Donkey Milk Study: Ultrasound, Heat, and Lab Curdling (Hou 2024)

Cow, goat, buffalo, and donkey milk were treated with ultrasound (200–1000 W) or heat (65–100 °C). In donkey milk, the β-sheet content of the proteins increased from 43.89% to up to 48.33%; its rennet gel was too weak for texture measurement.

The study in brief

• Original title: Impact of ultrasonic and heat treatments on the physicochemical properties and rennet-induced coagulation characteristics of milk from various species
• Authors: Hou W., Ma X., Yu Z. et al.
• Journal: Ultrasonics Sonochemistry, 111, 107084 (2024)
• Country: China · Study type: Laboratory study

How it was investigated

Raw milk from cows, goats, buffaloes, and donkeys from one farm each in China (donkey milk from Liaocheng, Shandong) was transported refrigerated and stored at −80 °C until the study. Three batches were prepared for each milk type: untreated control, ultrasound treatment in an ultrasound bath (20 min at 200, 400, 600, 800, or 1000 W, water temperature 20 ± 2 °C), and heat treatment in a water bath (65 °C/30 min, 80 °C/30 min, 90 °C/15 min, 100 °C/15 min). Turbidity, protein and fat globule size, zeta potential, protein secondary structure (FTIR, Amide I band), intrinsic fluorescence, and surface hydrophobicity were measured.

Key figures

• 4 – milk types compared (cow, goat, buffalo, donkey)
• 200–1000 W – five ultrasound levels, each for 20 minutes at 20 ± 2 °C
• 65–100 °C – four heat levels (65 and 80 °C for 30 min each, 90 and 100 °C for 15 min each)
• 287 ± 2.1 – protein particle size of untreated donkey milk, the highest value among the four milk types (unit not specified in the article)
• 371 ± 7.2 – highest protein particle size of donkey milk after heat treatment (unit not specified)
• 43.89% → 48.33% – β-sheet content of donkey milk proteins: control vs. 600 W ultrasound

Results

Untreated, donkey milk showed the lowest turbidity, the lowest absolute zeta potential, the lowest intrinsic fluorescence, and the lowest surface hydrophobicity among the four milk types, but the largest protein particles (287 ± 2.1; unit not specified). The authors attribute this to the composition of the caseins in donkey milk.

In donkey milk, turbidity increased significantly after heat treatment from 80 °C/30 min; the protein particles became larger after 80 °C/30 min and 100 °C/15 min (maximum 371 ± 7.2). The size distribution of the fat globules in donkey milk hardly changed with either treatment, while cow and goat milk showed a bimodal distribution after 100 °C/15 min.

Ultrasound at 400 and 600 W increased the β-sheet content in donkey milk (from 43.89% to 48.17% and 48.33%, respectively) and decreased α-helix (to 21.18% and 21.01%, respectively) and β-turn (to 17.39% and 17.38%, respectively); at 800 and 1000 W, the values were again close to the control. After heat treatment, the α-helix and random coil content decreased slightly, while the β-sheet content increased to 45.43% and 45.69% at 65 °C and 100 °C, respectively.

Limitations of the study

• Origin of the samples unclear: For each milk type, the milk came from a single farm; the number of animals, breed, lactation stage, and whether they were individual or pooled samples are not specified. The three replicates are parallel batches of the same milk. In addition, the milk was frozen at −80 °C before treatment, which can already change the protein structure and fat globules.
• Donkey milk only partially documented: For donkey milk, numerical values are only available for the secondary structure (Tables 1 and 2) and two particle sizes; turbidity, zeta potential, water-holding capacity, and rheology are only shown as figures. The unit for the particle sizes is missing, and the formula for the water-holding capacity is not printed in the text.
• Inconsistent control values: The untreated controls should be identical in Tables 1 and 2, but they differ: donkey milk β-turn 19.04 ± 0.12 vs. ± 0.13 and random coil 14.49 vs. 14.47%, buffalo milk β-turn 18.99 vs. 17.99%. The cohesion of the buffalo milk gels (6.52–9.49) is about an order of magnitude higher than the values for cow and goat milk (0.12–0.38), which suggests a unit or transcription error.
• Contradictions between text and method: The abstract mentions a heat level of 63 °C/30 min, the method 65 °C; in the rheology section, a maximum at "83 °C for 30 min" is described, but no such level existed. The conclusion speaks of increased modules due to the treatments, while the results section speaks of decreasing modules at higher intensity – the differences mainly concern the milk types compared to each other.
• Nominal ultrasound power: The treatment was carried out in an ultrasound bath with three samples side by side; only the nominal power of the device is specified. The energy actually delivered to the milk was not determined, which makes comparison with other studies difficult.

Level of evidence: Basic research (laboratory) – describes the properties of milk, no statement about effects in humans.

Our assessment

The study is a food technology laboratory investigation on rennet coagulation. It describes how protein structure and gel formation change under defined laboratory conditions; statements about nutritional value, effect, or the treatment of milk in the household cannot be derived from it.

Source: Hou W., Ma X., Yu Z., Bari L., Jiang H., Du Q., Fan R., Wang J., Yang Y., Han R. (2024). Impact of ultrasonic and heat treatments on the physicochemical properties and rennet-induced coagulation characteristics of milk from various species. Ultrasonics Sonochemistry. DOI: 10.1016/j.ultsonch.2024.107084 · Summarized by the DON DONKEY'S editorial team, as of October 2026.

End of the study summary. The study was conducted independently by DON DONKEY'S® and does not refer to our products. Our donkey milk is a food.