Introduction

Meat is an important source of proteins, fatty acids, vitamins, as well as minerals such as iron and zinc, making it an important food source for human beings1,2,3. Global meat consumption has been growing steadily for several decades, exceeding 300 million tons since 20184. However, meat is also prone to spoilage, due predominantly to the rich nutrient composition providing an excellent growth media for microorganisms, including food-born human pathogens and spoilage microorganisms. Thus, meat needs to be properly preserved and stored to ensure safe consumption.

Meat preservation is typically achieved by storing at low temperatures (4 °C within 4 h of slaughtering), or by freezing (<−20 °C) or super-chilling in a partially frozen state. In general, low-temperature methods have the advantage of maintaining better meat freshness compared to other techniques such as drying, smoking, and canning, or through chemical preservation (using salts, nitrites, acids, etc.), or biopreservation (using natural antimicrobial agents such as essential oils, nisin, and lysozyme, etc.)5,6. However, safely transporting meat requires energy to maintain the cold chain storage, which over long distance can be subject to disruption, risking spoilage and contamination with harmful microorganisms.

In this work, we demonstrate an ultra-high-temperature flash heating (UFH) method that can efficiently preserve meats while retaining their nutrient level and texture. The concept stems from a similar ultra-fast and high-temperature heating method, also known as flash Joule heating, that has been previously applied in materials science to sinter ceramics7, synthesize high entropy catalysts8, recycle plastics9, and convert waste (including food waste)10, etc. Here, we propose to adapt this technology to food preservation. In this approach, Joule heating is achieved by applying a high electric power (~1800 W) over a short time interval (<1 s) to a carbon substrate, creating a transient increase of the substrate temperature to >~2000 K. By placing the meat in direct contact with the carbon substrate, this rapid heating causes ultra-fast dehydration and inactivation of microorganisms on the meat surface, forming a thin (~100 μm) protective layer that also serves as a barrier to inhibit inward bacterial migration toward the bulk of the meat. As a result of surface dehydration and microorganism-inactivation after the application of UFH, we achieved an extended shelf life of 5 days at room temperature for beef storage without compromising the interior appearance, texture, and nutrition of the meat. The UFH method is essentially a food surface treatment technology, which could be cost-effective for the industrial-scale preservation of large carcasses. The approach could be used alone or in combination with low-temperature technologies to decrease the risk of temperature variation during transport and storage for improved food safety.

Results

UFH creates a microbe-inactivated and dehydrated thin layer on the meat surface

A conceptual illustration of the UFH for the surface treatment of meat is shown in Fig. 1. The carbon felt is connected to a DC power supply (not shown), and is heated to >~2000 K in less than 1 s. For demonstration purposes, a beef steak (Fig. 1a) is placed on the carbon felt; the beef surface is subjected to ultra-fast dehydration and inactivation of microorganisms during the current pulse and associated high-temperature flash heating. The entire UFH process recorded by a high-speed camera (200 fps) can be seen in Supplementary Movie S1. Figure 1b compares UFH with other heating methods, in terms of temperature and operating time11,12,13,14,15,16,17. In general, the temperature in the UFH is an order of magnitude higher than that in conventional food heating methods, such as boiling, baking, and grilling; and it is more than 3–4 orders of magnitude faster. A butane torch may reach a temperature that is close to the UFH, but the heat transfer is mainly through convection, which is still 2 orders of magnitude slower. The effect of the UFH is schematically shown in Fig. 1c–e using a 1 cm3 beef cube. The heating procedure is applied to all six faces of the cube. Compared to the fresh beef (Fig. 1c), the UFH-treated beef possesses a thin, dehydrated outer layer that seals the surface (Fig. 1d, e); allowing the bulk beef to remain fresh (Fig. 1e).

Fig. 1: Schematic of the UFH process for meat preservation.
figure 1

a Demonstration of the heating process. Carbon felt is used as the heating element, due to its high electrical resistance. The arrow marks the electric current direction. b Comparison of operating temperature and time of conventional food heating methods with the UFH method (error bar symbols represent value ranges in the reported data). c–e Schematic appearance of a beef cube before (c) and after heating treatment (d, e). After the UFH is applied to the entire surface of the meat, producing a pale appearance (d), while the interior meat remains fresh, as revealed in the cross-sectional view of (e). The thickness of the outer layer with respect to the bulk sample is not to scale.

Full size image

As a proof-of-the-concept experiment, we placed a 1 cm3 beef cube (prepared from sirloin beef stored at ~253 K before use) between two pieces of carbon felt, which served as the Joule heating element (Fig. 2a). When a transient current passed through the carbon felts, their temperature increased due to the Joule heating effect, causing the emission of electromagnetic radiation (Fig. 2b). The brightness of the carbon felt is associated with its temperature, which can be determined using a high-speed camera and color ratio pyrometry (Fig. S1)18. Fig. 2c shows a typical heating profile change for the carbon felt during the UFH process, where a temperature of ~1500 K at ~0.2 s and a peak temperature of 2091 K at ~0.7 s was observed. The period during which the carbon felt temperature exceeds 2000 K lasts for ~0.2 s.

Fig. 2: Analysis of the UFH treatment of meat.
figure 2

a Photograph of a beef cube of 1 cm3 sandwiched between two pieces of carbon felt that serve as the heating element. b The UFH treatment in operation. The light emitted by the carbon felts is due to the high temperature. c Temperature profile of the carbon felt during the UFH treatment. d, e Thermal analysis results of the evolution of the beef temperature (d) and water content (mass fraction) (e) over 10 s of the UFH treatment; the simulation sensors are along the central line normal to the carbon felt, from the heated beef surface across the entire sample. f, g Photographs showing a fresh (f) and UFH-treated (6 surfaces) beef cube (g). h Optical micrograph of the cross-section of a UFH-treated beef. i Histological micrograph of the surface of a UFH-treated beef (red line indicates the contour of the carbonized layer). j, k Scanning electron microscopy (SEM) image of the UFH-treated beef surface (j) and the corresponding energy dispersive spectroscopy (EDS) analysis (k). l Raman spectra of the surface and central part of the UFH-treated beef.

Full size image

We conducted a two-dimensional numerical simulation to investigate the transient temperature distribution of the beef sample during the UFH process. To simplify the simulation, we assumed that only one face of the beef was in direct contact with the carbon felt heating element, with the other faces being exposed to an air environment at 293 K. The initial bulk temperature was set at 253 K, which is typical for long-term food storage19. The simulation results indicate that the beef surface temperature increased from ~450 K at 0.1 s to the peak point of ~900 K in 1 s during the UFH, then quickly dropped down below ~400 K in 5 s, after the heating was stopped (Fig. 2d). Our simulation also showed that at distances of >~2 mm from the heated surface, the beef temperature was almost unchanged.

The overall appearance of fresh and the UFH-treated beef, as well as the cross-section of the UFH-treated beef, are shown in Fig. 2f–h. Image analysis (see Fig. S2 in the Supplementary Information for details) of the histological micrograph (Fig. 2i) reveals the thickness of the surface layer is ~100 μm. The dark appearance of the UFH-treated beef may be due to surface carbonization. Scanning electron microscopy (SEM) reveals the meat featured a rough surface after heating (Fig. 2j). Additionally, electron dispersion spectroscopy (EDS) analysis indicates that the dark regions consist mostly of carbon (Fig. 2k). Raman spectra reveal that the surface of the UFH-treated beef shows two enhanced peaks at 1374 and 1596 cm−1 as compared with the central region of the meat sample (Fig. 2l). We assign these two broad peaks to the D- and G-band of graphite, indicating the existence of carbon species in a poorly ordered form20,21.

Microbiological tests reveal a significant inhibition in the growth of microorganisms during storage, due to the UFH treatment

Meat appearance can be used as a quick check of the microorganism proliferation (Fig. 3a). The UFH-treated and untreated beef were stored at room temperature for 80 h, and their central regions were examined every 20 h. The visual appearance of the UFH-treated beef remained largely unchanged during storage under these conditions (top row in Fig. 3a). Meanwhile, the color of untreated beef became dark in the first 20 h and the surface appeared watery after 40 h due to microbial proliferation (bottom row in Fig. 3a).

Fig. 3: Safety assessment of the untreated and UFH-treated beef.
figure 3

a Photographs of the UFH-treated (top row) and untreated beef (bottom row) during storage. b, c Short-term (24 h) microbiological test for b APC and c EB. d–f Long-term (100 h) microbiological tests for d APC, e EB, and f YM. Solid lines are fits using the Baranyi–Roberts model. g Cell viability of 3T3-L1 (mouse embryonic fibroblast cell) and CCD18co (human colon tissue cell) at different concentrations of meat sampled from the surface of the UFH-treated beef. Current microbiological tests do not involve bacterial spores. Error bars indicate one standard deviation of uncertainty.

Full size image

To quantify the efficacy of the UFH method for inhibiting bacterial and fungal growth, we compared the content of microorganisms including total aerobic mesophiles (represented in aerobic plate count (APC)), enterobacteriaceae (EB), and yeasts and molds (YM), for the UFH-treated and untreated beef samples at room temperature (Fig. 3b–f). Microbiological tests were performed on a short-term (24 h) and long-term (100 h) basis; and the Baranyi–Roberts model was used to fit the microbial growth profile (solid lines in Fig. 3d–f)22. For the untreated beef, APC and EB counts increase rapidly in the first 24 h while the APC counts remain below the detection limit in the first 22 h and EB 24 h for the UFH-treated beef (the lower detection limit of APC and EB are 0.70 and 0.48 log CFU/g, see Methods section for details), as shown in Fig. 3b, c. Long-term tests reveal the complete microbial growth profiles for untreated beef (Fig. 3d–f for APC, EB, and YM, respectively). Note that data of short-term test for YM is not shown, as the YM count remains undetectable for both untreated and UFH-treated beef; but long-term test clearly shows a rapid increase in YM populations in the untreated beef after 20 h and reaching ~106/g (Fig. 3f) after 100 h. In contrast, no YM were detected in the UFH-treated beef during the entire 100 h period. At the late stages of storage, the untreated beef samples had completely spoiled, whereas the bacterial counts in the UFH-treated beef samples barely reached the critical limits23. For example, according to the USDA Agricultural Marketing Service (AMS), the APC critical limit for ground beef is 4.0 log CFU/g24. The APC counts of untreated beef stored at room temperature exceeds the critical limit after ~12 h (Fig. 3b), while that of the UFH-treated beef remain below the limit even after storing for 100 h. We also conducted the cytotoxicity assessment of the surface of the UFH-treated beef to address potential safety concerns. Compared to untreated beef, the cell viabilities of 3T3-L1 (mouse embryonic fibroblast cell) and CCD18co (human colon tissue cell) did not change when cultured with the addition of beef sample collected from the UFH-treated surface (Fig. 3g).

UFH treatment maintains beef quality during storage

Besides assuring safety, it is critical for a preservation method to maintain food quality as characterized by appearance, texture, and nutrient level. Texture is an important measure for meat quality, being directly associated with the mouthfeel during consumptionThermal and water content analysis during UFH

We use COMSOL MultiPhysics software, based on the finite element method, to study heat transfer and the change in the water content of frozen beef during the UHF process. Our model, illustrated in Fig. S10, is reduced to a two-dimensional problem, where only the surface of the frozen beef is heated, while the remaining parts are exposed to ambient conditions. The heating element is positioned at the bottom of the frozen beef, generating a high temperature of 2000 K for a duration of 1 s. This heat transfer increases the beef temperature, melts the ice, and causes water within the beef to evaporate. The simulation is performed in a time frame of 10 s, including both the heating and the cooling phases. Details such as heat transfer equations, basic assumptions, boundary conditions, as well as material property parameters used in simulation are given in the Supplementary Information.

Microbiological tests

The UFH-treated and control samples, 5 g each, were collected, immersed in 45 mL of sterile buffered peptone water (BPW), and stomached for 2 min using a stomacher (Seward type 80, UK). Additional four dilutions of 10−1, 10−2, 10−3, 10−4, and 10−5 were successively prepared. For the long-term test, five additional dilutions of 10−3, 10−6, 10−9, 10−12, and 10−15 were prepared.

Aerobic Plate Count (APC). All dilutions were evenly dispensed onto the APC plate (Petrifilm™, 3M). The prepared plates were incubated at 35 ± 1 °C for 48 h, and the colony number was then counted. The countable range was between 25 and 250 CFU.

Enterobacteriaceae (EB). All dilutions were plated following the same procedure for APC, using (Petrifilm™, 3M). The EB count plates were incubated at 35 ± 1 °C for 24 h, and the colony number was then counted. Red colonies associated with gas bubbles or colonies surrounded by yellow zones with or without gas were counted. The countable range was between 15 and 100 CFU.

Yeast and Mold (YM). All dilutions were plated following the same procedure for APC, using rapid yeast and mold count plate (Petrifilm™, 3M). The EB count plates were incubated at 25 ± 1 °C for 48 h, and the colony number was then counted. All red colonies were counted regardless of their size or color intensity. The countable range was between 10 and 100 CFU.

Chemical characterization

Total protein content. A nitrogen determinator (Leco TruMac N) was used to quantify the total protein content in beef samples. 1.0 ± 0.2 g of beef sample was loaded into a ceramic boat and was dried at 101 ± 1 °C for 45 min in a convection oven before the nitrogen analysis measurements. Blank boat weight was measured with a standard deviation of <0.002%. Ethylenediaminetetraacetic acid (EDTA) was used as a calibration standard. The furnace temperature was 1100 °C, with a lance flow of ≈1.8 L/min and purge flow of ≈4.2 L/min. A nitrogen-to-protein conversion factor of 6.25 was multiplied to report the protein content based on the measured nitrogen content.

Total volatile basic nitrogen (TVB-N). The Conway method was used to determine TVB-N. 10 g of minced beef, with 100 mL of DI water added, was homogenized for 30 min (Ultra-Turrax T25, IKA, Staufen, Germany).

Into the inner chamber of a Conway dish with a diameter, 3 mL of boric acid absorption solution (20 g/L) and 50 mL of pH indicator (mixture of methyl red and bromocresol green at 1:5 volume ratio) were added; and into the outer chamber 3 mL of potassium carbonate saturated solution and 1 mL of filtered liquid beef homogenate were added, successively.

The Conway dish was then sealed and incubated for 2 h at 37 °C. HCl (0.1 mol/L) was used to titrate against the boric acid-absorbed nitrogenous compounds. The TVB-N content \({c}_{{TVBN}}\)(mg/100 g) was quantified by

$${c}_{{TVBN}}=\frac{\left({v}_{1}-{v}_{2}\right)\times c\times 14}{m\times 0.05}\times 100,$$
(1)

where \({v}_{1}\) is the HCl titration volume for the tested sample (mL), and \({v}_{2}\) of the blank sample (mL); \(c\) is the actual concentration of HCl (mol/L); \(m\) is the weight of the beef sample (g).

Cytotoxicity assessment

3T3-L1 and CCD18co cell lines were obtained from the American Type Culture Collection (ATCC, CL-173). Cells were seeded and cultured in the Dulbecco’s modified Eagle’s medium, containing sodium bicarbonate (3.7 g/L), penicillin-streptomycin (1%), and bovine calf serum (10%). Adipocytes were developed after treating the post-confluent cells with fetal bovine serum (10%), 3-isobutyl-1-methylxanthine (0.5 mM), dexamethasone (1.0 mM), and insulin (1.67 mM) after two days.

Physical characterization

Scanning electron microscopy (SEM). Beef samples were freeze-dried before the SEM (Tescan XEIA FEG SEM, Brno, Czechia) experiments. The sample surface was coated with a platinum layer of 1 nm thick; and the SEM measurements were performed under 10 kV accelerating voltage. A 4pi Analysis System was used to acquire the images (Durham, NC).

Raman spectroscopy. Freeze-dried beef samples were used for the experiments, using a Horiba Jobin Yvon (Edison, NJ) confocal Raman spectrometer. The excitation laser wavelength was 532 nm and the objective lens magnification was 10×. Spectra were collected using a 600 g/mm grating. Neutral density filters were added to prevent sample damage if needed.

Texture analysis. A strain sweep was performed on the beef samples with a thickness of 2 mm at room temperature, using a rheometer in parallel plate geometry (TA Instruments). Plateau stress and strain at the end of the linear viscoelastic region were used to quantify the beef texture.

Histological microscopy. Beef samples were soaked in phosphate buffer (pH = 7.4) containing 4% (volume fraction) of formaldehyde first, and were then plunged into liquid nitrogen-cooled isopentane. A cryomicrotome (Leica Microsystems 3050S, Nussloch, Germany) was used for sectioning. Histological studies were performed using an optical microscope (Olympus VS-BX), and micrographs were registered by a digital camera (VC50).

Color analysis. Euclidian distances \({D}_{E}\) were computed based on 50 randomly picked pixels on a photograph of beef, using the following equation:

$${D}_{E}=\sqrt{{\left({R}_{2}-{R}_{1}\right)}^{2}+{\left({G}_{2}-{G}_{1}\right)}^{2}+{\left({B}_{2}-{B}_{1}\right)}^{2}},$$
(2)

where \(R\), \(G\), and \(B\) represent the three components, red, green, and blue, of the color at a given pixel. The subscripts (2 and 1) refer to two different pixels28.