Proteus® now quantify
Frequently Asked Questions
Your Questions. Answered Clearly.
Whether you're new to Proteus® Now Quantify or exploring how it fits into your workflow, our FAQ hub provides in-depth answers to the most common questions. Browse by topic to find the information you need.
Technical Questions and Functionality
Hardware, software, and compatibility – everything you need to run Quantify smoothly.
Understanding Quantify
Machine Learning (ML) is a subfield of artificial intelligence where algorithms learn from data to detect patterns and make predictions—without being explicitly programmed for each task. In our software, we use a combination of traditional ML models, like decision trees, and deep learning models such as neural networks to uncover complex interactions that would be difficult or impossible to detect using conventional methods.
Quantify is a cloud-based application and cannot be installed on a local PC.
Near real-time, thanks to optimized cloud algorithms.
If polymers outside the trained dataset are present, unexpected peaks will appear and are marked in the result. Quantification is not possible, but identification via Proteus®® Identify is recommended.
Yes, but only if the samples are homogeneous and show low variation. We recommend uploading individual curves for best results.
Model Training & Updates
Our model is trained on thousands of expert-curated datasets containing well-defined mixtures of representative polymers. This allows the model to detect subtle differences and even challenging compositions accurately.
No, only virgin polymer blends with defined mixing ratios typical for recycling streams are used. While recyclates behave slightly differently, the variance within a polymer class is greater than between virgin and recycled materials—so the model remains valid for recyclates.
No. Your data are never used to train or improve the model. All training data come from internal, expert-validated measurements.
Regularly. New polymer types are added continuously to expand functionality.
Measurement & Sample Requirements
10 K/min is a common standard in DSC analysis. While lower rates improve peak separation, this heating rate offers a balance of resolution and measurement speed and is optimized for Quantify’s performance.
Such components are not fully integrated into the model and may reduce accuracy. For best results, determine the inorganic content separately using TGA or muffle furnace ash analysis and adjust the sample weight accordingly.
Use homogenized pellets instead of flakes or ground material. Pellets are processed and homogeneous, while flakes or powders may vary locally.
- Heating/cooling rate: 10 K/min
- Sample weight: 10 ± 1 mg
- Valid calibration (sensitivity, temperature)
See our tutorial “How to Run a DSC Measurement for Quantify Analysis” for detailed instructions.
Results, Accuracy & Validation
Yes. We use standard validation techniques like cross-validation and provide RMSE values to indicate confidence in the prediction. The model is rigorously tested using separate training and test sets before deployment.
RMSE (Root Mean Square Error) gives an indication of model accuracy and is affected by the amount and quality of training data for each polymer class.
The model can detect contamination levels down to 1%. However, detection depends on the polymers and their compatibility. For example, 1% LLDPE in HDPE is usually not detectable due to co-crystallization.
Material Compatibility
Supported Polymers and Upcoming Material Expansions
In its initial release, Quantify supports common thermoplastic packaging polymers, such as:
- High-density polyethylene (HDPE)
- Low-density polyethylene (LDPE, LLDPE)
- Polypropylene (PP)
- Semi-crystalline polyethylene terephthalate (PET)
More materials (e.g., engineering plastics, amorphous thermoplastics) are under development and can be prioritized upon request.
? A detailed overview of all currently supported polymers as well as our development roadmap can be found on the page Supported Materials and Roadmap.
Quantify is trained on virgin polymer blends with defined mixing ratios that mirror typical recycling streams. Virgin materials provide clean, reproducible thermal fingerprints, and accurate mixing ratios are essential for building a reliable model. While recyclates differ in many properties such as viscosity, the thermal fingerprint is not significantly affected for most polymers. Moreover, the variation within a polymer class is larger than the difference between virgin and recycled grades, ensuring the model remains valid for recyclates while maintaining consistency and robustness in the database.
Quantify can still analyze your measurement and provide the best possible match with the available training data. In the case of semi-crystalline contaminants, an additional peak may appear in the DSC curve that is not highlighted by Quantify. This is a clear indicator that the results should be cross-checked using the NETZSCH Identify software for identification without quantification. For amorphous polymer contaminants, detection is more challenging, since their glass transition may not be clearly visible. In such cases, the contaminant effectively reduces the enthalpy value of the sample, leading to an erroneous prediction. Full support for filled systems is part of our long-term roadmap.
At present, filled polymers are not fully integrated into the Quantify model and may reduce accuracy. Because fillers do not produce a DSC thermal fingerprint, the inorganic fraction must be determined separately (e.g., by TGA or muffle furnace ash analysis) and subtracted from the sample weight for reliable results. Quantify can then correctly analyze the polymer fraction. Full support for filled systems is part of our long-term roadmap, and expanded functionality will be introduced in future updates.
Subscription & Pricing
Managing Your Plan, Payments, and Access
You can cancel anytime via the User Management section—no later than 3 months before the current subscription period ends. Cancellations take effect at the end of the billing period. Refunds for past payments are not possible.
You will retain read-only access to your data for 3 months. After that, data will be permanently deleted unless you renew your subscription.
Pausing is not possible. You may cancel at the end of the term and resubscribe later. Note that stored data will be deleted 3 months after the subscription ends.
Subscriptions renew automatically each year unless canceled. You’ll receive a reminder email 3 months before expiration.
Pricing is based on the number of DSC uploads per year. Packages can be upgraded anytime via the User Management section and are activated immediately after payment. Downgrades must be requested through your NETZSCH sales contact.
We accept bank transfers, SEPA direct debit, and purchase orders with invoicing.
Version 9.8 or higher is needed to export files for Quantify. Contact your local NETZSCH representative for upgrade options.
Yes, you can export analyses as PDF or CSV during an active subscription.
This is currently not possible. For each subscription you can have multiple users with their own account. In the future, it will be possible to add users and access rights in one account.
Yes, all data are stored on certified, GDPR-compliant servers in the EU. Communication is encrypted.
Only .ngb-quant files exported from Proteus® Analysis 9.8+ are supported. Other formats like .csv are not accepted.
Yes. A 6-month free trial is available, allowing up to 25 DSC uploads. No cancellation is required—the trial ends automatically.
No, Quantify requires internet connectivity. For offline environments, we recommend using our Proteus® desktop software.
Ready to Explore Quantify in Action?
Go beyond the FAQs and discover how Quantify can transorm your polymer composition analysis.
Learn more about the software’s key features, AI-powered evaluation, and how it supports recyclate producers, buyers, and test labs alike. The main product page offers a full overview, example use cases, and access to your free trial.