A simple method for fume cupboard performance assessment

PII: S0003-4878(99)00107-6

Ann. occup. Hyg., Vol. 44, No. 4, pp. 291±300, 2000 7 2000 British Occupational Hygiene Society Published by Elsevier Science Ltd. All rights reserved Printed in Great Britain. 0003±4878/00/$20.00

A Simple Method for Fume Cupboard Performance Assessment GRAHAM P. NICHOLSON*, RAYMOND P. CLARK, FRED GROVER and MERVYN L. DE CALCINA-GOFF University of Westminster, School of Communication, Design and Media, Watford Road, Northwick Park, Harrow HA1 3TP, UK The performance of a fume cupboard is determined by a complex interaction of factors which are time consuming and expensive to determine. This paper describes a simple and practical means of ranking, and assessing fume cupboard installations that can help to discharge managerial responsibility for a `safe' environment. The method also gives an economically viable and technically defensible system for assessing fume cupboard performance as part of upgrading exercises or performance audits. The assessment strategy uses ¯ow visualisation techniques and measurements of in¯ow air velocity as well as overall condition evaluation to rank performance and identify poor performing cupboards. The method has been used to carry out a condition and performance survey of 199 fume cupboards, both aerodynamic and boxtype designs, in an academic institution. The results of this survey are presented which not only highlight performance characteristics but also provide insights into user attitudes and knowledge of fume cupboard operation and performance. It is suggested that surveys such as this could be helpful in training programmes for laboratory workers to enable them to optimise the use of fume cupboards. 7 2000 British Occupational Hygiene Society. Published by Elsevier Science Ltd. All rights reserved Keywords: fume cupboard; face velocity; visualisation; ranking

INTRODUCTION

Fume cupboards are found in practically every science laboratory. British Standard BS 7258 (BSI, 1994) de®nes a general purpose laboratory fume cupboard as `a partially enclosed work space that limits the spread of fume to operators and other personnel. It is ventilated by an induced ¯ow of air through an adjustable working aperture that dilutes the fume, and by means of an extract system, provides for the release of fume remotely and safely'. There are di€erent types of cupboard ranging from simple `box' types to comparatively sophisticated `aerodynamic' designs. In recent years, and with the introduction of national regulations, emphasis has Received 8 December 1998; in ®nal form 23 September 1999. *Author to whom correspondence should be addressed: Centre for Disability Research & Innovation, University College London, Brockley Hill, Stanmore, HA7 4LP, UK. Tel.: +44 181 954 2300 (ext. 752); Fax: +44 181 385 7151. 291

been placed on establishing the performance levels of fume cupboards in relation to user safety. There is a large literature on the installation, use and maintenance of fume cupboards including methods for assessing their performance. A number of national standards exists such as ASHRAE 110 (ASHRAE, 1993) in the USA, DIN 12 924 (DIN, 1991) in Germany and BS 7258 (BSI, 1994) in the UK. Various guidelines have been available from organisations such as the British Occupational Hygiene Society (BOHS, 1976) and the Royal Society for Chemistry (RSC, 1990). Much information is available in papers, and several books (Cook and Hughes, 1986; Hughes, 1980; Saunders, 1993; Nicholson, 1997) summarise this literature. Performance assessment strategies di€er in type and complexity and range from simple face velocity measurement and ¯ow visualisation to various forms of quantitative containment testing. In BS 7258 emphasis is placed on a complex face velocity `type' test, with a subset of these measurements used for commissioning and a modi®ed method,

292

G. P. Nicholson et al.

with reduced complexity, for maintenance testing; testing for containment is a recommendation. Commissioning and maintenance tests indicate relative di€erences from the type test parameters. The literature gives di€ering guidance on performance. The term `performance' covers a wide range of parameters including mean in¯ow air velocity, variation of in¯ow over the aperture, pattern of air¯ow within a cupboard, and containment of speci®c tracers released in the working space. Common practice is not to specify a face velocity but to leave this to the purchaser/user. BS 7258 does not specify any minimum performance criteria and only includes a note that states: `In practice, with the sash set at the maximum working opening, it is unlikely that face velocities below 0.3 msÿ1 will give satisfactory containment. In some cases, face velocities of 0.5 msÿ1 or above may be necessary'. An average minimum face velocity of 0.5 msÿ1 has generally been regarded as giving the minimum performance that can resist environmental air movement disturbances which, in most workplaces, are of the order of 0.25±0.3 msÿ1 (BOHS, 1988). For satisfactory performance, it is also required that di€erences in velocity across the aperture should not vary by more than 220% from the mean (BS 7258). In establishing the type, condition and performance of a large number of fume cupboards there are limitations of time and expense which restrict the use of some available assessment strategies. An assessment strategy has to take into account: . the need for a quick determination of performance particularly when a large number of cupboards is involved; . the likelihood that current performance ®gures will not be available for the majority of cupboards; . an expectation that variation in performance will be large; and . the assessment may well be used to determine priority for replacement. Tests described in BS 7258 were considered inappropriate for the proposed assessment as the BS was written primarily for the design, installation and maintenance of new cupboards. The face velocity test method and gas containment tests were considered too complex and time consuming when large numbers of older cupboards are being assessed. Because of this a hierarchical test strategy was designed using ¯ow visualisation and face velocity measurements and was conceptually similar to the strategy adopted in British Standard BS 5726 (BSI, 1992) for microbiological safety cabinets. The criteria chosen for satisfactory performance are: . ¯ow visualisation showing in¯ow over the whole aperture;

. face velocity r0.5 msÿ1; and . variation in velocity at any point R20% from an average face velocity r0.5 msÿ1. In addition, an objective of this work was to classify the general design, construction and siting (with reference to the guidance given in BS 7258) and the `as found in use' condition of each fume cupboard surveyed.

ASSESSMENT METHOD

The assessment is carried out at a working sash height of 500 mm [measured from the top edge of either the bench or lipfoil (where ®tted) to the lower edge of the sash]. Where a wide cupboard has two or three sashes, all are positioned at the same working height during the tests, and each face is tested in turn. The performance tests are carried out for each fume cupboard in the condition in which it is found. No attempt should be made to alter the environmental conditions around the cupboard and people should be allowed to enter and leave the laboratory as they wish. The physical state of the cupboard should be recorded as satisfactory/poor, the cleanliness recorded as satisfactory/dirty and blockage recorded as cluttered/moderate/none. Notes concerning the environment around the fume cupboard and its position should also be recorded. Both of these factors when combined with performance assessment can have a signi®cant e€ect on upgrading and replacement policies. For air¯ow visualisation, water fog generated by an ultrasonic nebuliser (Kennedy, 1987) and used to visualise the direction of air ¯owing into and within the cupboard is useful although other methods using oil or other smoke are suitable. Flow visualisation should be carried out using a ¯exible tube attached to the tracer-aerosol-generating system. The tracer and gas carrier velocity at the tube should not exceed the mean in¯ow air velocity into the cupboard. A velocity of 0.1±0.2 msÿ1 is satisfactory. Tracer should be presented systematically to all parts of the front aperture from a maximum distance of 100 mm from the aperture plane. The tracer aerosol should next be manually presented to the aperture plane from inside the cupboard with tracer projected normally out towards the operator from the plane of the aperture from a maximum of 100 mm inside the fume cupboard. The air ¯ow visualisation should be recorded as: . Satisfactory when there is a clear inward ¯ow of water fog over the entire aperture. . Questionable when water fog is shown to linger around the lipfoil or sash handle.

Fume cupboard performance assessment

. Unsatisfactory when there is substantial leakage of tracer back into the room from the cupboard. The in¯ow face velocity is ideally measured with a 100 mm diameter rotating vane anemometer (for example Air¯ow Developments Ltd., Worthing, Sussex, UK) for one minute at six positions equally spaced in the plane of the aperture. The face velocity is expressed as the overall mean of the average velocity measured at each position. The variation in velocity across the face is expressed as the maximum percentage deviation of the average velocity at that position from the overall mean face velocity.

FUME CUPBOARD RANKING

In order to rank the fume cupboard performance, the mean face velocity, its variation across the aperture and ¯ow visualisation variables are scored as shown in Table 1. These scores are used in the following formula to give each fume cupboard a performance index:

necessity being a containment test to a speci®ed level of operator protection factor. Visual inspection of oil smoke or water fog is a very selective method of identifying good and bad ¯ow patterns at the front aperture of a safety cabinet or fume cupboard. Because of these considerations the scoring of ¯ow visualisation is emphasised by being squared in the index. SURVEY

We have conducted a survey of fume cupboards in a major university within the United Kingdom as part of the development of a cost-e€ective replacement programme. The survey included assessing 199 installed and `in-use' fume cupboards. These were located over several sites and included aerodynamic and simple `box' types, some dating back to the early 1900s (Table 2). Their use ranged from general solvent storage and handling to complex experimental set-ups, including radioactive and chromatographic work.

Index ˆ Velocity score  Variation score  …Water fog visualisation score† 2 The lower the index the better the performance. The ¯ow visualisation is squared to give weight to this factor because of its importance in the overall assessment of fume cupboard performance. Some justi®cation of this is necessary. BS 5726 (BSI, 1992) was the product of many years' experience of type, commission and maintenance testing of microbiological safety cabinets carried out by numerous organisations. One of the most important factors to emerge and to be widely recognised was that the measurement of in¯ow air velocity alone was not sucient to ensure satisfactory containment. Subtleties of ¯ow reversal and turbulence that were sucient to reduce containment could not be identi®ed from the measurements made with generally available anemometers that had insucient selectivity and sensitivity. For this reason, the concept of a hierarchy of assessments evolved with the ®rst and most basic requirement being a demonstration of adequate in¯ow over the whole front aperture. The next requirement was a satisfactory level of in¯ow air velocity and uniformity, the ®nal and ultimate

293

RESULTS AND DISCUSSION

Table 2 shows the number and type of fume cupboards tested. Some were aerodynamic and some were simple `box' types, having one, two or three sashes; a number had opposing sashes. These box cupboards had di€erent internal designs ranging from those with a single extract hole in the top to those with tapered extract ducts. Some had no baf¯e whilst others had full rear ba‚es of varying types. A number had been retro-®tted with improvised lipfoils in an attempt to smooth the air¯ow over the lower front lip. Materials of construction, physical state, cleanliness, equipment blockage and environment The materials used in the construction of the fume cupboards are shown in Table 3. The majority of the box types were constructed of wood with tiled work surfaces and wire reinforced glass. Most of the two-sided cupboards were made of PVC and used for radioactive material. The relatively modern aerodynamic fume cupboards were mainly made of composite materials. Some had stainless steel work interiors for radioactive work. The majority of cupboards were assessed as being

Table 1. Variable scores Velocity r0.5 msÿ1 < 0.5 > 0.3 msÿ1 < 0.3 > 0 msÿ1

Score

Velocity variation

Score

Flow

Score

1 2 3

Variation R20% and face velocity r0.5 msÿ1 Variation r20% and face velocityr0.5 msÿ1 Variation R20% and face velocity R0.5 msÿ1 Variation r20% and face velocity R0.5 msÿ1

1 2 3 4

Satisfactory Questionable Unsatisfactory

1 2 3

294

G. P. Nicholson et al. Table 2. Number and type of fume cupboards tested

Fume cupboard type

Number in survey

% of total

Aerodynamic Box type with single sash Box type with double sash Box type with triple sash Box type with single opposing sashes Total number tested

57 81 41 12 8 199

28.7% 40.7% 20.6% 6.0% 4.0%

in a satisfactory physical state. Those assessed as being in poor condition mainly had cracked or broken tiles lining the work surface and walls (Table 4). Some had obvious ®re damage, were badly corroded or the sash glass was cracked. Only a few aerodynamic cupboards were assessed as being in a poor condition. There was evidence that bunsen burners and other heat sources had damaged the PVC walls in a number of cases. The observations of fume cupboard cleanliness are shown in Table 5. There were as many satisfactory as dirty cupboards. One clear factor was that many users and supervisors paid little regard to their fume cupboards. A large number was used for teaching where the students, through lack of `ownership', appeared to have no real instruction or understanding of the `house-keeping' needed to maintain a cupboard in good condition. Very few of the cupboards were empty when tested (Table 6). A high percentage were severely cluttered with equipment, mainly bottles and glassware. Sometimes, these items obstructed the rear slot (Fig. 1) eliminating any scavenging air¯ow across the work surface. In other cases, large items of equipment such as gas chromatographs obstructed nearly all of the working volume (Fig. 2). A large percentage of cupboards were installed in poor positions in corners and near to windows and doors (Table 7). Within parts of the university there appeared to be no management control over the opening of either windows or doors. Some fume cupboards were positioned in main thoroughfares. In other cases the front of the cupboard was partially obstructed by furniture such as ®ling cabinets. Face velocity Thirty-®ve cupboards were found to be inoperative, sometimes a fact unknown to the user. Of the

working cupboards, 45.5% had average face velocities >0.5 msÿ1 (Fig. 3), and 17% <0.3 > 0 msÿ1. The overall range of average face velocities was from 0.07 to 1.31 msÿ1 suggesting that little attention was paid to the generally accepted face velocity of 0.5 msÿ1. Of the working aerodynamic cupboards tested, 70.4% had average face velocities >0.5 msÿ1 with a total range from 0.35 to 1.10 msÿ1. Three were found to have been switched o€ although two had a measurable air ¯ow at a lower sash height of 250 mm which was thought to be due to thermal updrafts. Of the working box fume cupboards tested, 32.7% had an average measured face velocity >0.5 msÿ1 with a total range from 0.07 to 1.31 msÿ1; 32 of these were inoperative (with no ¯ow). Of the cupboards with triple sashes, 77.8% had face velocities >0.5 msÿ1. Velocity variation Only 25% of the working fume cupboards tested had a velocity variation <20% at a face velocity >0.5 msÿ1 (Fig. 4); 35% had a velocity variation >20% at a face velocity <0.5 msÿ1. Of the aerodynamic cupboards that were working, 48.1% had a velocity variation <20% at a face velocity >0.5 msÿ1; 72.2% had a velocity variation <20% irrespective of face velocity. Only 5.6% were very poor. Of the box cupboards that were working, 13.6% had a velocity variation <20% at a face velocity >0.5 msÿ1; 32.7% had a velocity variation <20% irrespective of face velocity. As many as 50.9% were very poor. There were exceptions to the commonly held view that aerodynamic cupboards are more resistant to variation of velocity across the aperture than

Table 3. Fume cupboard exterior material (% of fume cupboard type) Material

Fibreglass/GRP/laminates Metal/stainless steel PVC/plastics Wood/glass/tiles

Aerodynamic

59.6% 28.1% 10.5% 1.8%

Box Single

Double

Triple

Two-sided

1.2% 1.2% 4.9% 92.7%

100%

100%

62.5% 37.5%

Fume cupboard performance assessment

295

Table 4. Fume cupboard condition (% of fume cupboard type) Condition

Satisfactory Poor

Aerodynamic

93% 7%

Box Single

Double

Triple

Two-sided

63% 37%

76% 24%

58% 42%

88% 12%

Table 5. Fume cupboard cleanliness (% of fume cupboard type) Cleanliness

Satisfactory Dirty Not recorded

Aerodynamic

54.4% 45.6%

Box Single

Double

Triple

Two-sided

35.8% 54.4% 9.8%

24.4% 75.6%

50.0% 50.0%

62.5% 37.5%

box cupboards; there was little di€erence in variation between the box cupboards with triple sashes and those of aerodynamic design. Flow visualisation Of all the fume cupboards tested 65% were assessed as being unsatisfactory, only 20% being satisfactory, while the rest were questionable (Fig. 5). Of the aerodynamic cupboards 31.5% were satisfactory; however, 35.2% were questionable because of reverse ¯ows around the lipfoil, where water fog was seen to accumulate in eddies on the lipfoil, or

because of `bounce-back' from equipment placed too near to the opening; 33.3% were unsatisfactory. Of the box cupboards 13.6% were satisfactory but 75.5% were unsatisfactory. This trend was similar for all types apart from those with three sashes.

Comparison of assessed variables as indicators of fume cupboard performance It was considered that overall fume cupboard performance (determined by the design, the face velocity, its variation across the aperture, obstructions and environmental conditions) could not be reliably

Fig. 1. Obstruction of the rear scavenging slot.

296

G. P. Nicholson et al. Table 6. Fume cupboard blockage (% of fume cupboard type)

Blockage

Cluttered Moderate None Not recorded

Aerodynamic

21.1% 75.4% 3.5%

Box Single

Double

Triple

Two-sided

24.7% 64.2% 4.9% 6.2%

39.0% 34.6% 24.4%

8.3% 83.4% 8.3%

37.5% 50.0% 12.5%

Table 7. Fume cupboard environment (% of fume cupboard type) Environment

In a corner By a window/door In a thoroughfare By LEV Cluttered Unrestricted

Aerodynamic

36.8% 40.4% 5.3% 5.3% 7.0% 26.3%

Box Single

Double

Triple

Two-sided

40.7% 43.2% 13.6% 4.9% 8.6% 11.1%

24.4% 29.3% 39% 4.9% 7.3% 19.5%

8.3% 1.7% 8.3% 0% 0% 58%

62.5% 0% 12.5% 0% 0% 25%

assessed from either measurements of face velocity or from ¯ow visualisation considered alone. Face velocity measurements showed more cupboards with satisfactory rather than unsatisfactory performance. However, the velocity variation and ¯ow visualisation showed the reverse. The velocity variation indicated that the majority of aerody-

namic cupboards had satisfactory performance criteria but the ¯ow visualisation indicated similar numbers with satisfactory and unsatisfactory performance criteria. Because of the hierarchial system adopted here, the ®rst condition is that ¯ow visualisation must demonstrate satisfactory in¯ow over the whole

Fig. 2. Obstruction of the front aperture by equipment within the cupboard and a large gas cylinder immediately adjacent to the sash.

Fume cupboard performance assessment

297

Fig. 3. The percentage of fume cupboards of a particular type with face velocities in the ranges >0.5, <0.5 > 0.3 and <0.3 > 0 msÿ1. Working aperture height is 500 mm.

aperture. From containment tests on safety cabinets and fume cupboards, indications are that poor ¯ow visualisation frequently means unsatisfactory containment whereas low velocity with a large variation may well mean satisfactory containment, at least in undisturbed conditions. Flow visualisation is therefore considered to be a more suitable ®rst indicator

of fume cupboard performance than face velocity measurements. It was clear from Fig. 6 that at a `normal' working aperture height of 500 mm there was an even spread of cupboards across the ranking order. Comparison of the aerodynamic and box-type cupboard indices (Fig. 7) demonstrated the greatest

Fig. 4. The percentage of fume cupboards of a particular type with a maximum variation (%) of velocity at a point from the mean face velocity (msÿ1) in the ranges <20% > 0.5 msÿ1, >20% > 0.5 msÿ1, <20% < 0.5 msÿ1 and >20% < 0.5 msÿ1. Working aperture height is 500 mm.

298

G. P. Nicholson et al.

Fig. 5. The percentage of fume cupboards of a particular type with air¯ow patterns visualised within the aperture plane subjectively assessed as satisfactory, questionable and unsatisfactory. Working aperture height is 500 mm.

di€erences in performance occurred at high and low values of the index; a high percentage of aerodynamic cupboards had the lowest index whereas none had the highest index. Existing literature shows that the bene®t of aerodynamic features to give improved performance over box cupboards can be eliminated by the e€ect of other variables such as poor installation, maintenance and untutored use

(Robertson and Bailey, 1980; Caplan and Knutson, 1982; Hughes, 1980; Saunders, 1993). In this survey this ranking system was used to identify the worst performing cupboards as those due for early replacement. As this was to be carried out over a period of time (during which the cupboards had to remain in service) an interim solution of limiting the aperture opening to improve the performance was recommended. Lowering the sashes

Fig. 6. Ranking of fume cupboard indices at working aperture heights of 500 mm (Q) and 250 mm (q).

Fume cupboard performance assessment

299

Fig. 7. Ranking of aerodynamic and box fume cupboard types at a working aperture height of 500 mm: aerodynamic (Q); all box types (q).

to a working aperture height of 250 mm substantially changed the ranking (Fig. 6). This increased those with the best performance from 8 to 38.7%. Of these the percentage of aerodynamic cupboards rose from 5.5 to 23.1% and the percentage of box cupboards from 2.5 to 15.6%. The assessment method described in this paper is not intended as a substitute for containment tests which are regarded as essential to ensure safety and which should be carried out at commissioning and routine maintenance periods. Although microbiological safety cabinets with poor air in¯ow patterns can produce inadequate containment performance, unsatisfactory containment can also be found in situations where air velocities are normal (Clark, 1997). The performance of many fume cupboard installations is often very poor when contrasted with the best safety cabinets and cupboards that are available. At the higher levels of performance that are currently available more sophisticated containment tests using sulphur hexa¯uoride or KI Discus are necessary (Nicholson et al., 1999) to accurately determine containment. The assessment method described in this paper is considered satisfactory for situations where large numbers of cupboards need to be assessed rapidly (particularly in relation to fume cupboard replacement programmes) and where there are expectations of relatively low performance. The fume cupboard assessment method described in this paper is easy to apply and the results are useful for planning management strategies. However, more work is needed (currently under-

way) before this approach is properly validated. Particularly signi®cant are the correlations between containment tests, ¯ow visualisation and face velocity. Once these relationships are established the tests described here, or variations of them, could provide a cheap, simple and convenient way of classifying cupboards.

CONCLUSIONS

. Using a simple assessment strategy and performance index (without the need for complex face velocity measurements or containment tests) showed a very large variation in fume cupboard performance. . Of the three variables assessed, ¯ow visualisation was judged to be more indicative of performance than either the measurement of face velocity or its variation across the aperture. It was therefore more heavily weighted than other factors in evaluating the performance index. . Used in a survey, this system of ranking fume cupboards with the performance index identi®ed the worst fume cupboards as those due for removal as part of a progressive replacement programme. . The majority of cupboards in the survey were in a satisfactory physical condition but it was clear that many users appeared to have little concept of the factors determining their safety. Departmental supervisors appeared to pay little regard to the need for managing the use of cupboards including aspects of assessment,

300

G. P. Nicholson et al.

house-keeping and cleanliness. . Some cupboards were switched o€ and did not work at all (unbeknown to the users). In some cases the variation in face velocity across the aperture was extreme and others had very low or very high face velocities. Some cupboards were cluttered with almost total obstruction of the air¯ow and in others the air¯ow was poor due to bad siting. Only 8% of the cupboards surveyed were considered to have satisfactory performance. . The overall range of average face velocities was from 0.07 to 1.31 msÿ1 suggesting that little attention was paid to the generally accepted face velocity of 0.5 msÿ1. . There were more satisfactory aerodynamic cupboards than poor ones whereas for box cupboards the reverse was true. However, there were some box types with better performance than aerodynamic types con®rming existing literature that the use of aerodynamic features to give improved performance could be eliminated by the e€ect of other variables such as poor installation, maintenance and use. . It was apparent that the majority of fume cupboards in this survey were installed and used in a way that was far from ideal. There appeared to be little understanding by the users of the principles of fume cupboard containment or of the e€ects of equipment blockage or personnel movement and environmental disturbances near to the aperture. The results presented here are probably representative of many academic institutions. Where a comprehensive testing programme and informed user awareness exists, better fume cupboard performance can be expected. From discussions with users during the survey reported here, it seems that proper microbiological safety cabinet use and testing is much better understood and practised. A European Standard for fume cupboards is being proposed and it is hoped that this will allow scope for the type of assessment described here which is a cheap, simple and quick way of assessing a large number of fume cupboards. REFERENCES ASHRAE (1993) Method of Testing Performance of Laboratory Fume Cupboards (ASHRAE 110). American

Society of Heating, Refrigeration and Air-Conditioning Engineers, Atlanta, GA. BOHS (1976) A Guide to the Design and Installation of Laboratory Fume Cupboards. In British Occupational Hygiene Society, Hygiene Technology Guide Series No. 1. H & H Scienti®c Consultants Ltd in association with Science Reviews Ltd, Northwood, UK. BOHS (1988) Controlling Airborne Contaminants in the Workplace. In British Occupational Hygiene Society, Technical Guide No. 7, eds T. Ogden and D. Hughes. H & H Scienti®c Consultants Ltd in association with Science Reviews Ltd, Northwood, UK. British Standards Institution (1992) BS 5726. Microbiological Safety Cabinets. Parts 1,2,3 and 4. British Standards Institution, London. British Standards Institution (1994) BS 7258. Laboratory Fume Cupboards. Parts 1,2,3 and 4. British Standards Institution, London. Caplan, K. J. and Knutson, G. W. (1982) In¯uence of room air supply on laboratory hoods. American Industrial Hygiene Association Journal 43, 738±746. Clark, R. P. (1997) Standards for safety cabinets. Nature Dec 11;390(6660):550. Cook, J. D. and Hughes, D. (1986) Fume Cupboards Revisited. In Occupational Hygiene Monograph No. 2, ed. D. Hughes. H & H Scienti®c Consultants Ltd in association with Science Reviews Ltd, Northwood, UK. DIN (1991) DIN 12 924, Part 1. Laboratory Furniture; Fume Cupboards; General Purpose Fume Cupboards; Types, Main Dimensions, Requirements and Testing. German Institute for Standardisation, Berlin. Hughes, D. A. (1980) Literature Survey and Design Study of Fume Cupboards and Fume Dispersal Systems. In Occupational Hygiene Monograph No. 5, ed. D. Hughes. H & H Scienti®c Consultants Ltd in association with Science Reviews Ltd, Northwood, UK. Kennedy, D. A. (1987) Water fog as a medium for visualization of air¯ows. Annals of Occupational Hygiene 31, 255±259. Nicholson, G.P. (1997) Studies on the performance of open fronted containment and `ultra-clean' ventilation systems. PhD thesis, King's College, University of London. Nicholson, G. P., Clark, R. P. and de Calcina-Go€, M. L. (1999) Theoretical and practical comparison of the potassium iodide tracer method (KI-Discus) for assessing the containment eciency of fume cupboards with the gas tracer method described in BS 7258: 1994: Part 4. Annals of Occupational Hygiene 43(4), 257±267. Robertson, P. and Bailey, P. V. (1980) Suggested improvements to prevent the escape of fume beneath the sash of a fume cupboard. Annals of Occupational Hygiene 23, 305±309. The Royal Society of Chemistry (1990) Guidance on Laboratory Fume Cupboards. In Technical Guide Series of the British Occupational Hygiene Society, eds T. L. Ogden and D. Hughes. H & H Scienti®c Consultants Ltd in association with Science Reviews Ltd, Northwood, UK. Saunders, G. T. (1993) Laboratory Fume CupboardsÐA User's Manual. Wiley and Sons, New York.