Flanders Interuniversity Institute for Biotechnology.

Our lab has its own Lab Manager, Gilles Velghe, who coordinates everything in the lab and is responsible for running the laboratory safely and efficiently. In addition, every member of our team got a (bio)safety and waste disposal training before performing our experiments in the lab. This training focused especially on the safe use of ethidium bromide, which we use as dye for DNA detection in gels. During the eventual execution of our experiments, we were supervised by our instructors who made sure we performed everything according to the safety regulations. In our lab, all glass work that has been in contact with biological material is disinfected with bleach and autoclaved before it is re-used, making sure no living micro-organisms are remaining. When applying bleach, we wear gloves as it can cause skin irritation. Also, benches and flows are regularly cleaned with Umonium, a broad activity spectrum disinfectant. Additionally any cultures that leave the lab are tested, in order to prevent contamination of other external laboratories.

Biosafety form.

Lab manager Gilles Velghe.