Production of Grain Whisky and Ethanol

from Wheat, Maize and Other Cereals

R.C. Agu 1,2, T.A. Bringhurst 1 and J.M. Brosnan 1

ABSTRACT Although the adoption of wheat has been very success-

ful in replacing maize, there are still some concerns re-
J. Inst. Brew. 112(4), 314–323, 2006
garding difficulties with handling this cereal in the distil-
Wheat replaced maize as the main cereal raw material for Scotch lery. These result from the higher viscosity, associated with
grain whisky production 20 years ago. However, other cereals wheat which can affect the efficient handling of process
might also have potential for use in grain distilleries and ethanol
streams, (cooking, mashing) and, more seriously, with the
production. Studies of the properties of wheat, maize, sorghum
and millet, showed that they had good potential for grain distill- recovery of co-products (spent wash /spent grains). Al-
ing and ethanol production at comparable nitrogen levels, and though the economics may be less favourable, maize is
had physiological processing characteristics within the range still considered to be superior to wheat in that it produces
accepted for wheat or maize. Rapid-Visco Analysis (RVA) stud- higher alcohol yields and presents fewer processing prob-
ies of low and high nitrogen wheat confirmed that, as well as lems. Cereals such as sorghum and millet are considered
influencing the amount of alcohol produced, the total nitrogen to have similar characteristics to maize 13. Further study of
content of the grain had a strong influence on its processing these cereals will help to elucidate their physiological
characteristics. In contrast, the alcohol yield potential of maize, properties in relation to their potential use for grain whisky
sorghum and millet appeared to be largely unaffected by the and ethanol production and may indicate if they can be
grain nitrogen levels. The study shows that, while it is possible
processed successfully compared with wheat and maize.
for wheat to produce similar alcohol levels to those previously
associated with maize, cereals other than wheat can potentially The general characteristics of maize, wheat, sorghum
be used without detriment to alcohol yield or processing per- and millet are well documented in the literature 13,18,19.
formance. These could be possible long term alternatives, if the These references show that whilst the cereals studied had
economic viability of wheat was to change. The extracted cereal similar levels of ␤-glucan (0.1%–0.5% w/w), the same
starches also showed significant differences from the original was not true of pentosans. Maize, sorghum and millet had
cereals, which had important implications for successful pro- similar levels of pentosan (ca 3.0% w/w), which was in
cessing, both in terms of cereal selection as well as cooking and contrast to wheat which had a pentosan content greater
fermentation performance. than 8.0% w/w 19. Pentosans, like ␤-glucans are known to
Key words: Ethanol, grain whisky, maize, millet, sorghum, cause processing problems5,7,11,20–22.
starch, wheat. Maize, sorghum and millet starches have higher gelati-
nisation temperature ranges (70°C–80°C)19. This possibly
INTRODUCTION reflects the tropical nature of these cereals (i.e. the cli-
matic conditions under which they are found in the field).
In Scotch grain whisky production, maize was the tra- This contrasts with the gelatinization temperature range
ditional grain of choice until it was effectively replaced by (52°C–54°C) of wheat starch, which is associated with
wheat during the 1980’s 10,24, primarily because of its more more temperate climates 1. The protein content of cereals
favourable economics. Bringhurst et al.8 have summarised can vary, depending on many factors including the grow-
the main features of raw materials employed in typical ing conditions in the field 3,17. The differences in pentosan
Scotch whisky grain distilleries. In this process, cereals, levels and gelatinization temperatures of wheat and maize
generally wheat, (although maize is still sometimes used), would also be expected to affect their processing charac-
normally are cooked under high temperature conditions to teristics.
release starch from the grain and liquefy it, so that it can Sorghum and maize have long been reported to have
be broken down efficiently by the enzymes from barley very similar characteristics and can be considered inter-
malt into fermentable sugars. These are converted to alco- changeable with regard to the brewing process 13. How-
hol by yeast during fermentation and the alcohol is col- ever, it is not clear how this similarity will extend to their
lected by distilling in continuous fractionating or Coffey processing characteristics in relation to distilling and etha-
stills. nol production. Assessment of wheat varieties by the dis-
tilling industry and dialogue with plant breeders has led to
1 The Scotch Whisky Research Institute, The Robertson Trust Build- improved varieties which are capable of producing alco-
ing, Research Park North, Riccarton, Edinburgh, EH14 4AP. hol yields equivalent to those obtainable from maize. The
2 Corresponding author. E-mail: [email protected] level of nitrogen taken up by wheat in the field is a key
factor in determining the performance of wheat in relation
Publication no. G-2006-1308-464 to its potential for alcohol production 9,23. However, the
© 2006 The Institute of Brewing & Distilling processing problems caused by the high residue viscosity

314 JOURNAL OF THE INSTITUTE OF BREWING

of wheat, in comparison with the low residue viscosity of ther filtered using a Buchner funnel prior to oven-drying
maize, are yet to be fully resolved and require further at 50°C to a moisture of approximately 5–8%.
study.
This paper describes work carried out at the Scotch Alcohol yield analysis
Whisky Research Institute to investigate more closely Alcohol yield from milled cereals. The method was
some of the properties of wheat and maize and other ce- based on that of Brosnan et al.9, which simulates the pro-
real grains (sorghum and millet), together with their asso- duction process conditions in a “typical” Scotch whisky
ciated starches, in order to gain a better understanding of grain distillery. Cereal flour, (30 g), obtained by milling
their processing characteristics in relation to their suitabil- the grains in a Buhler Miag disc mill (setting 0.2 mm) was
ity for distilling and ethanol production. Some of the in- transferred into a stainless steel mashing beaker and slur-
formation described below was presented as a poster at ried with water (81 mL), and 25 µL of Termamyl 120L (a
the Worldwide Spirit Drinks Conference, held in Edin- bacterial ␣-amylase supplied by Novozymes France S.A.)
burgh 18–22 September 2005 4. was added. The contents of the beaker were gradually
heated up to 85°C (temperature rise 2°C/min) in a water
bath, before pressure cooking at 142°C for 15 min in an
MATERIALS AND METHODS autoclave. The cooked slurry was cooled to 85°C and
given a second treatment with Termamyl (25 μL) for 30
Cereal samples min to prevent starch retrogradation. The mash was then
Cereal samples were obtained from a range of sources. transferred to a water bath at 65°C and mashed for 1h
High and low nitrogen wheat samples were obtained from with a calculated amount of high enzyme grain distilling
a plant breeder. Maize was obtained from a distilling com- malt grist (Miag setting 0.2 mm), equivalent to a malt
pany. Other maize samples, sorghum, and millet were inclusion rate of 20% (dwb) to 80% wheat. After cooling
obtained from various sources including Heriot Watt Uni- to room temperature, the mash was transferred to a suit-
versity and an African brewing company. able fermentation vessel, pitched with distiller’s yeast
(‘M’ type, supplied by Kerry Bioscience Ltd, or equiva-
Total nitrogen determination lent) at a pitching rate of 0.4% (w/w) pressed yeast and
Total nitrogen of the cereals was determined on Buchi adjusted to 250 g with water. The mash was then fer-
Kjeldahl apparatus, using a method similar to that for bar- mented at 30°C for 68 h and distilled to collect the alco-
ley, described in the Recommended Methods of Analysis hol. The alcohol yield was determined from the alcohol
of the Institute of Brewing 6. strength of the distillate, which was measured using a
Paar 5000 density meter. The alcohol yield (PSY) is
Physiological properties of the cereals – quoted as litres of alcohol per tonne (LA /tonne) on a dry
Rapid-Visco Analysis (RVA) weight basis (dwb).
Alcohol yield of starch. The alcohol yield of the starch
A study of the physiological properties of these cereals from the different samples was determined essentially as
was carried out using a Newport Scientific Rapid Visco described for the original cereals above, but using an ex-
Analyser ® (RVA) instrument supplied by Calibre Control. cess of Termamyl 120L (35 µL) to digest the starch.
The Rapid Visco Analyser is a rotational, continuously re-
cording viscometer, with heating, cooling and variable Residue viscosity
shear capabilities, specifically configured for starch-based The in-grains residue from the alcohol strength distilla-
materials. A slurry of cereal flour with water can be sub- tion was adjusted to 250 mL and centrifuged to separate
jected to a temperature programme to analyse its pasting the spent grains from the liquid. After filtering the super-
and gelatinisation properties. This method measures the natant through a Whatmans GF /A filter paper, its residue
starch properties which affect cereal handling during the viscosity was determined at 20°C using an Ostwald U-
cooking cycle. A weighed amount of sample (approxi- tube capillary viscometer (Size B), and expressed as milli-
mately 3.0 g), adjusted for moisture, was slurried with a Pascal per sec (mPas).
measured amount of water (total weight 28 g) and then
processed in the RVA analyser using a programme de- Colour determination
signed for unmalted cereals, as summarised in Table I. The colour of the distillation residue was determined
using a Lovibond 2000 comparator (Hellige and Co.,
Extraction of starch from wheat, maize,
UK)16.
sorghum and millet
Starch deriving from the different cereals was extracted
Table I. RVA programme for unmalted cereals for distilling.
using a traditional wet milling process, modified from
Watson 25. Grains (100 g) were steeped in water for at least Time Temperature Speed
50 h at room temperature, with a change of the steep (min:sec) (°C) (rpm)
water at 24 h intervals, to enable the grains to absorb suf- 0:00 50 960
ficient water. The steeped grains were milled into a slurry 0:10 50 160
1:00 50 160
using a Moulinex blender, and then washed with a large 4:42 95 160
amount of water through a 250 micron mesh (Endecotts 7:12 95 160
(test sieve); Griffin and George Ltd). After settling, the 11:00 50 160
excess water was decanted and the settled starch was fur- 13:00 50 End

VOL. 112, NO. 4, 2006 315

 Fig. 1. A typical RVA profile for distilling wheat.

RESULTS AND DISCUSSIONS (1.44%), which gave a much higher peak viscosity and
was characterised by the absence of a well defined ‘break-
down’ zone. The breakdown viscosity is normally re-
Initial comparisons of wheat, maize, sorghum
garded as a measure of the disintegration of the starch
and millet granules as they are heated 12, and during the holding
RVA analysis. Fig. 1 shows a typical example of an phase, the disintegration of the starch granules reaches an
RVA profile for distilling wheat generated by the pro- equilibrium which influences the degree of breakdown 12.
gramme which was used during the study, and explains Breakdown viscosity is also referred to as ‘shear thinning’,
the terminology used. The data from the RVA analysis of which results from the alignment of molecules of soluble
wheat, maize, sorghum and millet are shown in Table II. starch in the direction of stirring 14. The degree of RVA
The red sorghum had the highest peak viscosity, while breakdown is related to the solubility of the starch, and
maize had a higher peak viscosity than wheat. The lower the more soluble the starch, the more it will thin on shear-
nitrogen wheat had a much higher peak (2.6 fold) and ing 14. Hence the continuing increase in viscosity after
final (5.3 fold) viscosity than the higher nitrogen wheat. gelatinisation observed with maize indicates that the
Whilst the peak temperatures were similar for all the ce- starch in maize is different from wheat starch, and may
reals at above 94°C, variations were observed in their show a greater degree of cross-linking. The sample of red
pasting temperatures (Table II). The final viscosity was sorghum had higher peak and final viscosities than the
high for all the cereals except for the higher nitrogen white sorghum, as well as a slightly higher total nitrogen
wheat sample that gave the lowest peak and final vis- content (1.53% vs 1.48%). Both of these were similar to
cosities. maize, and the profile was different to that observed for
The overall trends are seen more clearly by looking at wheat (Fig. 2).
the RVA profiles shown in Fig. 2. While the pasting pro- It is evident in Fig. 3, that the RVA pasting properties
files of both low (1.45%) and higher (1.93%) nitrogen of the extracted starch of these cereals gave similar pro-
were similar, differing only in magnitude, these were both files, conforming to the patterns generally associated with
very different from that generated by the maize sample starch 12,15. The main differences between them were in

Table II. RVA results for the different cereals studied (flour of whole cereal).
Pasting temperature Peak time Peak temperature Peak viscosity Final viscosity
Sample (°C) (min) (°C) cP cP
Low N wheat 60.25 5.00 94.80 440 636
High N wheat 66.10 4.67 94.40 169 120
Yellow maize 67.65 6.60 94.90 712 1390
White sorghum 76.65 6.93 95.00 495 1145
Red sorghum 72.60 5.93 95.00 795 1652
Millet 70.05 4.87 94.90 722 917

316 JOURNAL OF THE INSTITUTE OF BREWING

 Fig. 2. RVA results for wheat, maize, sorghum and millet (overview).

Fig. 3. RVA Pasting curves for starch extracted from wheat, maize, sorghum and millet.

terms of the magnitude of the viscosities. For some ce- of maize and sorghum (Fig. 2), suggesting that the ex-
reals, such as maize and sorghum, the profiles shown by tracted starches behaved in a different way to the starch in
the starches were different from those given by the origi- the whole cereals for reasons not fully understood. Again
nal cereals (Fig. 2). For all the cereals, the extracted millet starch gave a profile which was intermediate be-
starches not only showed well defined peak viscosity, they tween maize and wheat starch.
also gave similar patterns of “breakdown” and “setback”. The numerical data generated by the RVA instrument
This contrasts with the absence of the distinct “break- from the extracted starches is shown in Table III. Very
down” pattern observed previously with the whole cereals high values for the peak and final viscosities of these

VOL. 112, NO. 4, 2006 317

starches, compared with the original cereals, especially in residues was also measured. The results are shown in
the case of maize and sorghum were observed. This con- Table IV. The low nitrogen wheat, maize, sorghum and
firms that the extracted starches had different properties millet gave higher alcohol yields than the higher nitrogen
from the original cereals. It is also possible that these dif- wheat. As we have seen from Table II, these gave a higher
ferences may also be partially explained by differences in RVA peak and final viscosity reflecting the presence of
the composition of the starch of these cereal grains. The higher amounts of gelatinized substrate. This was in con-
lack of a well defined breakdown zone (Fig. 2) and higher trast to the higher nitrogen wheat, which gave a lower
peak and final viscosities of the extracted starches of alcohol yield, but also had lower peak and final viscosity
maize and sorghum (Fig. 3) suggest that maize and sor- resulting from reduced levels of gelatinized substrate.
ghum starches may have a greater tendency for retrograda- These results suggest a link between the RVA viscosity
tion, than wheat and millet starches. This property is nor- and alcohol production, whereby a higher RVA peak and
mally associated with amylose 14. final viscosity are associated with a higher alcohol yield.
Alcohol yield and residue viscosity (Ostwald vis- The alcohol yield of 472 LA/tonne (dwb) obtained for
cometer). The full implications of the information gener- the low nitrogen wheat was unusually high for wheat and
ated by the RVA analysis for these cereals is not yet fully reflects careful management of the application of nitrogen
understood and the data must be considered in the context on the growing site. The alcohol yield obtained from the
of alcohol yield and residue viscosity. The cereals were higher nitrogen wheat sample, which was from the same
analysed for alcohol yield and the residue viscosities were variety, was approximately 20 LA /tonne lower than for
measured on the distillation residues. The colour of the the low nitrogen wheat. However, its alcohol yield of 450

Table III. RVA data from starch extracted from maize, wheat, sorghum and millet.
Pasting temperature Peak time Peak temperature Peak viscosity Final viscosity
Starch source (°C) (min) (°C) cP cP
Maize 64.50 4.93 94.80 2361 2402
Wheat 75.00 5.33 94.90 1552 1546
Sorghum 67.65 5.33 94.85 2168 2265
Millet 70.05 5.27 94.95 1828 2009

Table IV. Total nitrogen, alcohol yield, colour and residue viscosity of wheat, maize, sorghum and millet samples.
TN (%) PSY (dry) PSY (as is) Residue colour Residue viscosity
Sample (dry) LA /tonne LA /tonne °EBC (mPas)
Wheat (Low N) (SWRI) 1.45 472 414 11.0 1.60
Wheat (High N) (SWRI) 1.93 451 401 13.5 1.74
Maize (Yellow) (SWRI) 1.44 475 417 3.5 1.15
Sorghum (White) (HWU) 1.48 472 415 3.5 1.15
Sorghum (White) (Nigeria) 1.70 480 434 — 1.09
Sorghum (Red) (SWRI) 1.53 475 415 3.5 1.16
Sorghum (Red) (Nigeria) 1.53 471 421 — 1.08
Millet (Nigeria) 1.60 462 422 3.5 1.16

Table V. Total nitrogen, alcohol yield, colour and residue viscosity of additional maize samples.
TN (%) PSY (dry) PSY (as is) Residue colour Residue viscosity
Sample (dry) LA /tonne LA /tonne °EBC (mPas)
Maize (Yellow) 1.39 475 416 3.5 1.15
Maize (Yellow) 1.48 477 419 3.5 1.16
Maize (White) 1.41 468 418 3.5 1.16

Table VI. Total nitrogen, alcohol yield, colour and residue viscosity additional wheat samples.
TN (%) PSY (dry) PSY (as is) Residue colour Residue viscosity
Sample (dry) LA /tonne LA /tonne °EBC (mPas)
Low N Wheat (Variety A) 1.59 457 401 13.0 1.72
High N Wheat (Variety A) 1.84 450 399 14.0 1.83
Low N Wheat (Variety B) 1.62 464 408 12.5 1.63
High N Wheat (Variety B) 1.85 450 402 13.5 1.82
Low N Wheat (Variety C) 1.51 462 406 11.5 1.63
High N Wheat (Variety C) 1.82 446 395 13.5 1.78
Low N Wheat (Variety D) 1.62 455 400 13.5 1.71
High N Wheat (Variety D) 1.83 445 395 14.0 1.79
Low N Wheat (Variety E) 1.57 453 400 12.5 1.72
High N Wheat (Variety E) 1.82 440 395 13.0 1.77
Low N Wheat (Variety F) 1.67 452 400 13.0 1.68
High N Wheat (Variety F) 2.02 441 395 14.0 1.85

318 JOURNAL OF THE INSTITUTE OF BREWING

LA /tonne was still much higher than expected for a dis- ences observed in the residue viscosities of these cereals.
tilling wheat sample with a high nitrogen level approach- However, in comparison with other cereals such as maize,
ing 2%. This wheat variety is therefore capable of produc- sorghum or millet, wheat contains a much higher percent-
ing a high alcohol yield at different nitrogen levels. age of pentosans 13,19. These are known to cause processing
The results obtained for the low and high nitrogen problems, so it is highly likely that these will be major
wheat samples are consistent with the well established contributing factor to any processing problems encoun-
inverse relationship between nitrogen content and alcohol tered when processing wheat.
yield of wheat 9. Table V shows the distilling potential The results in Tables IV and VI show that both the low
analyses for additional samples of maize, while Table VI and high nitrogen wheat samples produced residues with
shows the distilling potential analyses for additional sam- higher viscosities than maize, sorghum or millet. The lower
ples of wheat, and together with the information on these nitrogen wheat sample however, produced a slightly lower
cereals from Tables V and IV suggest that in contrast to residue viscosity than the higher nitrogen wheat sample.
wheat, high alcohol yields were obtained from maize ir- The wheat samples also produced residues with darker
respective of nitrogen content. Agu and Palmer 2 have colour than maize or sorghum or millet. Again, the addi-
shown that a 0.2% (w/v) difference in nitrogen content tional wheat samples (low and higher nitrogen) analysed
has a significant effect on extract recovery in malted bar- further confirmed that low nitrogen wheat is associated
ley, and the results in Tables IV and VI confirm that dif- with higher peak and final RVA viscosities (Table VII).
ferences in nitrogen content will affect the alcohol yield
from wheat. It is also clear from these results that the al- Alcohol yield of extracted starch
cohol yields of maize and sorghum are not influenced by The extracted starch of maize and sorghum gave very
nitrogen levels to the same extent as wheat. The reason for high alcohol yields (Table VIII), in relation to the levels of
this apparent trend is not immediately clear, but further alcohol that were obtained when whole grains of maize
highlights significant physiological differences between and sorghum were processed (Table IV). In contrast, the
the cereals studied. starches of wheat and millet gave lower levels of alcohol
The residue viscosity data shown in Tables IV and VI yield but their whole grains gave higher levels of alcohol
are based on the analysis of the residues from the alcohol yield. While this may in part be attributed to the extrac-
distillations. These were analysed using an Ostwald vis- tion of the ‘pure’ starch, resulting in a lower proportion of
cometer, and are thus completely distinct from the RVA non-fermentable material, this further highlights that our
viscosity data discussed earlier. In contrast to the RVA understanding of the physiological differences between
analysis of the original cereals, the residue viscosity rep- the cereals and their starches is not complete. The high
resents the composition of the material remaining after alcohol yields of the starches extracted from maize and
fermentation of each of the cereal substrates was com- sorghum (over 500 LA/tonne dwb), in contrast to those
plete, and thus represents unfermentable material, such as obtained from wheat and millet (400 LA/tonne dwb),
pentosans and ␤-glucans, which survive the process 18. which are shown in Table VIII, may also suggest that the
High viscosity residues can cause severe distillery pro- starches extracted from maize and sorghum have greater
cessing problems. Since the ␤-glucan content of wheat, ability to withstand the effects of heating and shear
maize, sorghum and millet are known to be similar 13,19, it stresses associated with processing, than that from wheat
is unlikely that ␤-glucan is the major cause of the differ- and millet.

Table VII. RVA data for six pairs of low and higher nitrogen wheat varieties.
Pasting temperature Peak time Peak temperature Peak viscosity Final viscosity
Sample (°C) (min) (°C) cP cP
LN wheat (Variety A) 66.15 4.93 94.75 385 462
HN wheat (Variety A) 67.00 4.67 94.40 100 69
LN wheat (Variety B) 74.25 5.13 94.90 566 873
HN wheat (Variety B) 66.80 5.93 93.60 101 70
LN wheat (Variety C) 72.60 5.00 94.80 430 727
HN wheat (Variety C) 76.60 5.00 94.85 304 458
LN wheat (Variety D) 69.35 4.93 94.75 327 447
HN wheat (Variety D) 72.50 5.00 94.80 190 243
LN wheat (Variety E) 70.10 4.87 94.85 296 308
HN wheat (Variety E) 75.80 4.80 95.00 101 78
LN wheat (Variety F) 72.50 4.93 94.80 422 690
HN wheat (Variety F) 72.60 4.47 91.95 68 40

Table VIII. Alcohol yield, colour and residue viscosity of starches extracted from wheat, maize, sorghum and millet.
Moisture PSY (dry) PSY (as is) Residue colour Residue viscosity
Starch source (%) LA /tonne LA /tonne °EBC (mPas)
Wheat (Low N) 8.1 404 371 3.5 1.23
Maize 5.1 516 489 2.0 1.09
Sorghum (Red) 5.3 509 483 2.0 1.07
Millet 5.2 397 376 5.5 1.14

VOL. 112, NO. 4, 2006 319

 This is an important factor for many processes and the the yeast activity associated with these fermentations. On
time and temperature at which starch starts to paste (past- this basis one might speculate that in these fermentations,
ing temperature) is related to the conditions that are nec- nearly all of the free amino nitrogen required to support
essary to cook a given material 15. The data supports the yeast growth during fermentation is derived from hydroly-
view that maize and sorghum will therefore require inten- sis of the soluble nitrogen material from the malted barley
sive cooking to produce high alcohol yields compared used during the conversion (saccharification) process,
with wheat, and will have implications for both energy rather than from the main cereal substrate. In the Scotch
efficiency and the thermal stability of other components in grain whisky process, wheat is cooked at a high tempera-
the substrate, such as proteins and amino acids. More in- ture prior to mashing (conversion) and this cooking pro-
tensive cooking will also facilitate the breaking of cross- cess will denature the proteins in the wheat, thus inactivat-
linking bonds in resistant starch, and in turn, is likely to ing any potential proteolytic enzymes in the cereal which
allow the release of more sugars from the dextrins, and would contribute to the availably of free amino nitrogen 2.
hence lead to a higher potential for alcohol production, This observation is extremely important since it has se-
but may also result in increased losses due to Maillard rious implications for distillery efficiency. This empha-
(browning) reactions. sizes the need to ensure adequate malt inclusion during
In contrast to maize and sorghum, the starches associ- grain whisky production 8, particularly when low nitrogen
ated with wheat and millet, appear to be more readily ac- substrates are used, since the malt inclusion rate is a “criti-
cessible (see Fig. 2) and ideally may not require the inten- cal factor” or “process limiting factor” in achieving high
sive cooking process used in this study to give high alco- alcohol yield during grain whisky production.
hol yields. This work therefore suggests that wheat and
Further studies on the relationships between
millet might require a less intensive, processing (cooking)
regime compared with maize and sorghum. This is sup- RVA viscosity and alcohol production from
ported by the slightly higher colour value for millet starch high and low nitrogen wheat samples
which gave a darker residue than wheat, despite the resi- The hypothesis that low nitrogen wheat is generally as-
due from the whole grain having a much lower colour, sociated with higher RVA peak and final viscosities, re-
comparable with maize and sorghum (see Table VIII). sulting from larger amounts of gelatinized substrates, thus
It is worth noting that for the extracted starches derived producing higher alcohol yields, than the higher nitrogen
from all the cereals studied (wheat, maize, sorghum and wheat was investigated further. The hypothesis was tested
millet), the colours of the residues obtained from the pro- by further looking at the sets of low and higher nitrogen
cessing were much lower, in comparison to when whole wheat samples (Varieties A–F), representing nationally
wheat was processed. The darker colour associated with listed (NL) wheat varieties (Tables VI and VII). The RVA
the wheat residues may be partly due to the presence of pasting data in Table VII together with the data in Table
darker pigmentation, but could also result from an in- VI show clearly that in every case, the lower nitrogen
creased susceptibility to the effects of Maillard reactions. wheat samples have higher peak and final viscosities when
This observation is important because it suggests that compared with their respective higher nitrogen counter-
some cereals are more suited to high temperature /pressure parts. The results in Tables VI and VII are interesting be-
cooking than others. cause they confirm that a definite pattern is emerging,
which suggests that the hypothesis might be correct. The
The role of nitrogen during the processing alcohol yield results together with the Ostwald residue
of wheat, maize, sorghum and millet viscosity and the RVA data, confirmed that the lower ni-
The differences between the processed cereals and the trogen wheat samples which produced more alcohol than
extracted starches, are difficult to explain fully in the con- the higher nitrogen wheat samples were also associated
text of these experiments, but may be related to the prop- with higher viscosity levels, and that conversely, the lower
erties of the nitrogen components. In the extracted nitrogen wheat samples produced residues that have lower
starches, the removal of the protein fraction, might ac- viscosity than the higher nitrogen samples.
count for some of the differences in colour and processing To further test the validity of the hypothesis that the
characteristics which were observed, since this, provides RVA pasting properties are related to the alcohol yield,
essential substrates for Maillard reactions. Maillard or correlation analysis between the pasting viscosities (RVA
browning reactions take place as a result of thermal degra- peak and final viscosities) and alcohol yield was carried
dation of sugars and proteins and will have an adverse out. The results obtained are shown in Fig. 4 for alcohol
effect on alcohol yield, since some of these components yield versus RVA peak viscosity and in Fig. 5 for alcohol
will no longer be available for yeast to utilise during fer- yield against RVA final viscosity. Also, the relationship
mentation. between peak viscosity and final viscosity was examined
Additionally, the removal of the grain nitrogen compo- (Fig. 6). The correlations between peak viscosity (Fig. 4),
nents from the extracted starch, coupled with the fact that and final viscosity versus alcohol yield (Fig. 5), (80% and
nitrogen will be denatured during the cooking process 2 75%) were significant and give strong support to the hy-
would also potentially reduce the free amino nitrogen pothesis that there is an important relationship between
available from the starches of these cereals. However, these parameters. A much stronger correlation of 98% for
there were no fermentation problems during the labora- the relationship between RVA peak and final viscosities
tory processing of these starches. This suggests that under (Fig. 6) also provides additional support to the hypothesis,
these conditions, the nitrogen components of the cereal since this shows that the two viscosity parameters are
starches, on their own, were not important in influencing closely interlinked. Since both the RVA peak and final

320 JOURNAL OF THE INSTITUTE OF BREWING

 Fig. 4. Correlation between wheat alcohol yield and RVA peak viscosity for six
pairs of low and high nitrogen wheat samples (A–F).

Fig. 5. Correlation between wheat alcohol yield and RVA final viscosity for six
pairs of low and high nitrogen wheat samples (A–F).

viscosities are correlated with alcohol yield, they will be that for wheat, which was different from both maize and
linked to the properties of the substrates which are avail- sorghum. The resemblances between the profiles of maize,
able for alcohol production, and together, they will have a sorghum and millet indicate that sorghum and millet
strong impact on the distilling performance of the wheat would perform in similar ways to maize during processing
samples. in the distillery, and suggest that no unusual problems
would be expected with front end processing of these ce-
CONCLUSIONS reals (i.e. with cooking and conversion (saccharification)),
other than those associated with the mechanical handling
These studies have given us a better understanding of of small grains. The RVA properties for the extracted
how methods such as RVA can be useful in assessing ce- starches showed that the actual starches from each cereal
reals for distilling. The rheological (RVA) properties of gave similar profiles, but were different in the magnitude
wheat, maize, sorghum and millet, showed that the RVA of the viscosities resulting from each cereal starch. This
pasting curve for millet was similar in some respects to suggests that some other factors within the cereal grain

VOL. 112, NO. 4, 2006 321

 Fig. 6. Correlation between RVA peak and RVA final viscosity for six pairs of
low and high nitrogen wheat samples (A–F).

have a mediating effect on the RVA properties which were market conditions would be a distinct advantage. In the
observed for the whole cereals. current economic climate, Scotch whisky grain distillers
In the context of the production of Scotch whisky grain would prefer to use wheat, since it is easily and econom-
spirit, the cooking process which is currently used for ically obtained from the local market. However, the pro-
wheat was originally developed to handle maize, and on duction of ethanol is a worldwide activity and this paper
the basis of this work, should be able to efficiently pro- has demonstrated some of the advantages that might be
cess all of the cereals studied. The residue (Ostwald) vis- obtained by processing locally produced cereals available
cosities of sorghum and millet were similar to maize, indi- in other parts of the world, which might result in reduced
cating that distillery co-products deriving from these ce- transport costs and promote other local benefits for these
reals would perform in a similar manner to maize. The areas.
importance and complexity of the role that nitrogen plays
in determining both the distilling quality and processabil- ACKNOWLEDGEMENTS
ity of cereal grains, especially with wheat, was confirmed
in this study. Low nitrogen wheat gave higher RVA peak The authors would like to acknowledge the contribution of
and final viscosities and produced higher alcohol yield, Elizabeth Glasgow. We would also like to thank Dr Gordon
compared with high nitrogen wheat, and underlines the Steele, SWRI Director of Research, and the Board of the Scotch
Whisky Research Institute for permission to publish this paper.
importance of the nitrogen content in defining the quality
of distilling wheat. With careful management of nitrogen
application in the field during growing it is now possible REFERENCES
to produce locally grown wheat with low nitrogen levels 1. Agu, R. C. and Palmer, G. H., A reassessment of sorghum for
which can produce similar levels of alcohol as maize, as lager-beer brewing. Bioresource Technol., 1998, 66, 253–261.
well as an acceptable harvest yield. In the UK, this may 2. Agu, R. C. and Palmer, G. H., Some relationships between the
help to maintain the competitive edge that wheat may have protein nitrogen of barley and the production of amylolytic en-
over the other cereals studied. zymes during mashing. J. Inst. Brew., 1998, 104(5), 273–276.
The influence that the level of nitrogen has on alcohol 3. Agu, R. C. and Palmer, G. H., Pattern of nitrogen distribution in
barley grains grown in the field. J. Inst. Brew., 2003, 109(2),
yield production with regard to wheat did not seem to
110–113.
apply to the same extent to the other cereals (maize, sor- 4. Agu, R. C., Swanston, J. S., Bringhurst, T. A., Brosnan, J. M.,
ghum, and millet) because these cereals produced high Jack, F. R. and Smith, P. L., The influence of nitrogen uptake on
alcohol yields regardless of nitrogen content. Further the quality of distilling wheat cultivars. In: Proceedings of the
work is necessary to establish this unequivocally, but if 2005 World Wide Distilled Spirits Conference (Edinburgh). J.
this is the case, this highlights important physico-chemi- H. Bryce, J. R. Piggott, and G. G. Stewart, Eds., Nottingham
cal differences between wheat and these other cereals. University Press: Nottingham, UK, 2007, (in Preparation).
5. Ahluwalia, B. and Fry, S. C., Barley endosperm cell walls con-
The problems regarding the impact of high residue vis- tain a feruloylated arabinoxylan and a non-feruloylated ␤-glu-
cosity on downstream distillery processing still have to be can. J. Cereal Sci., 1986, 4, 287–295.
fully resolved and could still prove to be a disadvantage 6. Analysis Committee of the Institute of Brewing. Method 1.8.
for wheat. The ability of a distillery or ethanol producer to Total Nitrogen in Barley. Recommended Methods of the Insti-
process a range of cereal types depending on favourable tute of Brewing. 1997. Institute of Brewing: London.

322 JOURNAL OF THE INSTITUTE OF BREWING

 7. Bathgate, G. N. and Palmer, G. H., A reassessment of the chemi- 17. Orth, R. A. and Shellenberger, J. A., Origin, production and
cal structure of barley and wheat starch granules. Die Starke, utilisation of wheat. In: Wheat: Chemistry and Technology, 3rd
1972, 24, 336–341. edn. American Association of Cereal Chemists: St Paul, MN,
8. Bringhurst, T. A., Fotheringham, A. L. and Brosnan, J. M., 1988.
Grain whisky: raw materials and processing. In: Whisky: Tech- 18. Palmer, G. H., Adjuncts in brewing and distilling. In: The Pro-
nology, Production and Marketing (Handbook of Alcoholic ceedings of the Second Aviemore Conference on Malting Brew-
Beverages Series). I. Russell, C. Bamforth and G. G. Stewart, ing and Distilling. I. Campbell and F. G. Priest, Eds., Institute of
Eds., Academic Press: London, 2003, pp. 77–116. Brewing: UK, 1986, pp. 24–45.
9. Brosnan, J. M., Makari, S., Paterson, L. and Cochrane, M. P., 19. Palmer, G. H., Properties of ␤-glucan extracts of endosperm cell
What makes a good distilling wheat. In: Proceedings of the walls. In: Cereal Science and Technology. G. H. Palmer, Ed.,
Fifth Aveimore Conference on Malting, Brewing and Distilling. Aberdeen University Press: Aberdeen, 1989, pp. 80–86.
I. Campbell, Ed., Institute of Brewing: London, 1999, pp. 225– 20. Palmer, G. H., Gernah, D. I., McKernan, G. and Nimmo, D. H.,
228. Influence of enzyme distribution on endosperm breakdown
10. Brown, J. H., Assessment of wheat for grain distilling. In: Pro- (modification) during malting. J. Amer. Soc. Brew. Chem., 1985,
ceedings of the Third Aviemore Conference on Malting Brewing 42, 17–28.
and Distilling. I. Campbell, Ed., Institute of Brewing: London, 21. Preece, I. A. and MacDougall, M., Enzymic degradation of
1990, pp. 34–47. cereal hemicelluloses II. Pattern of pentosan degradation. J.
11. Coote, N and Kirsop, B. H., A haze constituting largely of pen- Inst. Brew., 1958 58, 353–362.
tosan. J. Inst. Brew., 1976, 82, 34. 22. Preece, I. A. and Garg, N. K. J., Enzymic degradation of cereal
12. Dengate, H. N., Swelling, pasting and gelling of wheat starch, hemicelluloses IV. Quantitative aspects of ␤-glucan degradation.
In: Advances in Cereal Science and Technology. Vol. 6. Y. J. Inst. Brew., 1961, 67, 267–273.
Pomeranz, Ed., Am. Ass. Cereal Chem.: St Paul, MN, 1984, pp. 23. Swanston, J. S., Newton, A. C., Brosnan, J. M., Fotheringham,
49–82. A., and Glasgow, E., Determining the spirit yield of wheat varie-
13. Haln, R. R., Sorghum as a brewing adjunct. Brewer’s Dig., ties and variety mixtures. J. Cereal Sci., 2005, 42(1), 127–134.
1966, 49, 70–74. 24. Walker E. W., Grain spirit – which cereal. In: Proceedings of the
14. Hoseney, R. C., Principles of Cereal Science and Technology, Second Aviemore Conference on Malting, Brewing and Distill-
2nd edition, American Association of Cereal Chemists: St Paul, ing. I. Campbell and F. G. Priest, Eds. Institute of Brewing: UK,
MN, 1998, pp. 29–64. 1986, pp. 375–380.
15. Newport Scientific, Application Manual for the Rapid Visco 25. Watson, S. A., Corn and sorghum starch: production. In Starch:
Analyser. Newport Scientific Pty. Ltd Warriewood, Australia, Chemistry and Technology, 2nd edn., R. L. Whistler, J. N. Be-
1998, pp. 19–26. Miller, and E. F. Paschall, Eds., Academic Press: London, 1984,
16. Ogu, E. O, Odibo, F. J. C., Agu, R. C. and Palmer, G. H., Malt- pp. 432–434.
ing studies of some selected brewing sorghum varieties. Tech.
Quart. Master Brew. Assoc. Am., 2004, 41(4), 386–389. (Manuscript accepted for publication December 2006)

VOL. 112, NO. 4, 2006 323