Bacterial contamination was performed as previously described [8, 13, 14]. At first four different membranes (1. Argonaut® porcine pericardium collagen membrane; Mem-Lok® RCM porcine collagen membrane; Mem-Lok® Pliable bovine Type I collagen membrane; PermaPro® synthetic non resorbable PTFE membranes; all Biohorizons biologics) that are clinically applied for GBR were cut into 5 × 5 mm pieces. At first, a sterile nutrient solution (10 ml Brain–Heart-Infusion Broth, Carl Rot, Karlsruhe, Germany) along with 100 µl of a bacterial culture of Enterococcus faecalis ATCC 29212 (Leibniz Institute DSMZ, Braunschweig, Germany) was incubated at 37 °C for 24 h Then day, a 10 µl sample of the incubated bacterial culture was transferred onto an agar plate (Sarstedt AG & Co. KG, Nümbrecht, Germany) The resulting E. faecalis colonies were then prepared for the bacterial contamination. All specimens were cultured with 2.5 ml of BHI broth and 2.5 µl of the E. faecalis overnight culture (1:1000 dilution). To ensure optimal bacterial growth, the plates incubated at 37 °C for 6 days in a shaker (Rotamax 120, Heidolph Instruments GmbH & CO. KG, Schwabach, Germany). BHI was changed every 2 days. All experiments were conducted using n = 4 specimen per experimental group. One specimen per experimental group was not colonized with bacteria and served as a negative control.

After bacterial colonization, specimens were washed using phosphate buffered saline and were then prepared for cold atmospheric plasma treatment. Specimen were placed in 6 well plates and then CAP (argon plasma) treatment was performed using a kINPen® MED (neoplas tools GmbH, Greifswald, Germany). One experimental group (n = 4) received 5 min CAP with 5 W and a second group (n = 4) received 10 min CAP. The time intervals were chosen based on a previously performed study, where a 5 min treatment interval reduced the bacterial load significantly by two third on collagen membranes [13]. A 10 min interval was additionally added in the present study to detect if further decontamination was possible with longer exposure to CAP. The distance from the device to the specimens was 1 cm (Fig. 1). A positive control group (n = 4) per membrane type received no treatment and served as positive control group.

Representation of plasma application using a kINPEN®

To quantify colony forming units of contaminated membrane types samples were washed with PBS and then placed in an ultrasonic bath (Branson 2210 R-MT Ultrasonic Cleaner, Branson Ultrasonics Corporation, Danbury/CT, USA) for 15 min to ensure bacterial detachment. Afterwards, a three-step dilution series was prepared for each membrane using SafeSeal reaction tubes (2 ml, Sarstedt AG & Co. KG, Nümbrecht, Germany). The first dilution (10⁻2) was prepared by adding 10 µl of the bacterial solution to 990 µl of sterile NaCl solution and thoroughly mixing it using a vortex shaker. The second dilution (10⁻3) was produced by mixing 100 µl of the first dilution with 900 µl of sterile NaCl solution. Finally, the third dilution (10⁻4) was made by mixing 10 µl of the second dilution with 900 µl of NaCl solution. From each dilution, 3 × 25 µl of the respective dilutions were applied to prepared BHI agar plates (Sarstedt AG & Co. KG, Nümbrecht, Germany). These plates were then incubated at 37 °C for 24 h in an incubator. After the incubation period, the colony-forming units were counted using a colony counter (BZG 25, Bender & Hobein AG, Zurich, Switzerland) to determine the number of viable, colony-forming bacteria.

In addition to CFU quantification, Live/Dead staining using a BacLight™ Bacterial Viability Kit (L7012) according to manufacture’s instructions. Briefly, Samples were washed in PBS and then placed on microscopy slides. To each 5 µl of dye mixture was applied. Then samples were covered with a cover slip. The prepared slides were then incubated in the dark at room temperature for 10 to 15 min and immediately prepared for fluorescence microscopy.

Bacterial colonization on all different membrane types was semi quantitatively evaluated using scanning electron microscopy (SEM). After incubation in a humified atmosphere of 37 °C at 5% CO2, the culture medium was removed and bacteria were fixed with glutaraldehyde (Sigma, St. Louis, USA) 3% in PBS at a pH value of 7.4 for 24 h. After removal of the glutaraldehyde solution, the samples were dehydrated in an ascending alcohol dilution for 300 s for each series. After that, drying with hexamethyldisilane for 1 min (Sigma-Aldrich, St. Louis, USA) and a gold vapor deposition with a thickness of 15 nm (SCD 500, CAL-Tec, Ashford, UK) were performed and SEM analysis (Jeol, Freising, Germany) was conducted at a voltage between 10 and 15 kV.

All experiments were conducted in quadruplicates (n = 4 per experimental group). Power analysis was performed with SPSS (IBM) to determine the smallest sample size that is suitable to detect the effect of a given test at the desired level of significance. All data were tested for normality using Shapiro–Wilk test. Statistical analysis of CFU data was performed using Graph Pad prism 7 program (San Diego, CA, USA). One-way analysis of variance (ANOVA) with Bonferroni’s multiple comparisons was used to compare means among the independent experimental groups. If there was no normal distribution, the Kruskal–Wallis test was used with Dunn’s multiple comparison. Differences were considered significant if p ≤ 0.05. Quantitative data in the text are presented as mean and standard deviation (SD).