Sterilization Methods Comparison Microbiology: Options Compared

Q: Can heat-sensitive liquids be sterilized?

Yes, by filtration through a 0.22 micrometer or 0.1 micrometer membrane with pre-use and post-use integrity testing. Filtration removes bacteria but not viruses, so viral safety depends on separate validated steps.

By Dr. Zubair Khalid, DVM, MS, PhD · 2026-11-10

Sterilization Methods Comparison Microbiology: Options Compared

Sterilization destroys or removes every viable microorganism on an item, including bacterial spores, while disinfection only reduces the microbial load and leaves spores and other resistant forms behind. The rule of thumb for choosing between methods is straightforward: pick the least aggressive method that can reach every surface of the load with a validated cycle, then prove the cycle worked with a biological indicator instead of trusting the gauge on the panel.

That rule sounds simple until you put it against real items. A stainless steel bone reamer, a flexible endoscope, a bottle of antibiotic solution, and a bench top contaminated with a spill all demand different processes, different validation, and different failure analysis. A sterilization methods comparison in microbiology is therefore not a ranking of technologies from strongest to weakest. It is a comparison of how each method delivers lethality to a surface, how it interacts with materials, and how you can demonstrate that it worked.

This guide walks through the main options, compares them on the criteria that actually decide a laboratory or clinical choice, provides worked scenarios, and closes with the mistakes that show up most often on bench audits.

Definitions that make the comparison meaningful

A few terms have to be settled before any comparison holds up.

Sterilization is a validated process that renders an item free of viable microorganisms. It is usually expressed as a sterility assurance level, or SAL, of 10⁻⁶, meaning a probability of fewer than one viable microorganism per million processed units. Textbook microbiology treats SAL as a statistical property of a validated cycle, not a statement about a single item.

Disinfection is a reduction in the number of pathogenic microorganisms on a surface or device. High-level disinfection kills all vegetative bacteria, mycobacteria, fungi, and most viruses but does not reliably kill large numbers of bacterial spores. Intermediate and low-level disinfection are progressively weaker categories.

The Spaulding classification links the two concepts to device use. Critical items enter sterile tissue or the vascular system and require sterilization. Semicritical items contact mucous membranes and require at least high-level disinfection. Noncritical items contact intact skin and require low-level disinfection. Most reprocessing errors trace back to a device being treated at the wrong category.

Bioburden is the number and type of microorganisms present on an item before processing. It matters because every sterilization cycle is validated against an expected bioburden, and a load that arrives with dried blood or soil carries far more organisms than the validation assumed.

D-value is the time at a given temperature or concentration required to reduce a population by 90 percent, or one log. F₀ is the equivalent exposure time at 121 °C summed across a cycle. Both are used to design and compare thermal processes.

Biological indicators are carriers, usually paper strips or glass discs, inoculated with a defined spore population, most commonly 10⁶ spores. Geobacillus stearothermophilus is the standard organism for steam and hydrogen peroxide systems. Bacillus atrophaeus is used for dry heat and ethylene oxide. A biological indicator is the closest thing a bench has to a direct readout of lethality.

Prions are infectious misfolded proteins that resist standard steam cycles and many chemical disinfectants. They are a separate problem from bacterial spores and drive specific, longer, more aggressive protocols in facilities that handle neurological tissue.

Sterilization methods comparison microbiology: the options side by side

The table below places the main families next to each other. Parameters are typical textbook cycle ranges. Always use the cycle written in the sterilizer manufacturer's validated instructions and in the device manufacturer's reprocessing instructions, because those documents, not the table, define what is allowed.

Method	Killing mechanism	Typical cycle	Kills spores	Best suited to	Main limitation
Moist heat (steam)	Denatures and coagulates proteins under saturated steam pressure	121 °C at 15 psi for 15 to 20 minutes, or 132 to 134 °C for 3 to 4 minutes	Yes	Stainless steel, glass, heat-stable plastics, media, liquids, waste	Heat and moisture damage, air trapping, no prion inactivation
Dry heat	Oxidizes cell components slowly at high temperature	160 °C for 2 hours, 170 °C for 1 hour, or 180 °C for 30 minutes	Yes	Glass, metal, powders, oils, sharp instruments	Very long cycles, poor penetration, damages many materials
Ethylene oxide	Alkylates nucleic acids and proteins	37 to 63 °C for several hours, plus aeration	Yes	Heat-labile complex devices, catheters, implants	Toxicity, flammability, long aeration, residual limits
Hydrogen peroxide gas plasma or vapor	Oxidizes through free radicals	45 to 55 °C, cycles of roughly 30 to 75 minutes	Yes	Heat-labile instruments, some lumens	Lumen length and diameter limits, material compatibility
Liquid chemical sterilants	Oxidize or cross-link microbial structures	Label contact times from minutes for high-level disinfection to hours for sterilization	Yes at sterilant contact time	Heat-labile immersible devices	Residuals, staff exposure, contact time errors, biofilm protection
Ionizing radiation	Breaks DNA strands	25 kGy typical gamma dose, or lower doses for electron beam	Yes	Single-use devices, packaging, industrial scale	Polymer degradation, capital equipment, dose mapping
UV-C light	Cross-links nucleic acids	Minutes of direct exposure at 254 nm	Partial	Air, water, exposed surfaces	Line-of-sight only, shadowing, no crevice penetration
Filtration	Physically retains organisms	0.22 µm or 0.1 µm membranes	Yes for bacteria	Heat-labile liquids, media, sera, antibiotics	Does not remove viruses, clogging, integrity testing required
Chlorine dioxide	Strong oxidation	Gas or liquid cycles set by generator or label	Yes in validated systems	Surfaces, chambers, some foods	Bleaching, corrosion, generation safety
Non-thermal physical methods	Cavitation, light pulses, pressure, volumetric heating	Varies widely	Variable	Food matrices, packaging, surfaces	Often pasteurization-level, not sterilization-level

Option by option: what each method actually does

Moist heat (steam autoclave)

Steam under pressure is the default sterilization method in microbiology for one reason: it is fast, it penetrates well, it leaves no toxic residue, and it is easy to monitor. Saturated steam contacts the load and condenses, releasing latent heat that denatures proteins. The critical word is saturated. Dry steam heats slowly and unevenly, and superheated steam behaves like hot air.

Two cycle designs dominate. Gravity displacement autoclaves let air escape through a port as steam enters from the top. Prevacuum autoclaves pull a vacuum first to remove air from the chamber and the load, which is why they can run short, high-temperature cycles and why they need an air-removal test, such as the Bowie-Dick test, at the start of the day.

Air is the most common cause of steam failure. Air trapped inside a lumened instrument, a tightly closed bottle, or a crowded pouch creates a pocket that steam cannot reach. Loading patterns, packaging, and container design matter as much as the pressure gauge.

There is an important distinction between killing organisms and removing material. In a study of surgical instruments that retained bone, cement, and fatty tissue debris after previous use, autoclave sterilization at 121 °C and 1 atmosphere for 20 minutes was repeated 350 times across five instrument types. No positive cultures were detected in any repetition [1]. The organisms were dead. The debris was not gone. Retained material still provides a surface for biofilm formation, can shield future bioburden from steam, and may carry endotoxin or prion risk that the cycle does not address.

Steam limits are equally clear. It corrodes carbon steel, dulls some fine edges after repeated cycles, and damages heat-sensitive plastics and electronics. It does not inactivate prions under standard conditions. It is poor for anhydrous materials such as oils, powders, and grease, which require dry heat.

Dry heat

Dry heat kills by oxidation. Because there is no moisture to conduct heat and no latent heat release, the process is slower and needs higher temperatures. Standard textbook cycles are 160 °C for 2 hours, 170 °C for 1 hour, and 180 °C for 30 minutes, with the time counted from the moment the entire load reaches temperature.

Dry heat suits glassware, metal instruments, powders such as talc, oils, and items that must stay dry. It penetrates poorly through dense loads and reaches surfaces unevenly in a forced-air oven without adequate circulation. It destroys most plastics, rubber, and any material that oxidizes or warps.

Monitoring uses Bacillus atrophaeus biological indicators, and the long cycle times make dry heat a poor choice for high-throughput work. Its niche is the loads that steam cannot touch.

Ethylene oxide

Ethylene oxide is an alkylating gas. It penetrates packaging, paper, and many plastics, and it sterilizes at low temperature, which is why it became the standard for heat-labile medical devices.

The tradeoffs are substantial. Ethylene oxide is a known irritant and carcinogen, it is flammable, and loads require aeration to remove residuals before use. Cycles are measured in hours rather than minutes, and aeration can add many more. Devices need moisture and temperature conditioning before the gas phase, so packaging and load configuration are constrained. Small lumens are difficult to sterilize reliably.

Ethylene oxide remains the practical answer for complex heat-labile devices that cannot tolerate hydrogen peroxide or moisture, and it is monitored with Bacillus atrophaeus indicators.

Hydrogen peroxide gas plasma and vaporized hydrogen peroxide

Hydrogen peroxide systems generate reactive oxygen species that attack membranes, proteins, and nucleic acids. The gas plasma version keeps the load at roughly 45 to 55 °C, needs no aeration because the residue breaks down into water and oxygen, and runs cycles of about 30 to 75 minutes.

These systems are clean and fast, which is why they spread into hospitals and laboratories as an alternative to ethylene oxide. Their weakness is geometry. Hydrogen peroxide vapor does not penetrate long, narrow lumens well, and device manufacturers specify maximum lumen length and minimum diameter. Materials that absorb peroxide, such as cellulose and some nylons, interfere with the cycle. Load density and packaging can block vapor access.

Liquid chemical sterilants

Liquid chemical sterilants include glutaraldehyde, ortho-phthalaldehyde, peracetic acid, and hydrogen peroxide formulations. They require complete immersion, displacement of air from lumens, correct temperature, correct concentration, and strict adherence to the labeled contact time. High-level disinfection typically takes minutes, and sterilization claims usually require many hours of immersion.

Performance differences between agents can be large even when both carry the same disinfection claim. In a comparison of automated flexible endoscope reprocessors for nasopharyngoscopes, 374 devices processed with ortho-phthalaldehyde were compared with 302 processed with hypochlorous acid under the same cleaning and machine protocol. The hypochlorous acid group had significantly lower colony-forming units after disinfection and a higher proportion of devices reaching zero or low colony counts [2]. Both agents were used correctly. The results differed.

Residuals are a separate hazard. Every liquid sterilant needs a defined rinse step, usually with sterile or filtered water, and a drying step. Residual glutaraldehyde and ortho-phthalaldehyde are tissue irritants and respiratory sensitizers. Residual peroxide damages tissue and inhibits cell growth in culture work.

Radiation: gamma, electron beam, and X-ray

Ionizing radiation breaks DNA and is used at industrial scale for single-use devices and packaged supplies. Gamma irradiation from cobalt-60 at a typical dose of 25 kGy penetrates full pallets, which is why it supports terminal sterilization of devices already in their final packaging. Electron beam delivers a similar effect at higher dose rates and lower penetration. X-ray sits between the two.

Radiation is a cold process relative to steam, but it is not gentle. Many polymers yellow, embrittle, or cross-link, and some drugs and biological materials lose activity at sterilizing doses. Dose mapping, dosimeter placement, and validation of dose uniformity across the load are essential. Radiation sterilization is not a bench method in most laboratories.

UV-C light

UV-C at 254 nm damages nucleic acids and is useful for air, water, and exposed surfaces. It does not penetrate, it is blocked by dust and organic film, and it cannot reach crevices, lumen interiors, or shadowed areas. A UV cabinet can reduce surface contamination on a clean bench. It cannot sterilize a complex instrument, and it should not be labeled as doing so.

Filtration

Filtration is sterilization of a fluid rather than of a surface. A 0.22 µm membrane retains essentially all vegetative bacteria, and a 0.1 µm membrane is used where mycoplasma or small bacteria are a concern. Viruses and bacteriophages pass through both, which is why viral clearance in biopharmaceutical production uses dedicated orthogonal steps rather than a single filter.

Filtration is the only practical option for heat-labile liquids such as antibiotic solutions, vitamin supplements, sera, and some protein-containing media. It requires a pre-sterilized housing, an integrity test before and after use (bubble point or diffusion), and a fluid path that does not leach or bind the product.

Chlorine dioxide and other gaseous oxidants

Chlorine dioxide is a strong oxidant with a lower pH dependence than hypochlorite and fewer chlorinated byproduct concerns. In a systematic review of disinfection methods for edible sprouts, chlorine dioxide was identified as a promising alternative to sodium hypochlorite and calcium hypochlorite, with higher oxidizing power and fewer carcinogenic byproducts, while hypochlorite and hydrogen peroxide reduced microbial load but raised concerns about seed viability, chemical residues, and limited action against internalized organisms or biofilms [3].

That same logic applies at the bench. Gaseous or liquid chlorine dioxide can reach surfaces that liquids miss, but it bleaches, corrodes stainless steel at the wrong concentration, and requires careful generation and monitoring. It is a disinfectant in most applications and a sterilant only inside validated, purpose-built equipment.

Non-thermal physical methods

Microwave, radiofrequency, and infrared heating deliver volumetric energy that can reduce pathogens while also driving off moisture. A review of microbiological safety in dehydrated foods noted that these electromagnetic technologies achieve log-scale pathogen reductions and improve drying efficiency through volumetric heating, and that hybrid approaches coupling a physical sterilization step with precision drying reached reductions in the 2.3 to 4.7 log range [4].

That number is the point. A 4-log reduction is a pasteurization-level process, not sterilization. High-pressure processing, pulsed light, cold plasma, and ultrasound sit in the same territory. They are valuable for foods, packaging, and heat-sensitive surfaces, and they should not be described as sterilization unless a validated process has demonstrated a 10⁻⁶ SAL for that specific matrix.

Comparing on the criteria that decide real choices

Material compatibility

Start with the item, not the technology. Stainless steel and glass tolerate steam and dry heat. Polypropylene tolerates steam. Polycarbonate, some silicones, and many electronics do not. Optically coupled devices, adhesives, and batteries frequently rule out steam and dry heat and push the choice toward hydrogen peroxide, ethylene oxide, or a liquid process. Radiation degrades specific polymers. Any method can be wrong for a material, so the device manufacturer's instructions are the controlling document.

Penetration and access

Sterilization happens at the surface, so access is everything. Steam needs air removed and condensate drained. Ethylene oxide penetrates packaging but struggles in long narrow lumens. Hydrogen peroxide vapor has defined lumen limits. Liquid sterilants need every channel filled and every bubble displaced. UV is line-of-sight only. Filtration works only where fluid actually passes through the membrane.

Turnaround and workflow

Steam wins on speed. A prevacuum cycle can be under an hour from load to ready, with drying included. Dry heat takes hours. Ethylene oxide adds aeration time. Hydrogen peroxide plasma is intermediate. Liquid chemical sterilization can take many hours of immersion plus rinse and dry. Throughput decisions usually come down to how many items are needed and how fast.

Residuals and neutralization

A sterilant that stays on the item is not a finished process. Contact lens care systems illustrate the principle. In a comparison of a 0.05 percent povidone iodine system and a 3.42 percent hydrogen peroxide system, iodine reached a peak concentration of about 0.05 percent within 2 minutes, held that level for 5 minutes, and was neutralized by 20 minutes, and it exceeded the standardized disinfection criteria against all tested strains within 1 minute. The peroxide system relied on a different neutralization profile [5]. The lesson transfers directly to bench work: the contact time and the neutralization step are part of the method, and skipping either changes the outcome.

Load complexity and failure modes

Simple, unwrapped, single-item loads fail in simple ways. Complex loads with lumens, hinges, mated surfaces, and mixed materials fail in ways that are hard to see. The failure modes differ by method. Steam fails through air pockets and wet loads. Ethylene oxide fails through inadequate humidity, overloaded chambers, and insufficient aeration. Hydrogen peroxide fails through lumen geometry and absorbent packaging. Liquid sterilants fail through incomplete immersion, diluted concentration, and shortened contact time.

Validation burden

Every method needs a monitoring hierarchy. Physical monitoring covers time, temperature, pressure, and concentration. Chemical indicators confirm exposure and, in higher classes, penetration. Biological indicators confirm lethality. Routine validation includes installation, operational, and performance qualification, plus revalidation after repairs or process changes. The burden scales with the complexity of the method.

Biofilm changes the answer

Planktonic bacteria in a broth are easy to kill. Bacteria inside a mature biofilm are not. Biofilm is a structured community embedded in an extracellular matrix of polysaccharides, proteins, and DNA. The matrix limits diffusion, the cells inside are metabolically slow, and the underlying surface is protected from mechanical removal. A sterilant that performs well in suspension may fail against the same species in a biofilm.

Endodontic models make this visible because they use a single species in a confined space. In a root canal biofilm model with Enterococcus faecalis, all irrigation activation methods tested produced significant reductions from baseline, but the activation method determined how much was removed [6]. A separate comparison of sonic activation, passive ultrasonic irrigation, and Er:YAG laser activation with photon-induced photoacoustic streaming found differences in both bacterial reduction and smear layer removal when all groups received 2.5 percent sodium hypochlorite [7]. A third study comparing laser and conventional protocols found the Er:YAG group had the highest clearance efficiency for E. faecalis, ultrasonic irrigation was lower than Er:YAG, and sonic irrigation, photodynamic therapy, and conventional syringe irrigation did not differ significantly, while all five methods cleared Candida albicans well [8].

Two conclusions transfer to general sterilization work. First, chemistry alone often underperforms when the biofilm is protected, and mechanical energy that disrupts the matrix changes the result. Second, species matter. In a dual-species biofilm model, E. faecalis persisted across all intracanal medicament groups with no significant differences, while C. albicans responded differently between groups [9]. Intrinsic tolerance is a property of the organism, not just of the disinfectant.

Cleaning is the rate-limiting step

The most consistent finding across reprocessing literature is that cleaning, not the sterilization cycle, determines whether the process succeeds. Soil protects organisms, neutralizes disinfectants, and blocks vapor and heat.

This shows up in cleaning verification data. A rapid protein detection kit evaluated for endoscope channels detected 5 micrograms or more of bovine serum albumin, equivalent to 0.4 micrograms per square centimeter, and recovered an average of 68 percent of the initial soil. Under real-use conditions it identified 51 percent of endoscopes as non-compliant, compared with 14 percent by a standard flush sampling method [10]. The devices were the same. The measurement method changed the answer, and the more sensitive method found more failures.

Cleaning hardware also matters. A comparison of a pulsed vacuum cleaning and disinfection machine against a spraying cleaning and disinfection machine for rigid endoscope instruments found that the pulsed vacuum group had a longer sterilization time but a significantly lower instrument damage rate and endotoxin positive rate, and significantly higher ATP bioluminescence and protein residue pass rates [11]. Better cleaning changed downstream metrics.

The practical rule is that a sterilization cycle should never be asked to compensate for inadequate cleaning. If an item arrives with visible soil, the cycle is a heat treatment of dirty material, not a sterilization event.

Worked scenarios

Scenario 1: Stainless steel surgical instruments in a small practice

A veterinary practice runs a benchtop autoclave for dental elevators, scalers, and surgical packs. The correct process is pre-cleaning and rinsing to remove visible soil, ultrasonic or manual cleaning of hinges and ratchets, drying, inspection, pouching with a chemical indicator, and a validated prevacuum or gravity cycle at 132 to 134 °C for the time specified by the sterilizer manufacturer. Wrapped packs need a drying phase and must be stored so that the packaging is not compressed or punctured.

Monitoring includes a physical readout, a chemical indicator in every pack, and a biological indicator at the frequency set by the practice's policy, typically at least weekly and with every load containing implants. When a cycle fails, the load is not released. It is repackaged and reprocessed after the cause is identified.

Scenario 2: A heat-labile flexible endoscope

A flexible nasopharyngoscope cannot be steam sterilized. It is a semicritical device that requires high-level disinfection after thorough cleaning. The process runs through an automated reprocessor with a validated disinfectant, and the choice of disinfectant affects results, as the hypochlorous acid versus ortho-phthalaldehyde comparison showed [2]. After disinfection, the channel is flushed and the device is dried and stored vertically in a way that prevents retained moisture. Storage conditions and drying are frequent failure points.

Where reprocessing capacity or compliance is unreliable, single-use devices change the risk profile. A systematic review and meta-analysis comparing disposable and reusable endoscopes found a 25 percent higher complication risk with reusable devices (risk ratio 1.25, 95 percent confidence interval 1.08 to 1.44) and a much lower fever incidence in the disposable group (risk ratio 0.38, 95 percent confidence interval 0.20 to 0.72) [12]. That is a system-level signal about reprocessing reliability, not a claim that every reusable scope fails.

Scenario 3: A heat-labile culture medium with antibiotics

A laboratory needs 2 liters of a selective medium containing an antibiotic that degrades above 60 °C. Steam is not an option. The sterile path is filtration through a 0.22 µm membrane into a sterile receiver, with an integrity test before and after filtration. The basal medium is autoclaved separately and cooled, and the two are combined aseptically. The filter housing must itself be sterilized, and the membrane must be compatible with the solvent and pH of the solution.

Scenario 4: A spill on a laboratory bench

A spill of an overnight bacterial culture is a disinfection problem, not a sterilization problem. The surface is cleaned first, then treated with an appropriate disinfectant at its labeled concentration and contact time. Chlorine dioxide and hypochlorite both work as oxidants, with the tradeoffs in byproduct formation and material damage described in the sprout disinfection review [3]. The bench does not become sterile, and it does not need to be. It needs to be safe to use.

Scenario 5: A drain or sink reservoir in a clinical area

Drains are a well-documented reservoir for resistant organisms. A systematic review of drain decontamination found that peracetic acid and other chemical and thermal disinfection methods can remove carbapenemase-producing Enterobacteriaceae from hospital drains, but only for short periods, and that the drainage system as a whole cannot be eradicated [13]. The lesson is that a single terminal disinfection is a temporary measure against a persistent reservoir, and repeated cycles plus engineering controls are part of the method.

Common Mistakes and Limitations

Treating disinfection as sterilization. A high-level disinfectant with a 12-minute contact time is not a sterilant, even if the same active ingredient appears in a sterilant product at a much longer contact time. The label claim, concentration, temperature, and time define the process.

Confusing visible cleanliness with sterility, or sterility with cleanliness. A sterilized instrument with retained debris is sterile in the culture sense and unacceptable in the clinical sense [1]. Both properties have to be verified separately.

Overloading the chamber. Steam and ethylene oxide both depend on space for circulation. Dense packs, closed containers, and stacked pouches create barriers that no gauge will detect.

Skipping air removal on prevacuum cycles. A failed Bowie-Dick test means the cycle should not be used for wrapped goods. Air removal is a daily check, not an annual one.

Using the wrong biological indicator. Geobacillus stearothermophilus for steam, Bacillus atrophaeus for dry heat and ethylene oxide. A mismatched indicator gives a false pass or a false fail.

Assuming a log reduction equals sterilization. A 4-log reduction is a 99.99 percent reduction. A 10⁻⁶ SAL process is a different specification entirely, and the difference matters for implants, sterile body sites, and cell culture.

Trusting ATP or protein tests as sterility tests. ATP bioluminescence and protein residue assays measure cleaning, not lethality [10][11]. They are valuable, and they answer a different question.

Ignoring residuals and neutralization. Peroxide, iodine, glutaraldehyde, and ortho-phthalaldehyde all need a defined removal step. A device that is sterile and still carrying sterilant is not ready for use [5].

Reusing a single-use device without validation. Reprocessing a device labeled single-use requires validation of cleaning, sterilization, and functional performance. Without that work, the process is unverified.

Assuming prions behave like spores. Standard steam cycles do not inactivate prions. Facilities that handle high-risk neurological tissue need specific, more aggressive protocols and dedicated instrument sets.

Treating a persistent environmental reservoir as a one-time problem. Drains, sinks, and water systems recolonize. Decontamination without a repeated schedule and without attention to the reservoir as a whole gives a short window of improvement [13].

Individual items, loads, and facility constraints still need to be matched to a validated protocol written for that specific device and that specific sterilizer. The comparison above narrows the options. It does not replace the manufacturer's instructions, the sterilizer's validated cycle, or the facility's own quality program.

Frequently Asked Questions

What is the difference between sterilization and disinfection?

Sterilization kills or removes all viable microorganisms including spores, and it is validated to a defined sterility assurance level. Disinfection reduces microbial numbers without a guarantee against spores.

What is the best sterilization method for a small clinic or laboratory?

Steam is the best default for heat-stable instruments, glassware, and media because it is fast, penetrates well, leaves no toxic residue, and is easy to monitor. Heat-labile items need hydrogen peroxide, ethylene oxide, or a validated liquid process instead.

Does autoclaving kill everything?

Steam inactivates bacteria, fungi, viruses, and bacterial spores when the cycle is validated and the load is properly prepared. It does not reliably inactivate prions, and it does not remove endotoxin or residual debris.

Can heat-sensitive liquids be sterilized?

Yes, by filtration through a 0.22 µm or 0.1 µm membrane with pre-use and post-use integrity testing. Filtration removes bacteria but not viruses, so viral safety depends on separate validated steps.

How do I know a sterilization cycle actually worked?

Use three levels of monitoring together: physical parameters such as time, temperature, and pressure, chemical indicators that confirm exposure and penetration, and a biological indicator with the correct spore species for the method. A cycle is released only when all three agree.

Why does cleaning matter if sterilization kills everything?

Soil shields organisms from heat and chemicals, neutralizes disinfectants, and blocks vapor access to surfaces. A sterilization cycle applied to a soiled item is a heat treatment of protected bioburden, and the outcome is not predictable from the cycle parameters alone.