Liquid smoke was produced in a pyrolysis reactor as explained in previous research ^{1, 16} . As much as 3 kg of dried durian peel was placed in the reactor set at 340°C and 380°C. The resulting smoke was condensed to produce tar, charcoal and grade 3 liquid smoke. The next step was distillation at 190°C to separate the liquid smoke from tar, resulting in production of grade 1 liquid smoke. The composition of liquid smoke grade 1 can be found elsewhere ¹ . The liquid smoke grade 1 mainly contains acetic acid, phenol and small amount of other compound such as ketones, aldehydes, and carboxylic acids. While the liquid smoke grade 3 could not be used as preservative because still contain a toxic compound such as tar ¹ . The liquid smoke grade 1 was then used as the preservative by soaking mackerel in different concentrations (0.5%, 1%, 2%, and 3%) for 60 hours (untill end of preservation time). The following analysis tests were conducted during storage: measurement of TVB, organoleptic quality, antibacterial activities and total number of bacterial counts. The analyses were conducted every six hours.

TVB-N is a method to determine the freshness of fish based on its spoilage due to microbial growth and loss of fats or proteins ²⁰ . The TVB-N was determined according the procedure established by Susanto et al. with a slight modfication ²¹ . TVB-N measurement was carried out by placing a Conway petri dish sideways with its lid half open inside an incubator at 35°C for 35 minutes. The petri dish contained the liquid smoke used to soak the fish, and K ₂BO ₃, and H ₃BO ₃ in each partition. After incubation, the dish was covered and shaken, before further incubation at 35°C for 8 hours. Afterwards, 0.1 ml boric acid was exposed to every indicator and left for 2 hours. Titrations with HCl (0.01 N) were performed until the color turned pink.

The organoleptic tests involved examining the samples using the senses of volunteer panelists, including examining the color, smell, taste and texture of the fish meat. These tests were carried out in order to identify how much people liked the liquid smoke preserved mackerel and to determine how long the fish would last. Testing was carried out in compliance with organoleptic testing standard manuals SNI 01-2346-2006 ²² . The number of panelists used in this research was 30 people, consisting of 23 non-standard subjects (people who were not trained in performing organoleptic assessment/testing, recruited from the pool of chemical engineering students at Syiah Kuala University) and seven standard subjects (people with high sensitivity towards testing product quality, and possessed knowledge and experience in assessing product quality, recruited from the Health Laboratory, Banda Aceh). The panelists were given briefing and training prior to performing the tests. The resulting values were then processed using hedonic tests.

Average quality value: x = ∑ x i n

Where x = average quality scores, x _i = value of organoleptic quality testing; panels i, and n = number of panelists

The statistical analysis of standard deviation was also performed on each data of the organoleptic test results for color, flavor, aroma and texture of fish during storage. For statistical analysis, One-Way ANOVA with Least Significant Different (LSD) test using the SPSS ver.22 for window was used. The authors presented the quantitative data as mean ± standard deviation (SD). Normally distributed quantitative data should be summarized as mean. Here, the SD refers to the variation in the values of the variable within the sample. The larger the SD, the greater the variability within the sample.

Flavor testing on fish was carried out after the samples had been steamed (at 90–100°C for 15 min) without changing the flavor. Table 1 describes the scale used to determine the flavor of fish samples.

The antibacterial activity testing was carried out to identify the activity of durian peel liquid smoke against Escherichia coli and Staphylococcus aureus bacteria. The method used was the disk diffusion (Kirby–Bauer) assay, as described by Tendencia ²³ , involving the use of Mueller Hinton (MH) media (Merck, KgaA, Germany), performed by the Health Laboratory (Banda Aceh, Indonesia). This method was chosen because of its simplicity and the ability to see the formation of inhibition zones as clear areas around the antimicrobial disks.

The total bacterial counts of mackerel was determined by Plate Count Agar (PCA) according to Indonesian standard manuals SNI 2897:2008 ²⁴ . The analysis was performed by the Health Laboratory (Banda Aceh, Indonesia)

The TVB-N values in fish after being treated with liquid smoke are shown in Figure 1 and Figure 2. The higher the concentration of liquid smoke, the lower the TVB-N value and the greater the antibacterial inhibition produced. The lowest TVB-N value was observed with 3% liquid smoke. Figure 1 shows that the TVB-N values resulting from the use of liquid smoke produced at 340°C at 0.5%, 1%, 2%, and 3% concentrations were low at 2.814, 1.407, 1.407 and 1.407 mg nitrogen/g (mgN/100g), respectively. In the meantime, liquid smoke produced at 380°C ( Figure 2) at the same concentrations resulted in the following TVB-N values: 2.814; 1.407; 1.407 and 0.7035 mgN/100 g within 6 hours. Within 54 hours, the TVB-N values for each soaking time were still within the safe limits allowed. After 60 hours, the TVB-N values increased to levels greater than 33 mgN/100g, which is above the acceptable consumer standard of 30 mgN/100g. These results are comparable to those obtained in previous research ¹⁶ using liquid smoke from oil palm kernel shells. A TVB-N value of 0–30 mgN/100g in fresh produce signifies good quality in compliance with National Indonesian Standards for food (SNI 01-2729-1992). TVB-N values increase due to a bacterial enzyme which degrades proteins into amino acids, and short peptide bonds resulting in production of a number of bases including, amine, ammonia, and trimethylamine which produce foul odors in foods ²⁵ . Longer soaking periods will result in greater bacterial activities, which in turn produce more bases and increase TVB-N values ¹⁶ .

Color. Results from organoleptic testing on color of samples showed that mackerel soaked in various concentrations of liquid smoke changed color depending on the soaking duration ( Table 2). With 0.5–3% liquid smoke concentrations produced at 340°C pyrolysis, the color changed from pale white to yellowish–cream in 36 hours. The same concentrations for 380°C pyrolysis caused the color to change from pale white to yellowish cream in 42 hours. A food product that has high nutritional value, good flavor and good taste will have little interest from customers if the product does not also have an attractive color. Fish soaked in 0.5–3% liquid smoke produced at 380°C had the most optimum results, probably due to a high content of phenol (1.73 wt.%) and acetic acid (8.51wt.%) compared to that of liquid smoke produced at 340°C (phenol=0.79 wt. % and acetic acid=3.40 wt. %) ¹ , which maintained the freshness of the fish ^{16, 19} . For comparison, fish samples that did not receive liquid smoke treatment turned a cream color after only 8 hours. Raw data are available on Zenodo ²⁶ .

Flavor and aroma. The hedonic test results for flavor and aroma are shown in Table 3 and Table 4. High concentrations of liquid smoke slowed down the occurrence of foul smells and bad flavors in mackerel for up to 30 hours ( Table 3 and Table 4). With regards to aroma, the use of 2–3% liquid smoke produced at 380°C maintained a desirable aroma for up to 48 hours of soaking, although the aroma grew thick (similar to the smell of liquid smoke). The smoked scent seeped into the mackerel and grew stronger due to a reduced content of acetic acid in the fish. Production of foul odors can also be used to indicate food spoilage caused by oxidation and fat oxidation lead to production of foul odors in fish ²⁷ . With regards to taste, liquid smoke produced a smoky smell in the fish but without smoke treatment, the deterioration in changes in taste and smell would have occurred in less than 12 hours. Saloko et al. ¹⁷ stated that the use of 5% liquid smoke in a nanocapsule from chitosan and maltodextrin had the potential to maintain the freshness of mackerel for up to 24 hours.

Texture. Texture tests can be carried out orally as well as by touching with hands aimed at feeling the texture of a food product. Table 5 shows that the best texture of mackerel occurred when 3% liquid smoke produced at 380°C was used for up to 48 hours. The fish texture was still quite chewy up to 42 hours in fish treated with 2–3% liquid smoke produced at 340°C. At both pyrolysis temperatures, the fish texture became rigid after 48 hours. At a low concentration (0.5%), the texture of the fish started to change within 36 hours. The present of acids and phenolic compounds in liquid smoke might affect the flavor, aroma, and texture of the fish ¹⁶ . The change in texture was influenced by the speed of bacterial growth. Fish quality decreased when the texture became tender due to the effects of cathepsin and collagenase enzymes on muscle tissues. Cathepsin in fish degrades protein and causes the meat to become tender, while collagenase degrades polypeptide bonds when protein is not denatured ²⁷ .

The results of SD calculation showed that a constant reproducibility SD was only obtained at 18 (eighteen) hours of preservation time on the organoleptic test results for color, flavor, aroma and texture of fish during storage. For preservation time from 24 hours until 60 hours, the standard deviation of the organoleptic test results for color, flavor, aroma and texture had fewer and more scattered data ranging of 0.00 to 0.50. These small standard deviation data indicated that generally the all-organoleptic test results are acceptable.

Table 6 shows the effects of liquid smoke on bacterial growth ( E. coli and S. aureus). A concentration of 0.5% liquid smoke produced at 340°C did not inhibit bacterial growth; however, bacterial inhibiting properties were shown at a concentration of 1%. The same antibacterial properties were seen in liquid smoke produced at 380°C even at lower concentrations. Kim et al. ²⁸ stated that the use of 0.1–1% liquid smoke produced from rice husks inhibited Salmonella growth while Milly et al. ²⁹ demonstrated that 1.5–9% liquid smoke produced from cinnamon can also inhibit bacteria. In addition, Saloko et al. ¹⁷ showed that 5% liquid smoke from chitosan and maltodextrin in a nanocapsule inhibited E. coli and P. fluorescens growth . The type of wood biomass used to produce liquid smoke will result in production of different phenol, carbonyl and acid compounds, which in turn will determine the antibacterial properties, as well as the sensitivity of the pathogenic bacteria to the liquid smoke raw material ¹⁴ .

The total number of bacterial counts in the mackerel will determine whether or not the product is acceptable for human consumption. Plate count agar (PCA) was used to determine the total bacterial counts on mackerel samples. The number of microbes present after soaking must fall within the safe limits for consumption, namely 5×10 ⁵ colonies/g, in accordance with SNI 02-2725-1992 ³⁰ . Table 7 shows that 12 hours of soaking did not result in any significant changes in the number of counts; levels ranged from 3.2×10 ⁵ to 3.5×10 ⁵ colonies/g, indicating that the products were considered to be safe. However, after 42 hours the number of colonies increased to levels of 5.02 × 10 ⁵ colonies/g, making the fish unsafe for human consumption. Liquid smoke can inhibit the growth of bacteria due to the presence of phenols, acids and carbonyl compounds working together to inhibit degradation and spoilage. In particular, acetic acid can penetrate the cell membrane and neutralize the pH gradient ³¹ .