Antifungal activity of essential oils evaluated by two different application techniques against rye bread spoilage fungi


K.I. Suhr, Mycology Group, Biocentrum-DTU, Technical University of Denmark, Søltofts Plads, Building 221, 2800 Kgs. Lyngby, Denmark (e-mail:


Aims: To study how antifungal activity of natural essential oils depends on the assay method used.

Methods and Results: Oils of bay, cinnamon leaf, clove, lemongrass, mustard, orange, sage, thyme and two rosemary oils were tested by two methods: (1) a rye bread-based agar medium was supplemented with 100 and 250 μl l−1 essential oil and (2) real rye bread was exposed to 136 and 272 μl l−1 volatile oil in air. Rye bread spoilage fungi were used for testing. Method 1 proved thyme oil to be the overall best growth inhibitor, followed by clove and cinnamon. On the contrary, orange, sage and rosemary oils had very limited effects. Mustard and lemongrass were the most effective oils by the volatile method, and orange, sage and one rosemary showed some effects. Oil compositions were analysed by gas chromatography-mass spectrography.

Conclusions: Antifungal effects of the essential oils depended on the application method. Larger phenolic compounds such as thymol and eugenol (thyme, cinnamon and clove) had best effect applied directly to medium, whereas smaller compounds such as allyl isothiocyanate and citral (mustard and lemongrass) were most efficient when added as volatiles.

Significance and Impact of the Study: This study proves that the method used for screening essential oils as potential antimicrobials should correspond with the application sought.


Antimicrobial properties of herbs and spices have been recognized and used since ancient times for food preservation and in medicine (Zaika 1988; Conner 1993). Scientific reports on natural antimicrobial agents also date to back more than a century, e.g. Chamberlain reported in 1887 the action of essential oil vapours on anthrax spores, as cited by Maruzzella and Sicurella (1960). A renewed interest in ‘natural preservation’ appears to be stimulated by present food safety concerns, growing problems with microbial resistance, and a rise in production of minimal processed food joined with ‘green’ image policies of food industries.

Numerous studies have documented the antifungal (Thompson 1989; Mishra and Dubey 1994; Cox et al. 1998; Özcan 1998; Cosentino et al. 1999; Aligiannis et al. 2001; Elgayyar et al. 2001) and anti-bacterial (Lachowicz et al. 1998; Wan et al. 1998; Canillac and Mourey 2001) effect of plant essential oils. Examinations of indigenous local herbs and plant material have also been reported from around the world e.g. India (Ahmad and Beg 2001), Australia (Cox et al. 1998), Argentina (Penna et al. 2001) and Finland (Rauha et al. 2000). Screening experiments with 13–52 essential oils and major active components against five to 25 micro-organisms (Conner and Beuchat 1984; Deans and Ritchie 1987; Smith-Palmer et al. 1998; Hammer et al. 1999; Dorman and Deans 2000) have reported thyme, clove, cinnamon, bay, oregano, garlic and lemongrass to be some of the best broad spectrum candidates for inhibition of food-borne pathogens and spoilage organisms. However, comparisons of studies that have used different methodologies are difficult, especially regarding minimal inhibitory concentrations (MIC), and the need for uniform and reliable procedures when testing activity has been emphasized (Davidson and Parish 1989; Smith and Navilliat 1997; Mann and Markham 1998). The extensively used method is diffusion (from discs, drops, wells or ‘double plate’) (Zaika 1988; Sivropoulou et al. 1996; Mangena and Muyima 1999) or agar and broth dilution (mixing the agent into the growth media) (Paster et al. 1990; Cosentino et al. 1999; Özcan and Boyraz 2000). Unfortunately, results obtained by the different methods using the same oils do not always compare (Hili et al. 1997). Solubility and diffusion rates of the oil compounds in the aqueous agar media are of paramount importance, and can lead to misinterpretations in diffusion tests. Emulsifying agents (Tween-20 or Tween-80) are often added to ensure uniform mixing in dilution tests (Smith and Navilliat 1997), but emulsifiers and solvents have been shown to decrease the activity (Remmal et al. 1993; Hili et al. 1997; Mann and Markham 1998). A 0·2% agar solution has subsequently been recommended as the least disturbing detergent (Remmal et al. 1993).

A third method for testing essential oils is the microatmosphere method (Maruzzella and Sicurella 1960) or ‘cup method’ (Zaika 1988) and fumigation (Paster et al. 1995; Arras and Usai 2001). This method brings the vapours of essential oils into contact with growth media and micro-organisms. Thus, the problem of inadequate mixing is circumvented, and instead distribution in the air-phase of the active compounds is important. Evaporation inside Petri plates (ca. 8 mm distance between agar surface and oil) can generate inhibition zones (Maruzzella and Sicurella 1960), indicating a gradient and non-uniform air-phase distribution of vapour compounds. The concentration of compounds in the head-space has also been reported to be different from the composition of the liquid oil, reflecting differences in volatility of the compounds (Benjilali et al. 1984). Thus, the dynamics of the volatile system are very important and conflicting results between volatile methods and diffusion or dilution methods could be expected.

To make the most appropriate assessments of potential antimicrobials, the future application should be considered when choosing a screening method. For active packaging (Nielsen and Rios 2000) and fumigation of storage crops this would be a volatile screening method, whereas substitutes for synthetic preservatives like sorbates, propionates and benzoates should be assayed by dilution methods.

Composition of media, exposure time and environmental factors (temperature, pH, aW) are important for all test methods. The lack of congruence between agar media assays and real food studies (Arras and Usai 2001; Ultee and Smid 2001) emphasizes the role that fat and protein content of the food play in solubility and chemical reactions with the active compound (Juven et al. 1994; Smith-Palmer et al. 2001), and encourages the use of food product analogues instead of standard laboratory media. This study was performed in a model system for bread, because bread is a staple diet worldwide that spoils fast when not preserved. In Denmark, rye bread is a very important bread type. The main spoilers of rye bread are Penicillium roqueforti, P. corylophilum and Eurotium ssp. (Lund et al. 1996). Also yeasts e.g. Endomyces fibuliger (Legan and Voysey 1991) and Aspergillus flavus are commonly found.

Materials and methods

The fungi

For inoculation, P. roqueforti (IBT 21319), P. corylophilum (IBT 6978), Euro. repens (IBT 18000), A. flavus (IBT 21323) and Endo. fibuliger (IBT 605) were obtained from the culture collection of Biocentrum-DTU, Kgs. Lyngby, Denmark. All originated from spoilage of rye bread and other bakery products.

The fungi were transferred to fresh media and incubated for 7 days, repeated twice, immediately before the experiment. Czapeck yeast autolysate extract agar was used for P. roqueforti, P. corylophilum and A. flavus, Dichloran 18% glycerol agar for Euro. repens and Yeast malt agar for Endo. fibuliger (Samson et al. 2000). All chemicals used were from Difco Laboratories (Detroit, MI, USA).

Suspensions of 106 spores or yeast cells per millilitre were made in water containing 0·5% Tween-80 and 0·5% agar for inoculation of media and rye bread.

The rye bread media

Two per cent (w/v) rye bread granulate (Q-Brod, Smørum, Denmark), 1% KH2PO4, 0·3% (NH4)2SO4, 0·1% (v/v) trace metal solution (according to Samson et al. 2000), 1% mineral solution (Samson et al. 2000), 21·2% (w/v) glycerol (J.T. Baker, Deventer, Holland) and 2% (w/v) agar (Bie & Berntsen A-S, Rødovre, Denmark) were added to double distilled water. All chemicals used were from Merck, Darmstadt, Germany, unless otherwise stated in text. The pH was adjusted to 6 and 4·5 with HCl and NaOH. Essential oils were diluted in 96% ethanol, filtered (0·22 μm), and aliquots were added to the autoclaved and cooled (ca. 50°C) medium in final concentrations of 100 or 250 μl l−1 medium. The stock solutions of essential oils in ethanol were made in 10 and 25 μl l−1, and were thus subsequently diluted 100-fold upon mixing with the medium. Final aW of the media was 0·95.

Extracts of bay, cinnamon leaf, clove, lemongrass, mustard, orange, sage, thyme and rosemary were used. All extracts were from Copeland and Sons Ltd, London, UK, except mustard which was from Extract Mex, San Luis Potosi, Mexico, and one of the two rosemary extracts, rosemary 2, from Danisco Cultor, Grindsted, Denmark.

Inoculation and incubation

Each fungus was inoculated in three points from the spore suspensions on three plates of each oil and pH. Plates were packed in polyethylene bags (LR Plast, Glostrup, Denmark) for maintaining water activity, and incubated in the dark at 25°C. Colony diameter was measured after 2, 4, 7 and 14 days.

The volatile system

Dark rye bread slices (Møllens Brød, Helsingør, Denmark) were punched out with a circular mould of 49 mm inner diameter. Each cut piece (8·11 ± 0·73 g) was placed on a 40-mm watch glass, convex side down, inside a 220-ml glass jar (Gratwohl, Glostrup, Denmark). Essential oil was added in the bottom of the jar, having no direct contact with the bread. After applying the oil the jar was closed airtight. Doses of oil were 30 μl, equivalent of 136 μl l−1 air, and 60 μl, equivalent of 272 μl l−1 air of each, except mustard oil, which was added in 0·25 μl (ca.μl l−1 air) and 0·50 μl (ca.μl l−1 air) doses.

Inoculation and incubation

Each piece of bread was inoculated with 10 μl of 106 spores per millilitre or cells per millilitre. Tests were made in triplicate and incubated in the dark at 25°C.

Growth or absence of growth was monitored visually the second and fourth day after inoculation. Colony diameter was measured on the seventh day and, if no growth had occurred, on day 14.

GC-MS analysis of essential oils

Components of the oils were separated in a gas chromatograph (Hewlett-Parckard 5890 Series II, Birkerod, Denmark) equipped with a low polar DB 1701 column (J & W Scientific, Folsom, CA, USA) 30 m long, 0·25 mm internal diameter and a phase thickness of 1 μm. Column flow rate was 1·1 ml min−1 He at 35°C. Five microlitres of 0·1% diluted oil in dichlormethan (Lab-Scan, Dublin, Ireland) was injected and the temperature was held at 35°C for 1 min after injection, followed by an increase to 175°C by 4°C min−1, and then to 260°C by 10°C min−1, which was held for 5 min. Mass spectra of the compounds were obtained (Hewlett-Packard, 5972 series), and subsequent compound identifications were based on comparison with library database (NBS 75K, Hewlett Packard). All oils were examined in triplicate.

Data analysis

Partial least squares regression (PLS-R) analysis was used for descriptive modelling of data in the multivariate data analysis software program Simca-P (version 9·0; Umetrics AB, Umeå, Sweden). Growth data from both experiments were quantified as the percentage of growth compared with control treatment without oil, using the equation %Inhibition=100 − (T/C × 100), where T is colony diameter after treatment with oil and C is control, with ethanol added to media or no oil in the volatile testing system.

Error bars on graphs represent the s.d. of triplicate samples.


Inhibitory effect of essential oils added to rye bread media

A comprehensive overview of growth inhibition at days 4, 7 and 14 from the media experiment was modelled with PLS-R. Figure 1 shows the correlations between oils, dosing, pH, and inhibition of fungi. The two Principle Components (PC) of the model explained 69% of the variation of the data. Selected raw data from this analysis is presented in Fig. 2.

Figure 1.

Partial least squares regression (PLS-R) model of fungal growth inhibition measured day 4, 7 and 14 on rye bread medium supplemented with essential oils. The explained variance for each component is given in parenthesis, and the independent variables are shown in bold. Conc., Concentration of oils; A.fla, Aspergillus flavus; E.fib, Endomyces fibuliger; E.rep, Eurotium repens; P.cor, Penicillium corylophilum; P.roq, P. roqueforti ; 4, day 4; 7, day 7; 14, day 14

Figure 2.

Growth inhibition of fungi on rye bread medium, pH 4.5, supplemented with 250 μl l−1 essential oils. Day: 7 (▪), 14 (bsl00039)

The first PC models primarily a measure of inhibition, i.e. factors and essential oils placed further to the right had the strongest inhibitory effect. Thyme, clove, cinnamon, mustard and lemongrass were all clearly inhibitory towards the fungi (and all positive on PC1). Overall the pH difference from 4·5 to 6·0 did not affect inhibition significantly, as the factor pH was located near origo in the model. Concentration of oil was the most important factor, positively correlated with growth inhibition of fungi on the first PC. Obviously, higher dosing (250 μl l−1 vs 100 μl l−1 medium) gave better inhibition. Orange oil, sage and the two rosemary oils were poor inhibitors, negatively correlated with inhibition of fungi. Maximum growth inhibition obtained during days 4–7 for these oils was only ≤15% for orange, <30 % for sage and for rosemary 1, and <40 % for rosemary 2, and inhibitory effects were lost after the seventh day. The effect of bay was also negligible, but relatively stronger than orange, sage and the rosemarys: growth inhibition between 5 and 15% was recorded after 14 days for P. roqueforti, P. corylophilum, A. flavus and Endo. fibuliger (predominantly at pH 6).

The PC2 models differences in inhibitory effect and divides the inhibitory oils into two groups: (A) thyme, clove and cinnamon and (B) lemongrass and mustard. Generally, the responses (inhibition of fungi) moved in the direction of PC2 with increased incubation time, from group B towards A. Thus thyme, clove and cinnamon (A) exhibited a stronger long-term inhibitory effect, whereas lemongrass and mustard (B) were most effective initially but lost effect over time relatively faster than group A oils. However, differences between fungal species are notable (Figs 1 and 2). Euro. repens, A. flavus and P. roqueforti obtained the highest positive values (>0·2) on PC2, corresponding to greater loss of inhibitory effect of mustard and lemongrass with time by these fungi than P. corylophilum and Endo. fibuliger that maintained values below 0·2 on PC2. Thus, mustard and lemongrass exhibited a more sustained inhibition of P. corylophilum and Endo. fibuliger in comparison with the other fungi.

Inhibitory effect of essential oils added as volatiles to rye bread

Volatile mustard oil inhibited growth of all fungi completely (Fig. 3). Lemongrass was nearly as efficient, but at low concentration growth of P. roqueforti and Endo. fibuliger occurred at day 14. The two rosemary oils showed different inhibitory properties by this method. Rosemary 1 was effective, whereas rosemary 2 had no inhibitory effect.

Figure 3.

Growth inhibition of fungi on rye bread with essential oils added as volatiles in an airtight system. 270 μl 1−1 oil in air phase: (▪) Day 7, ( inline image) Day 14, 135 μl 1−1 oil in air phase:(bsl00039) Day 7, (□) Day 14 (mustard oil dosing was 2 μl−1 and 1 μl−1, respectively)

The yeast Endo. fibuliger and the two Penicillia were the most resistant fungi. Only mustard and lemongrass were able to inhibit growth of these fungi for the whole period. Oil of orange, sage, thyme, bay and rosemary 1 were also considerably inhibitory for the first 7 days, as opposed to cinnamon, clove and rosemary 2 (Fig. 3). Orange oil did not significantly inhibit the yeast at day 7 but ‘time before growth was observed’ was increased from day 2 for the control to day 7 for orange (Table 1).

Table 1.  Number of days before visible mould growth was observed on rye bread treated with essential oil vapours. Mu, mustard; Lg, lemongrass; Or, orange; Th, thyme; Ci, cinnamon leaf; Cl, clove; Sa, sage; Ro1, rosemary type1; Ro2, rosemary type2; Bay, bay oil; a, 135 μl l−1 air; b, 270 μl l−1 air; ng, no growth
Penicillium roqueforti2ng14ng24222222424222
Endomyces fibuliger2ng14ng77242224744222
Penicillium corylophilum4ngngng7144444441447444
Aspergillus flavus4ngngngngng7ng771414ng77477
Eurotium repens4ngngng14ngngng44414ng14ng447

Eurotium repens and A. flavus were the most sensitive of the fungi tested, and rosemary 2 was the only oil without clear effect (Fig. 3).

Rye bread was an excellent substrate for the major spoilers P. roqueforti and Endo. fibuliger. Growth was visible 2 days after inoculation. As opposed to this A. flavus grew slowly on rye bread, and particularly poor on bread treated with essential oil vapours (Table 1).

Comparison of methods

In order to compare the application methods, growth inhibition in the seventh day from both methods was modelled by PLS-R (Fig. 4). The analysis divided the oils into three groups: (A) the overall best inhibitors by both methods – mustard, lemongrass and thyme, (B) inhibitors upon direct addition to media – thyme, clove and cinnamon and (C) inhibitors upon vaporous contact – orange, rosemary 1 and sage.

Figure 4.

Partial least squares regression (PLS-R) model comparing the two assay methods by the measurement of fungal growth inhibition by the 10 oils at day seven. The explained variance for each component is given in parenthesis, and the independent variables are shown in bold writing. A.fla, Aspergillus flavus; E. fib, Endomyces fibuliger; E. rep, Eurotium repens; P. cor, Penicillium corylophilum; P. roq, P. roqueforti. The suffix: −V, volatile system

The effect of higher dosing was most important for the media experiment.

Composition of essential oils

A summary of the main compounds of the essential oils identified by GC-MS is listed in Table 2. All major compounds of the oils were monoterpenes, except allyl isothiocyante of mustard oil. The compositions of clove and cinnamon leaf oil were very similar with a high content of eugenol, 85 and 72%, respectively. Eugenol also made up the major part of bay oil (32%). It was not possible to detect any volatile compounds from rosemary 2 by the GC methodology used; only the carrier compound, propylene glycol, was identified.

Table 2.  Composition of the essential oils obtained by GC-MS
Essential oilMajor compoundsContent (%)
Cinnamon leafCinnamaldehyde1·88
 Benzyl benzoate5·53
 Geranial (citral a)31·93
 Neral (citral b)45·99
MustardAllyl isothiocyanate99·00
 Eucalyptol (1,8-cineole)24·93
 Eucalyptol (1,8-Cineole)19·59
Rosemary2Propylene glyco91·68
 Not identified 


The composition of essential oils varies significantly because of different species and chemotypes (Buttery et al. 1974; Tucker and Maciarello 1986; Knobloch et al. 1989; Tantaoui-Elaraki et al. 1993), geographical origin (Chalchat et al. 1993; Perry et al. 1999), season (Senatore 1996; Palá-Paúl et al. 2001; Santos-Gomes and Fernandes-Ferreira 2001) and extraction procedure (Schaneberg and Khan 2002). In this study we used a thymol-chemotype of thyme oil (Cosentino et al. 1999; Echeverrigaray et al. 2001) and a Spanish type of rosemary (Chalchat et al. 1993). The compositions of the other oils were also similarity to previous reports (Buttery et al. 1974; Perry et al. 1999; Jirovetz et al. 2000; Schaneberg and Khan 2002).

Thyme, clove and cinnamon proved to be the overall best long-term inhibitors upon direct addition to media (Fig. 1). Thymol, the major compound in thyme (Table 2), and eugenol, in cinnamon and clove, are both phenolic compounds. Bay oil also contained eugenol (32%), but not in levels comparable with clove (85%) and cinnamon (72%). The inhibitory effect of bay was accordingly much weaker than clove and cinnamon, but nevertheless close to the B-group of media-inhibitors in the PLS-R analysis, Fig. 4. Several authors have reported high antimicrobial activity in phenolic structured compounds (Sivropoulou et al. 1996; Cosentino et al. 1999; Dorman and Deans 2000), followed by aldehydes, ketones, alcohols and other hydrocarbons (Shelef 1984; Özcan and Boyraz 2000). Such statements have been based on assays using agar diffusion or dilution methods, and are consequently in agreement with the results of our media experiment.

An increased inhibitory effect of rosemary oil at lower pH (4·5–7) towards Staphyloccocus aureus has been reported (del Campo et al. 2000), although the pH effect disappeared when aW was lowered by 30 g l−1 NaCl in the reported experiment. Because antimicrobial compounds from essential oils are believed to exert their effect by hampering the plasma membrane (Helander et al. 1998; Brul and Coote 1999; Cox et al. 2000), electric charge and polarity must be important factors for activity. In theory, changing pH of the substrate could influence this. But the change of pH from 4·5 to 6 in rye bread media did not have a significant effect on activity in our work.

Rosemary 2 was a natural anti-oxidant product and contained propylene glycol as a carrier of the active compounds, which presumably were larger and less volatile compounds such as flavonoids and diterpenoids. Rosemary 2 can be expected to be active by direct addition. However, none of the two types of rosemary oils showed considerable effect in the media at the concentrations tested. Only a limited short-time effect was recorded, and the highest effect of these was rosemary 2 towards A. flavus (Fig. 2).

Allyl isothiocyanate (AITC) is a low molecular volatile of mustard oil, with high antimicrobial potential (Lin et al. 2000). It was fungicidal to all fungi at 1 μl l−1 air in the volatile and airtight system (Fig. 3). However, when added to agar media at 250 μl l−1 medium the mustard oil was not able to maintain its inhibitory effect for 2 weeks (Fig. 2). The reason why the inhibitory effect of AITC was lost after the first week could be evaporation out of the Petri plate and PE bags, or decomposition upon reaction with substances in the medium. Eurotium repens and A. flavus were the most resistant fungi towards AITC in the media. As Aspergillus is the imperfect (asexual) form of Eurotium some mutual properties of conidospore and mycelium structure or metabolic capacity could be the reason for their similar resistance. Great differences in the effect of mustard/AITC have also been reported by others, e.g. MIC ranging from 0·016 to 600 ppm were found by Delaquis and Mazza (1995) from their own and the studies of others, which confirm our findings of the great impact methodology has on results.

Citral isomers are (non-aromatic) aldehyde monoterpenes and the major constituents of lemongrass oil. This oil also showed a much stronger effect in volatile form. However, the yeast Endo. fibuliger was different from the moulds in the media experiment by being most inhibited by lemongrass of all oils tested.

Like mustard and lemongrass oil, orange oil, showed notable difference in activity between the two assay methods. Orange oil contained mainly the volatile component d-limonene. No activity was recorded in the media, but in the volatile system the two Penicillia, A. flavus and Euro. repens, were inhibited.

Sage and rosemary1 had many major compounds in common (α-pinene, eucalyptol, camphor and limonene; Table 2), and their inhibitory properties were also similar in both methods. An inhibitory effect was lacking upon addition to medium, but the effect in vaporous form was considerable (Fig. 3P. corylophilum, A. flavus and Euro. repens).

There was no visible sign of inhibitor gradients building up in the bread, compared with the reported inhibition zones from micro-atmosphere methods in Petri plates. The greater distance from bread to oil disposal, and the small amounts of oils used, which ensured that the oils evaporated completely inside the airtight system, can be the reason for this.

Benjilali et al. (1984) reported that headspace analysis above essential oils gave a higher concentration of the lightest compounds (e.g. α-pinene, camphene) compared with the heavier ones (e.g. phenolic compounds) than was found in analysis of the liquid oil. Differences in the specific volatility of the active compound can explain the better effect of mustard, lemongrass, sage, rosemary 1 and orange observed in the volatile system compared with the media in this study. However, the composition of the media/bread was not identical in the two systems, although pH and aW was somewhat comparable (low pH media). The content of water was lower in the bread than agar media, and the sourdough-fermented rye bread was a more complex medium than the 2% rye bread media. Aspergillus flavus and Euro. repens showed, in particular, poorer growth on bread than on medium.

Despite the differences in media between the two methods, our results showed that smaller and non-phenolic volatile compounds, such as AITC, citral and limonene, were most effective when applied through the air phase. Oils composed of larger phenolic compounds, such as thymol and eugenol, had the best use as direct additives. Thymol, from thyme, is a smaller and more volatile molecule than the ether-containing eugenol from clove and cinnamon. Thyme was also generally more effective than clove and cinnamon in the volatile experiment (Fig. 4). Findings of Inouye et al. (2002) support our results, as they reported that lemongrass was absorbed by mould mycelium and media in greater amounts compared with thyme and cinnamon bark oil, when these were applied as vapours.

Most often essential oils are screened for antimicrobial effect by direct addition and diffusion in media. Our work has shown that screening methods should reflect the future application. Thus a volatile screening method should be employed for fumigation or active packaging purposes.


Financial support from EU, Fair PL98-4075, is gratefully acknowledged. Anne Winter Hinsby and Jeorgos Trihaas are acknowledged for their excellent technical assistance.