Phase I. Design and Evaluation of Prototypes in the Laboratory

The features of the biosensor prototypes were selected in accordance with annotations made by ^{Valentín (2015)}. The size of the prototype support was 4 x 4 cm, and three types of material support were chosen: linen, crepe paper, and goose feathers based on the bibliography consulted and the ease of obtaining them ^{(Valentín, 2015, p. 348}; ^{Urkullu, 2001, p. 58}; ^{Assis et al., 2020, p. 5983)}.

We chose three types of culture medium: Sabouraud broth, which is widely used to recover filamentous fungi due to its composition of peptones and glucose ^{(Guinea, Peláez, Alcalá & Bouza 2005, pp. 333-334)}; DG18 broth (dicloran-chloramphenicol glycerol), with a low availability of water (Aw), that favors the growth of xerophilous microorganisms, such as certain filamentous fungi reported in museum and library environments ^{(Manrique, Patiño & Gutiérrez 2012, pp. 5-6)}, and lastly, nutrient broth as a medium used to grow microorganisms with low nutritional requirements, including fungi and bacteria.

Another characteristic of the biosensor prototype design was a pH indicator; four were proposed with the criteria of turning from neutral to slightly alkaline in a range of 7-8 or slightly acidic in a 6-6.5 pH range ^{(Martín & Villegas, 2021, p. 111)}. The chosen indicators were bromothymol blue, bromocresol purple, litmus blue, and methyl red. An experiment in test tubes was conducted for all the indicators in order to identify the most appropriate. Each test tube had 3 ml of culture medium with one of the four indicators, to which were added 0.5 ml of a solution of Penicillium sp. conidia containing 4.8 x 10.5 (10^5-10⁵) conidia/ml. The test tubes were incubated for a week at 25° C. The characteristics of the previously selected biosensors were combined, as seen in Figure 1.

Prepared in triplicate, in addition to a control, the linen, paper, and feather supports were sterilized in the autoclave, along with culture mediums with and without indicators, and the Petri dishes (121° C, 17 psi). They were subsequently oven dried at 70° C for 40 minutes, until the liquid culture medium had dehydrated (Figure 2).

(Photograph: Lizeth P. Russy-Velandia, September 10, 2021)

Figure 2 Design of the biosensor prototypes in Petri dishes a) Biosensor Prototype LINEN:DG18, b) Biosensor Prototype PAPER:CS, c) Biosensor Prototype FEATHER:DG18 

We prepared a total of 108 biosensor prototypes for use in the Archivo Central de la Universidad Militar Nueva Granada (UMNG) in the Cajica, Cundinamarca campus. They were all placed in the center of the storage area, approximately one meter above ground, except for the control; the Petri dishes containing biosensors were exposed to the environment for one hour, so the microorganisms suspended in the air could be sedimented onto the supports. After the period of exposure, the Petri dishes were closed and taken back to the laboratory where they were separated into two sets of 54 prototypes each to begin the tests under different conditions of temperature and relative humidity.

The first set was exposed in a curing chamber in the UMNG Materials Laboratory under constant conditions of 90% HR and 23° C for a period of four weeks, with weekly follow-up to photographically register and quantify the colony forming units (CFU) in the prototypes. The second was placed on a table in the UMNG Multiple Laboratory I at ambient temperature and relative humidity. The monthly average of these environmental variables was obtained by placing the prototypes close to each other in an Onset HOBO Datalogger; the same photographic follow-up was carried out over a period of four weeks.

Following this, a qualitative and quantitative description was made of the colonies that grew on the prototypes in each set. In the case of filamentous fungi, morphotypes and even genus were identified by means of tinction with lactophenol blue, observation under the 40X optic microscope, and follow-up of taxonomic codes ^{(Barnett & Hunter, 1998, pp. 6-197}; ^{Franco et al., 2012, pp. 137,138, 184 and 434-437)}. The prototypes from both sets that did not present CFU growth after four weeks were rehydrated with a saline solution of 0.9% p/v ^{(Valentín, 2015, p. 348)} and incubated at 25° C to identify whether there were spores or conidia which had been unable to develop in the conditions they were under.

The best biosensor prototypes were selected based on three criteria: 1) time of appearance of the CFU of filamentous fungi over the course of four weeks, registering the first time colony growth had been observed on the biosensors with the formula:

CFU_H being the number of colonies counted on the biosensor; 2) total CFU of filamentous fungi; and 3) diversity of filamentous fungi in relation to the different number of isolated morphologies for each support. The frequency of appearance (F_a) was calculated over the total number of isolated filamentous fungi in each prototype with the formula:

where N° CFU_Morphotypes is the number of times a certain morphotype appeared on the same support during the identification process, and the Total CFU_HS is the type of CFU that grew on each support. The results were analyzed by means of descriptive statistics to compare those parameters in the evaluated prototypes.

Phase II. Evaluation of the Prototypes in Environments of Collection Storage

The prototypes that showed the least time for the appearance of CFU in Phase I, a significant number of accumulated CFU, and a greater diversity of morphologies were prepared once again and, along with a control, placed in triplicate next to a Datalogger in the storage areas of selected collections in the cities of Bogota, Mede llin, and the municipality of Cajica, Colombia. The biosensors were deployed over several weeks, with a weekly photographic registry to monitor the growth of CFU and filamentous fungi. The prototypes that showed no growth of CFU after four weeks were taken back to the laboratory, rehydrated with a saline solution of 0.9% p/v ^{(Valentín, 2015, p. 348)}, and incubated at 25° C for two weeks. The results obtained were correlated with the variables of relative humidity and temperature (RH and T°) in each of the storage areas to analyze the prototypes’ performance.

Phase I. Design and Evaluation of the Prototypes in the Laboratory

Time of Appearance and Quantity of CFU/cm² of Filamentous Fungi

As mentioned earlier, these criteria were only evaluated for the first set of prototypes that had microorganism growth. Through photographic follow-up over four weeks, we obtained an average of growth replicas of CFU /16 cm²; that is to say, for the total area of supports (4 x 4 cm²) exposed to 90% RH and 23° C, we calculated the standard deviation (Figure 3), which allowed us to see how scattered our data was around the average value of the three tests ^{(Lee et al., 2015, pp. 221-222)}.

*CS:Sabouraud broth, CN:nutrient broth, DG18:dichloran-glycerol chloramphenicol broth, ABT:bromothymol blue

(Table: Lizeth P. Russy-Velandia, April 15, 2022)

Figure 3 Standard deviations obtained from the average of CFU/16 cm² repetitions, with each biosensor prototype under conditions of 90% RH-23° C in the curing chamber. 

High values for standard deviation in certain prototypes indicate that, in most cases, the data extends to a wider range with regard to the median ^{(Barde & Barde, 2012, p. 113)}, which could be linked to the positioning of prototypes on shelves during the test in the curing chamber and its interaction with the surrounding water vapor.

Keeping in mind the criteria of time of appearance of CFU, quantity of CFU /cm² of filamentous fungi so as to select the best biosensors, we identified that prototypes made of PAPER:CS, FEATHER:CS, and FEATHER:CS:ABT had the shortest time for CFU appearance (as of the first week) and showed the highest accumulated quantity of CFU/cm² of filamentous fungi by the end of the fourth week (Figure 4).

*CS:Sabouraud broth, CN:nutrient broth, DG18: dichloran-glycerol chloramphenicol broth, ABT: bromothymol blue

(Table: Lizeth P. Russy-Velandia, May 18, 2022)

Figure 4 Number of CFU/cm² for each biosensor prototype over the four-week test period under conditions of 90% RH-23° C in the curing machine. 

In the case of prototypes made with a linen support, we observed a growth of 0.02 CFU/cm² during the first week in the Sabouraud culture medium; however, by the end of the test, there was greater isolation of filamentous fungi in the DG18 medium, indicating that although it takes a week longer to start detecting the microorganism, it is possible to isolate xerophile microorganisms that tend to grow with low water requirements. Based on the analysis presented above, the best biosensors were PAPER:CS, FEATHER:CS, QQ total of 18 morphotypes with the most frequent being Cladosporium sp. morphotype No. 1 with 22.2% (Figure 5).

(Table: Lizeth P. Russy-Velandia, May 30, 2022)

Figure 5 Percentage of frequency of morphotypes identified on the linen support. 

The prototypes with the greatest diversity of morphotypes on a linen support were the prototypes LINEN:CS and LINEN:CS:ABT, with four morphotypes each, and LINEN:DG18 and LINEN:DG18:ABT, presenting three and four morphotypes (Figure 6). Based on the selection criteria, it was determined that the prototype with the greatest detection capacity for the linen support was LINEN:DG18, with a cumulative average of 1.83 CFU/cm². Although the appearance of filamentous fungi was slowin the first week, by the end of exposure in optimal conditions it produced the highest number of colony growth (Figure 7).

(Table: Lizeth P. Russy-Velandia, June 21, 2022)

Figure 6 Presence of morphotypes on the linen support according to the three culture mediums, with and without indicator. 

(Photograph: Lizeth P. Russy-Velandia, June 2, 2022)

Figure 7 a) Biosensor Prototype LINEN:DG18, b) Support of the same prototype sown in culture medium dicloran-glycerol chloramphenicol: the growth of Penicillium sp. morphotype No. 2 can be observed. 

Regarding biosensors made of paper, a total of 22 morphotypes were obtained, with the most frequent being Cladosporium morphotype No. 1 (27.3%) and Penicillium sp. morphotype No. 1 green (18.2%) (Figure 8).

(Table: Lizeth P. Russy-Velandia, June 23, 2022)

Figure 8 Percentage of frequency of morphotypes identified on the paper support. 

The greatest diversity of morphotypes were found on prototypes made of crepe paper (PAPER:CS, with six morphotypes) (Figure 9). Based on the selection criteria, it was determined that among the crepe paper supports, the prototype with the greatest detection capacity was PAPER:CS, with a cumulative of 2.31 CFU/cm² and a brief time before the appearance of CFU, even during the first week, showing a greater cumulative growth of fungi colonies when compared to other prototypes at the end of the test.

(Table: Lizeth P. Russy-Velandia, June 23, 2022)

Figure 9 Presence of morphotypes on the paper support, according to the three culture mediums, with and without indicator. 

The prototypes made of goose feathers (Figure 10) presented the following frequency over a total of 21 morphotypes: 23.8% Cladosporium morphotype No. 1, 23.8% Aspergillus flavus, and 19.0% Aspergillus niger.

(Table: Lizeth P. Russy-Velandia, May 30, 2022)

Figure 10 Percentage of frequency of morphotypes identified on the feather support. 

The greatest diversity of morphotypes were found on prototypes FEATHER:CS and FEATHER:CS:ABT, with five morphotypes each (Figure 11). Based on the selection criteria, it was determined that the goose feather support prototype with the best detection capacity was FEATHER:CS, with a cumulative of 1.25 CFU/cm² and rapid detection of colonies of filamentous fungi as of the first week of the test.

(Table: Lizeth P. Russy-Velandia, June 25, 2022)

Figure 11 Presence of morphotypes on the feather support, according to the three culture mediums, with and without indicator. 

During the course of the tests, it was concluded that the pH indicator was not crucial for making biosensors since the results obtained were very similar with or without it and did not affect the growth of CFU/cm², while the changes of color with the bromoethanol blue indicator were imperceptible.

As for the results obtained in the curing chamber, the biosensor prototypes made with nutrient culture isolated eight morphotypes of bacteria. Given that these were not visible to the naked eye, the growth of bacteria could only be observed once the supports were transferred and planted in Petri dishes with the same culture medium used to create the biosensors and incubated in the laboratory at 37° C for a week; hence they were not taken into consideration for the study.

Phase II. Evaluation of Prototypes in Environments of Collection Storage

The same biosensor prototypes were set out in two different Colombian cities in order to compare them. Bogota has an ambient temperature range between 14.5 and 22.5° C and a percentual relative humidity between 77 and 83 points ^{(Natarajan, Rodríguez & Vellei., 2015, p. 244)}, while Medellin presents a range between 23.3 and 27.2° C and 62.6% RH ^{(Restrepo-Betancur, Peña-Serna & Martínez-González, 2019, p. 315)}. Figure 12 summarizes the locations of the biosensors, the type of collection stored, as well as the average monthly temperature and relative humidity.

The prototype placed in the Colonial Museum of Bogota did not show any detection of CFU or filamentous fungi. The same occurred with the prototypes exposed in the UMNG, a result that is linked to the temperature and relative humidity registered in the three storage environments (Figure 12): on the one hand, in every collection, it does not go above 20° C, and although average monthly relative humidity can reach 62% in the José Ricardo Cure Hakim Zoological Collection, no microorganism growth was detected there either. In the city of Medellin, the registered average monthly temperature range was between 22.7 and 19.5° C, with a relative humidity of 58.7-57.3% (Figure 12); that is to say, quantities under 60%, which prevented the rehydration of the culture medium in those environmental conditions.

Since it did not display growth of CFU of filamentous fungi in any of the storage environments located in the different cities, the prototypes were taken to the laboratory to be rehydrated and incubated for a week at 25° C; thus, we proved that the prototypes had sedimented spores which did not grow until they were placed in optimal temperature and humidity conditions. Fungi of the genera Cladosporium, Penicillium, and Aspergillus grew, which are typically reported as biological contaminants in indoor sites ^{(Khan & Karuppayil, 2012, pp. 406-414}; ^{Cepeda et al., 2019, pp. 42-44)}. The microorganisms found in those environments and recovered from the biosensors match the results obtained by ^{Valentín (2015, p. 349)}, who reported contamination by environmental micro organisms corresponding to Cladosporium sp. and Penicillium chrysogenum (2017, p. 349) as well as by ^{Borrego, Herrera, and Paneque and Penicillium in storage environments (2021, pp. 5-6)}, who reported the predominance of Aspergillus, Cladosporium.

The reported optimal conditions for the growth of microorganisms are between 25 and 30° C and 70 to 75% RH ^{(Mallo et al., 2017, p. 621)}; however, none of the storage environments in this study reached an average monthly temperature over 25° C, and the highest percentage of relative humidity was 62 points. When environmental conditions are not favorable, the microorganisms’ capacity to begin reproductive processes diminishes ^{(Mallo et al., 2017, p. 621)}. These results concur with those obtained in Phase I of the tests.