The quicker, the better

Control of production processes in the food industry has always focused on examination of end products. However, feedback of test results to the production process is generally not possible because it takes too long before the results of the analyses are known. Moreover, high numbers of samples have to be analysed to obtain statistically reliable results and inspection of end products only enables defects to be observed; it cannot establish their cause.

Because of this, analysis at the end of the process has shifted to control of the process by the introduction of GMP and HACCP systems. The use of a continuous, preferably in-line, monitoring system is necessary to make sure that the critical points in the process are controlled. This enables rapid detection and correction of slight deviations of process parameters yielding increased productivity and profitability. In addition, large margins that are used e.g. in heat treatments to prevent safety issues, can be minimised to improve quality aspects such as nutritional value and taste.

At present, sensors and instrumentation are commercially available and applied in the food industry for in-line and on-line measurement of physical and physical-chemical properties of products, for example temperature, pressure, flow and levels in tanks. However, sensors are not yet often applied to determine product composition and quality and safety aspects such as complex textures, concentrations of chemical compounds and micro-organisms. By means of a few cases this article illustrates the opportunities of technologies for process control in the dairy industry by rapid measurements of product quality and safety related parameters.

In-line monitoring of texture changes

A powerful tool for in-line monitoring of texture changes is based on diffusing wave spectroscopy (DWS). DWS is an optical measuring technique based on multiple scattering of laser light by colloids and polymers in turbid systems. The non-destructive method is applicable in a broad temperature range and suited to determine the mobility of the colloidal particles (oil droplets, protein particles)1.

DWS-based systems can be used for monitoring, for example acid-induced gelation of milk2 and in-line curd formation during cheese manufacturing3. The system measures gel strength and can be applied to automatically determine the cutting point of the gel during cheese manufacturing. If the curd is too firm at cutting time, a small decrease in loss of fat and curd fines content is observed. If the curd is too soft, the loss of fat and curd fines in the whey also increases. The optimal renneting time is approximately 30 DWS units. Decreasing from 30 to 20 DWS-units (approximately 2 minutes shorter renneting time) increased the curd fines content by a factor of six and the loss of fat by a factor of one and a half (Figure 1). It can be calculated that increasing the cheese yield using DWS in a plant processing 300.000.000 L milk annually results in an economic benefit of around €150.0004.

The technology has also been applied to follow changes in the rheology characteristics of turbid systems such as desserts, drinks, sauces and the production of yoghurt5. In summary, the DWS system can detect deviations from the norm and process trends can be detected at an early stage contributing to improved gelation, stabilisation and fermentation processes. This results in, for example, improved quality control of ingredients, increased production capacity, constant quality of final products and minimisation of product loss2,3,4,5.

In-line detection of micro-organisms

Microbiological testing in the dairy plant is critical to ensure the quality of raw milk and final products. The traditional methods usually require several hours or days of culturing, but provide relatively low limits of detection (in the range of 1 cfu/ml or lower). Fast methods can provide results within approximately ten minutes. However, the limits of detection are in the range of 104 to 105 cells per ml. Despite extensive efforts to develop reliable and rapid analysis technology, no commercial methods are available as yet.

NIZO food research investigated the potential of rapid detection methods for in-line control of pasteurisation. The running time of continuous-flow process equipment such as pasteurisers is limited, mainly by thermoresistent Streptococci (TRS). TRS are not completely inactivated by pasteurisation and grow at temperatures between 30 and 50°C, i.e. in the regenerative section of heat exchangers. During processing, the attached TRS are released into the product flow. Depending on the initial level in the raw milk, running times of approximately 4 to over 11 hours can be realised before the critical level of 105 TRS/ml is reached and the pasteuriser must be cleaned. Bactoscan (a technique combining staining of live and dead bacteria and microscopic detection) and ATP measurements (detecting bacterial activity of living cells) were tested for indication of reaching the critical level. The results showed that, owing to the detection limits (104 to 105 cells per ml), both systems are only suited as an ‘emergency break’, but not as a timely indicator that cleaning is necessary. If equipped with a suitable filtration-concentration unit, the technology can probably reach the target detection limit6.

Optimisation of cleaning-in-place procedures

Cleaning and disinfection are essential to assure quality and safety in the food industry. Most dairy processes require at least daily cleaning. The applied procedures are often based on experience. Large margins are chosen for the intensity and length of the cleaning steps to ensure food safety. With production batches getting smaller and product diversity increasing, flexibility in CIP processes gains in relevance. Strategies based on in-line and on-line monitoring of cleaning steps can save energy and time, and decrease consumption of water and raw materials.

NIZO food research developed a monitoring system, called OPTI-CIP, based on in- and at-line measurements of removal of deposits and cleaning agents and parameters such as temperature, flow, conductivity and valve settings. Cases in the dairy industry demonstrated that the efficiency of cleaning can be improved by reducing the cleaning time by 50 per cent (Figure 2)7.

However, the current device measures removal of fouling at-line. In-line monitoring would offer advantages, especially to production facilities with a large variety of products and short run times. Therefore, the application of a turbidity sensor for monitoring of organic fouling removal was tested during a two-step cleaning process in an evaporator.

It was found that in-line and off-line measurements provided comparable results: the peaks were observed at the same time. The in-line method measured higher levels of turbidity, i.e. organic fouling removal. This is due to higher background levels and foaming of the cleaning solution affecting the sensor measurements. When the in-line results were corrected for this both methods were comparable.

It can be concluded that real-time monitoring of cleaning procedures is possible. For application in production plants more research is needed with respect to the robustness of the system7.

Food Sense: new generation gas analyser

Early detection of smouldering by measurement of carbon monoxide can be used to prevent disastrous explosions in spray dryer installations used for production of, for example, milk or coffee powders. For extremely large spray drying installations, background interference renders the signal-to-noise ratio inadequate. The application of a dedicated mass spectrometer – the Food Sense, based on a new generation of gas analysers developed by V&F (www.vandf.com), is a clear improvement of the detection system.

The basis of the Food Sense is a unique and patented ionisation concept resulting in high sensitivity and selectivity. This allows real time gas analyses with very low detection limits (<< ppb level) for volatile compounds in a mobile unit.

The technique offers many opportunities for R&D and process control in the food industry. Illustrative for its power as a tool in product development and R&D is the sensitive detection (ppb level) of the sulphur allyl volatiles (e.g. dimethyl sulphide, H2S etc.) characteristic for flavours such as onions, cheese and garlic.

An example, illustrating the value of this technology in quality control of liquid products, is the immediate detection of unwanted compounds (e.g. chloroform) generated by residuals from cleaning agents. Data show that chloroform can be measured at 2 ppb level in liquids which relates to residuals of chlorine containing cleaning agents of 0.007% of active chlorine. By implementing this gas analyser in a manufacturing process, the risk of introducing cleaning agents in liquid foods can be minimised.

What for the future?

In addition to the already applied physical techniques, new technologies, including (bio)sensors, will become more and more available for in/on-line application in the food chain for screening of raw materials, process control and final product analysis of quality and safety parameters. Furthermore, significant progress is made in the field of nanotechnology, high throughput systems and micro-arrays combined with smart data handling. This enables detection of specific, single or multiple (bio)chemical and microbial components. However, before successful implementation major bottlenecks such as suitability of the techniques in the production environment, sensitivity and accuracy and costs should be solved. If this is achieved in-line/on-line detection methods are important tools that can contribute to assuring high, constant quality and safety of foods.

References

1. Grotenhuis E. ten, Berg G. van den, Kruif C.G. de, (1999a) ‘Science evolving into a versatile industrial tool: The diffusing wave spectroscopy story’, Eur. Dairy Mag. 2, 44-46.
2. Vasbinder A.J., Mil P.J.J.M. van, Bot A., Kruif C.G. de (2001), ‘Acid-induced gelation of heat-treated milk studied by diffusion wave spectroscopy’ Coll. and Surf. B:Biointerfaces, 21, 245-250.
3. Grotenhuis E. ten (1999b) ‘Prediction of cutting time during cheese production’, Eur. Dairy Mag., 11, 40-41.
4. Straatsma H., Hoven G. van den, Kanning M. & Grotenhuis E. ten (2002) ‘New optical tool for monitoring of gelation processes’, Eur. Dairy Mag.,1 , 31-33.
5. Grotenhuis E. ten, Paques M., Aken G. van (2000) ‘The application of diffusing wave spectroscopy to monitor the phase behaviour of emulsion-polysaccharide systems’ J. Coll Interf. Sci. 227, 495-504.
6. Giffel M.C. te, Meeuwisse J., Jong P. de (2001) ‘Control of milk processing based on rapid detection of micro-organisms’ Food Control, 12, 305-309.
7. Asselt A.J. van, Houwelingen G. van, Giffel M.C. te (2002) ‘Monitoring system for improving cleaning efficiency of CIP processes in dairy environments’ Trans. Inst. Chem. Eng., 80(Part C), 276-280.
8. Steenbergen A.E., Houwelingen G. van, Straatsma J. (1991), System for early detection of fire in a spray drier, J. Soc. Dairy Technol. 44, 76-79.