Application of Multivariate Analysis and LC-MS for the Detection of Counterfeit Cosmetics

Significance of the Topic

Counterfeit cosmetics pose significant health risks and economic losses to consumers and manufacturers. Rapid and reliable methods for distinguishing genuine products from fakes are essential for customs enforcement, quality control laboratories, and brand protection.

Objectives and Study Overview

This work evaluates a non-targeted approach combining high-resolution LC-MS and multivariate statistical analysis to differentiate authentic and counterfeit cosmetics. The study focuses on three product types (cream, lotion, serum) sourced from a US manufacturer and visually identical samples obtained from an online retailer in Asia.

Methodology and Instrumentation

Sample Preparation:

• Dilution of cosmetic formulations in tetrahydrofuran (5 mg/mL)

Chromatography:

• Waters ACQUITY UPLC I-Class system
• CORTECS UPLC C18 column (2.1 × 100 mm, 1.6 µm, 40 °C)
• Gradient elution (0.1 % formic acid in water/methanol; 20 %–99 % B over 2.5 min, hold, re-equilibration)

Mass Spectrometry:

• Waters Xevo G2-XS QTof
• ESI in positive and negative modes
• MS range 50–1200 m/z, desolvation 450 °C, capillary 2.0–3.0 kV

Data Processing:

• UNIFI Scientific Information System for peak detection and marker table generation
• EZInfo for PCA, PLS-DA, OPLS-DA, and S-plot analysis

Main Results and Discussion

Multivariate models (unsupervised PCA, supervised PLS-DA) achieved clear separation of authentic versus counterfeit samples, particularly in negative ionization mode. OPLS-DA and S-plot analysis identified two consistent markers unique to all counterfeit products (m/z 151.0409 at 2.59 min; m/z 179.0725 at 3.64 min). Automated elemental composition and fragment matching suggested these compounds are methylparaben and propylparaben—preservatives avoided in high-end authentic cosmetics.

Benefits and Practical Applications

This workflow enables rapid screening of cosmetics at points of entry or in manufacturing QC, even when authentic formulations are not fully disclosed. It can be extended to other consumer goods (food, beverages, pharmaceuticals) requiring quality and authenticity assessment.

Future Trends and Opportunities

• Expansion of spectral and chemometric libraries to cover a broader range of ingredients
• Integration of machine learning models for automated anomaly detection
• Development of portable LC-MS platforms for on-site screening
• Real-time data sharing between customs agencies and manufacturers

Conclusion

The combination of high-resolution LC-MS with multivariate statistical tools provides a powerful, non-targeted approach for the rapid identification of counterfeit cosmetics. Key chemical markers were successfully identified, demonstrating the method’s potential for routine authenticity testing and brand protection.

Reference

1. OAMI. Observatory news: Counterfeit cosmetic losses. 2015.
2. CCAP Congress Brussels. Fast screening of counterfeit goods. 2012.
3. U.S. FDA. Cosmetics ingredients. 2014.
4. Business Insurance. The truth about cancer-causing cosmetics. 2013.
5. FIT BEAUT. Paraben-free cosmetics discussion. 2012.