Introduction

The use of 3D printing technology has brought innovations to several fields including engineering, medicine, aviation, art, education, aerospace, and food processing [1]. 3D printing technology, also known as additive manufacturing (AM), was first applied to plastics, photopolymer materials, metals, and ceramics, and has been used in food processing for over a decade [2,3,4]. 3D printing technology has advantages such as not requiring special equipment during processing, maintaining product quality, making it possible to design the product shape and texture, and allowing to change the composition of the product according to the desired nutritional qualities, aroma, and color [2, 5]. Research indicates that liking is a crucial motivation factor for food choices in most cases [6]. 3D food printing allows the production of customized foods with novel and attractive designs [7]. Moreover, 3D food printers enable the utilization of by-products and the reduction of food waste and allow fast, automated, and repeatable processes. 3D food printing can be used to develop customized products with enhanced nutritional value [8].

Trimmings and offcuts with different compositions and quality are generated during the processing of a beef carcass or special meat cuts according to the requirements of the consumers; these by-products are sold at a low price or are considered waste [9]. Thus, 3D food printers can reduce food waste by allowing the utilization of meat trimmings, deformed fruits and vegetables, and seafood by-products to produce novel foods for human consumption [2, 10].

Printability is a crucial parameter in the 3D-printing process. Extrusion-based 3D printers are suitable for complex food ingredients because of their wide compatibility with the rheological properties of materials [7]. Food products such as cheese, chocolate, and hummus are natively printable as they are composed of hydrogel-forming substances such as starch and protein [11]. However, rice, meat, fruit, and vegetables are not natively printable and thus require pre-processing [5]. Hydrocolloids (e.g., gellan gum, carob gum, pectin, carrageenan, and xanthan gum) can be added to the formulation to improve printability by changing the rheological properties of the non-printable foods [12, 13]. Some additives and components in food formulations can cause physicochemical changes. Lipton et al. [14] improved the texture of scallop and turkey mince by adding transglutaminase and bacon fat for 3D printing, and the final meat products retained their initial shapes after cooking. Sun et al. [3] determined the effects of the printing parameters on the geometry and design of unbaked cookies. Moreover, chocolate [15], cereal-based products [16, 17], cheese [18], fruit-based products with a high nutritional value [1], plant-based meat analogs [19, 20], chicken meat-based snacks, and fiber-enriched chicken meat [21,22,23] have been produced by 3D printing. Research indicates that the development of printable formulations, the optimization of printing parameters, and the post-processing conditions are critical for 3D printing.

The shear-thinning behavior of the formulation and the ability of the product to retain its shape are crucial parameters for 3D printing of foods [12]. Furthermore, the heat resistance of the gel and the post-processing conditions should be considered to maintain the desired textural properties after storage and/or cooking. The regulation of internal structures and the void fraction (porosity) changes the textural properties of the products [9, 24].

In this study, the printability of meat pastes prepared with different water and salt contents was assessed, and the optimal printing parameters were determined. So far, there has been no comprehensive study on the effects on printability of the salt concentration in meat products. Moreover, the effects of the cooking methods on the final 3D meat products were determined. It is expected that this study will enable the use of less valuable beef cuts and provide guidelines for manufacturing a high-quality meat product using an extrusion-based 3D food printer and will fill the gap in the 3D printing technology of meat products.

Materials and Methods

Sample Preparation

M. semimembranosus muscle from two bovine carcasses was purchased from a local market in Antalya, Türkiye. Fat and visible connective tissues were removed, and the meat was cut into small pieces and mixed thoroughly. After packaging (300 g), the meat was frozen at ˗18 °C and stored until use. The moisture and fat content of the meat were 74.41 ± 0.02% and 4.71 ± 0.01%, respectively. Edible salt (NaCl purity > 99.0%) was purchased from a local market.

Before printing, the frozen meat was thawed at 4 °C for 3 h. Then, the meat was minced two times with a meat grinder (Bosch MFW3520W, using a 2.5 mm mincer plate) while in a chilled state for easy comminution [9]. To obtain extrudable and stable 3D-printed products, meat mince was mixed with four different ratios of water: control = 0%, 5 W = 5%, 10 W = 10%, 15 W = 15%, and 20 W = 20% (w/w). Water enhances the miscibility of other additives in the meat paste. After the most printable sample was identified, the paste was mixed with four different NaCl levels: 10W0.5 S = 0.5%, 10W1S = 1%, 10W1.5 S = 1.5%, and 10W2S = 2.0%. Salt improves the stability and taste of the 3D-printed meat product and increases the solubility of proteins. Three different batches were produced for each sample. The paste was kept below 5 °C during the process. All samples were homogenized (Retsch GM200, GmbH, Haan, Germany) for 1 min. Before printing, all samples were refrigerated at 4 °C for 20 min.

Rheological Characterization

The rheological properties of the samples were characterized using a rheometer (MCR 102e series, Anton Paar Co. Ltd., Graz, Austria). Meat paste samples were placed in a 25 mm parallel plate and a platform with a 1 mm gap. The apparent viscosity was recorded at a shear rate range between 0.01 s−1 and 100 s−1.

The viscoelastic properties of NaCl-added meat pastes were determined by oscillatory tests. The frequency sweep was from 0.1 to 100 rad/s at 0.1% constant strain within the linear viscoelastic region. The storage modulus (G′), loss modulus (G′′), and loss tangent (tanδ = G′′/G′) were measured [22]. All measurements were conducted at 25 °C (similar conditions to the 3D-printing process) and each test was performed in triplicate.

Back Extrusion Test

The back extrusion test was performed by immersing the probe twice at a certain speed and depth according to the compression force principle. A TA.XT-ExpressC texture analyzer (Stable Micro Systems TA.XT2, Godalming, UK) equipped with a cylindrical probe with a diameter of 35 mm was immersed to a depth of 25 mm at a speed of 1 mm/s. The analyses were performed with three replications for each paste, and the hardness, consistency, cohesiveness, and viscosity index of the samples were determined.

Texture Profile Analysis

Texture profile analyses of the meat paste (prior to printing) and final products were conducted using a TA.XT-ExpressC texture analyzer equipped with a 5 kg load cell. Texture profile analysis (TPA) was performed through a compression test using a 35 mm diameter cylindrical probe. The compression test was performed to a depth of 5 mm with two sequential compressions (test speed = 2 mm/s). The six replicate analyses were carried out in parallel, and hardness, springiness, gumminess, cohesiveness, chewiness, and resilience were evaluated.

Printability Assessment and Optimization of the Printing Parameters

An extrusion-based Foodini 3D food printer (Natural Machines, Barcelona, Spain) was used. The nozzle diameter of the capsule was 1.5 mm. The printability of different water- and salt-added pastes was determined according to the line thicknesses of the printed designs. The optimal formulation was chosen according to the printability, accuracy, and stability of the printed samples.

The model was designed as a square prism with a size of 20 × 20 mm (different layer heights were applied). The initial printing parameters were print speed of 2500 mm/min, ingredient flow speed of 4, fill factor of 1%, first ingredient hold of 4.2 mm, first layer nozzle height of 1 mm, ingredient hold of 3 mm, turning speed factor of 0, and distance between layers of 1.5 mm.

After identifying the most printable meat paste, the ingredient flow speed (3, 3.5, 4, 4.5, and 5), fill factor (1.2%, 1.3%, 1.4%, 1.5%, and 1.6%) and distance between the layers (1.2, 1.4, and 1.6 mm) were optimized by performing one-factor-at-a-time experiments (OFAT). Different printing speeds were evaluated to shorten the printing time. As a result, 2500 mm/min was found to be the optimal printing speed and was used for all printing processes.

The samples were printed as a single layer in the ingredient flow speed trial. After evaluation of the line thickness, the meat samples were printed with different fill factors. To determine the most suitable fill factor, which is crucial for ensuring the shape accuracy and robust printing of the products, the samples were printed at two different infill levels: two lines and four lines. Then, the line thickness and inner and outer diameter of the samples were evaluated according to the desired shape.

The optimum distance between the layers was identified to ensure the stability of the products, particularly those with a greater layer height. Thus, 3D-printed products with four lines infilled at different heights (one, two, three, and four layers) were printed and the 3-D structures were evaluated to determine the optimal value.

Determination of the 3D Structure and Line Thickness

Photos of the 3D-printed samples were taken in a lightbox with 6500 K light and 806-lumen brightness from different sides. The line thickness, length, width, and height of the printed samples were measured using the ImageJ image-scanning software [25,26,27] (Online Resource 1). Six replicates were used to determine the printing performance and the main dimensional properties.

Post-Processing of the 3D Products

To determine the post-processing stability of the 3D products, different cooking methods were applied. The samples printed with four and full layer infilled and 4 layers height were deep-fried (180 ± 5 °C for about 2 min using sunflower oil), baked in an oven, or pan-fried. For baking, a conventional oven (Unox XBC 405E, Italy) was used at 120 ± 2 °C for 25 min; the samples were placed on aluminum trays and covered with aluminum foil [28]. All cooking conditions were determined by trials according to the desired cooking level of the products until achieving a core temperature of 75 °C. The effects of post-processing conditions on the shrinkage, moisture retention, fat content, and textural properties of the samples were analyzed.

Moisture Content and Moisture Retention

The moisture contents of the raw and cooked samples were determined according to the official method described in AOAC 950.46a [29]. Approximately 2 g of sample was placed in a glass Petri dish and dried using a vacuum oven (VWR-VENTI-Line, EC) at 100 °C until the samples reached a constant weight (12 h). The moisture content was calculated as percent moisture (%). The moisture retention of the samples cooked using different methods was calculated according to El-Magoli et al. [30] (Eq. 1).

$$Moisture\:retention\:\left(\%\right)=\frac{Cooked\:weight \times percent\:moisture\:in\:cooked\:samples}{Raw\:weight \times percent\:moisture\:in\:raw\:samples} \times 100$$
(1)

Fat Content

The fat content of the 3D samples was determined according to the Soxhlet reference method (960.39b) [31]. The sample (4 g) was weighed and dried in an oven (VWR-VENTI-Line, EC) for 6 h at 100 °C, then extracted with petroleum ether using a Soxhlet apparatus for 6 h. The extracts were evaporated using a rotary evaporator and dried in an oven for 6 h at 100 °C. After reaching room temperature, the samples were weighed.

Shrinkage

To examine the post-processing characteristics of the samples, shrinkage (%) values were determined using Eq. (2) [9].

$$Shrinkage\:\left(\%\right)=\frac{\left(Raw\:length-Cooked\:length\right)+\left(Raw\:width-Cooked\:width\right)+(Raw\:height-Cooked\:height)}{Raw\:length+Raw\:weight+Raw\:height} \times 100$$
(2)

Statistical Analysis

The effects of water (control, 5 W, 10 W, 15 W, and 20 W) and salt (10W0.5 S, 10W1S, 10W1.5 S, and 10W2S) on the printability and textural properties of the meat paste were investigated by performing OFAT experiments. After the optimal water content for printability was determined, experiments on the salt content were conducted. Subsequently, the printing parameters (ingredient flow speed (3, 3.5, 4, 4.5, and 5), fill factor (1.2%, 1.3%, 1.4%, 1.5%, and 1.6%), and distance between the layers (1.2, 1.4, and 1.6 mm)) were assessed based on the line thickness and 3D shape observations. The results were analyzed using the general linear model of SAS University Edition (Statistical Analysis System, Cary, NC, USA) and expressed as mean ± standard error. Data were subjected to analysis of one-way variance (ANOVA) at a confidence interval of 95%, followed by Duncan’s multiple comparison tests for pairwise comparisons with a level of significance of 5%. All 3D printing experiments were performed in triplicate. Water, salt treatments and printing parameters (ingredient flow speed, fill factor and distance between layer) were considered fixed terms and selection of meat samples and the replication of experiments were treated as random effects.

Analysis of the rheology, back extrusion, moisture, and fat were conducted in triplicate, whereas six replicates were used for the analysis of the texture profile, line thickness, and 3D structure.

Results and Discussion

Rheological Properties of the Meat Paste

The printability of foods is influenced by various factors such as the temperature, printing parameters, and rheological properties of the food material [32]. Because of its fibrous structure and complex matrix, meat has poor printability, especially when using small-diameter nozzles. Therefore, water and salt improved the textural and rheological properties of the samples, resulting in good printability (Fig. 1). At a 100 s−1 shear rate, the viscosity of samples was as follows: control = 7.8 Pa.s, 10 W = 2.5 Pa.s, 10W0.5 S = 2.56 Pa.s, 10W1S = 3.4 Pa.s, 10W1.5 S = 6.62 Pa.s, and 10W2S = 6.68 Pa.s.

Fig. 1
figure 1

Rheological behaviors of meat pastes at different salt levels. a Viscosity (Pa.s). b Loss factor (tanδ). c Storage modulus (G′). d Loss modulus (G′′)

Full size image

Liu et al. [33] pointed out that the rheological behavior of the material is a crucial parameter in extrusion-based 3D printers. Moreover, the rheology reflects the self-supporting ability of the product [12]. Hence, the viscosity of the samples was monitored according to the increased shear rate to determine the ease of extrusion. G′, G″, and the phase angle were characterized by frequency sweeps to assess the self-supporting ability of the meat pastes.

Viscosity indicates the fluidity of the material and affects printability [7]. The optimal apparent viscosity allows both easy printing with a small-diameter nozzle and the deposition of layers by stacking [24]. Furthermore, extrusion-based 3D printers require non-Newtonian fluids that exhibit thinning behavior to maintain the desired shape [34]. Here, the viscosity of the meat pastes decreased with an increased shear rate. Water and salt concentrations play a crucial role in the viscoelastic properties of the meat paste. While the addition of water decreased the viscosity of the samples, the addition of salt increased viscosity. The increase in the viscosity of the meat paste by salt addition facilitates the incorporation of fat to obtain stable batters [35]. Thus, the highest salt concentration corresponded to the highest viscosity. In addition, the viscosity of the control group was higher than that of the 10 W, 10W0.5 S, and 10W1S groups. This difference was the result of the addition of water. At low shear rates (from 0.1 to 10 s−1), the control and 10W2S group had a dramatically high viscosity compared to other groups. However, the viscosity of all samples at high shear rates (from 10 to 100 s−1) was similar. Yang et al. [7] determined that, at low shear rates (0.1–0.3 s−1), NaCl reduces viscosity but at high shear rates, viscosity increases. Here, the addition of 10% water changed the viscoelastic properties of the samples and led to a decrease in viscosity. Dick et al. [12] reported that the viscosity of pork pastes decreased with increasing shear rates with shear-thinning behavior. Wang et al. [36] determined the effects of different salt ratios and printing parameters on the 3D printing of surimi. The addition of NaCl improved the rheology and 3D structure of the surimi products. Moreover, NaCl, xanthan gum, and guar gum have been used to obtain a more extrudable structure to manufacture printed meat products for dysphagia patients [12].

The storage (elastic) modulus (G′), loss (viscous) modulus (G′′), and loss tangent (tanδ = G′′/G′) are indicators of the elastic and viscous behavior and post-printing stability of the samples. G′ represents the recovered or accumulated energy and G′′ represents the amount of dissipated energy in each deformation cycle. The loss tangent (tanδ) value indicates the viscoelasticity of the samples. A tanδ < 1 value indicates a predominantly elastic behavior, whereas tanδ > 1 indicates a predominantly viscous behavior [33]. A tanδ value close to 1 suggests that the material exhibits fluidity, whereas a tanδ value close to 0 suggests that the material is fairly solid [7].

The frequency sweep curves of the meat samples are displayed in Fig. 1c and d. The storage modulus and loss modulus of the samples increased with increasing angular frequency. The lowest storage modulus corresponded to the 10 W sample. NaCl addition increased the storage modulus. Because of water addition, the storage modulus of the 10W0.5 S and 10W1S samples were below that of the control group. However, the 10W1.5 S sample exhibited similar results to the control. The 10W2S group had a higher G′ value compared to other groups. It was reported that the G′ and G′′ values of chicken pastes increased with an increase in the frequency and NaCl level, and the G′ value was higher than the G′′ value for all NaCl-added samples [7]. Similar results were obtained in this study, where the storage elastic modulus was higher than the loss modulus.

The loss tangent of all samples was below 1, suggesting that all the samples exhibited a predominantly elastic behavior. Yang et al. [7] showed that salt addition increased tanδ value of the chicken pastes and rheological properties establish a balance between extrusion and stability. In this study 1.5% and 2% salt added samples had higher G′, G′′, and tanδ values than other samples.

Back Extrusion Test

Back extrusion tests of the samples were conducted to estimate the printability and optimum firmness of the samples for printing. As a result—similar to TPA analysis—the firmness of the water-added samples decreased significantly (P < 0.05) as the water content increased (Table 1). The consistency, cohesiveness, and work of cohesion of the samples decreased as the water content increased. Nonetheless, the 10W1.5 S and 10W2S groups had similar firmness values. By contrast, salt addition made the samples firmer. Thus, the consistency, cohesiveness, and work of cohesion of the samples increased significantly (P < 0.05) with increased salt concentration. Similar to firmness, the consistency and cohesiveness values of the 10W1.5 S and 10W2S samples exhibited no significant differences. The firmness and consistency values of the 10W1.5 S and 10W2S samples were lower than those of the control sample. In addition, it was reported that a 1.75% salt concentration is sufficient to produce acceptable perceived saltiness and firmness in lean meat products [37]. Water content of samples may affect protein-water interaction [38]. Also, it suggested that salt addition increase the solubilization and extraction of myofibrillar proteins and protein interactions [39] by enhancing hydrogen bonding and hydrophobic interactions between proteins with forming salt bridges [40, 41] thus, forming harder and firmer structure [39]. In this study, when evaluated with the printing experiments, the increasing protein matrix with the increase in salt concentration may have enhanced the flexibility of the paste and resulted in better extrusion. The work of cohesion defines the viscosity index at the same time. Thus, the work of cohesion analysis supports the viscosity results.

Table 1 Back extrusion analysis results of meat pastes
Full size table

Texture Profile Analysis (TPA)

The textural parameters of the meat pastes were determined by texture profile analysis. The addition of different amounts of water significantly (P < 0.05) affected the hardness of the meat paste (Table 2). The gumminess, chewiness, and resilience values of the samples decreased as the amount of added water increased. A 10% water content was selected for the 3D-printed meat samples. Next, the effect of different salt concentrations on the textural parameters was assessed. Salt addition significantly (P < 0.05) increased the hardness of the meat paste. The hardness of the printing material is strongly correlated with the extrusion force required and is also related to printability [22]. Water and protein interactions have crucial effects on the functional and technological properties of meat products [42]. Salt (NaCl) affects the water binding and gel formation capacity of meat products. Moreover, it can enhance myofibrillar swelling, water-binding capacity, salt-soluble protein extraction, and binding properties of the proteins, which together improve texture. A higher salt-soluble protein content ensures the formation of a more stable gel, which results in decreasing cooking losses [35, 43]. Furthermore, it was reported that a ≤ 1.5% salt concentration can change the ionic strength, resulting in a stronger gel [44].

Table 2 Texture profile analysis results of meat pastes
Full size table

Here, the decrease in chewiness, gumminess, and resilience with the increase in the amount of water was found to be significant (P < 0.05). Moreover, higher salt content in the meat pastes significantly (P < 0.05) increased hardness. The addition of 1.5% and 2% salt resulted in meat pastes with a higher adhesion value. In particular, the 10W2S sample was stickier than other samples. Meija et al. [45] found that the hardness and cohesiveness of meat emulsions increased as the salt content increased from 0.70 to 1.30%. The springiness of meat pastes may affect the printing accuracy and dimensional printing deviation of the printed products [9]. However, in this study, no significant differences were observed regarding the springiness of the samples. By contrast, the gumminess and chewiness of the salt-added samples significantly (P < 0.05) increased with an increase in the salt concentration.

Printability of the Meat Pastes

The quality of the final printed food products, including the accuracy of the 3D structure, depends on several parameters such as the nozzle speed, nozzle height (layer height), nozzle diameter, extrusion rate, and infill percentage of the printed product [15]. The selection of the nozzle diameter affects the accuracy of the printed meat paste. A nozzle diameter < 2 mm results in a more accurate and complex structure; thus, the formulation is the most important parameter for printing [9]. Hence, a 1.5 mm nozzle was chosen to print the products. The printability of the meat paste was determined by the line thickness and shape accuracy of the printed products. While printing the control samples and 5% water-added samples, the desired shapes could not be printed due to the lack of breaking behavior. In particular, the corners of the designs could not be printed sharply (Table 3). On the other hand, the 15 W and 20 W samples showed greater line thickness because of the flow behavior of the meat paste. Thus, the higher amount of water added may have made the samples more fluid, and this lowered their printability; for instance, the line thickness of these samples was not printed uniformly. No significant differences (P < 0.05) were observed between the line thickness values of these samples (Table 3); however, the samples with 10% water were printed uniformly according to the desired shape (20 × 20 mm).

Table 3 Line thickness of printed meat pastes at 5 water (0, 5, 10, 15, 20%(v/w)) and 4 NaCl (0.5, 1, 1.5, 2%(w/w)) level
Full size table

Although the meat paste was initially homogeneous, the 15 W and 20 W groups showed water release during printing. This may be due to the deterioration of the homogeneous structure by the pressure applied to the nozzle during printing. However, adding 5% water did not allow complete extrusion of the paste, and the lines broke during printing. The control group was more viscous, the printed lines were not uniform, and the desired shape was not achieved. However, no significant changes in the line thickness and outer length values of the water-added samples (P < 0.05) were observed. The lines of the 10 W samples were printed more uniformly and the outer length of the group was determined to be 20.25 ± 0.09 mm, which was the closest to the desired shape. During printer extrusion, materials are exposed to relatively high pressure and mechanical shear forces. Therefore, the rheological properties of products are important indexes to evaluate printability [46]. The better printability of the 10 W product gave information about other parameters of the product to be used in the printer. Similarly, as the amount of water increased, the firmness and consistency values ​​decreased. The water content of the meat paste is an important parameter for shape accuracy. The addition of water lower than 10% prevented the printability with higher firmness and viscosity index. Also, higher than 10% water provided lower firmness, viscosity index and unstable shape accuracy. Therefore, the 10% water-added samples were chosen for further experiments.

The 10W2S samples exhibited a greater inner and outer length than the 10W1.5 S samples. Moreover, the 10W2S group exhibited greater stickiness, and a less uniform product was obtained during printing. Hence, the 10W2S samples had an inadequate emulsion structure. The line thickness of the products was improved with salt addition. The 10W0.5 S and 10W1S samples had a similar line thickness to the 10 W sample. The most suitable printing process was carried out and the desired shape was observed with meat paste containing 1.5% salt. As shown in Table 3, the shape and lines were more uniform in the 10W1.5 S group. Considering back extrusion and TPA analysis, the better printability of the 10W1.5 S sample, which had similar firmness values ​​to the control sample, showed the importance of the gumminess, chewiness, and resilience values ​​of the products in printability studies. It was determined that the 10W1.5 S sample, which had the best printing accuracy, had higher gumminess, chewiness, and resilience values ​​than the control sample (P < 0.05). The addition of water reduced the viscosity and firmness values of meat and provided continuous and accurate printing. Moreover, the addition of salt not only increased the firmness, consistency, and viscosity, but also improved the gumminess, chewiness, and resilience values of samples. Therefore, meat samples with 10% water and 1.5% salt were considered to be optimal for printing.

Optimization of the Printing Parameters

Various studies have shown that the printing parameters affect the printability of food materials and thus the quality of the final product. Printing settings (nozzle speed, nozzle height, nozzle diameter, extrusion rate, and infill percentage) are critical parameters to obtain a final product with an accurate shape. All critical parameters can be rearranged to obtain the desired shape [9, 15].

The optimum nozzle height determines the shape accuracy and dimensions of the 3D-printed product. Thus, it is recommended that the optimal nozzle height be equal to the nozzle diameter. If the nozzle height is lower than optimal, it may cause scattering of the deposited stream and result in an expanded product compared to the original design because of the springiness of meat paste [9]. A nozzle height of 1 mm was used in this study based on trials. At higher nozzle heights (1.4, 1.3, 1.2, and 1.1 mm), the meat paste was not fixed on the printing mat, and it was impossible to print the desired shape. Moreover, the layers could not be printed uniformly.

The ingredient flow speed (i.e., the extrusion rate) has a linear relationship with the printing speed (nozzle movement). A higher printing speed requires a higher ingredient flow speed [22]. Thus, different ingredient flow speeds were tested at a 2500 mm/min constant print speed. As shown in Table 4, the line thickness of the samples increased significantly (P < 0.05) with increasing ingredient flow speed. Hence, the outer length of the samples increased, whereas the inner length decreased. For instance, a speed of 2.5 (which was not considered in the statistical analysis) led to inconsistent printing results. When the printing process was conducted with a speed of 3, the lines were more uniform and the line thickness was 2.50 ± 0.06 mm. Accordingly, the optimal ingredient flow speed was 3 as it provided the fastest and most accurate printing.

Table 4 Line thickness, outer and inner length of 3D printed samples at different ingredient flow speed levels
Full size table

The fill factor is the amount of space between the horizontal print lines. If the fill factor is smaller, the printed lines get closer. Conversely, a decreased fill factor leads to gaps or holes between the print lines. This causes the layers not to be placed properly while the product is printing in layers vertically, thus preventing the formation of a 3D structure.

Two lines and four lines of infilled objects were used for fill factor optimization. The line thickness values of the two levels were increased significantly (P < 0.05) with increasing fill factors (Table 5). The line thickness was expected to be close to 5.00–10.00 mm. Lines were too close and overlapped in samples printed with fill factors of 1.2 and 1.3. By contrast, fill factors of 1.5 and 1.6 resulted in gaps between the lines (Fig. 2). Thus, the optimal line thickness corresponded to a fill factor of 1.4, which resulted in no gaps and no overlaying.

Fig. 2
figure 2

Images of 3D printed meat paste with two and four lines infilled at different fill factor

Full size image
Table 5 Line thickness, outer and inner length of 3D printed meat pastes at different fill factor and infill levels
Full size table

The inner lengths of the samples significantly (P < 0.05) decreased with an increasing fill factor. The greatest inner lengths corresponded to the two- and four-line infilled designs with a 1.2 fill factor.

After optimization of the infill factor, the “distance between layers” was optimized. This parameter is the distance between two layers that start with the beginning of the first line and the second line. When this value is low, the nozzle prints by dipping into the previous layers, or excess product may accumulate around the nozzle. When it is high, the layers may not be aligned properly and collapse may occur. To determine the effects of different distances (1.2, 1.4, and 1.6 mm, based on preliminary experiments) on the 3D structure, the samples were printed with one, two, three, and four layers with four lines infilled. There were no significant (P < 0.05) differences in the distances between the layers in all layer heights (Table 6). Thus, 1.4 mm which was recommended as the distance between layers by the manufacturer was chosen.

Table 6 Heights and images of 3D samples at three distance between layer levels (1.2, 1.4, 1.6 mm)
Full size table

Finally, different layer heights with different infill levels were applied to validate the optimized parameters. Square prism shaped with one line infilled with a 28-layer height and two lines infilled with a 36-layer height were successfully printed using the optimized parameters (Online Resource 2). Also, images of 3D printed different shapes are presented in Online Resource 3.

Post-Processing Conditions and Characteristics

To determine the effects of the post-processing conditions, 3D meat samples with two infilled levels (full-infilled and four-line-infilled) and a four-layer height were baked, pan-fried, or deep-fried.

To evaluate the changes in the cooked meat, shrinkage measurements can be used to estimate the quality characteristics. Cooking causes protein denaturation and structural changes like shrinkage of the muscle fibers and connective tissues [47]. Tornberg [48] reported that during cooking, the transverse shrinkage of the meat products occurs mainly at 40–60 °C and widens the gap between the fibers and the surrounding endomysium. At 60–70 °C, both the connective tissue and muscle fibers longitudinally shrink, and the rate of shrinkage increases with temperature. The shrinkage of the baked samples was higher than that of other cooked samples. Nonetheless, the baked samples retained the initial design, whereas the full-infilled pan- and deep-fried samples exhibited a bulge in the middle, and the edge lengths shortened. Despite the lower shrinkage results, the full-infilled deep- and pan-fried samples did not retain the initial design. By contrast, the four-line infilled pan- and deep-fried samples retained the initial design (Table 7). Lipton et al. [14] determined the effect of deep-frying on space shuttle-shaped 3D-printed transglutaminase-added scallop meat. They found that the extremely thin parts of the products were deformed. Most of the water in muscle is contained within the myofibrils. Major changes in water distribution within myofibrils result from changes in these cavities. Cooking causes structural changes that reduce the water holding capacity of meat. Shrinkage during cooking causes the highest water loss at 60–70 °C, and it’s reported that water is expelled from the meat by the pressure exerted by the shrinking connective tissue on the extracellular space aqueous solution [48]. The type of material, additives, and post-processing methods affected the final product. In addition, pan and deep frying caused crust formation. The reason of the shape changes, especially in the full-line infilled fried samples, may have been the pressure of the water in the myofibrils that couldn’t be expelled due to crust formation. Dick et al. [28] concluded that longer cooking times and lower temperatures can prevent thin crust formation. The fact that the baked samples had thinner crust than fried samples may allow the moisture to be expelled more easily, thus providing shape stabilization during baking.

Table 7 Shrinkage of Baked, pan and deep fried four line and full infilled 3D meat pastes
Full size table

Because of the thermal denaturation of meat, some water and fat losses occur. The most common dimensional change during heat treatment is the decrease in patty diameter because of shrinkage. In a study, shrinkage was in the range of 12–22% for different types of meat during frying. Meat with the highest fat content showed the greatest shrinkage. When frying beef burgers, the water- and fat-releasing mechanism is related to pressure-driven mass transfer due to shrinkage [49]. Accordingly, the inability to transfer the liquid and oil in the sample because of the formation of a crust during frying may result in deformation. The higher moisture retention of the pan- and deep-fried samples observed here supports this conjecture.

The moisture retention of the cooked 3D-printed meat products is presented in Table 8. Different cooking methods significantly (P < 0.05) affected the moisture retention of the products. Pan-fried 3D-printed meat had higher moisture retention than other cooked products. Moreover, the fat content of the samples was significantly affected by the cooking conditions; deep-fried samples had the highest fat content.

Table 8 Moisture retention, fat retention of Baked, pan and deep fried meat products
Full size table

The results of the texture profile analysis of printed four line infilled and full infilled raw sample are as follows; hardness (N): 8.72 ± 0.09, 25.61 ± 1.15; adhesiveness (g.sec): -1308.16 ± 253.59; -1396.10 ± 113.06; springiness : 0.18 ± 0.01, 0.39 ± 0.07; cohesiveness: 0.69 ± 0.07, 0.64 ± 0.03; gumminess (N): 8.39 ± 1.16, 16.32 ± 1.37; chewiness (N): 1.49 ± 0.31, 6.30 ± 0.60; resilience: 0.10 ± 0.01, 0.11 ± 0.00, respectively.

Cooking methods significantly affected hardness value of the samples (P < 0.05). Highest hardness value determined in baked samples (Fig. 3A). Wen et al. [50] indicated that cooking time and temperature is critical for protein denaturation, cross-linking, aggregation, and shrinkage which leads to water loss and changes in textural properties of meat. Also, they reported that shrinkage caused by cooking resulted in a harder product with lower springiness value. Dick et al. [12] pointing out that higher water retained samples contribute lower penetration force as well as flexible and softer texture. In this study baking time was longer than other cooking methods with lower temperature. Moreover, longer cooking time may have led to higher moisture loss and harder texture. Despite the crust formation (in pan-fried samples), the short cooking time provided higher moisture retention and lower the hardness value.

Fig. 3
figure 3

The (A) hardness, (B) springiness, (C) cohesiveness, (D) gumminess, (E) chewiness, (F) resilience of cooked 3D meat pastes using different methods

Full size image

Only the four-line infilled pan-fried samples were softer than the full-infilled pan-fried samples. The amount of infill ratio did not affect the hardness value in the other cooking methods. Springiness results showed that pan fried samples were physically backward better after deforming with first compression during TPA analysis (P < 0.05). Four-line infilled samples were more flexible than full infilled samples except deep fried samples. Similar springiness value was observed in full infilled and four-line infilled deep-fried samples (Fig. 3B). Deep fried samples showed higher cohesiveness and resilience than other samples (P < 0.05) (Fig. 3C and F).

No differences were found in gumminess values between all baked samples and full-infilled deep-fried samples. Similar to hardness value four-line infilled pan-fried samples had lowest gumminess and chewiness value (P < 0.05) (Fig. 3D and E). These lower parameters showed that lower energy required to swallow for four lines infilled pan-fried samples.

Conclusion

One of the functions of 3D food printers is to reduce food waste, as they enable the use of meat leftovers, deformed fruits and vegetables, and seafood by-products to manufacture products for human consumption. The shear-thinning behavior of the formulation as well as the self-supporting ability of the materials are crucial parameters for 3D food printing. In this study, a printable meat formulation using water and salt was developed. Salt is the simplest food additive. To determine the printability of the meat, paste line thickness, textural, and rheological evaluations were conducted. Salt was found to increase the firmness of the meat paste and thus improve the printability and line thickness of the printed samples. While a 1.5% salt concentration improved the printability, textural, and rheological properties of the product, a 2% salt concentration caused the deterioration of the emulsion structure. The 3D food printer parameters the ingredient flow speed, fill factor, and distance between the layers were optimized for meat paste. The final 3D-printed meat pastes and printing results revealed that the optimized formulation and printing parameters resulted in enhanced extrudability, accuracy, and forming performance.

Post-processing results demonstrated that the baked samples retained the initially designed shape despite the greater shrinkage of the full-infilled samples. Therefore, this study can be used as a guide in the application of 3D printing in meat products and the development of novel methods for producing 3D-printed meat products.