Non-Abrasive Cleaning Methods for Turbine Components: Preserving Tolerances and Coatings
A technical assessment of cleaning methods that maintain surface integrity on precision turbomachinery
Executive Summary
Modern gas turbine components rely on advanced coating systems and tight dimensional tolerances to achieve optimal aerodynamic performance and service life. Thermal barrier coatings (TBCs), MCrAlY bond coats, and chromide diffusion layers are particularly sensitive to cleaning methods that can alter surface profiles, remove protective material, or create stress concentration points. This paper examines why surface preservation is critical for turbine components, what damage conventional cleaning methods can cause, and how non-abrasive alternatives compare in terms of coating compatibility and dimensional integrity.
The paper is intended for maintenance engineers, quality assurance professionals, and turnaround planners responsible for turbine overhaul decisions. The key takeaway is straightforward: cleaning method selection directly affects component service life and should be documented as part of QA/QC records, with non-abrasive methods warranting serious consideration for coated and precision-tolerance components.
Key takeaways:
- TBC, MCrAlY, and diffusion coating systems are sensitive to both abrasion and chemical attack during cleaning operations.
- Abrasive blasting can thin coatings, alter surface profiles, and introduce stress risers that initiate fatigue cracks.
- Non-abrasive methods (dry ice blasting, CO2 snow, ultrasonic) preserve coating integrity and dimensional tolerances.
- Cleaning method selection should be documented in QA/QC records, with OEM guidance consulted first.
1. Introduction: Why Surface Integrity Matters
Gas turbine performance depends fundamentally on blade geometry. Even minor surface deviations can significantly impact aerodynamic efficiency. Research indicates that surface roughness changes in the transitionally rough regime (approximately 106 micrometres) decrease normalized turbine efficiency by approximately 4% with roughened stator or rotor blades, and up to 8% when both are affected [1]. In the fully rough regime (400 micrometres), efficiency losses can reach 19% with roughness on both stator and rotor blades [1].
These figures represent substantial economic consequences. For a 100 MW gas turbine operating at base load, even a 1% efficiency reduction translates to significant additional fuel consumption annually. Beyond efficiency, surface damage creates stress risers that initiate fatigue cracks during cyclic thermal and mechanical loading [2].
The maintenance paradox is clear: cleaning is essential to restore performance degraded by fouling, but inappropriate cleaning methods can cause permanent damage that accelerates component degradation. This is particularly true for modern turbines that rely on sophisticated coating systems operating at the limits of material capability.
2. Turbine Coating Systems and Their Sensitivity
2.1 Coating Types and Functions
Modern hot-section turbine components employ multi-layer coating systems, typically comprising:
Thermal Barrier Coatings (TBCs): These 100 micrometre to 2 mm thick ceramic layers (usually yttria-stabilized zirconia) provide thermal insulation, allowing higher gas path temperatures while protecting the metallic substrate [3]. TBCs are applied via air plasma spray (APS) or electron beam physical vapor deposition (EB-PVD), creating characteristic columnar or lamellar microstructures.
MCrAlY Bond Coats: Applied beneath TBCs, these metallic coatings (where M represents nickel, cobalt, or iron) provide oxidation resistance and mechanical compliance between the ceramic top coat and superalloy substrate [4]. Typical thickness ranges from 75 to 150 micrometres.
Diffusion Coatings: Aluminide and chromide coatings are formed by diffusing aluminum or chromium into the substrate surface. These coatings become integral to the base material structure, making them particularly sensitive to aggressive stripping or cleaning methods [5].
2.2 Coating Failure Mechanisms
TBC systems fail through several mechanisms that can be accelerated by inappropriate cleaning:
- Thermally grown oxide (TGO) thickening: The alumina layer that forms between the bond coat and ceramic top coat is beneficial in thin layers but leads to spallation when it exceeds critical thickness [3].
- Interfacial delamination: Loss of adhesion between coating layers, often initiated by mechanical damage or contamination at interfaces.
- Erosion and CMAS attack: Particle impact and calcium-magnesium-alumino-silicate contamination degrade the ceramic layer.
The literature indicates that coating failure modes including surface cracking and interfacial delamination are sensitive to both thermal cycling and mechanical stress [3]. Cleaning methods that introduce additional mechanical stress or alter the coating microstructure can accelerate these failure mechanisms.
3. Damage from Conventional Cleaning Methods
3.1 Abrasive Blasting
Traditional grit blasting, while effective at contaminant removal, presents several documented concerns for coated turbine components:
Material removal and thinning: Grit blasting can result in uneven material removal and thinning of the base metal [5]. For components with walls that have become thinner through service oxidation, this additional material loss may render parts unserviceable.
Surface profile alteration: Abrasive media create surface profiles that differ from the original machined or as-coated finish. Studies indicate that grit blasting can leave surface finishes of approximately 160 microinches (4 micrometres Ra), which may exceed allowable roughness specifications for aerodynamic surfaces [6].
Stress concentration introduction: Surface irregularities created by abrasive impact act as stress risers. Research confirms that corrosion pits and surface irregularities exceeding critical size are likely to initiate fatigue cracks [2]. The same principle applies to abrasive-induced surface damage.
Interface contamination: Grit blasting can re-contaminate the surface and destroy bond interface integrity, complicating subsequent coating operations [5].
Practice note: On components that will be re-coated after cleaning, some OEMs specify controlled grit blasting to achieve the required anchor profile for bond coat adhesion. The concerns above apply specifically to cleaning operations where the intent is to preserve existing coatings, not prepare surfaces for re-coating.
3.2 Chemical Cleaning
Chemical methods offer material-selective cleaning but carry their own risks:
Intergranular attack (IGA): Acid stripping, particularly of MCrAlY coatings, can result in stress corrosion cracking, pitting, and alloy depletion [5]. Many OEMs and operators limit repair cycles because of IGA concerns.
Coating dissolution: Aggressive chemical solutions may attack not only contaminants but also protective coating materials. The literature notes that cleaning in acid often leads to deeper corrosion on grain boundaries, which can be unfavourable under cyclic mechanical loading [7].
Residue concerns: Chemical cleaning requires thorough rinsing to remove all traces of cleaning solutions. Any remaining chemical residue can attack coating materials and components during subsequent service [8].
3.3 Waterjet Methods
High-pressure water cleaning and abrasive waterjet (AWJ) processes represent intermediate approaches:
Traditional hydroblasting introduces moisture that can lead to oxidation concerns and, for certain coating types, may cause damage through thermal shock or hydraulic penetration into porous coating structures [9].
Precision AWJ with controlled parameters can remove coatings without compromising base metal integrity and avoids intergranular attack [5]. However, the process still involves abrasive particle interaction with surfaces and requires careful parameter control.
4. Non-Abrasive Cleaning Alternatives
4.1 Dry Ice Blasting
Dry ice blasting uses solid carbon dioxide pellets accelerated by compressed air to clean surfaces. The process is fundamentally different from abrasive methods due to several characteristics:
Low hardness: Dry ice has a Mohs hardness of approximately 1.5 to 2.0, making it softer than virtually all metallic substrates and coating materials [10]. For comparison, aluminum has a Mohs hardness of 2.5-3, while ceramic TBC materials and metallic bond coats are considerably harder.
Sublimation on impact: Dry ice pellets undergo phase change from solid to gas almost instantaneously upon impact. Very little kinetic energy transfers into the substrate surface, making the process inherently non-abrasive [10].
No secondary waste: Unlike abrasive media that require collection and disposal, dry ice sublimates completely, leaving only the removed contaminant as waste.
Adjustable aggressiveness: Blast pressure, pellet size (from micro-particles up to 3 mm diameter), and feed rate can be adjusted to match cleaning requirements, from aggressive carbon deposit removal to delicate sensor cleaning [10].
Research on dry ice blasting effects on metal surfaces indicates surface roughness changes limited to 8-12%, substantially below abrasive methods [11]. Critically, dry ice blasting does not create a surface profile but rather reveals the existing surface texture, which is essential for maintaining as-machined or as-coated finishes.
Field experience indicates that dry ice blasting is effective for removing carbonaceous deposits, oil films, and particulate contamination from turbine components without altering blade profiles or damaging coating systems. However, the literature is thin on quantitative before/after coating thickness measurements and long-term fatigue performance data. This represents a gap where additional published research would benefit the industry.
Practice note: When dry ice blasting coated turbine components, start with reduced blast pressure (2-3 bar / 30-45 psi) and increase only as needed. The adjustability of the process is its primary advantage for coated surfaces — there is no reason to use full pressure on a lightly fouled TBC surface.
4.2 CO2 Snow Cleaning
CO2 snow cleaning uses liquid carbon dioxide expanded through a nozzle to create a mixture of solid CO2 particles (snow) and gaseous CO2. The process operates at lower velocities than pellet blasting and is particularly suited for precision cleaning applications.
The cleaning mechanism combines mechanical momentum transfer, thermal effects (CO2 snow temperature of -78.5 degrees Celsius freeze-fractures organic contaminants), and solvent action of liquid CO2 on hydrocarbon-based residues [12].
Studies indicate that CO2 snow cleaning removes more than 95% of micrometre-sized contamination from test surfaces, compared to 88% for ultrasonic cleaning and 74% for high-speed air jets [12]. The process produces no surface profile change and is widely used in semiconductor and precision optics manufacturing where nanometre-scale surface integrity is required.
For turbine applications, CO2 snow cleaning is most applicable to smaller components, instrumentation, and final cleaning operations where residual particle contamination must be minimized.
4.3 Ultrasonic Cleaning
Ultrasonic cleaning uses high-frequency sound waves (typically 28-80 kHz, up to 100-120 kHz for delicate components) to generate cavitation bubbles in a liquid bath [13]. Bubble implosion near surfaces creates localized cleaning action that reaches internal passages, blind holes, and complex geometries inaccessible to line-of-sight methods.
Advantages for turbine components include:
- Effective cleaning of cooling passages and film cooling holes
- Non-contact cleaning mechanism that does not alter surface profiles
- Compatibility with various cleaning chemistries tailored to specific contaminants
- Repeatable, documented process parameters for QA/QC compliance
Frequency selection matters: lower frequencies (28 kHz) provide more aggressive cavitation for heavy deposits, while higher frequencies (80-120 kHz) are gentler for delicate geometries [13].
The primary limitation is that components must be immersion-compatible, and certain coating types may be sensitive to extended ultrasonic exposure. Additionally, cleaning chemistry must be selected for compatibility with all materials present, including base metal, coating systems, and any brazed joints.
4.4 Laser Cleaning
Laser cleaning represents an emerging technology for precision contaminant removal. Research demonstrates that continuous wave fiber lasers can remove CMAS (calcium-magnesium-alumino-silicate) contaminants from TBC surfaces without damaging the underlying yttria-stabilized zirconia coating [14].
The process offers precision targeting of contaminants, minimal thermal input to the substrate, and no media consumption. However, equipment costs are higher than conventional methods, and the technology is not yet widely adopted for routine turbine maintenance. Current applications focus primarily on high-value aerospace components where coating preservation justifies the investment.
5. OEM Guidance and Industry Standards
5.1 OEM-Approved Methods
Major turbine OEMs including GE, Siemens, and Pratt & Whitney maintain approved lists of cleaning methods and chemistries for their equipment. Key principles from available OEM repair guidance include:
- Chemical cleaners and stripping agents must be evaluated for compatibility with engine materials and coating systems before use [15]
- Cleaning solutions should not chemically attack protective blade coatings — particularly TBCs, bond coats, and diffusion coatings [15]
- Demineralized water rinsing should follow all cleaning cycles to remove chemical residue [15]
The literature indicates that cleaning chemicals approved by one OEM may not be approved by another, emphasizing the importance of consulting engine-specific component maintenance manuals rather than assuming universal applicability.
5.2 Surface Finish Standards
ASME B46.1-2019 is the primary standard in North America for specifying and measuring surface texture parameters including roughness average (Ra), waviness, and lay [16]. The standard provides measurement methodology but does not prescribe specific Ra limits for turbine components; these limits are defined by individual OEMs in their engineering specifications.
Published research indicates that turbine blade surface roughness values in operational studies range from Ra/c ratios of 7.6 x 10^-6 to 7.9 x 10^-5 (where c is chord length) [17]. For a blade with 100 mm chord, this corresponds to Ra values from approximately 0.76 to 7.9 micrometres. However, specific acceptance criteria vary by component type, stage location, and OEM requirements.
For metallographic evaluation of coatings, ASTM E3 (specimen preparation), ASTM E407 (microetching), and ASTM E1920 (thermal spray coating preparation) provide relevant methodology [18].
5.3 Aerospace Quality Standards
AS9100 and related aerospace quality management standards require documented procedures, traceability, and evidence of conformance for maintenance processes including cleaning [19]. Key requirements applicable to turbine cleaning include:
- Process parameter documentation (time, temperature, chemistry, pressure)
- Traceability to specific work orders and component serial numbers
- Calibration records for measurement equipment
- Qualification records for personnel performing critical operations
The literature emphasizes that cleaning is not an isolated task but part of the maintenance ecosystem, with documentation requirements ensuring that cleaning process decisions are traceable to quality records [19].
6. Cleaning Method Selection: A Decision Framework
Selecting appropriate cleaning methods requires balancing multiple factors:
6.1 Contaminant Type
- Light oils and greases: Solvent-based cleaning, CO2 snow, or aqueous degreasing
- Carbonaceous deposits: Dry ice blasting, ultrasonic with appropriate chemistry
- Particulate fouling: Dry ice blasting, ultrasonic, water wash
- CMAS contamination: Laser cleaning (emerging), specialized chemical treatment
- Oxide scale: Chemical methods with careful coating compatibility assessment
6.2 Component Characteristics
- Coated surfaces: Prioritize non-abrasive methods; verify chemical compatibility
- Cooling passages: Ultrasonic or flow-through methods that access internal features
- Thin-wall sections: Avoid aggressive mechanical methods that risk material removal
- Brazed joints: Non-abrasive methods that do not undermine braze material
6.3 Documentation Requirements
For QA/QC traceability, cleaning records should document:
- Cleaning method employed and equipment identification
- Process parameters (pressure, temperature, duration, chemistry)
- Personnel qualifications
- Inspection results confirming contaminant removal
- Evidence of coating/surface integrity preservation (photographs, measurements where applicable)
- Reference to OEM-approved methods where required
Field note: Photographing components before and after cleaning with a calibrated scale bar provides low-cost evidence for QA/QC records. Pair with surface roughness measurements at marked locations when coating integrity is critical.
7. Recommendations and Best Practices
Based on the reviewed literature and industry guidance, the following practices are recommended for cleaning coated and precision-tolerance turbine components:
1. Consult OEM documentation first. Component maintenance manuals and service bulletins specify approved cleaning methods. Deviation from OEM guidance requires documented engineering justification.
2. Default to non-abrasive methods for coated components. Dry ice blasting, ultrasonic cleaning, and CO2 snow cleaning preserve coating integrity where abrasive methods may cause damage.
3. Match cleaning aggressiveness to contamination. More aggressive cleaning is not better cleaning if it damages substrate or coating materials. Adjustable parameters (pressure, frequency, chemistry concentration) allow optimization.
4. Document cleaning method selection rationale. QA/QC records should capture why a particular method was selected, particularly when multiple methods are acceptable.
5. Verify surface integrity post-cleaning. Visual inspection, surface roughness measurement, and coating thickness checks (where baseline data exists) confirm that cleaning has not caused damage.
6. Consider total lifecycle impact. A cleaning method that is slightly less effective at contaminant removal but preserves coating integrity may extend component life and reduce total maintenance cost.
8. Conclusion
Turbine component cleaning is a maintenance activity with significant implications for component service life and equipment efficiency. Modern coating systems and tight dimensional tolerances require cleaning methods that remove contaminants without altering the engineered surface condition.
Non-abrasive methods including dry ice blasting, CO2 snow cleaning, and ultrasonic cleaning offer demonstrated capability to clean turbine components while preserving surface integrity. Dry ice blasting is particularly notable for its inherent non-abrasiveness (Mohs hardness approximately 1.5), complete sublimation leaving no secondary waste, and adjustability across a wide range of cleaning requirements.
The industry would benefit from additional published research quantifying coating preservation across cleaning methods, including coating thickness measurements, surface roughness data, and fatigue life correlation studies. Such data would strengthen the basis for cleaning method selection and support development of consensus industry guidelines.
For maintenance organizations, the immediate actions are clear: review current cleaning practices against OEM guidance, ensure documentation supports QA/QC traceability requirements, and consider non-abrasive alternatives where coating and tolerance preservation are priorities.
References
[1] Y. Yun and J. Park, “Performance Degradation due to Blade Surface Roughness in a Single-Stage Axial Turbine,” Semantic Scholar, 2015.
[2] Allied Power Group, “Introduction To Gas Turbine Blade Damage,” 2024.
[3] A. G. Evans et al., “Mechanisms controlling the durability of thermal barrier coatings,” Progress in Materials Science, 2001.
[4] TWI Global, “What is an MCrAlY coating?”
[5] Power Engineering, “Removing Turbine Component Coatings Without Damage,” 2018.
[6] STACOAT, “Surface preparation metal: Cleanliness, Roughness and Profile,” Technical Document.
[7] U. Krupp, K. Wackermann, H.-J. Christ, et al., “Intergranular Oxidation Effects During Dwell-Time Fatigue of High-Strength Superalloys,” Oxidation of Metals, Vol. 88, pp. 3-14, 2017.
[8] “Evaluation of the aluminide coating on cleaned internal passages of used gas turbine blades,” Surface and Coatings Technology, 2016.
[9] T. Bergs et al., “Pure waterjet controlled depth machining for stripping ceramic thermal barrier coatings on turbine blades,” Procedia CIRP, Vol. 85, pp. 261-265, 2019.
[10] A. Dzido and P. Krawczyk, “Abrasive Technologies with Dry Ice as a Blasting Medium — Review,” Energies, Vol. 16, No. 3, Article 1014, 2023.
[11] “Influence of dry ice blasting process properties on surface roughness and residual stresses,” Procedia CIRP, 2022.
[12] “CO2 snow cleaning of miniaturized parts,” Microelectronic Engineering, 2017.
[13] Kaijo Shibuya, “Advanced Ultrasonic Cleaning for Aerospace Manufacturing,” 2024.
[14] “Laser cleaning process of high-pressure turbine blade,” Surface and Coatings Technology, 2023.
[15] R.C. Bauer et al., “Gas Turbine Repair Technology,” GE Power Systems, GER-3957B, 2001.
[16] ASME, “B46.1-2019 Surface Texture (Surface Roughness, Waviness, and Lay),” 2019.
[17] “Surface Roughness Effects on Turbine Blade Aerodynamics,” Journal of Turbomachinery, Vol. 127, 2005.
[18] ASTM International, “Aerospace Material Standards,” 2024.
[19] Graymills, “Optimizing Parts Cleaning in Part 145 Aviation Maintenance,” 2025.