# Spore Removal Techniques for Milk Purification

Topic radar · 45 documents · 2021-01-01–2026-09-22 · updated 2026-10-26

Canonical: https://slate.greyb.com/radar/spore-removal-techniques-for-milk-purification/innovations

## In one minute

### The separator is the only machine sold on spores specifically
A bactofuge is a centrifuge run hot enough, usually 50–60 °C, to thin the milk so spores and cells migrate outward faster than the cream. Tetra Pak rates its range at 5,000 to 55,000 litres an hour with up to 99% reduction of incoming bacterial levels, naming aerobic spores such as Bacillus cereus and anaerobic spores as the target [1]. Alfa Laval sells the same class of machine on the same claim [2].

The filings show what the design variables actually are. Huangshi's process runs a purifier centrifuge at 30,000 litres an hour below 50 °C, states removal of more than 98% of anaerobic and more than 95% of aerobic spores, and then evaporates, fills aseptically and freezes the concentrate to −18 °C [3]. An independent trial on preheated raw milk measured 99.5% total bacterial reduction and 95% aerobic spore reduction, which puts the vendor claim in the right range [4].

The goat-milk family attacks the same lever from the other end. Instead of running the separator harder, it feeds raw milk in below the bowl's capacity and runs the bowl 20% to 200% above the feed rate, so the milk stays in the disc stack longer. That buys a product at or below 500 spores per millilitre and more than one log of spore reduction, and the family is filed in six jurisdictions [5] [6].

1. [tetrapak.com](https://www.tetrapak.com/solutions/integrated-solutions-equipment/processing-equipment/separation/tetra-pak-bactofuge-unit)
2. [alfalaval.com](https://www.alfalaval.com/products/separation/centrifugal-separators/separators/dairy/dairy-separation-solutions/)
3. [CN108142563B](/?paper=CN108142563B&utm_source=radar&utm_medium=dashboard)
4. [ojs.uel.br](https://ojs.uel.br/revistas/uel/index.php/semagrarias/article/view/43980)
5. [NZ782247A](/?paper=NZ782247A&utm_source=radar&utm_medium=dashboard)
6. [US12667113B2](/?paper=US12667113B2&utm_source=radar&utm_medium=dashboard)

### The removal step is already sold, and it is sold on other grounds
Bactofugation is a catalogue product from at least three vendors, ranked on throughput and specified on bacterial reduction. GEA positions its version as two bacteria-removing separators dropped into an existing pasteurisation line, sold on low investment and running cost. Used 25,000 litre-per-hour machines trade secondhand.


### Every process patent in this field belongs to a dairy, and three of the four families are Chinese
The filings that describe how a plant should be run come from operators, each naming a product the company already makes. Huangshi Group South China Dairy claims a low-temperature separation sequence that freezes the concentrate instead of drying it. Guangxi Royal Dairy Industry claims a buffalo-milk line built around a 0.09-second hold at 144 °C. Xinjiang Fengtuoyuan Dairy Technology claims three filter tanks in series for camel milk, with a cleaning circuit into every stage. Dairy Protein Coop Food claims the bowl-residence trick for goat milk, filed in New Zealand and granted in the US.


### Membranes get more logs, and the pore size is the setting that decides how many
Cold ceramic microfiltration at 1.2 micrometres removed 4.57 log of Bacillus licheniformis spores from skim milk, against 2.17 log at 1.4 micrometres, while Geobacillus spores came out almost completely at either pore size. Tetra Pak states up to 99.95% removal of bacteria and spores ahead of UHT.


### The approvals reaching market cover pathogens, and spores survive them
The clearest commercial clearance of the period is Tamarack Biotics' TruActive ultraviolet platform, accepted by the FDA and confined to powdered dairy ingredients such as whey and milk protein concentrate, with liquid raw milk targeted for 2027. The claim behind it is log reduction of pathogens, five logs of the most UV-resistant organisms and seven including Mycobacterium avium paratuberculosis. No spore inactivation is reported for it. The route to market is licensing the platform to dairy ingredient producers.


### The separator makers patent the drum, and their filings are older
GEA Westfalia Separator's 2024 application claims a distributor with second openings that feed part of the product outward into the drum, described as a bacteria-removing separator for sterilising milk. Tetra Laval's 2019 filing stacks a sub-stack of microorganism-clarification discs and a sub-stack of milk-fat separation discs in one bowl, sharing a distribution channel. GEA Mechanical Equipment's 2017 separator places a clarifying agent inside an inner drum and has drawn twelve forward citations across eleven jurisdictions.


### Two complete processes exist, and the newer one puts the milk through 144 °C
Only two filings describe a full sequence rather than a unit operation.


### The one regulation that moved was a border certificate on a contaminated ingredient
On 24 February 2026 the European Commission required consignments of arachidonic acid oil from China to carry an official certificate and laboratory results showing absence of cereulide, with identity and physical checks on half of consignments and a transitional window to 26 April. It followed recalls that began in December 2025 and reached 99 countries and territories, with 144 suspected and confirmed cases across ten countries. Nestlé, Danone and Lactalis were among the companies recalling, with losses reported in millions of dollars of sales.


## Overview


Spore removal is engineered, catalogued and priced. Every working route reaches the same answer: take the spore out of the milk physically. Tetra Pak specifies up to 99% reduction of incoming bacterial levels on its bactofuge range and up to 99.95% of bacteria and spores ahead of UHT on its membranes.


## Research

### The 2026 design work has moved to plumbing, not pores
The newest filing treats staging as the invention. Xinjiang Fengtuoyuan's utility model for camel milk puts three filter tanks in series, each with its own element and a three-way valve that routes the first two stages onward and the third to output. A separate feed, pump and discharge assembly connects to the rinse inlet and outlet of every tank, so each stage can be cleaned on its own.


### The removal step works, and what happens after it decides the shelf life
The corrective comes from outside the filings. A study of 287 packages of microfiltered, pasteurised extended-shelf-life milk found 15% spoiled, and in 55% of the spoiled packages the culprit was a pure culture of Bacillus cereus at above 6 log per millilitre. Bulk-tank spore levels were low and the dominant raw-milk species was never detected in the finished product, meaning the barrier had done its job and the plant had recontaminated the milk afterwards.


### The combined train has been described on paper, and not demonstrated at scale
No source reports a plant running a separator and a membrane bank together and measuring the result. A European filing does describe that sequence: bactofuge treatment to remove microorganisms, fractionation into cream and skim, microfiltration of the skim to take out what remains, and pasteurisation at 65 °C to 80 °C with no prior high-temperature step.


## Market

### Spore limits arrive as customer specifications, so the buying base is narrow
Grade A rules govern raw milk sanitation ahead of pasteurisation and set no spore number for fluid milk. Powdered infant formula codifies limits for Cronobacter and Salmonella. Where spore numbers do bind, they bind as a buyer's tolerance: thermophilic spores are the main concern in dairy powders, and processors in that work were targeting customer limits below 1,000 per gram.


### The published market size overstates the spore opportunity
The filtration figures disagree badly enough that they cannot be one number. One report puts membrane filtration systems for dairy at $2.4 billion as a current level; another puts membrane filtration for dairy at $15.1 billion by 2033; a third forecasts the equipment segment to 2032 without stating a value. A present level, a future target and an unscoped forecast are being read as one market.


### Capacity is shifting to China, where both complete process designs are filed
The dairy processing equipment market is projected at USD 12.73 billion in 2026 rising to USD 17.36 billion by 2031, about 6.4% a year, according to MarketsandMarkets. China alone is estimated at USD 7.2 to 7.8 billion in 2026, with liquid processing and aseptic filling lines taking 55 to 60% of demand and import dependence of 30 to 35% for high-speed aseptic fillers and membrane filtration systems.


### The market has not yet shown a buyer, and one customer has gone the other way
No named dairy buys a removal-only spore barrier at plant scale in this record. The goat-milk process carries no launch, order or price, and its stated value pool is goat protein for infant formula, where a spore specification is a condition of sale. Novozymes replaced a depth-filtration system with UV-based raslysation, citing energy cost and safety, and planned to repeat the switch at other sites.


## Strategy

### The 2026 filings claim plumbing and sequence, not a new way to take spores out
No recent filing in this field claims a new physical mechanism. The buffalo-milk family claims the order of barriers: centrifugal pretreatment, a very short hot step, homogenisation after sterilisation, ISO 5 filling. The camel-milk utility model claims staging, three-way valve logic and per-tank cleaning. The goat-milk family claims a feed rate below bowl capacity with the bowl running above it, which is an operating regime.


### Inactivation research has not produced a spore capability, and its effort stays exploratory
The kill-side routes remain at laboratory or pilot scale across the whole evidence base. Cold plasma inactivated E. coli O157:H7, Salmonella and Listeria in skim milk and left Bacillus cereus spores intact. Ultra-high-pressure homogenisation reached over 5 logs on Bacillus pumilus and Bacillus subtilis at 250 MPa on the bench. Pulsed electric field licensing covers vegetable oils, juices, wine and potatoes from April 2026, with dairy absent.


### The farm-side work is a separate community solving a different part of the problem
Upstream effort is published as protocols and cost models by university groups, and none of it is patented. Cornell's package of chlorine, drying and staff training was run on five New York farms, then costed per farm and modelled as a choice between paying farmers on raw-milk counts and buying a plant-side technology. A 2025 simulation then tied the payoff to how often a plant recontaminates milk after pasteurisation.


### The engineering direction is settled, and the effort now goes into arrangement
Removal before heat, more barriers in series, shorter hot steps, and value captured in validated trains and cleaning circuits. That is the direction the 2026 filings describe, and none of them needs a new physical mechanism to get there.


## Areas

- Nonthermal Processing & Preservation · 10 documents
- Upstream Prevention & Farm-Level Control · 9 documents
- System Integration & Fractionation · 4 documents
- Downstream Physical Separation & Purification · 3 documents
- Market, Regulatory & Commercialization · 8 documents
- Cross-Chain Reviews & Spoilage Analysis · 4 documents