Introduction
When processors evaluate low-cost, high-availability feedstocks for biodiesel or HVO production, palm oil mill effluent sits near the top of the list for good commercial reasons. POME is abundant wherever palm oil is produced, it carries a strong sustainability argument as a waste-derived input, and its cost profile compared to virgin vegetable oils or even used cooking oil can be genuinely attractive. Many fuel projects in Southeast Asia and beyond have been built, at least in part, around access to POME as a primary or secondary feedstock.
But POME is not a feedstock that rewards assumptions. It is one of the most compositionally complex and variable waste streams that a fuel plant can be asked to process, and the distance between its raw state and the clean, stable input that a transesterification reactor or a hydrotreating unit can work with is considerable. The controls required to cross that distance are not a preamble to the real process. They are a core part of it. How well the pretreatment system is designed, and how rigorously it is operated, has a direct and measurable effect on three things that define whether the plant is commercially viable: yield, uptime and downstream performance.
Understanding what POME actually demands of a pretreatment system — and why the demands are different in kind from what cleaner feedstocks require — is the starting point for designing a plant that can deliver reliably on all three.
What POME actually contains, and why it is not like other feedstocks
To understand why POME pretreatment is demanding, it helps to compare it against a cleaner feedstock. A degummed, neutralised vegetable oil — say a refined palm oil or degummed rapeseed oil entering a conventional biodiesel pretreatment unit — carries a relatively predictable impurity profile. Free fatty acids may be present but within a manageable range. Phospholipid and gum content is reduced. Moisture has been controlled. The main pretreatment task is to push residual FFA, phosphorus and trace contaminants below the thresholds that protect catalyst performance, and to do so consistently.
POME as a raw feedstock arrives in a fundamentally different condition. It is an aqueous emulsion of water, oil and fine solids generated during the palm oil milling process — specifically from sterilisation condensate, separator sludge and hydrocyclone wash water streams. In its untreated state, POME typically contains water at levels well above 90% by weight, along with suspended solids, colloidal particles, residual palm oil, free fatty acids, phospholipids, proteins, carbohydrates and a complex mix of metals including iron, calcium, magnesium and potassium. The oil fraction itself, once recovered, commonly carries FFA levels ranging from around 5% to over 15%, and in some cases higher depending on the mill and the freshness of the stream.
That is a very different starting point from refined or semi-refined oils. Every one of those components — the water, the solids, the metals, the high FFA — is a problem for a downstream fuel conversion process if it is not addressed upstream. And in POME, they arrive together, not individually.
Variability: the control challenge that never stands still
If POME's impurity load were fixed and predictable, designing a pretreatment system for it would be a defined engineering task. The real complication is that POME varies. Across sources, across mills, across seasons and across the time that passes between generation and processing.
Seasonal variation in palm fruit composition translates into changes in the oil content and FFA profile of the effluent. The age of fresh fruit bunches, harvesting practices, sterilisation parameters and the specific configuration of the palm oil mill all influence what ends up in the effluent stream. A POME source that delivers consistent feed quality over several months may shift meaningfully when the harvest period changes or when upstream milling conditions are adjusted. Processors who have designed their pretreatment systems around an average feed specification frequently find that the average is not what they actually receive on a consistent basis.
Storage and handling introduce additional variability. POME that has been held in lagoons — common practice for biological treatment before disposal — undergoes partial degradation. FFA levels rise as triglycerides hydrolyse. The microbial activity that makes lagoon treatment effective for wastewater purposes also accelerates the deterioration of the oil fraction. A plant receiving POME directly from the mill is working with a different material than one receiving lagoon-aged effluent, even from the same source.
Metal contamination levels can also shift with mill equipment condition, maintenance cycles and process water quality. Iron contamination in particular — introduced through contact with processing equipment — is a recurring issue that affects downstream catalyst performance in both biodiesel and HVO pathways. Its level is not static.
What this means for pretreatment design is that the system cannot be optimised for a single feed condition and left to run. It must be designed with the range of variability in mind, with control logic that can respond to changes in feed quality rather than simply processing at fixed parameters regardless of what arrives.
Yield: how pretreatment losses compound and where they are recovered
In a fuel plant processing POME, yield losses occur at multiple points in the pretreatment sequence, and their interaction is often underestimated at the design stage.
The first major yield consideration is oil recovery from the aqueous emulsion itself. Before any of the conventional pretreatment chemistry can be applied, the oil fraction must be separated from the water and solids matrix. Incomplete separation at this stage means that oil is carried out of the system in the water effluent or retained in solids — losses that do not show up immediately in the product quality but reduce the effective yield from the feedstock before any reaction has taken place. Recovery efficiency at this stage depends on temperature management, residence time, and whether the separation equipment is correctly sized and controlled for the actual feed viscosity and emulsion stability it is receiving.
Once the oil fraction is recovered, the high FFA content of POME-derived oil creates a second set of yield implications. In a conventional alkali-catalysed biodiesel transesterification route, free fatty acids above roughly 0.5 to 1% react with the sodium methylate catalyst to form soap rather than methyl esters, consuming catalyst and creating an emulsion that reduces both yield and separation efficiency. The higher the FFA entering the reaction stage, the more severe this effect. Without effective deacidification upstream, a plant feeding POME-derived oil directly to transesterification will not only lose yield on the FFA fraction itself — it will contaminate the main reaction with soap that compromises the entire batch or run.
This is why deacidification — whether through steam stripping in a packed column, through glycerolysis to convert residual FFAs into usable oil, or through a combination of both — is not optional in POME pretreatment for biodiesel production. It is a yield-protection step. Every percentage point of FFA that enters the reactor without being addressed upstream represents feedstock value that is converted into a waste or rework stream rather than into product. At the volumes involved in commercial fuel production, those losses are financially significant.
For the HVO pathway, where POME-derived oil is processed through catalytic hydrotreatment rather than transesterification, the FFA issue does not manifest as soap formation. But it still matters. High FFA levels contribute to higher hydrogen consumption during hydrotreatment, since free fatty acids require processing to remove the carboxyl group. More importantly, the metals and phospholipids that accompany POME-derived oil are potent catalyst poisons in the hydrotreatment environment. Phosphorus deactivates hydrotreatment catalysts by blocking active sites. Iron causes deposits that increase pressure drop and reduce catalyst life. These effects are cumulative and accelerate with POME's typically higher contamination levels compared to cleaner feedstocks.
Uptime: what the plant loses when pretreatment controls fail
The connection between pretreatment performance and plant uptime is often framed in terms of catalyst life, but the relationship is broader than that.
In a biodiesel plant, inadequate removal of gums and phospholipids leads to fouling in heat exchangers, reactors and separation equipment. Phospholipids that are not removed during degumming form deposits under heat that restrict flow, reduce heat transfer efficiency and eventually force cleaning shutdowns. The severity of this fouling scales with the phospholipid content of the incoming oil — and POME-derived oil, even after initial recovery, typically carries higher phospholipid loads than degummed virgin oils. A pretreatment system that is designed to handle that load under average conditions but not under peak or variable conditions will foul progressively until a planned or unplanned shutdown is required.
Metals, particularly iron and calcium, present a related but distinct uptime risk. Iron contamination leads to oxidative instability in the processed oil, which manifests as colour bodies, sediment and filter plugging downstream. It also accelerates catalyst deactivation in both biodiesel and HVO processes. Calcium and magnesium form deposits in high-temperature sections. In a HVO pretreatment plant, where the downstream hydrotreating unit operates under high pressure and temperature, these deposits can create serious reliability problems that are expensive to resolve once they are established in the catalyst bed.
Water content control is a third uptime lever that is particularly relevant for POME. Residual moisture entering a biodiesel transesterification reactor promotes hydrolysis of the methyl esters already produced — running the reaction in the wrong direction — and can cause emulsification that disrupts phase separation. In a hydrotreating unit, high moisture content in the feed increases hydrogen consumption and can cause water knockout issues that affect both the process and the plant's mechanical reliability. Vacuum drying to reduce moisture to below 0.05% before the reaction stage is a standard control, but it requires correctly sized and operated equipment. A dryer that is undersized for POME's high initial water content, or operated without attention to the actual water load it is receiving, will not reliably deliver the feed quality the downstream process depends on.
The pattern across all of these failure modes is the same: the controls that protect uptime are not independent. Fouling is worse when moisture is high. Catalyst deactivation accelerates when metals and phospholipids are not both addressed. Separation problems are compounded when FFA is high and soap has formed. A pretreatment system that performs well on some parameters but not others does not deliver proportionally partial protection — it delivers compounding risk.
Downstream performance: why clean feed defines output quality
In both biodiesel and HVO production, the quality and consistency of the final fuel product is heavily influenced by what the pretreatment system allows into the conversion stage. This is true for any feedstock, but it is especially true for POME, where the distance between raw feed condition and the required reactor input specification is large.
In biodiesel production, the connection between pretreatment and final product quality runs through several channels. Residual metals in the feed — even at trace levels — can appear in the finished methyl ester and create problems with oxidative stability, filter blocking and cold flow behaviour. Colour bodies that are not removed during bleaching contribute to off-specification product appearance. Soap contamination from incomplete FFA removal before transesterification produces a glycerol phase that is difficult to separate cleanly, increasing total glycerol content in the biodiesel and making the glycerol by-product harder to refine to commercial value.
For HVO, the downstream performance relationship is more concentrated around catalyst management. HVO is produced through catalytic hydrotreatment — a process in which the oil feedstock is reacted with hydrogen over a noble-metal or base-metal catalyst under elevated temperature and pressure. The catalyst is the most capital-intensive single component of the HVO conversion system, and its active life determines a significant portion of the plant's operating economics. Phosphorus, sulphur, metals and other contaminants that are not removed in pretreatment do not simply reduce product quality — they progressively and irreversibly deactivate the catalyst. The cost of catalyst replacement, combined with the downtime required to do it, means that pretreatment quality directly controls one of the plant's largest operating cost variables.
This is why the specification targets for POME pretreatment — FFA below 0.1%, phosphorus below 1 to 2 ppm depending on the downstream process, moisture below 0.05%, metals at or near detection limits — are not conservative safety margins. They are the thresholds at which the downstream conversion process can perform as designed. Feed that arrives close to but not within these thresholds does not produce proportionally acceptable results. It produces downstream processes that are compensating for upstream failures, with all the yield, efficiency and reliability consequences that compensation entails.
The control sequence: what a well-designed POME pretreatment system does
Given the complexity of POME as a feedstock, a robust pretreatment system is not a single-step operation. It is a staged sequence in which each step prepares the oil for the one that follows, and in which the combination of steps together brings the feed from its raw condition to the required specification.
Initial oil recovery from the aqueous POME stream — typically through settling, decanting and mechanical separation — is the first stage, and the one that determines how much of the feedstock's value is available to the rest of the process. This step requires attention to temperature, residence time and equipment sizing relative to the actual emulsion characteristics of the feed being received. Gravity decanting may be sufficient for some POME streams; others require centrifugal separation to achieve adequate recovery, particularly where the emulsion is tight or solids content is high.
Washing and drying follows, removing water-soluble impurities and reducing moisture to levels that allow downstream processing to proceed without hydrolysis risk. Water and oil are brought into counter-current contact to maximise impurity transfer into the aqueous phase, and the washed oil is then dried under vacuum to strip residual moisture. The vacuum dryer must be sized for the actual moisture load presented by POME-derived oil, which is substantially higher than for most other feedstocks.
Degumming and bleaching together address the phospholipid, metal and colour contaminant load. Acid treatment — typically with phosphoric acid — conditions the phospholipids for removal and also precipitates metal contaminants. Bleaching earth addition adsorbs the conditioned gums, residual soaps, colour bodies and remaining trace metals. Filtration then removes the spent earth and the contaminants it carries. The bleaching step requires careful control of acid dosing, retention time, temperature and earth addition rate — parameters that need to track the variability of the incoming POME-derived oil rather than being fixed at a single design-point setting.
Deacidification — steam stripping in a packed column operating under vacuum — removes free fatty acids to below the thresholds required for the downstream conversion process. For POME-derived oil with its high initial FFA content, this step carries a significant processing load. Where the FFA level is very high, glycerolysis — a reaction between the fatty acids and glycerol at high temperature under vacuum — can convert a portion of the FFA fraction into usable triglyceride oil rather than simply stripping it off as fatty acid vapour. This improves yield recovery from the high-FFA POME fraction while reducing the load on the steam stripping column.
Across all of these steps, the monitoring and control framework matters as much as the equipment itself. Feed quality testing on receipt, in-process monitoring at key stages, and the ability to adjust dosing, residence time and operating parameters in response to feed variability are what distinguish a pretreatment system that performs consistently across the full range of POME quality from one that performs well only when conditions are close to the design assumption.
POME across both pathways: same discipline, different tolerances
One of the practical realities of POME pretreatment is that the same feedstock is increasingly being processed across both the biodiesel and HVO pathways, and sometimes in plants designed to serve both. The pretreatment requirements overlap substantially, but the downstream conversion processes impose different sensitivity profiles on the cleaned feed.
In biodiesel transesterification, FFA and moisture are the primary quality drivers, since they directly affect catalyst performance and soap formation in the reaction stage. Metal contamination matters but the tolerances are somewhat more forgiving than in hydrotreatment, and the catalyst — typically sodium methylate — is a relatively low-cost consumable compared to a hydrotreatment catalyst.
In HVO hydrotreatment, catalyst protection becomes the overriding concern. Phosphorus and metals are the critical specifications, because even trace contamination in the feed can cause measurable and cumulative catalyst deactivation over the operating cycle. The hydrotreatment catalyst operates at elevated temperature and pressure, and its active sites are more sensitive to contamination than the catalyst environment in a transesterification reactor. This means that while the pretreatment process steps are broadly the same across both pathways, the target specifications — particularly for phosphorus and metals — are tighter for HVO, and the bleaching and degumming sections must be designed and operated to reflect that.
Plants designed for multi-feedstock or multi-pathway operation — which is increasingly the commercial model for flexible fuel production — need pretreatment systems that can meet the stricter HVO specification without over-engineering the system to a degree that adds unnecessary cost and complexity for biodiesel operation. That balance is a design decision, not a default.
Where process design capability is the deciding factor
The controls that protect yield, uptime and downstream performance in a POME pretreatment system are not complicated in concept. They are demanding in execution — precisely because POME is variable, because the processing steps interact, and because the consequences of inadequate control show up not just in one place but across the whole plant.
A pretreatment system designed around a single, ideal feed specification will perform acceptably when POME arrives close to that specification and will underperform progressively as it diverges from it. The seasonal shifts, the source variability, the handling and storage effects that characterise real POME supply chains mean that divergence from ideal is the norm, not the exception. Building a system that accommodates that range — through appropriate equipment sizing, through control logic that responds to variability rather than assuming it away, and through the integration of each pretreatment step with the ones before and after it — is where the process design work is most important.
This is the same principle that applies across pretreatment in general: that process engineering depth matters more than equipment specification alone. The individual units — settlers, dryers, bleachers, columns, filters — are not the difficult part. What is difficult, and what determines whether the plant actually delivers, is how they are configured, sequenced, controlled and integrated as a system against the real feed conditions the plant will encounter over its operating life.
For processors working with POME as a feedstock — whether for biodiesel, HVO, or both — the right technical conversation begins with an honest assessment of what the feed actually contains, how variable it is likely to be, and what the downstream process genuinely requires. Kumar's experience across biodiesel pretreatment and HVO pretreatment provides a foundation for that assessment, and for the process design work that follows it. Whether the requirement is a new POME-fed plant, an upgrade to an existing pretreatment system, or an evaluation of feedstock strategy, that conversation is the place to start.
