990 SANDERSON ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 86, NO. 5, 2003 FOOD COMPOSITION AND ADDITIVES
Photostimulated Luminescence Detection of Irradiated Herbs, Spices, and Seasonings: International Interlaboratory Trial DAVID C.W. SANDERSON, LORNA A. CARMICHAEL, and SAFFRON FISK Scottish Universities Research and Reactor Centre (SURRC), Scottish Enterprise Technology Park, Rankine Ave, East Kilbride, G75 OQF, UK Collaborators: P. Brown; Y. Burton; S.D. Clemenson; G. Hart; H. Nootenboom; S. Pinnioja; G.A. Schreiber; U. Wagner; and C. Wiezorek
An interlaboratory trial was conducted to validate photostimulated luminescence (PSL) methods for herbs, spices, and seasonings. Forty products (11 herbs, 17 spices, and 12 seasonings) were purchased from a local commercial source, and randomly selected samples were irradiated with 10 kGy. Four blended products were prepared at Scottish Universities Research and Reactor Centre, mixing varying proportions of irradiated material with the untreated product. Precharacterization against a predefined threshold identified low sensitivity products (black and white peppers) and products with high natural signals (thyme, sage, parsley, and mixed herbs), both of which might be susceptible to misclassification. Precharacterization also revealed whether calibration was likely to resolve overlap between classification categories. Eight sets of screening data and 5 sets of calibrated data were returned by participants. Of the 840 samples sent, 1593 screening measurements and 788 calibrated measurements were received from 662 samples. In screening mode, participants reached definitive conclusions in 87% of cases, 99% of which were correct. Of the remaining 13%, calibration to identify low-sensitivity resolved 60% of cases. Overall, 94% of samples were correctly identified by either screening alone, or screening plus calibration; 6% remained unclassified and therefore required further investigation by thermoluminescence. The results confirm the validity of the PSL method for herbs, spices, seasonings, and blends, and emphasize the need for calibration to identify low-sensitivity samples. This method has now been adopted by the European Committee for Standardization (CEN) and the Codex Alimentarius Commission.
n international interlaboratory trial was conducted to validate photostimulated luminescence (PSL) meth-
A
Received February 27, 2003. Accepted by SG May 23, 2003. Corresponding author’s e-mail:
[email protected].
ods for herbs, spices, and seasonings in 1997. The development of PSL techniques for detecting irradiated foods, where energy to release trapped charge carriers is provided optically (1–3), has been aimed at extending the availability of luminescence techniques to users for which thermoluminescence (TL) methods were unsuitable. The lengthy preparation and need for irradiation in all cases limits TL to a small number of laboratories. In 1992 (4–8), a low-cost instrument was developed for high-sensitivity PSL measurements from food samples, using highly radiation-specific UV-Vis luminescence signals which can be stimulated by IR sources, together with pulsed lock-in techniques for background suppression. Samples are introduced in disposable Petri dishes with minimal preparation, are undamaged by the measurement process, and can thus be calibrated following irradiation without drastic changes of sensitivity. Initial investigation of a set of herbs, spices, and seasonings, presented both nonirradiated and irradiated, confirmed the viability of this technique for measuring unprepared whole samples (9, 10). More than 90% of irradiated samples could be identified against predefined thresholds (1, 8) using intensity measurements, with a small overlap between high-sensitivity nonirradiated samples and low-sensitivity irradiated samples. Irradiating samples to a known dose and re-reading the PSL signals (CalPSL) allowed the sensitivity of the sample to be estimated, which resolved the overlap from samples of pure irradiated or nonirradiated materials. Thus, 2 modes of operation were defined: screening mode, where the luminescence intensity detected from the sample is used for preliminary classification into negative, intermediate, or positive bands, and calibrated PSL (CalPSL) measurements which can distinguish between low- and high-sensitivity samples. Since 1992, more than 2000 samples have been analyzed at Scottish Universities Research and Reactor Centre (SURRC) with PSL methods. In 1996, a national surveillance exercise conducted in the United Kingdom on behalf of the Ministry of Agriculture, Fisheries and Food (MAFF; 11) included luminescence testing at SURRC. At this stage, a combined PSL/TL approach was adopted. Any sample exceeding a given threshold during screening was also analyzed by TL, as were a random 10% of samples not exceeding that threshold. TL confirmed all PSL results. The exercise identified as irradiated a paprika, which was included as a standard, and 4 other spices, 3 of which were identified as mixtures of irradiated and
SANDERSON ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 86, NO. 5, 2003 991 Table 1. Decision-making charts as sent to participants Screening decision scheme Negative Less than Threshold 1 No evidence of irradiation
Intermediate
Positive
T1 < >T2
Greater than Threshold 2
Possibly irradiated
Likely to be irradiated
Calibration classification chart Second PSL reading (after irradiation to known dose) First PSL reading Negative Intermediate Positive
a b c
Negative
Intermediate
Positive
Second reading > original reading
Indeterminatea
Indeterminatea
No evidence of irradiation
No evidence of irradiation
Likely to be irradiated
May contain an irradiated componentb
May contain a minor irradiated componentb
Irradiatedc
Irradiated
Contains a minor irradiated componentb
Indicates low sensitivity to irradiation. Further investigation by TL recommended. Further investigation by TL recommended to assess the contribution from irradiated material. Indicates a normalization dose smaller than the initial dose in a material with relatively low sensitivity.
nonirradiated material. The surveillance exercise also included shellfish. The bulk of PSL analyses conducted at SURRC have been submitted for routine quality assurance testing by industry, with many blends (samples containing both irradiated and nonirradiated material) identified in the process. Blends may either be compound products, e.g., curry powders, or a single product blended to achieved consistent quality. There has been less commercial demand for CalPSL. In 1996, when the herb, spices, and seasoning samples were being prepared for the trial, there were 10 laboratories with PSL capabilities (there are now more than 15). The trial aimed to test PSL performance in an interlaboratory context and, where possible, to compare screening and calibrated procedures. For this PSL trial, a selection of 40 products was purchased from Lion Foods Ltd., consisting of 11 herbs, 17 spices, and 12 seasonings. Samples were randomly selected for irradiation under commercial conditions at Isotron (Swindon, UK). The irradiated samples were given a maximum dose of 10 kGy. All 10 laboratories agreed to participate. Each laboratory received 2 pots of each product, in a random allocation of irradiated and nonirradiated samples. In addition, each laboratory received 4 spice blends prepared at SURRC to address what is becoming an increasingly important issue. Each participant was asked to analyze a total of 84 samples, using screening and, if possible, CalPSL. Of 840 samples sent, 1593 screening measurements and 788 calibrated measurements were returned from 662 samples. Experimental Protocol Instructions were sent to participants giving details of sample preparation, screening, and calibrated measurements, as-
sessment, and reporting of results. To help classify results as nonirradiated and irradiated, a screening decision scheme and a decision-making chart for those laboratories calibrating their samples were enclosed (Table 1). Screening decisions were to be based on 2 thresholds (1, 8): T1 (700 counts in 60 s) and T2 (5000 counts in 60 s). These thresholds were those used routinely at SURRC for commercial analysis; the results of the trial and ongoing statistical work on the SURRC database suggest the possibility of refining these thresholds (e.g., product-specific thresholds). For calibrated PSL results, participants were asked to compare calibrated results with their screening results and identify samples with low sensitivity to irradiation or samples containing irradiated components. Precharacterization Precharacterization testing was performed at SURRC on all of the products as purchased for the trial. This was to establish whether any samples were already irradiated and to assess whether there was sufficient sensitivity. Duplicate aliquots were measured using both screening and calibrated modes. The 2 sets of measurements were repeated on the irradiated material on its return from Isotron. Choice of Samples and Predistribution Treatment Forty samples comprising 11 herbs, 17 spices, and 12 seasonings were purchased from Lion’s Foods Ltd., with enough extra material to provide replacement samples, if necessary, and for future work. All products were supplied in retail packs. The 11 herb samples purchased were basil, chives, Mediterranean mixed herbs, mint, mixed herbs, oregano, parsley, rosemary, sage, tarragon, and thyme. The 17 spices purchased were ground black pepper, whole black pepper, cayenne, ground cinnamon, ground coriander, ground cumin, chopped garlic, minced garlic, garlic puree, ground
992 SANDERSON ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 86, NO. 5, 2003
Figure 1. Precharacterization calibrated versus screening data for herbs.
Figure 3. Precharacterization calibrated versus screening data for seasonings.
ginger, ground mixed spice, ground nutmeg, whole nutmeg, Spanish paprika, whole pickling spice, turmeric, and ground white pepper. The 12 seasonings purchased were barbecue seasoning, chicken seasoning, hot chili powder, curry powder, mild korma curry powder, hot madras curry powder, medium madras curry powder, singha curry powder, tandoori curry powder, extra hot vindaloo curry powder, salad seasoning, and steak seasoning. Sample packs were randomly divided on arrival; control samples (nonirradiated) were placed in storage and those samples selected for irradiation were black-bagged in their original containers and sent to Isotron for a commercial irradiation dose of 10 kGy (nominal). Participants received 84 blind samples comprising 2 each of the 11 herbs, 17 spices, and 12 seasonings, plus 4 blended products. The blends were prepared at SURRC by mixing irradiated (10 kGy) and nonirradiated material. Participants received 2 curry powder blends (1 and 10% irradiated) and 2 paprika blends (1 and 5% irradiated). An equal number of irradiated and nonirradiated pots of each product were distributed, allocated to each participant either as 2 irradiated pots, 2 nonirradiated pots, or one of each category. Every participant received 4 blends, described only as curry powder or paprika, as appropriate. Preparation of Blind Samples Figure 2. Precharacterization calibrated versus screening data for spices.
Sample preparation was very simple; samples were dispensed into 50 mm Petri dishes to cover the base in a thin layer, ca 5 g. It was recommended that samples be dispensed in subdued lighting to minimize bleaching and under a lami-
SANDERSON ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 86, NO. 5, 2003 993 Table 2. Qualitative results for herb samplesa Nonirradiated
Irradiated
Samples
N
P
L
L
P
N
Thyme
3
8
Mint
5
1
0
5
0
0
0
10
0
0
Basil
8
0
Oregano
6
0
0
8
0
0
0
10
0
0
Tarragon
9
Sage
7
0
0
7
0
0
2
0
7
0
0
Parsley Rosemary
1
9
0
6
0
0
6
0
0
10
0
0
Mixed herbs
4
4
0
7
0
0
Mediterranean mixed herbs
7
0
0
6
0
0
Chives Total a
8
2
0
5
1
0
64
26
0
81
1
0
Eight laboratories returned classification decisions for 172 blind coded samples (4 irradiated samples were not measured) based on the classification scheme in the protocol. N = No evidence of irradiation; P = possibly irradiated; L = likely to be irradiated.
Table 3. Qualitative results for spice samplesa Nonirradiated Samples
Irradiated
N
P
L
L
P
N
Ground black pepper
9
1
0
6
0
0
Whole black pepper
7
0
1
3
5
0
Ground white pepper
4
1
0
0
9
1
Cayenne
1
7
0
8
0
0
Ground cinnamon
6
0
0
9
1
0
Ground coriander
8
2
0
6
0
0
Ground cumin
5
0
0
11
0
0
Ground ginger
10
0
0
6
0
0
Ground mixed spice
8
0
0
8
0
0
Ground nutmeg
8
1
1
5
1
0
Whole nutmeg
8
0
0
5
1
0
Spanish paprika
5
1
0
9
1
0
Whole pickling spice
7
1
0
9
0
0
Turmeric
3
3
0
8
0
0
Minced garlic
8
0
2
6
0
0
Garlic puree
6
0
0
8
2
0
Chopped garlic Total a
9
1
0
6
0
0
112
17
4
113
20
1
Eight laboratories returned classification decisions for 267 blind coded samples (should be 272; however, 3 nonirradiated and 2 irradiated samples were not measured) based on the classification scheme in the protocol. N = No evidence of irradiation; P = possibly irradiated; L = likely to be irradiated.
994 SANDERSON ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 86, NO. 5, 2003 Table 4. Qualitative results for seasoningsa Nonirradiated
Irradiated
Samples
N
P
L
L
P
N
Barbecue seasoning
8
0
0
8
0
0
Chicken seasoning
2
4
0
8
0
0
Salad seasoning
3
1
0
10
0
0
Steak seasoning
5
6
0
6
0
0
Mild korma curry powder
4
0
1
10
0
0
10
0
1
6
0
0
Hot madras curry powder
9
0
0
7
0
0
Hot Vindaloo curry powder
9
0
0
6
0
0
Curry powder
9
0
0
8
0
0
Tandoori curry powder
4
2
0
10
0
0
Singha curry powder
6
1
1
8
0
0
Hot chili powder
4
4
0
8
0
0
73
18
3
97
0
0
Medium madras curry powder
Total a
Eight laboratories returned classification decisions for 191 blind coded samples (should be 192; however, one irradiated sample was not measured) based on the classification scheme in the protocol. N = No evidence of irradiation; P = possibly irradiated; L = likely to be irradiated.
nar flow cabinet to minimize cross-contamination. All laboratories were asked to make screening measurements on duplicate aliquots, followed by irradiation with 1 kGy, then a second measurement (CalPSL) after overnight storage. In practice, only 5 of the 8 laboratories that returned results were able to irradiate their samples. PSL Measurement and Recording Participants were asked to conduct measurements for 60 s, with signals recorded every second and as a cumulative terminal count. Results were to be recorded as PSL (recording each separate count) and summary (terminal counts only) files on disk for both screening and calibrated measurements, leading to classification. Classification of samples using screening mode (Table 1) was based on a consensus of the 2 aliquots. A sample that gives 2 positive (>T2) results should therefore be classified as irradiated. For a sample with one aliquot exceeding either T1 or T2 and the other not exceeding the same threshold, 4 further aliquots should be measured and the classification based on the 2 highest results. Samples which gave intermediate signals for both aliquots should be classified as intermediate and requiring further investigation.
of the screening signals from nonirradiated material fell below the lower threshold (T1); however, one nonirradiated sample produced signals between the lower and upper thresholds (intermediate). This suggests there may be some scope for product-specific thresholds. For the irradiated samples, most gave screening signals greater than the upper threshold, with a few cases producing intermediate signals. Calibration of the samples produced complete separation between the 2 categories, thus differentiating between high-sensitivity nonirradiated and low-sensitivity irradiated samples. For the spice samples, good separation between nonirradiated and irradiated material was observed in screening mode, with the exception of 2 irradiated pepper samples which produced signals below T1. Calibration again produced complete separation except for these 2 pepper samples (4 aliquots). The separation between irradiated and nonirradiated seasoning samples was exceptionally good. All the nonirradiated products gave signals below the lower threshold, and all the irradiated products gave signals in excess of the upper threshold. These seasonings contain salt, which has a very high sensitivity to irradiation. Participants’ Results
Results and Discussion Precharacterization Precharacterization of the products provided a means of checking the irradiation status of the samples as purchased for the trial and of assessing the sensitivity of each product (CalPSL). Figures 1–3 show precharacterization screening versus calibrated data for herbs, spices, and seasonings, respectively. In screening mode, good separation was observed between the irradiated and nonirradiated herb samples. Most
Initial screening results were returned from 8 laboratories, of which 5 also returned calibrated results. Participants returned data in the form of initial and calibrated PSL counts for all the blind samples and a qualitative decision for each sample using the guidelines provided. One participant did not perform duplicate measurements on each sample. (a) Qualitative screening results.—The true status of each sample was compared with the results reported by the participant. Tables 2–4 summarize the participants’ quali-
SANDERSON ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 86, NO. 5, 2003 995 Table 5. Qualitative results for blendsa Irradiated Samples
L
P
1% Curry powder blend
8
0
0
1% Paprika blend
6
2b
0
5% Paprika blend
7
1b
0
8
0
0
29
3
0
10% Curry powder blend Total a
b
N
Eight laboratories returned classification decisions from a total of 32 blind coded samples based on the classification scheme in the protocol. N = No evidence of irradiation; P = possibly irradiated; L = likely to be irradiated. Intermediate was recorded with classification as positive, i.e., inter (+ve).
tative results for herbs, spices, and seasoning samples, respectively. A total of 172 blind herb samples (90 nonirradiated and 82 irradiated) were analyzed by the participants. From these samples, no false-positive instrumental results or incorrect classifications were returned. Of the total number of blind samples, 84.3% were correctly identified on the basis of positive or negative instrumental results alone. The remaining 15.7% produced intermediate signals; most were from nonirradiated samples such as thyme, parsley, and mixed
herbs, which were neither incorrect, nor correct. Participants classified these as requiring further investigation. From a total of 267 blind spice samples (133 nonirradiated and 134 irradiated), correct identification of 84.3% was achieved, based on positive or negative instrumental results alone, and 13.9% of the total produced intermediate results requiring further investigation. Participants recorded one false negative and 4 false positives, one of which could be explained by instrumental error. Of a total of 191 blind seasoning samples (94 nonirradiated and 97 irradiated), 89% were classified correctly from screening and 9.4% gave intermediate results, but there were 3 false positives. There were no false positives in the precharacterization. Correct identification of 100% was achieved for the irradiated samples. Thirty-two blind blended samples, prepared at SURRC, were sent to the participants. Table 5 displays participants’ results. Correct classification of 100% was achieved from 29 positive and 3 intermediate results. For the curry powder blends, all participants obtained positive instrumental results, even with only 1% of irradiated material present. With the paprika blend, some intermediate results were produced for both the 1 and 5% dilutions. The intermediate results were classified by the participants as the sample having been irradiated, contrary to the protocol. (b) Calibrated results.—The sensitivity of samples to irradiation can be assessed by calibration, which provides additional information for interpreting screening results, particularly in the case of intermediate signals. A total of 788 calibrated measurements on 400 samples (199 nonirradiated and 201 irradiated) were performed by 5 of the 8 laboratories. Quantitative results from the 5 participants are shown in
Figure 4. Calibrated versus screening results for participants’ data for herbs, spices, and seasoning samples.
996 SANDERSON ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 86, NO. 5, 2003
Figure 4 for herbs, spices, seasonings, and Figure 5 for blends, plotting calibrated versus screening results. For herb samples, the screening and calibrated results show that calibration distinguishes between high-sensitivity nonirradiated and low-sensitivity irradiated samples. Participants’ results support the classification criteria described in the decision-making chart incorporated in the protocol (Table 1). For spices with intermediate screening results, there was generally less resolution by calibration than for herb samples. There was, however, evidence of contamination and instrumental problems for one laboratory. Positive calibrated results were obtained from both irradiated and nonirradiated samples which had given intermediate screening results, and intermediate calibrated results were mostly, but not exclusively, from irradiated samples. This implies that a double intermediate signal suggests irradiated material, but further investigation is still advisable, whereas further investigation is essential for the intermediate positive combination. From the precharacterization data, black pepper (both whole and ground) were seen to display low sensitivity, and no participants produced positive calibrated signals for these products; all these should have been recommended for further investigation. One laboratory routinely used a lower screening threshold for peppers and was able to correctly distinguish irradiated samples. For seasoning samples, 3 false-screening positives were rejected because there was evidence of laboratory contamination. All the seasoning samples contained salt, which has an exceptionally high sensitivity to irradiation. As a result, all the irradiated samples were easily identified by screening alone, leading to complete separation of the 2 categories. Calibration of the nonirradiated products confirmed that salt was present. The blends prepared at SURRC were mixtures of irradiated and nonirradiated portions of the same material. All the curry powder samples should therefore have had very similar sensitivities, as should all the paprikas. Therefore renormalization should affect all the samples of each product in the same way, and any spread in the calibrated measurements is attributable to instrumental differences and sample handling, with a minor contribution from the inherent sensitivity variation. Figure 5 shows a tight grouping of calibrated results consistent with this.
from nonirradiated samples; 79% of the nonirradiated samples could be classified by screening alone, compared with 93% of the irradiated samples. Table 2 shows that 21 of the 26 (81%) intermediate signals recorded from nonirradiated herbs came from thyme, parsley, and mixed herbs, which is consistent with the high natural sensitivity frequently observed during routine testing at SURRC. For spices (Table 3), 41% of intermediate results from nonirradiated material were for cayenne and 18% for turmeric, again consistent with experience. Whole black pepper and ground white pepper contributed 25 and 36%, respectively, of the intermediate screening results from the irradiated samples, confirming that these products often display low sensitivity. Where nonirradiated seasoning samples produced intermediate screening results, this was likely to be related to the composition of the mixture. These findings support the case for product-specific thresholds. A total of 400 samples (201 irradiated and 199 nonirradiated) were also calibrated, producing 788 measurements. These samples included 55 of the 85 which had given intermediate screening results; 22 intermediate and 33 positive measurements were recorded after calibration. Calibration therefore resolved 60% of the intermediate cases. Overall, 94% of samples were either correctly classified by screening alone, or resolved by calibration. The remaining 6% of samples required further investigation by TL.
Conclusions An international interlaboratory collaborative trial was conducted to establish validated PSL screening methods for herbs, spices, seasonings, and blends. Eight laboratories participated and analyzed the herb, spice, and seasoning samples under blind conditions, and 5 of these laboratories also conducted calibrated measurements renormalizing the samples to 1 kGy. A total of 1593 screening measurements were performed on 662 samples (345 irradiated and 317 nonirradiated); 577 (87%) qualitative classifications were returned, of which 569 (99%), based on positive or negative instrumental readings, correctly identified the irradiation status of the product. Eight incorrect identifications were received, most of which were attributable to operator or instrumental error; 85 samples (13%) produced intermediate signals requiring resolution by calibration. The majority of these intermediate signals derived
Figure 5. Calibrated versus screening results for participants’ data for the blends.
SANDERSON ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 86, NO. 5, 2003 997
Calibration was shown to be most valuable where samples gave intermediate screening results and identified those samples requiring further investigation because of low sensitivity. For irradiated products, screening classification was so good that calibration made very little difference. With nonirradiated material, CalPSL enabled the classification of an additional 8% of the samples. Analysis of SURRC blends showed that even 1% irradiated material was detectable. These blends, however, were prepared from irradiated and nonirradiated portions of the same material. Ingredients in commercial blends, whether compound foods or derived from several batches of a single product, may exhibit differing sensitivities. This will affect the detection limit for irradiated components; a new MAFF-funded study for the detection of blends with varying sensitivities is in progress. Classifications in this study were based on predefined thresholds. Study results indicate that the upper threshold (T2) adequately separates the irradiated products. There is, however, a case for reviewing the lower threshold (T1). Low sensitivity products such as peppers might benefit from product-specific thresholds. The large number of intermediate signals recorded from nonirradiated material suggests that the general threshold could also be altered to optimize separation. The case for validation of the PSL screening method for herbs, spices, and seasonings is supported by results in excess of 85% correct overall. Optimization of the lower threshold would increase the percentage correct for nonirradiated samples. The use of calibration must be recommended, however, for all samples producing intermediate results and for all products likely to have low sensitivity, which produce nonpositive screening results. CalPSL increased the percentage of correct classification to 94% overall. Further investigation (analysis by TL) is recommended for all cases where calibration does not resolve the issue (6% in this study). Acknowledgments Support for this work by MAFF under contract number 1B073 is gratefully acknowledged. We are also grateful to the following participants for the work and time they contributed to the study, without which it would not have been possible: P. Brown and Y. Burton, Lincolne, Sutton and Wood Ltd., Norwich, UK
.
S.D. Clemenson, East Anglian Food Ingredients Ltd., Essex, UK G. Hart, McCormick UK plc, Buckinghamshire, UK H. Nootenboom, Food Inspection Service, Nijmegen, The Netherlands S. Pinnioja, University of Helsinki, Helsinki, Finland G.A. Schreiber and U. Wagner, BGVV, Berlin, Germany C. Wiezorek, Chemisches Landes und Staatliches, Veterinarintersuchungsamt, Munster, Germany References (1) Sanderson, D.C.W. (1991) in Potential New Methods of Detection of Irradiated Food, J. Raffi & J.J. Belliardo (Eds), EUR 13331, Luxembourg, pp 159–167 (2) Sanderson, D.C.W., & Clark, R.J. (1994) Radiat. Meas. 23, 633–639 (3) Clark, R.J., & Sanderson, D.C.W. (1994) Radiat. Meas. 23, 641–646 (4) Sanderson, D.C.W. (1993) Detection of Irradiated Samples, UK Patent No. 93-8542930424 (5) Sanderson, D.C.W. (1997) Detection of Irradiated Samples, UK Patent No. 2,291,707 (6) Sanderson, D.C.W., Carmichael, L.A., & Naylor, J.D. (1996) in Detection Methods for Irradiated Foods, C.H. McMurray, R. Gray, E.M. Stewart, & J. Pearce (Eds), Royal Society of Chemistry, Cambridge, UK, pp 140–148 (7) Sanderson, D.C.W., Carmichael, L.A., & Naylor, J.D. (1995) Food Sci. Technol. Today 9, 150–154 (8) Sanderson, D.C.W., Carmichael, L.A., & Naylor, J.D. (1996) in Detection Methods for Irradiated Foods, C.H. McMurray, R. Gray, E.M. Stewart, & J. Pearce (Eds), Royal Society of Chemistry, Cambridge, UK, pp 124–138 (9) Sanderson, D.C.W. (1990) in Food Irradiation and the Chemist, D.E. Johnston & M.H. Stevenson (Eds), Royal Society of Chemistry, Cambridge, UK, pp 25–56 (10) Sanderson, D.C.W., Carmichael, L.A., Ni Riain, S., Naylor, J.D., & Spencer, J.Q. (1994) Food Sci. Technol. Today 8, 93–96 (11) MAFF Working Party on Food Authenticity (1997) Undeclared Irradiation of Foodstuffs, Ministry of Agriculture, Fisheries and Food, Draft Report, June 1997, London, UK
