Tempe part 1: traditional fermentation, fungal trials, and regional seeds

Added on by Bernat Guixer.

by Bernat Guixer

This project investigates applications of tempe mould (primarily Rhizopus spp.) in the kitchen.
The results will be spread over two blog posts. The first is devoted to introducing and contextualising tempe as a food product: its origin, its raw materials, and the key points of its fermentation process. We provide the protocols used here at NFL for bean and wheat tempe production. Moreover, we describe the use of some Nordic legume varieties for tempe production. The second part will cover how we further developed our preliminary tempe into a completely different and exciting product by harnessing the moulds’ metabolic enzymes to produce different sensory qualities.


Tempe is a fermented food from Indonesia, originally based on soy beans. It is produced through a solid state fungal fermentation process leading to a mycelia-knitted compact cake of beans. The key microorganism leading the process is a fungus from the Rhizopus genus.

Traditional tempes from Indonesia

Traditional tempes from Indonesia

Beyond soy beans, there are several primary ingredients used for tempe production. We can divide tempe into five general categories based on its raw materials: legumes, grains, grains and legumes, presscake residues and non-legume seeds (Nout & Kiers 2005, Feng 2006). This is a quite generic classification, and certainly there are several other factors that greatly  influence the final product—for example, the fermentation process.

Tempe is an Indonesian staple food, and originated hundreds of years ago in central and east Java Island (Shurtleff & Aoyagi 1979). Nowadays, much tempe is made with whole soybeans (Nout & Kiers 2005), yet earlier forms of tempe have been made from okara, the residue from soymilk extraction for tofu making. The earliest mention of tempe is 'Sambal Lethok', a dish made of overripe tempe, recorded in a 17th-century Centhini (old Javanese) inscription. The first occurrence of the word for soybean in ancient Javanese is 'kadele' ('ka' meaning 'bean', and 'dele meaning 'black') documented in the Sri Tanjung from the 12th-13th century. Thus it is believed that tempe was initially made with black soybeans (Astuti, 2001).

Traditional tempe taste testing

Traditional tempe taste testing

Tempe is a highly nutritious, easily digestible and tasty product. In a world where more and more people are eating too much animal meat, by any measures of health, nutrition, and the environment, developing additional delicious, satisfying foods from proteinous raw materials can help to address over-consumption of resource-intensive foods such as animal meat. Soybeans, and thus soybean tempeh, contain all essential amino acids, those that must be supplied to the organism since our body is not able to synthesize itself. It is low in saturated fats and free of cholesterol. It is highly digestible and generally low in cost.

All these features make tempe an interesting product to investigate for its gastronomic potential.

Tempe fermentation at a glance

As with many fermented foods, tempe likely developed over the course of years, through chance discovery, trial and error, and knowledge transfer generation over generation, in particular geographic and climatic conditions.

These conditions favour the development of certain plant, microbe, and other species that become the building blocks for human-microbe-environment interactions, and the resulting fermented foods. Through careful observation craft producers were able to modify production conditions during fermentation to incrementally alter the final product.

Tempe is a good example of a fermented food whose transformation is advantageous not primarily for preservation purposes, but rather for the increased nutritional properties of the final product (Shurtleff & Aoyagi 1979). Fresh tempe has a relatively limited shelf life due to its microbial enzymatic activity. When stored, freshly-produced tempe will begin to turn brown in a matter of days or even a few extra hours, the firm and spongy texture of the mycelia collapses, and a characteristic ammoniac odour develops. Yet such a product, after being either intentionally stored for too long or deliberately over-ripened, can still be used to produce strongly flavoured sauces, toppings or cookies (Wijaya & Gunawan-Puteri 2015, Nout & Kiers 2005). Despite its rather short shelf life, the fermentation process contributes other great benefits: it increases the digestibility and bioavailability of proteins, carbohydrates, lipids and minerals; enhances the vitamin content of the final product; significantly decreases the amount of so-called ‘antinutrients’ (trypsin inhibitors, phytic acid, etc.) and flatulence-related sugars which might be contained in the raw materials; and significantly shortens the cooking time (and thus energy consumption) required for the primary ingredients.

Production protocols at NFL

Tempe is relatively simple to produce. However, controlled stable conditions during the production process are required to achieve a satisfactory result, particularly in a place such as Denmark where the climate differs strongly from that of Indonesia. It is not our purpose here to systematically review tempeh production protocols—there are several excellent bibliographic sources with many details on tempe production processes (Shurtleff & Aoyagi 1979, Nout & Kiers 2005, Steinkraus 2004).

With this project we wanted to see what this Indonesian technique would yield with Nordic raw materials. As a starting point we selected brown beans from Øland in Sweden as the primary material for our tempe. We also experimented with other Nordic legumes and grains to expand the varieties previously reported for tempe production (Nout & Kiers 2005, Feng 2006).

The Swedish brown Øland bean

The Swedish brown Øland bean

We did not aim to fully optimise this protocol since we were more interested in focusing on other features besides the production method, which will be detailed in part 2.

Bean tempe protocol
Yield: around 1 kg

Øland brown beans, or any kind of beans you desire
Water for soaking and boiling
Rhizopus spp. spores (commercially available)
Ziplock bags
An incubator or any chamber able to maintain a temperature of around 30ºC and stable humidity conditions, also having some space for airflow.

1. Perforate ziplock bags with holes approximately .5-1mm in diameter spaced approximately 1 cm apart, to to allow oxygen to penetrate the bean cake for proper fungal development.
2. Place 600g of dry beans in water and bring to a boil.
3. Cool down water with ice and let soak overnight.
4. Strain beans and dehull them.
5. Add vinegar to water until it reaches a pH of 4.8, and cook beans in this solution for 12 min.
6. Strain beans and spread them in a towel for drying and cooling (approximately 20 min).
7. Inoculate with fungal spores following ratio indicated by culture supplier (Top Cultures: 3g of spores; Raprima: 1g of spores).
8. Place the sporulated beans in perforated ziplock bags The thickness of the filled package should not exceed 2cm.
9. Incubate at 30˚C for 24-30 hours. Humidity largely self-regulates within the ziplock bag.

Fermenting tempe in the oven.

Fermenting tempe in the oven.

In the course of developing this protocol, we tested three different pure cultures corresponding to different Rhizopus species and strains.

A)    Rhizopus oryzae (Top Cultures)
B)    Rhizopus oligosporus (Top Cultures)
Rhizopus oligosporus (Raprima culture)—Culture broadly used in Indonesia for most tempe producers not using wild cultures for tempe inoculation. Michael actually brought this culture back to Copenhagen directly from Indonesia.

We observed dramatic differences between these Rhizopus species, both in their behaviour during fermentation and on the sensory features of the final product.

Left to right: NFL Tempe trials on Øland brown beans with fungal cultures A, B, and C.

Left to right: NFL Tempe trials on Øland brown beans with fungal cultures A, B, and C.

Successful trials of NFL tempe with Øland brown beans and Raprima culture

Successful trials of NFL tempe with Øland brown beans and Raprima culture

R. oryzae generates a surprisingly high amount of alcoholic and organic solvent aromas. The flavour was slightly acidic depending on the fermentation time: mild at 24 hours, well-balanced with interesting and surprising sour notes at 28 hours, and overwhelming sourness at 30h or more. The mycelia was white and remains so during this window (24-30h). Mycelia reaches its best consistency at approximately 28h, although it was much lighter than (not as dense or well-structured as) that expected from traditional tempe.

R. oligosporus sporulated rapidly, generating black spots (spores) around the perforations (and thus most aerated areas) as early as 24 hours. By 30h most of the mycelia covering the bean cake developed a grey colour with black spotted areas. The aroma was soft and mushroomy and the flavour was mild with little differences from the beans themselves. Dark spores are not hazardous, but they generally deter the unfamiliar eater.

Raprima culture produced a nice compact white mycelia after 30 hours of fermentation, when it reached what we judged its best characteristics. Sporulation appeared much later on the developing pattern of this mould. The aroma and flavour were mild and after being tested by an Indonesian palate (our team member Dwi) we concluded this sample to be closest to commercial tempe in Indonesia.

Since the Raprima culture behaved differently from the other R. oligosporus commercial culture we tested, they must differ at the strain level.

Øland wheat tempe protocol
Yield: around 1 kg

1. Perforate ziplock bags with holes approximately .5-1mm in diametre, to to allow oxygen to penetrate the bean cake for proper fungal development.
2. Rinse 450g of Øland wheat with water.
3. Add vinegar to water until it reaches a pH of 4.8, and boil the wheat in this solution  for 35 minutes.
4. Drain grains and crack them in a thermomix (power 10, 10 seconds) and let them cool down in a towel for 5 min.
5. Inoculate with spores following ratio indicated by culture supplier (Top Cultures: .6g/200g dry grain; Raprima: .3g/200g dry grain).
6. Place the sporulated beans in perforated ziplock bags. The thickness of the filled package should not exceed 2 cm.

7. Incubate at 30˚C for 40-46 hours.

Counterclockwise from left: NFL tempe trials on Øland wheat, using cultures A, B, and C

Counterclockwise from left: NFL tempe trials on Øland wheat, using cultures A, B, and C

Øland wheat tempe developed an extraordinary umami taste. This was its dominant sensory feature. Raprima culture developed compact, white, appealing mycelia. In terms of flavour and aroma, we judged Raprima to  produce the best tempe of the three cultures tested. R. oryzae developed a weak white mycelia, and R. oligosporus sporulates rapidly. We observed that Øland wheat tempe needs to be consumed straight away, otherwise an astringent flavour arises and rapidly spoils the product.

Heritage varieties of Danish legumes from Muld farm

Heritage varieties of Danish legumes from Muld farm

We had the opportunity to work with some regional varieties of fava beans and peas from Muld farm, which participates in the Pometet program (we have worked with them before) for preserving plant species at risk of disappearing. From those varieties we made tempe using R. oryzae and we got interesting results after 24 hours of fermentation.

NFL tempe made on different heritage varieties of Danish legumes from Muld farm

NFL tempe made on different heritage varieties of Danish legumes from Muld farm

We report some informal tasting notes:

Figure 1. Informal tasting notes for tempe made on different varieties of Danish legumes

Figure 1. Informal tasting notes for tempe made on different varieties of Danish legumes

A closer look at tempe fermentation

A deeper understanding of the fermentation process helps us influence different characteristics of the final product. Here are some of the main variables.

Mould metabolism

R. oligosporus and R. oryzae are highly versatile species. They feed on the lipids in soybeans, and are also able to adjust their metabolism to use oligosaccharides and proteins depending on variable rates of available substrates (Nout & Kiers 2005, Shurtleff & Aoyagi 1979).

Studies on the structural and functional effects of R. oligosporus on soybean tempe reported that the amount of total free amino acids formed after 24 hours of fermentation significantly increases, while the pH of the fermented soybeans stays neutral (Handoyo & Mortia 2006). Proteins (albumin, globulin, alkaline soluble) are rapidly degraded to low-molecular weight peptides and amino acids. The cell structure of the soybeans is disorganised during fermentation, loosening the bean’s initial firmness.

The germination of fungal sporangiospores starts several hours after the inoculation of the beans, leading to the development of mycelia that penetrates several layers of cells into the bean. This is why it is crucial to dehull the beans to allow the mycelia to penetrate the bean cotyledon. Disruption of the cell wall is essential to facilitate the diffusion of fungal extracellular enzymes and the degradation of the solid substrate into soluble fragments for the proper growth and development of Rhizopus spp.(Varzakas 1998).

Abiotic factors

Moisture content is described to be a crucial factor for tempe production, affecting mould activity and subsequently the quality and characteristics of the product (Shurtleff & Aoyagi 1979). Moisture of the beans has been reported to range from 55 to 70%, and while there is not much consensus in the bibliography on relative humidity preferences during fermentation (between 45-85% reported), it seems most likely to be around 80%, the typical relative humidity in Indonesia.

For proper tempe production the mould requires oxygen, and most of the strains are unable to survive anaerobic conditions. Due to low rates of diffusion into the packed tempe cake, oxygen concentration decreases during the active stages of fermentation, while CO2 concentration increases (Sparringa et al. 2002).

Furthermore, fungal metabolic activity during fermentation is an exothermic process that releases considerable heat. Due to poor heat transfer in the stationary phase (tempe cake) the temperature might increase to unsustainable levels (40-50ºC)—the optimum temperature for Rhizopus metabolism is 37ºC (Nout & Kiers 2005).

Besides abiotic aspects, another growth limitation factor is ammonia production by degradation of nitrogenous compounds (Sparringa & Owens 1999) that eventually inhibits the mycelia and fungal development.

To palliate aforementioned limiting factors (oxygen, heat and ammonia) it is convenient to reduce the thickness of the tempe cake to improve heat transfer and ventilation to the inner parts of the cake.

From metabolism to flavours and aromas

From the mould’s point of view, tempe fermentation transformations are the result of its efforts to secure its survival. The mould creates a variety of tools to help it make most use of available substrates, like the enzymes it uses to break down complex molecules (proteins, lipids and carbohydrates) into simpler molecules that they can then take up for use in growth and proper metabolic performance. These simpler molecules also become the source for increased taste—sugars lend sweetness, amino acids contribute umami, etc. These metabolic processes also produce secondary molecules which bring other flavours and aromas—alcohols, organic acids, aldehydes, ketones, lactones, esters. Moreover, each mould species and strain makes different use of different nutrients, thus producing significant variability the end product.

Sorting through many tempe trials

Sorting through many tempe trials


Tempe might apparently seem a simple product, but as most fermented foods, digging deeper yields a complex process of many overlapping factors.
In the upcoming second part, we ask: and now, with this new technique, what else can we make?


We would like to acknowledge Ane Hofmeyer from Muld farm, who kindly provided the legume samples, and Louise Windfeldt from Nationalmuseet who provided useful information for this project.


Astuti, M. (200). History of Development of Tempe. In Agranoff. J. 2001. The complete Handbook of Tempe. The Unique Fermened Soyfood of Indonesia (2nd ed., pp. 2-15). Singapore: American Soybean Association Southeast Asia Regional Office.

Feng, X. (2006), Microbial dynamics during barley tempeh fermentation. Doctoral thesis. Swedish University of Agricultural Sciences, Uppsala.

Handoyo, T. & Morita, N. (2006), Structural and functional properties of fermented soybean (tempeh) by using Rhizopus Oligosporus. Int. J. Food Prop. 9, 347-355.

Nout, MJR. & Kiers JL. (2005), Tempe fermentation, innovation and functionality: update into the third millennium. J. Appl. Microbiol. 98, 789-805.

Shurtleff, W & Aoyagi, A. (1979), The Book of Tempeh. A super soyfood from Indonesia. Professional edition. Harper & Row, New York, NY.

Sparringa, RA. Et al. (2002), Effects of temperature, pH, water activity and CO2 concentration on growth of Rhizopus oligospours NRRL 2710. J. Appl. Microbiol. 92, 329-337.

Sparringa, RA. & Owens, JD. (1999), Inhibition of the tempe mould, Rhizopus oligosporus, by ammonia. Lett. Appl. Microbiol. 29, 93-96.

Steinkraus, KH. Et al. (2004), Handbook of indigenous fermented foods. Marcel Dekker. New York, NY.

Varzakas, T. (1998), Rhizopus oligosporus mycelial penetration and enzyme diffusion in soya bean tempe. Process Biochem. 33, 741-747.

Wijaya, CH. & Gunawan-Puteri, MDPT. (2015), “Tempe Semangit”, the overripe tempe with natural umami taste. Channywijaya’s blog. Visited: 25th September 2015. http://channywijaya.staff.ipb.ac.id/2015/06/05/tempe-semangit-overripe-tempe-natural-umami-taste/