Simultaneous Analysis of Fat-Soluble Vitamins A, D, and E in Food Using ACQUITY Arc Two-Dimensional Liquid Chromatography

1

INTRODUCTION
Vitamins A, D, and E are fat-soluble vitamins essential for the body to 
maintain normal metabolism and functions. Vitamin A, also known as retinol, 
plays an important role in promoting body growth, maintaining the integrity 
of the epidermis, etc. Vitamin D includes two major forms, i.e. vitamin D

2 

(ergocalciferol) and vitamin D

3 (cholecalciferol), which promote calcium and 

phosphorus metabolism and bone formation in mammals. Vitamin E includes 
tocopherols and tocotrienols. There are 8 active forms of vitamin E due to the 
variation of methyl substitution on the parent tocopherol and tocotrienol ring, 
including α-, β-, γ-, and δ-tocopherols. Among them, α-tocopherol is usually 
singled out in food science as it has the highest activity and antioxidative 
and anti-aging properties.1 Infant formula and adult nutritional products 
and animal feeds are two important forms of fat-soluble vitamin fortified 
products. In actual samples, vitamin A and vitamin E can be quantified 
directly because their content levels are relative high and matrix interference 
is negligible; however, vitamin D is generally added in a small amount, 
has relatively low UV absorption, and suffers severe interference from the 
matrix, so liquid chromatography-mass spectrometry or semi-preparative 
normal-phase clean-up is specified for separation and assay of vitamin D in 
the current national standards.2 However, the liquid chromatography-mass 
spectrometry is relatively expensive; while, the normal-phase purification is 
susceptible to mobile phase conditions and its process is relatively tedious 
and time-consuming.

This application is based on Waters' existing solutions of two-dimensional 
ultra-high performance liquid chromatography (UPLC 2D).3 Using ACQUITY 
Arc (UHPLC) 2D technology under reversed-phase conditions, the separation 
and assay of vitamin A, α, β, γ, δ-tocopherol, vitamin D

2, and vitamin D3 can 

be completed simultaneously with one sample injection, and the entire assay 
only takes 15 min. Vitamin A and α, β, γ, δ-tocopherol are separated and 
quantified on the 1st dimension column, while vitamin D is heart-cut into the 
trap column after preliminary separation and clean-up on the 1st dimension 
column, followed by transfer into the 2nd dimension column for further 
separation and assay to achieve baseline separation of vitamin D

2 and D3.

Through the matrix spiking tests done on actual samples provided by relevant 
companies and the analysis of QC samples, the results showed good linearity, 
excellent correlation coefficient and excellent repeatability. For five replicated 
injections, the RSD of retention time was <0.5%, the RSD of peak area was 
<2%. The limit of detection of vitamin D (D

2 and D3) was as low as 0.5 ug/kg.

Simultaneous Analysis of Fat-Soluble Vitamins A, D, and E in Food Using 

ACQUITY Arc Two-Dimensional Liquid Chromatography
Qin Yuhong, Huang Defeng, and Chen Jing 
Waters Technology (Shanghai) Co., Ltd.

WATERS SOLUTIONS
ACQUITY™ Arc™ 2D technology

Empower™ 3 Chromatography Data Software

XSelect™ HSS C

18SB Column  

(3.0 × 150 mm, 3.5 µm)

XBridge™ BEH C

18 Direct Connect HP Trap 

Column (2.1 mm × 30 mm, 10 µm)

KEYWORDS
Two-dimensional liquid chromatography, 
heart-cutting, vitamins ADE

APPLICATION BENEFITS

■

The two-dimensional column switching-
ultraviolet detection method can separate 
vitamins A, D, and α/β/γ/δ-tocopherol in a 
single sample injection, greatly increasing 
the efficiency of analysis;

■

With the heart-cutting technique, the 
vitamin D is cleaned-up on the first 
dimension column, and then separated 
into vitamin D

2 and D3 peaks. This 

approach helps to eliminate co-elution, 
which reduces the interference from 
complex matrix for the vitamin D

2 and  

D

3 determination.

2

Simultaneous Analysis of Fat-Soluble Vitamins ADE in Food Using ACQUITY Arc Two-Dimensional Liquid Chromatography

EXPERIMENTAL

SAMPLE PREPARATION
In this application, the test samples were actual samples of milk powder, and 
sample supply and preparation were provided by our collaborating laboratory. 
The sample preparation steps were performed using the “GB 5009.82-2016 
Assay of Vitamins A, D, and E in Food”2 as a guide. The specific process is  
as follows:

EXPERIMENT
LC conditions
Instrument system:  Waters ACQUITY Arc 2D 

system:  

QSM1 (quaternary pump  

1) + QSM2 (quaternary    

pump 2) + FTN injector + 

Column Manager  

(CM-A) equipped with 

2-position 6-port valve +   

2998 PDA Detector

Mobile phase:  

A: water; B: acetonitrile;  

C: methanol

1st dimension 

analytical column:   Poroshell 120 PFP  

(4.6 × 100 mm, 2.7 µm)

2nd dimension 

analytical column:   XSelect HSS C

18SB  

(3.0 × 150 mm, 3.5 µm)

Trap column:  

XBridge BEH C

18 Direct   

Connect HP Trap 

Column (2.1 mm ×  

30 mm, 10 µm)

Column temp.:  

1st dimension column  

35 °C, 2nd dimension

column 40 °C

Detection  

wavelength:  

1st dimension VA:  

325 nm (0–5.5 min); 

VD: 264 nm (5.5–6.8 min); 

VE: 294 nm (6.8–10 min); 

 2nd dimension VD2  

and VD3:   

264 nm (10–15 min)

Injection volume:     10 µL
Gradient method:  

1st dimension 

Time Flow rate 

(mL/min)

A (%)

B (%)

C (%)

Curve

0

1

20

0

80

6

9

1

0

0

100

6

10

1

0

0

100

6

11

1

20

0

80

6

15

1

20

0

80

6

2nd dimension

Time Flow rate 

(mL/min)

A (%)

B (%)

C (%)

Curve

0

0.5

0

100

0

6

4

0.5

0

100

0

6

10

0.5

0

20

80

6

10.5

0.5

0

100

0

6

15

0.5

0

100

0

6

Valve switching time:

0 min

5.74 

min

6.04 

min

8.5  

min

15  

min

Right valve 

position

2

1

2

2

2

Left valve 

position

1

1

1

2

1

Figure 1. Sample preparation process.

Saponification

n

  An amount of 5–10 g of homogenized solid samples (or 50 g of liquid samples) was weighed 

and put into a 150 mL flat-bottom flask, for solid samples, 20 mL of warm water was added 

into the flask, mixed well;

n

  For samples containing starch: 0.5–1 g of amylase was added and the flask was placed in a 

thermostat water bath at 60 °C, shaken for 30 min in the dark before the flask being removed 

from the water bath;

n

  1.0 g of ascorbic acid and 0.1 g of BHT were added into the above treated solution and mixed 

well. Then, 30 mL of absolute ethanol and 10 mL of potassium hydroxide were added to 

saponify the samples in an 80 °C water bath for 30 min, it was cooled to room temperature 

after saponification.

Extraction

n

  The saponified solution was transferred into a 250 mL separatory funnel using 30 mL of water, 

50 mL of mixture of petroleum ether-diethyl ether was added, and the solution was extracted 

by shaking for 5 min;

n

  The lower layer solution was transferred to another 250 mL separatory funnel,  50 mL of 

extracting ether mixture solvent was added and the solution was extracted again;

n

  The extracts obtained from the above two steps was combined.

Note: If only vitamin A and α-tocopherol are to be assayed, petroleum ether can be used as the extractant.

Concentration

n

  The washed ether layer was dried with anhydrous sodium sulfate (about 3 g), then filtered into 

a 250 mL rotary evaporation flask or a nitrogen evaporator tube, the separatory funnel and 

anhydrous sodium sulfate was rinshed twice with 15 mL of petroleum ether, and the rinsing 

fluid was pooled into the evaporation flask;

n

  The samples were concentrated by reduced-pressure distillation in a 40 °C water bath or  

by nitrogen evaporation to about 2 mL, and the samples were dried under a gentle blow  

of nitrogen;

n

  The dried residue was transferred to a 5 ml volumetric flask with methanol as reconsitution 

and transfer solvent, where multiple rinse may be needed for the transfer, then the solution 

was diluted to volume with methanol. The solution was filtered through a 0.22 µm organic filter 

membrane and the samples were injected for testing.

Note: When the vitamin D content in the samples is low and an increased injection volume is required, 

  the solvent strength can be appropriately diluted to reduce the solvent effect (such as using the    

  initial mobile phase to transfer the residue and make up the volume).

Washing

n

  The ether layer was washed with about 100 mL of water and the washing step was repeated 

about 3 times until the ether layer became neutral (the pH value of the lower layer solution 

can be tested with a pH test paper);

n

  The lower aqueous phase was removed.

[ APPLICATION NOTE ]

3

Simultaneous Analysis of Fat-Soluble Vitamins A, D, and E in Food Using ACQUITY Arc

RESULTS AND DISCUSSION

METHOD DEVELOPMENT
The goal of the 1st dimension separation is to ensure that the vitamin A and the four vitamers of vitamin E 
are separated and quantified without any interference  from matrix. In addition, the vitamin D peak needs to 
be completely separated from the vitamin A and vitamin E with peak width as narrow as possible, so that it 
can be completely heart-cut and trapped without any extra amount of matrix background being also cut and 
transferred to the 2nd dimension, which could affect the further separation of vitamin D on the 2nd dimension 
LC. The advantage of this method’s connection and configuration is that by using only one detector, one can 
flexibly adjust the heart-cutting time window based on the retention time of vitamin D in different samples on 
the 1st dimension LC, so as to avoid any loss of the target substance during the heart-cut process.

The goal of the 2nd dimension LC is to not only separate the vitamin D peak from matrix interferences, but 
also separate the vitamin D

2 and the vitamin D3 at baseline resolution in order to achieve qualification and 

accurate quantitation purposes. Also, the vitamin D

2 and D3 peaks should be as far away as possible from 

the pressure disturbing peaks that is caused by valve switching, as well as other major interfering peaks to 
ensure the accuracy of quantitation. 

From the separation chromatogram of the standards in Figure 3 and the chromatogram of the actual spiked 
sample in Figure 4, it can be seen that good separation of vitamin A, four vitamin E vitamers, and impurities 
can be achieved on the 1st dimension column, and good separation of two vitamin Ds and impurities can be 
achieved on the 2nd dimension column.

INSTRUMENT CONNECTION AND CONFIGURATION
Vitamin A and vitamin E were measured on the 1st dimension LC, and vitamin D was measured on the 2nd 
dimension LC after heart-cutting. The key to the method is that the 1st dimension LC should be able to 
separate the vitamin A and vitamin E from the matrix inferference peaks, and it also should have excellent 
retention time repeatability to ensure the accuracy of heart-cutting with minimal cutting window. In 
order to more intuitively determine whether the cutting is accurate, and to allow a timely and convenient 
adjustment in the retention time when it is needed, two 2-position 6-port valves were used to connect 
a detector, so that both the 1st dimension and the 2nd dimension peaks were collected on the same 
chromatogram. The method was as follows: Vitamins A, D and E were passed through the 1st dimension 
column and the detector, and all peaks showed up. At the same time, vitamin D was subjected to heart-
cutting so that it was separated again on the 2nd dimension column and the peaks were detected again.

The specific configuration used in this application is shown in Figure 2.

Figure 2. Connection and configuration of the heart-cutting system in the ACQUITY Arc 2D two-dimensional system.

[ APPLICATION NOTE ]

4

Simultaneous Analysis of Fat-Soluble Vitamins A, D, and E in Food Using ACQUITY Arc

METHOD PERFORMANCE

Repeatability
The results of repeated injections (n = 5) of the standards at a 5 µg/L concentration are shown in Figure 5.  
The RSD for the retention time of vitamin A and four vitamin E vitamers on the 1st dimension LC and the 
RSD for the retention time of vitamin D

2 and vitamin D3 on the 2

nd

 dimension LC was all <0.5%, and the 

RSD for their peak area was <2%, which ensures the accuracy of the heart-cutting method.

Figure 3. Chromatogram for the standard of vitamin A, four isoforms of vitamin E, vitamin D2, and vitamin D3.

Figure 4. Chromatogram of the actual milk powder spiked sample.

Figure 5. Chromatogram of repeated injections (n = 5) of the standard at a 5 µg/L concentration.

[ APPLICATION NOTE ]

5

Simultaneous Analysis of Fat-Soluble Vitamins A, D, and E in Food Using ACQUITY Arc

Linearity
In this application, the linearity of vitamin A, four vitamin E vitamers, vitamin D

2, and vitamin D3 was 

examined on the basis of the requirements specified in the national standards. The results are shown 
in Figure 6. Different vitamins all maintained good linear relationships in different linear ranges, and the 
coefficient of determination R2 was greater than 0.998.

Figure 6. Standard curves in different linear ranges: Vitamin A (0.02–10 mg/L), 

four vitamers of vitamin E (2–60 mg/L), vitamin D

2 , 

and vitamin D

3 (0.001–1 mg/L); the injection volume was 250 µL.

[ APPLICATION NOTE ]

6

Simultaneous Analysis of Fat-Soluble Vitamins A, D, and E in Food Using ACQUITY Arc

LIMIT OF DETECTION (LOD) AND LIMIT OF QUANTITATION (LOQ)
In actual samples, the contents of vitamin A and vitamin E are generally high and can be accurately 
quantified, while the content of vitamin D is generally low and is prone to matrix interference, and 
hence the quantitation of vitamin D is relatively difficult. Therefore, the LOD of vitamin D is an important 
parameter in the evaluation of this instrument method. Detection sensitivity was investigated in this 
work. When the injection volume was 250 µL, the measured S/N ratio of vitamin D

2 and vitamin D3 was 

11 at an injection concentration of 0.5 µg/L, as shown in Figure 7. The refore, 0.5 µg/L was used as the 
LOQ of vitamin D. The LOD of vitamin D was calculated at 0.17 µg/L using a S/N of 3. This LOQ meets 
the assay requirement for the lowest concentration point of 50.0 µg/L on the standard curve in liquid 
chromatography methods specified in national standards. The LOD and LOQ results are shown in Table 1. 
Therefore, if the content of vitamin D in the sample is relatively high, an injection volume of 10 µL can be 
used to meet the requirements described in national standards.

ASSAY OF ACTUAL SAMPLES AND EVALUATION OF MATRIX EFFECT
Using the method established in this application, milk powder samples provided by our clients were 
analyzed. Sample preparation was performed by the client according to the preparation method 
described in this document. The results show that the vitamins added to this batch of samples are  
mostly vitamin D

3 and α-tocopherol (Table 2).

The magnitude of the matrix effect has a direct impact on the quantitative results. This application also 
evaluated the matrix effect at the same time. The above samples provided by the client were also spiked 
and the test results were used as the basis for evaluating the matrix effect. The test results of the actual 
samples and the evaluation of the matrix effect are shown in Table 2.

Table 1. LOD and LOQ

Figure 7. LOQ of vitamin D

2 and D3 (0.5 µg/L).

LOD 

(µg/L)

Milk powder sample 

(µg/100 g)

LOQ 

(µg/L)

Milk powder sample 

(µg/100 g)

Vitamin D

2

0.17

0.017

0.50

0.050

Vitamin D

3

0.17

0.017

0.50

0.050

Note: The sample size was 10 g.

[ APPLICATION NOTE ]

Table 2. Test results of actual samples and the matrix effect

Note: ND means not detected.

Sample 1

Sample 2

Sample 3

Sample 4

Background 

concentration 

(mg/kg)

Spiked  

concentration 

(mg/kg)

Matrix effect  

(%)

Background 

concentration 

(mg/kg)

Spiked  

concentration 

(mg/kg)

Matrix effect  

(%)

Background 

concentration 

(mg/kg)

Spiked  

concentration 

(mg/kg)

Matrix effect  

(%)

Background 

concentration 

(mg/kg)

Spiked  

concentration 

(mg/kg)

Matrix effect  

(%)

Vitamin A

7.3

1.0

-18

8.2

1.0

-14

3.3

1.0

-5.0

9.5

1.0

-8.0

δ-tocopherol 

ND

1.0

-17

5.6

1.0

-10

ND

1.0

-13

ND

1.0

-7.0

β-tocopherol

ND

1.0

-15

ND

1.0

-12

ND

1.0

-11

ND

1.0

-9.0

γ-tocopherol

ND

1.0

-10

3.2

1.0

-20

2.1

1.0

-18

ND

1.0

-7.0

α

-tocopherol

29

1.0

-12

50

1.0

-19

49

1.0

-7.0

54

1.0

-8.0

Vitamin D

2

ND

1.0

-3.0

ND

1.0

-5.0

ND

1.0

-14

ND

1.0

-4.0

Vitamin D

3

0.067

1.0

-10

0.079

1.0

-13

1.2

1.0

-10

2.3

1.0

-9.0

CONCLUSION
In this application, a fast analytical method is established based on the two-dimensional liquid 
chromatography to simultaneously assay the contents of vitamin A, α/β/γ/δ-tocopherol, vitamin D

2,  

and vitamin D

3 in infant milk powder. The method is simple, fast, highly automated, precise and accurate. 

It is suitable for the assay of vitamin A, α/β/γ/δ-tocopherol, vitamin D

2, and vitamin D3 in formula milk 

powder or other milk products.

References
1.  Mcmahon A, Christiansen S, Shinel, et al. J AOAC Int, 2013, 96 (5): 1073.

2.  GB 5009.82-2016. Assay of Vitamins A, D, and E in Food.

3.  ACQUITY UPLC Systems with 2D Technology Product Solution, 720003899EN

Waters, ACQUITY, Arc, UPLC, XSelect, XBridge, Empower, and The Science of What’s Possible are 

trademarks of Waters Corporation. All other trademarks are the property of their respective owners.

©2020 Waters Corporation. March 2020 720006655EN APAC-PDF

Waters Corporation 
34 Maple Street 
Milford, MA 01757 U.S.A. 
T: 1 508 478 2000 
F: 1 508 872 1990 
www.waters.com