Screening of Aquatic and Wetland Plant Species for Phytoremediation of

Explosives-contaminated Groundwater from the Iowa Army Ammunition Plant

E. P. H. BEST," M. E. ZAPPl: H. L. FREDRICKSON: S. L. SPRECHER,b S. L. LARSON,h AND M. OCHMAN"

OAAScl Corporation, 3402 Wisconsin Avenue Suite 5

Vicksburg, Mississippi 39081 'US. Army Corps o,f Engineers Waterways Experiment Station

3909 Halls Ferry Road Vicksburg, Mississippi 39180-61 99

Munitions material such as 2,4,6-trinitrotoluene (TNT) and hexahydro-1,3,5- trinitro-1,3,5-triazine (RDX) and their combustion and decomposition products can enter the environment from production activities and field usage and disposal.',* The presence of these substances is of concern because of their potential toxicity and mutagenicity?

The purposeful use of plants for clean-up of the environment has received relatively little attention despite the fact that plants, like microorganisms, play an important role in nature in sustaining and restoring environments. The capabilities of plants to absorb, accumulate and metabolize, directly or indirectly, various organic substances suggest their use in the remediation of contaminated environments (phytoremediation).

In the aquatic environment, both TNT and RDX can disappear rapidly from water owing to photolysis because they are sensitive to irradiance above 290 nm (ultraviolet (UV) and visible light). Adsorption to sediment is not significant. Bio- transformation of TNT in water by microorganisms is potentially important, since rather rapid degradation rates have been f o ~ n d . ~ The most common initial degrada- tion step of TNT was recently confirmed to be reduction of one or more nitro groups (products: 2-amino-dinitrotoluene (2ADNT), 4-amino-dinitrotoluene (4ADNT)) followed by irreversible binding of the products to organic matter.5 Only under strict anaerobic conditions was mineralization to COz demonstrated. Another, aerobic, degradation pathway is described in which one or more nitro groups are removed (products: 2,4-dinitrotoluene (2,4DNT), 2,6-dinitrotoluene (2,6DNT), 2-nitro- toluene (2NT), 4-nitrotoluene (4NT), 3-nitrotoluene (3NT).6 The biodegradation rate of RDX in water is slower than that of TNT. Biodegradation of RDX is proposed to proceed under anaerobic conditions via sequential reduction of the nitro groups to a point where destabilization and fragmentation of the ring occurs.' Degradation products include: hexahydro-l-nitroso-3,S-dinitro-l,3,S-triazine, hexa- hydro-1,3-dinitroso-S-nitrohydrazine, 1,2-dimethylhydrazine, formaldehyde and methanol.*

179

180 ANNALS NEW YORK ACADEMY OF SCIENCES

Recently it has been suggested that TNT disappears rapidly from water in the presence of several algae and submersed macrophytes, and that TNT derivatives with reduced nitro groups are early biodegradation products; no mineralization to COz was demonstrated? This was largely confirmed by studies similar to that de- scribed in this paper, evaluating submersed'" and emergent macrophytes." Degrada- tion of TNT by freshwater sediments has been seen to originate largely from aquatic plant enzymes;" however, reduction activity was so low that it could be attributed also to the presence of the very few bacteria in the extracts tested. Screening studies'O~ll indicated slow disappearance of RDX in the presence of aquatic mac- rophytes where plant-stimulated activity of microorganisms inherent to the explo- sives-contaminated water may have contributed to removal.

The Iowa Army Ammunition Plant (IAAP) encompasses a 26 square-mile area near Burlington, Iowa (longitude 91" 20' W, latitude 40" 48' N) and has on-going munition manufacturing activities. Explosives contamination has been detected at several locations. The U.S. Army Engineer District, Omaha, in conjunction with the U.S. Army Environmental Center has investigated various options for the removal and remediation of explosives contamination within both soil and ground- water matrices. Two sites, Lines 1 and 800, are scheduled for excavation of contami- nated soils and groundwater seeping into the excavation pit will be treated by constructed wetlands.

This paper presents the results of a screening study to quantify the ability of ten submersed and emergent macrophytes, adapted to lentic habitats in Iowa, to phytoremediate local explosives-contaminated groundwater. Species evaluated under hydroponic batch conditions were the submersed Potamogeton nodosus Poir. (American pondweed), and Ceratophyllum demersum L. (coontail), and the emer- gent AZisma subcordaturn Raf. (water-plantain), Sagittaria latifolia Willd. (common arrowhead), Carex uulpinoidea Michx. (fox sedge), Scirpus cyperinus (L.) Kunth (wool-grass), Eleocharis obtusa (Willd) (Blunt spikeursh), Phalaris arundinacea L. (reed canary grass), Typha angustifolia L. (narrowleaf cat-tail). Parrot-feather (Myriophyllum aquaticum (Vell.) Verdc.) was included in the evaluation to provide a comparison between the present study and other similar evaluations. The effects of native and heat-inactivated sediments on the explosives in the aqueous phase were also examined. The impact of a decrease in temperature from 25 to 10°C on contaminant disappearance was assessed in three species.

MATERIAL AND METHODS

Plant Material, Water and Sediment

The nine submersed and emergent species evaluated for ability to degrade explosives in IAAP groundwater were selected from 42 local native species shown to possess nitroreductase enzyme activity on the basis of USEPA immuno- assay testing in the field at IAAP, August 1995.13 Ecological traits important in field deployment were also considered, including perenniality, high biomass production, adaptability to water depth and extensive rootirhizome systems. Attention was given to the ecological habitat range of the species, including their presence in wet meadow, marsh and pond sites. Plants for evaluation were collected during the first two weeks of September 1995 from commercial sources or government agencies.

BEST et aL: PHYTOREMEDIATION OF GROUNDWATER 181

TABLE 1. Chemical Characteristics of the Iowa Army Ammunition Plant Groundwater Composite

~~ ~

Characteristic Value

pH Macro-, micronutrients (mg L-')

Alkalinity Bicarbonate Carbonate Total Dissolved Solids Kjeldahl-N Ni tritc/nitrate-nitrogen Ammonium-nitrogen Total-phosp horus Phosphate-phosphorus Sulfate Calcium Iron Manganese

Octahydro-1,3,5,7-tetranitro-l,3,5,7-tetrazocine (HMX) 2,6-Diamino-4-nitrotoluene (2,6DANT) 2,4-Diamino-6-nitrotoluene (2,4DANT) Hexahydro-l,3,5-trinitro-l,3,5-triazine (RDX) 1,3,5-Trinitro-benzene (TNB) 1.4-Dinitro-benzene (1,4DNB) l,3-Dinitro-benzene (I ,3DNB) Nitrobenzenc (NB) 2,4,6-Trinitrotoluene (TNT) 2-Amino-dinitrotoluene (2ADNT) 4-Amino-2,6-dinitrotoluene (4ADNT) 2,4-Dinitrotoluene (2,4DNT) 2,6-Dinitrotoluene (2,6DNT) 2-Nitrotoluene (2NT) 4-Nitrotoluene (4NT) 3-Nitrotoluene (3NT)

Explosives (pg L-')

7.5 rt 0.1

170 rt 1.4 170 i 1.3 0.6 5 0.2 839 t 58 1.6 2 0.2

0.15 * 0.09 0.17 2 0.01

4 . 2 0.185 ? 0.020 56.35 2 0.72

121.16 i 3.43 0.18 i 0.04 0.06 i 0

NA 596 rt 36 23 i 14

12785 t 1744 1422 2 88

20 i- 4 15 i 3

681 i 70 60 t 4 41 i- 2

6 i 7

-

-

-

- -

Mean values and standard deviations (N = 3). NA, not analyzed; -, below detection.

Groundwater samples collected from IAAP Monitoring Wells G19 and G20, mixed approximately 50: 50, were analyzed (TABLE 1). RDX was the primary con- taminant with TNT and TNB also being of concern.

IAAP (Stump Lake) sediment was collected at the same time as the groundwater samples. This sediment is low in organic matter (77 g kg DW-') and cation exchange capacity (CEC; 34 mgeq 100 g DW-I), and it is expected that adsorption of explosives is OW.^^.^^

Experimental Procedures

The treatment effect of individual plant species, sediments, and groundwater alone on the TNT and RDX components of IAAP groundwater over a 10-day incubation was evaluated in triplicate using a total test volume of approximately 3.4 L in 15 X 15 X 37.5 cm glass aquaria (reactors). Two incubation temperature regimes were evaluated: 25°C and 10°C. The 25°C experiment was carried out in

is2 ANNALS NEW YORK ACADEMY OF SCIENCES

an environmental chamber equipped with high pressure sodium halide lamps which provided a full photosynthetic spectrum of 400 to 500 pE m-? sec-' at 22.5 cm above the water surfaces. The 10°C incubations were performed in cooled, cabinet growth chambers equipped with fluorescent cool-white (TL and TLC) lamps which provide 300 to 650 p E m-' sec-' 22.5 cm above the water level within the reactors. The latter light source provided a similar illumination but a different photosynthetic spectrum compared to the former. Auto-timers were used to maintain 14 hours a day of artificial sunlight within incubation chambers.

All plants were evaluated in hydroponic culture without sediment to segregate sediment-based explosives transformations from plant-based ones. Submersed spe- cies were planted without support using 15 cm apical shoots at a density of 9.0 g fresh weight per liter of groundwater, yielding a total plant mass of approximately 30.4 g per reactor. Emergent plants, except parrot-feather, were supported using stainless steel hardware cloth. Approximately half of the biomass of each emergent plant was in contact with the water, therefore, these plants were planted at twice the density of the submersed ones. The water in the reactors was not aerated in order to prevent vertical mixing and mimic the low-flow conditions expected in a constructed wetland system.

Water from each reactor was sampled at 1, 4, 12, 24, 48, 96, and 240 hours for explosives analysis; 1 00-ml samples were first concentrated using solid phase extraction (Waters RDX cartridges catnr 47220). Subsequently, explosives were eluted from the cartridges using acetonitrile. The samples were evaporated almost to dryness using N2, redissolved in a 2 mL mixture of acetonitrile :water (50: 50 v : v), and subsequently analyzed using high performance liquid chromatography (HPLC) lollowing a modification of USEPA Method 8330." Explosives detection level was 0.1 pg L-'. Prior to sampling, contents of the reactor were mixed with a glass rod to collect a representative sample. Initial groundwater and water samples composited from the 25°C triplicate reactors at the end of the ten-day incubations were analyzed for pH, alkalinity, total dissolved solids, sulfate, Kjeldahl nitrogen, nitratehitrite, ammonia, total phosphate, ortho-phosphate, chemical oxidation dc- mand, calcium, iron, magnesium and manganese. After the final water sampling, oxygen concentration was measured within the reactors of one 25°C replicate. Azoxy compounds were measured only in the ten-day water samples of one 25°C replicate due to the lengthy procedure required for their analysis. Details are given else- where."

After completion of the incubation, plant material from selected reactors was quick-frozen in liquid nitrogen and ground to fine powder. This plant material was extracted with acetonitrile, and analyzed via HPLC using a modification of USEPA Method 8330 (S. L. Larson, US. Army Engineer Waterways Experiment Station, Vicksburg, MS, personal communication). Sediments were analyzed similarly, with- ou t grinding.I3

Data Analysis

Statistical summary and analysis of explosives concentrations in the groundwater were carried out on TNT and RDX separately, using STATGRAPHICS Plus (Ver- sion 7; Statistical Graphics Corporation, Bitstream Inc., Cambridge, MA) to perform analysis of variance (ANOVA), regression analysis and multiple range tests. Signifi- cance was tested at the 95% confidence level. Where log transformation was required of data on nitrobody concentration that were below detection, the detection levcl was used in place of zero.

BEST el al.: PHYTOREMEDIATION OF GROUNDWATER 183

RESULTS AND DISCUSSION

TNT Concentration in Water: Effects of Plant Species and Temperature

At 2S°C, plant incubations achieved 94 to 100% TNT removal in 10 days while controls ranged from 62 to 85% (TABLE 2). All of the submersed and thrcc of the eight emergent species removed TNT to levels below the detection limit (of 0.1 pg L ') within the 10-day incubation period. This indicates that plants provide addi- tional TNT sinks over those in sediment and groundwater alone. Increased removal may be attributed to increased adsorption sites on the plant mass, or enzymatic degradation of the TNT on and within the plant. Chcmical analysis of the plant mass, discussed later, indicated that enzymatic degradation was the likely cause of TNT removal. It is important to note that the additional adsorption sites added by the plant mass are not additive upon the sediment sorption sites since the plant incubations were done without sediment.

Of the controls, groundwater alone removed the lowest percentage of TNT, 62%. This was probably due to photolysis and biodegradation by groundwater- inherent microorganisms. Volatilization is not a factor because TNT is not volatile. Both sediment incubations yielded approximately 20% more TNT removal than groundwater alone. Since it is assumed that autoclaving eliminates sediment-based microorganisms and most enzymatic activity, the removal achieved is probably due to adsorption and sediment nutrient-based stimulation of groundwater-inherent microorganisms. The actual mechanism cannot be determined from this study.

Significant differences among species and controls in TNT concentration in groundwater at 25 and at 10°C over the whole incubation period are shown in TABLE 3 with reed canary grass, coontail and pondweed ranking as the most active. At 10"C, all three species were more effective than groundwater alone.

To examine kinetics differences among treatments, two types of regression were performed. An exponential regression model, Y = exp(a + bX), where Y = concentration and X = time, with a negative slope, was used to describe TNT decrease over time. These regression statistics were used to calculate first order rate constants (slope of the regression; pg TNT L-' h-'), half-life of TNT in hours, t,,2 = -ln2/b, and to extrapolate hydraulic retention time (HRT required to reach a TNT clean-up level of 2 pg L-I). While half-life statistics are independent of concentration and intercept, this time to target concentration does involve these data; therefore, different species with the same half-life do not necessarily provide the same cleanup kinetics.

On the basis of the first order rate constants (TABLE 2), reed canary grass had the most rapid TNT removal rate (0.038 pg L-' h-'), followed closely by pondweed, arrowhead, fox sedge (0.037 pg L-' h-I), and by coontail 0.033 p g L-' K'). This ranking is consistent with the chemical removal results mentioned above indicating that these plant species performed best in removing the compounds analyzed during this study. Estimated HRTs for the plant species varied from 4.9 days for reed canary grass to 19.8 days for parrot-feather. The groundwater alone control required the longest HRT (61.4 days) for removing TNT of all systems evaluated. When rate constants were normalized for total plant dry mass (KJ, coontail, pondweed, parrot-feather and wool-grass were most effective.

TABLE 4 illustrates the impact on removal of the lower incubation temperature for the treatments evaluated. The plant systems were impacted more by the tempera- ture decrease than were the controls. It is estimated that at 10°C it will take reed canary grass four times longer to remove TNT to target treatment levels than at 25°C (22 versus 5 days). It will take controls only one to two times longer. Although

TA

BL

E 2

. C

urve

Fit

Stat

istic

s fo

r T

NT

Con

cent

ratio

ns i

n G

roun

dwat

er o

ver

10-d

av In

cuba

tions

with

Pla

nt S

Dec

ies

and

Con

trol

s ~

~~

Initi

al M

ass

Incu

bate

d Y

= e

xp(a

+ bX

) (g

DW

) In

terc

ept

Slop

e H

alf-

D

ays

to

K(p

g L

-' In

Con

tact

T

reat

men

t a

b R?

lif

e (h

) 2

pg

L '

h-'

g D

W I

) w

ith W

ater

To

tal

25°C In

cuba

tions

Su

bmer

sed

P. n

odos

us

6.40

7 -0

.037

0.

95

18.7

6.

4 - -

0.00

9 3.

73

3.73

C.

dem

ersu

m

5.49

3 -0

.033

0.

42

21.0

6.

1 -0

.01

1 3.

03

3.03

Em

erge

nt

A. s

ubco

rdat

urn

S. la

tifol

ia

C. v

ulpi

noid

ea

S. c

yper

inus

E.

obt

usa

P. a

rund

inac

ea

T. a

ngus

tifol

ia

M. a

quat

icum

C

ontr

ois

Gro

undw

ater

Se

dim

ent

Aut

ocla

ved

sedi

mc.

10

" Inc

ubat

ions

6.67

8 6.

493

6.53

6 6.

360

6.63

6 5.

166

6.47

0 6.

400

6.59

1 6.

587

:nt

6.54

0

-0.0

30

-0.0

37

-0.0

37

-0.0

16

-0.0

20

-0.0

38

-0.0

20

-0.0

12

-0.0

04

-0.0

08

-0.0

08

0.89

0.

99

0.99

0.

96

0.68

0.

65

0.68

0.

92

0.94

0.

95

0.92

23.1

18

.7

18.7

43

.3

34.6

18

.2

34.6

57

.7

173.

2 86

.6

86.6

8.3

6.5

6.6

14.7

12

.4

4.9

12.0

19

.8

61.4

30

.6

30.5

-0.0

02

-0.0

02

-0.0

01

-0.0

04

-0.0

005

- 0.

001

-0.0

01

-0.0

07

-0.0

0005

-0

.000

05

4.64

19

.05

22.6

2 26

.47

20.7

4 21

.05

7.47

1.

83

0 15

2.84

15

2.82

12.3

4 22

.34

28.9

9 43

.15

35.8

3 36

.83

16.2

6 5.

51

0 15

2.84

15

2.82

Emer

gent

E.

obt

usa

6.53

5 -0

.008

0.

88

86.6

30

.4

-0.0

005

15.1

6 16

.65

P. a

rund

inac

ea

6.44

2 -0

.011

0.

88

63.0

21

.7

-0.0

007

14.8

2 20

.40

M. a

quat

icum

6.

536

-0.0

0s

0.80

13

8.6

48.6

-0

.002

1.

87

5.61

Gro

undw

ater

6.

645

-0.0

03

0.65

23

1.0

82.6

0

0 C

on tr

uls

Sedi

men

t 6.

639

-0.0

05

0.85

13

8.6

49.5

-0

.000

03

152.

80

152.

80

Dat

a no

t no

rmal

ized

for

pla

nt d

ry w

eigh

t. K

,, ra

te c

onst

ants

nor

mal

ized

for

mas

s. I

nitia

l m

ass,

mea

ns o

f tr

iplic

ates

. Ini

tial

TNT

conc

entr

atio

n in

gr

ound

wat

er:

681

pg

L-I

.

bl

BEST et al.: PHYTOREMEDIATION OF GROUNDWATER 185

TABLE 3. Treatment Effects on TNT Concentration in Groundwater over 10-dav Incubation

Homogeneous Treatment LS Mean Groups

25°C Incubations P. arundinacea 271.91 a C. demersum 301.52 ab P. nodosus 348.63 bc S. latifolia 362.52 cd C. vulpinoidea 374.28 cde S. cyperinus 390.74 cdef T. angustifolia 414.65 defg A. subcordatum 420.22 efgh

E. obtusa 464.16 h Autoclaved sediment 527.42 i Sediment 550.75 i

M. aquaticum 434.65 gh

Groundwater 601.84 j 10°C Incubations

P. arundinacea 463.85 a E. obtusa 526.50 b M . aqualicum 563.73 b Sediment 625.45 C

Groundwater 682.54 d

Multiple range analysis by treatment. ANOVA showed that treatment and time affected the TNT concentration significantly ( p < 0.001; blocks as covariates). Difkrent letters indicatc significant differences between treatments.

this extends clean-up time to 83 days with water alone, the temperature effect is an important consideration when designing a wetland system expected to be opera- tive year-round. It is expected to be even greater when plants are under natural low-temperature conditions when lower mass and metabolic activity occur.

TABLE 4. Impact of Temperature on TNT First Order Rate Constants and Hvdraulic Residence Times to Meet Treatment Goal (2 U P L-')

Emergent E. obtusa -0.020 0.40 2.45 P. arundinacea -0.038 0.29 4.43 M. aquaticum -0.012 0.42 2.45

Groundwater -0.004 0.75 1.35 Sediment -0.008 0.63 1.62

Controls

RDX Concentration in Water: Effects of Plant Species and Temperature

RDX removal was slower than that observed for TNT at 25°C (TABLE 5). Significant differences among species and controls in RDX concentration over the

6

TA

BL

E 5

. C

urve

Fit

Stat

istic

s for

RD

X C

once

ntra

tions

in G

roun

dwat

er o

ver

10-d

ay In

cuba

tions

with

Pla

nt S

peci

es a

nd C

ontr

ols

Y=

a+

bX

In

itial

Mas

s In

cuba

ted

(g D

W)

Inte

rcep

t Sl

ope

Tre

atm

ent

a b

R*

25°C

Inc

ubat

ions

Su

bmer

sed

P. n

odos

us

C. d

emer

sum

A. s

ubco

rdat

um

S. la

tifol

ia

C. u

ulpi

noid

ea

S. c

yper

inus

E

. obt

usa

P. a

rund

inac

ea

T. a

ngus

tifol

ia

M. a

quat

icur

n

Emer

gent

cont

rois

G

roun

dwat

er

Sedi

men

t A

utoc

lave

d se

dim

ent

10"

Incu

batio

ns

Em

erge

nt

E. o

btus

a P.

aru

ndin

acea

1269

4 -3

.714

0.

06

1315

0 -7

.461

0.

13

1259

7 9.

654

0.24

12

699

-6.9

85

0.15

12

799

1233

7 12

546

1248

5 12

894

1256

9

1287

2 12

562

1265

7

-13.

053

-2.5

62

-13.

336

1.53

2

0.88

9 3.

635

-5.8

36

-10.

211

-13.

452

0.51

0.

03

0.01

0.

24

0.01

0.

04

0.07

0.

29

0.42

1216

3 0.

685

0.01

11

742

-0.5

16

0.01

M

. aqu

atic

um

1256

9 4.

315

0.19

Gro

undw

ater

12

594

0.42

8 0.

01

Sedi

men

t 12

45 1

-4.5

60

0.13

Con

trol

s

Hal

f-

Day

s to

%

(pg

L-'

In C

onta

ct

with

Wat

er

Tot

al

life

(h)

2 pg

L-'

h-'

. g D

W-')

1696

14

2 -0

.995

3.

73

3.73

90

5 73

-2

.462

3.

03

3.03

-

-

902

75

490

40

2320

20

0

456

39

-

-

-

-

1110

91

60

4 51

46

5 39

-

-

1036

6 94

7

-

-

1328

11

3

-

4.64

12

.34

-0.3

62

19.0

5 22

.34

-0.5

77

22.6

2 28

.99

4

-0.0

97

26.4

7 43

.15

3 F -

20.7

4 35

.83

-

7.47

16

.26

-

1.83

-0.6

34

21.0

5 36

.83

5.51

3 4

-0

.067

15

2.84

15

2.84

-0

.088

15

2.82

15

2.82

A

4 0

4

-

15.1

6 16

.65

-0.0

35

14.8

2 20

.40

-

1.87

5.

61

5 G 0

OO

-

0 0

$ -0

.030

15

2.80

15

2.80

Dat

a no

t no

rmal

ized

for

pla

nt d

ry w

eigh

t. K

,, ra

te c

onst

ants

nor

mal

ized

for

dry

mas

s. I

nitia

l mas

s, m

eans

of

trip

licat

es. I

nitia

l R

DX

con

cent

ratio

n in

gro

undw

ater

: 12

785

pg L

-I.

3 m

v1

BEST et at.: PHYTOREMEDIATION OF GROUNDWATER 187

whole incubation period are shown in TABLE 6. Only treatment with reed canary grass was significantly more effective than water alone at 25°C. At IOT, reed canary grass and spikerush were more effective. Most plants achieved no more removal of R D X than did controls (TABLE 5 ) . In three of the incubations with emergents, an increase in R D X was seen over the incubation period. This anomaly may result from interfercnce of some plant constituent that has a similar retention time when eluted from the HPLC column used, and would contribute to an explanation of the variability seen in the R D X data. None of the plants indicated a potential for rapid R D X removal within the 10-day incubations evaluated; however, reed canary grass, fox sedge, arrowhead and coontail achieved 27%, 26%, 15% and 12% R D X removal, respectively. Failurc by planted reactors t o achicvc at least the same level of removal as the groundwater alone may be due to decrease in photolysis caused by shading.

For the incubations with groundwater alone, photolysis and biotransformation by groundwater-inherent microorganisms apparently were minor mechanisms of R D X removal. Low photolysis is supported by the relatively long half-life of 46 days found in the present study compared to 13 days in surface water in winter estimated by Spanggord and coworker^.'^ The increased RDX removal in sediment incubations indicates that mechanisms for removal other than photolysis are active in sediment. Actual mechanisms are not known, but are speculated to be adsorption and sediment-nutrient-based stimulation of groundwater-inherent microorganisms.

A linear regression model, Y = a + bX, where Y = concentration, X = time, and b is negative over time, gave the best fit to changc in R D X over time (TABLE

TABLE 6. Treatment Effects on R D X Concentration in Groundwater over 10-day Incubation

I S Homogeneous Treatment Mean Groups

25°C Incubations P. arundinacea Autoclaved sediment Scdiment C. vulpinoidea S. cyperinus S. lutifolia P. nodosus Groundwater E. obtusa C. dernersum M. aquaticurn T. angustifolia A. subcordaturn

10°C Incubations €'. arundinacea E. obtusa Sediment Groundwater M. aquuticurn

11776.60 1 1952.83 1201 9.58 12105.70 12200.94 12328.78 12496.37 12578.76 12627.42 12753.64 12788.65 12941.22 13110.22

11715.1 9 12 199 .56 12208.78 12617.40 32786.03

a ab abc abcd abcde abcde abcdef bcdef hcdef

cdef def

ef f

a b b

C C

Multiple rangc analysis by treatment. ANOVA showed that treatment affected the 'I'NT concentration significantly (25°C p = 0.008; 10°C p < 0.001), as did time ( p < 0.001; blocks as covariates). Diffcrent leltcrs indicate significant differences between treatments.

188 ANNALS NEW YORK ACADEMY OF SCIENCES

TABLE 7. Impact of Temperature on RDX Zero Order Rate Constants and Hydraulic Residence Times to Meet Treatment Goal (2 pg L-')

Treatment K&g L-' h- ' ) K i a K r HRTiol HRT25

Emergent E. obtusa NR - NRlNR P. ariindinacea -13.336 0.04 24.28 M. aauaticum NR NRlNR -

con trois Groundwater - - - 5.836 NRI-5.836 Sediment -10.21 1 0.44 2.22

NR = No removal of RDX observed.

5) . These statistics were used to calculate zero order constants (slope of the regres- sion; pg RDX L-' h-'), half-life of RDX and HRT. However, fit was poor and incubations with four emergents had non-negative slopes. Rate constants in sedi- ment controls were higher than in groundwater alone.

Reed canary grass, fox sedge and autoclaved sediment achieved the most rapid removal. Rate constants normalized for total plant dry mass (K,) indicated that coontail, pondweed, reed canary grass and fox sedge were the most effective species.

Decreased temperature appears to be more detrimental for RDX removal than for TNT (TABLE 7). The predicted HRT for reed canary grass, the only plant to remove RDX to 2 pg L I in the 10°C incubations, increased 24 times at 10°C. The same decrease in temperature only increased estimated HRT for TNT by approximately four times (TABLE 4). At 10°C RDX removal was essentially elimi- nated except in the presence of sediment.

Thus, while decreasing temperature resulted in slower removal of TNT within the planted systems, it had a far more dramatic impact on the removal of RDX. Considerable differences in performance were observed among the control systems, and it is possible that differences in irradiance in the growth chamber and cabinet may have contributed to differences in photolytic removal. Lamps in the 25°C growth chamber provided a better photonic flux (higher similarity to sunlight spectrum) than in the 10°C growth cabinet. Too few observatiohs were made to correlate the RDX disappearance rate with oxygen concentration. It was significantly correlated to oxygen concentration in a similar study, however, indicating that RDX clean-up period increases with increasing O2 or that RDX is more rapidly degraded under less oxygenated conditions.'" This suggests the involvement of facultative and/or obligate anaerobic microorganisms in RDX removal.

Azoxy Compounds

No azoxy compounds were detected in the 10-day samples incubated at 25°C in any treatment.

Explosives and TNT Degradation Products in Plant Material and Sediments

Plant tissue of selected species from the groundwater incubations underwent HPLC analysis to assess the ultimate fate of explosives in contact with plants.

BEST et aL: PHYTOREMEDIATION OF GROUNDWATER 189

Unexposed reference plants were also analyzed, to assess whether natural chemical constituents co-elute with targeted contaminants.

In incubated plants TNT concentrations were usually near detection level (0.1 p g L in extract) and slightly higher in the aerial portions of water-plantain and arrowhead (FIG. 1). TNT did not accumulate. RDX concentrations were substantial (to 200 pg g DW-') in the plant portions in contact with water, and often far higher in the aerial portions (to 1600 pg g DW-'), indicating translocation. The aerial portion of arrowhead at 200,000 p g L-' RDX was 15X higher than the final concen- tration in the incubation groundwater (13,000 pg L-'). Apical transport of RDX has been reported for terrestrial plants.18 Condensation and crystal formation on the outside of the plants is also feasible, since evapotranspiration was substantial.

Submerged 25OC 0 Reference lncubuted

J

i-LLKd- PN CO AS SL CV SC EO PA TA MA SLD ASEO

15 I 14 - 13 - 12 - 11 - 10 - 9 - 8 - 7 - 6 - 5 , T Is 3

0 Aerial 25 C 0 Reference

lncubuted

2 1

PN CD AS SL CV SC EO PA TA MA SEO ASEC

2000

I aoo -- 1600 '3

1400

3 1200

1000

aoo e c c 8 600 ::

400

200

PN CO AS SL N SC EO PA TA MA SED ASCD ' PN CD AS SL CV SC EO PA 1A MA SFD ASE

Treotrnent Treatment

FIGURE 1. Concentrations of TNT and RDX in aerial and submersed portions of explosives- contacted plants and sediments (mean values and standard deviations, N = 3) and in reference plants (single values); 25°C incubations. Abbreviations: PN, P. nodosus; CD, C. clemer.sum; AS, A. subcordaturn; SL, S latifolia; CV, C. vulpinoidea; SC, S. cyperinus; EO, E. obtusa; PA, P. antndinacea; TA, T. angustifolia; MA, M. aquaticum; SED, sediment; ASED, auto- claved scdiment.

190 ANNALS NEW YOHK ACADEMY OF SCIENCES

Reference plants appeared to have very little material that co-eluted with RDX (FIG. 1).

Concentrations of TNT-metabolites and TNT-photolytic products in plants were generally low (TABLE 8)" Of note are: 1) the occurrence of substantial 2ADNT and 4ADNT only in arrowhead, and 4ADNT in pondweed; in contrast, these metabolites were present in the groundwater initially and after incubation with most plant species and controls; 2 ) the absence of 2,6DANT and 2,4DANT in all plant tissue, metabolites that had increased significantly in water incubated with some of the emergents; 3) the absence of products derived from TNT by removal of one or more nitro-groups in all plants and incubation water (only 2,4DNT shown in TABLE 8; other NTs were below detection); 4) the presence of TNB in most plant species, with higher concentrations in aerial than in submersed plant portions. In general, the aerial portions of arrowhead had the highest level of T N T metabo- lites.

The concentrations of explosives and TNT degradation products in the sediments were near detection level, consistent with expected explosives adsorption based on low C E C and low organic matter

Plant Health and Growth

Plant growth was minimal over 10-day incubation. Visual inspection during the incubations indicated that most plants recovered from transplant shock to groundwater after 7 days. Coontail and arrowhead did not appear to recuperate in that period. Water-plaintain, wool-grass, cat-tail and parrot-feather exhibited posi- tive weight gains, while coontail and arrowhead exhibited the largest weight loss (data not shown). Light levels were considered adequate for the submersed plants being close to photosynthetic saturation (600-800 pE m-* s-').'' To the emergent plants, however, light levels were approximately half of typical, but the amount supplied is considered adequate for maintaining activity.

Low biomass increase in the present study was attributed to the lateness of the growing season, lack of carbon availability and/or light and nitrogen, and the short term of the study. Growth inhibition by the explosives concentration cannot be ruled out. Although inhibition by TNT is not likely, since its concentration was lower than that reported as toxic (2 to 15 ppm) for aquatic plants and inhibition by RDX concentration is feasible. No data on toxicity of R D X to aquatic plants are available to date. Toxicity of the IAAP water is suggested by the slow removal rate of parrot-feather in the present study compared to that documented in a similar study.'" However, differences in parrot-feather strains may have contrib- uted to the contrasting reactions.

General Discussion

Disappearance rates seen here for TNT and R D X from groundwater, with or without plants, are rather low relative to literature values. Realistic comparison is difficult, however, because of non-uniformity among studies. Similar submersed plant screenings at WES show T N T removal in 4 to 5 days in most specie^.",'^ The longer half-lives of both T N T and R D X found in the present study are attributed to low photolysis, with lack of UV light. If the contribution of the plants to the disappearance of explosives per se is linked directly to plant growth, this influence is expected to be low at the end of the growth season in autumn and far higher in

TABL

E 8.

Con

cent

ratio

ns o

f R

DX

and

TN

T, T

NT

-Met

abol

ites

and

TNT-

Phot

olys

is P

rodu

cts

(pg

g D

W-')

in

Plan

t Tis

sue

and

Non

/Aut

ocla

ved

Sedi

men

t af

ter

10-d

ay In

cuba

tion

~ ~~

TN

T-P

hoto

lysi

s T

NT

-Met

abol

ites

prod

ucts

In

Con

tact

Pl

antlS

edim

ent

with

Wat

er

RD

X

TN

T

2AD

NT

4A

DN

T

24D

NT

T

NB

D

NB

Su

bmer

sed

P. n

odos

us

C. d

emer

sum

21

2 18

1 51

.9

0.3

0.9

0.8

1.6 -

+ + -

0.3

Em

erge

nt

A. s

ubco

rdai

um

1557

21

8 99

8 14

4 64

5 10

2 49

2 17

9 43

8 18

2 28

7 12

8 15

9 22

0 63

1 21

7

3.9

0.3

7.9

0.7 -

-

-

0.7

0.7 -

-

-

1 .o - -

0.6

2.6

1.5

18.5

1.

1 0.

8

1.3

0.5

1.5

0.3

-

-

-

2.7

11.3

0.

6 1.

2 1.

5

1.3

1.9

1.3

1.4

0.9

0.8

0.5

-

-

-

-

1.5

2.1

11.2

4.

5 2.

9 0.

7 0.

6 0.

5 0.

9 1.

0 4.

0 2.

3 0.

8 4.

4 2.

0 6.

0

2.1

0.5

3.7

0.2

1.6

1.2

0.1

0.3 -

-

-

-

-

-

1.1

0.8

-

2.2 2.4

0.2

0.5

0.5 -

-

-

S. la

tifol

ia

C. v

ulpi

noid

eu

S. c

yper

inus

E. o

btus

a

P. a

rund

inuc

eu

T. a

ngus

tifol

ia

M. aq

uatic

um

8

1.7

3.5

Con

trol

s Se

dim

ent

0.2

0.3

0.1 -

+ i-

-

0.2

-

0.1

Aut

ocla

ved

sedi

men

t

Mea

ns o

f tr

iplic

ates

. -,

belo

w d

etec

tion.

192 ANNALS NEW YORK ACADEMY OF SCIENCES

spring and summer. However, this contribution can also be linked indirectly to plants, e.g., to the microorganism-activity-stimulating effects of leachates. In case of the latter, highest effects are expected either during the growth season when soluble carbohydrates are leached (about 10% of net primary production), or in autumn and winter when phenols are exuded to enter the environment during plant senescence and decomposition.22

The generally negligible TNT-metabolite concentrations in plant material versus those in the incubation water suggest that either most TNT degradation occurs in the plants early in the incubation period and is followed by TNT metabolite leaching from the plants to the water, or that most TNT is degraded outside the plants by microorganisms in the water but is stimulated by plant leachates. A combination of both mechanisms cannot be excluded. Only reductive pathways appeared to be active. Evidence for reduction in plants was found in arrowhead (2ADNT and 4ADNT) and in pondweed (4ADNT) but not for the presence of a pathway involving the removal of one or more nitro groups from TNT. The presence of reductive pathways of TNT and absence of nitro group removal from TNT in plants agrees with studies on terrestrial plant^.*^-*^

The lack of significant differences in RDX decrease rates between plant species and groundwater suggests that breakdown by microorganisms was the dominant mechanism, probably stimulated by plant leachates. Translocation and accumulation of RDX into aerial plant portions agrees with studies on terrestrial Since no RDX metabolites were screened, the degradation route of this compound cannot be assessed.

While RDX disappearance rate in the present study was not statistically corre- lated with oxygen concentration, due to unreplicated observations, a similar study showed RDX disappearance was significantly correlated to oxygen concentration with a positive coefficient, indicating that RDX clean-up period increases with increasing 02, or that RDX is more rapidly degraded under less oxygenated condi- tions.'" This suggests the involvement of facultative and/or obligate anaerobic micro- organisms in RDX removal.

SUMMARY

The results of this study indicate that the presence of plants did enhance TNT and TNB removal from IAAP groundwater. Most effective at 25°C were reed canary grass, coontail and pondweed. Groundwater and plant tissue analyses indicate that in presence of the plants tested TNT is degraded to reduced by-products and to other metabolites that were not analyzed. TNT removal was best modeled using first order kinetics, with rate constants at 25°C incubations ranging from 0.038 pg L-' h-I for reed canary grass to 0.012 pg L-' h-l for parrot-feather. These kinetics predict hydraulic retention times (HRTs) ranging from 4.9 days to 19.8 days to reach a TNT concentration of 2 p g L-I. Decreasing incubation temperature to 10°C affected reed canary grass more than parrot-feather, increasing estimated HRTs by factors of four and two, respectively.

The plant species tested showed a far lower potential for RDX removal from the IAAP groundwater. Most effective at 25°C were reed canary grass and fox sedge. Analyses of plant material indicated the presence of RDX in under-water plant portions and in aerial plant portions, and RDX accumulation in the latter. RDX removal was best modeled using zero order kinetics, with rate constants for the 25°C incubation ranging from 13.45 pg L-' h-' for reed canary grass to no

BEST ef al.: PHYTOREMEDIATION OF GROUNDWATER 193

removal in four species. Based on these kinetics, estimated HRTs to reach 2 p g L-' RDX increased from 39 days. Decreasing the temperature to 10°C increased HRT 24-fold for reed canary grass.

By using the biomass-normalized K value, submersed plants are identified as having the highest explosives-removing activity ( p g explosive L-' h-' g DW-I). However, biomass production o f submersed plants is normally five to ten times less than that of emergent plants per unit area, and, thus, in plant selection for wetland construction, both, explosives removal potential and biomass production are important determinants.

1.

2.

3.

4.

5.

6.

I .

8.

9.

10.

11.

12.

13.

14.

REFERENCES

SMALL, H. J. & D. H. ROSENBLAIT. 1974. Munitions production products of potential concern-Phase 11. Technical Report 7404. AD91 9031. U S . Army Medical Bioengi- neering R&D Laboratory. Fort Detrick, Frederick, MD.

SPANGGORD, R. J., W. R. MABEY, T. MILL, T. W. CHOIJ, J. H. SMITH, S. LEE & D. RORERTS. 1983. Environmental fate studies on certain munitions wastewater constituents. Phase IV. Lagoon model studies. Final Report Phase IV. DAMD17-78-C-8081. SRI Interna- tional, Menlo Park, CA.

MARVIN-SIKKEMA, F. D. & J. A. M. DEB0N.r. 1994. Degradation of nitroaromatic com- pounds by microorganisms. Appl. Microbiol. Biotechnol. 42: 499-507.

SPANOGORD, R. J., T. MILL, T. W. Ciiou, W. R. MAREY, J. H. SMITH & S. LEE. 1980. Environmental fate studies on certain munitions wastcwater constituents. Phase I. Literature review. Final Report. Contract No. DAMD17-78-C-8081. SRI International, Menlo Park, CA.

RIEGER, P. G. & H. J. KNACKMUSS. 1995. Basic knowledge and perspectives on biodegra- dation of 2,4,6-trinitrotoluene and related nitroaromatic compounds in Contaminated soil. In Biodcgradation of Nitroaromatic Compounds. J. C:. Spain, Ed.: 1-18. Plenum Press, New York.

VORBECK, C., H. LENKE, P. FISCHER & H. J. KNACKMIJSS. 1994. Identification of a hydride Meisenheimer complex as a metabolite of 2,4,6-trinitrotoluene by a Mycobacterium sp. J. Bacteriol. 176 932-934.

M ~ O R M I C K , N. G., J. H. CORNELI. & A. M. KAIJLAN. 1981. Biodegradation of hexahydro- 1,3,5-trinitro- I ,3,5-triazine. Appl. Environ. Microbiol. 42: 817-823.

M ~ O R M I C K , N. G . , J. H. CORNELL & A. M. KAPLAN. 1984. The anaerobic biotransforma- tion of RDX, HMX, and their acetylated derivatives. Technical Report NaticklTR- 85/007. 36 pp.

SCHNOOR, J. L., L. A. LICHT, S. C . MCCIJTCHEON, N. L. WOLFE & L. H. C A I ~ R ~ ~ I I M . 1995. Phytorcmediation of organic and nutrient contaminants. Pilot and full-scalc studies are demonstrating the promise and limitations of using vegetation for remediat- ing hazardous wastes in soils and sediments. Environ. Sci. Tcchnol. 29: 318-323.

BEsr, E. P. H. & S. L. SPRECHER. 1996. Phytoremediation of explosives-contaminated groundwater using constructed wetlands. Phase 1 report: Plant screening study- submersed plant species. Letter Report Environmental Laboratory, IJSAE Walerways Experiincnt Station, Vicksburg, MS.

TVA. 1995. Screening of wetland emergent species for remcdialion of explosivcs-contam- inated groundwater. Tennessee Valley Authority Report. 29 pp.

VAN BEELEN, P. & D. R. BURRIS. 1995. Reduction of the explosive 2,4,h-Trinitrotoluene by enzymes from aquatic sediments. Environ. Toxicol. Chcm. 14 21 15-2123.

BEsr, E. P. H., M. E. ZAPPI. H. L. FREDRICKSON, S. L. SPRECHEK & J. MILLER 1997. Screening of aquatic and wetland plant species for phytoreinediation of explosives contaminated groundwater from the IAAP. U.S. Army Corps of Engineers Technical Report EL-97-2.

PENNING.I.ON, J . & W. PATRICK. 1990. Adsorption and desorption of2,4,6-Trinitrotoluene by soils. J. Environ. Qual. 19: 559-567.

194 ANNALS NEW YORK ACADEMY OF SCIENCES

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

PENNINGTON, J.. C. HAYES, K. MYERS, M. OCHMAN, D. GUNNISON, D. FeLr & E. MCCOR- MICK. 1995. Fate of 2,4,6-Trinitrotoluene in a simulated compost system. Chemosphere

JENKINS, T. F., P. H. MIYARES, K. F. MYERS, E. F. MCCOKMICK & A. I3. STRONG. 1995. Comparison of solid phase extraction with salting-out solvent extraction for preconcentration of nitroaromatic and nitramine explosives from water. Anal. Chim. Acta 289 69-78.

SPANGGORD, R. J., T. MILL, T. W. CHOIJ, W. R. MABEY, J. H. SMITH & S. LEE. 1980. Environmental fate studies on certain munition wastewater constituents. Phase 11. Laboratory studies. Final Report AD A099256. SRI International, Menlo Park, CA.

CATALDO, D. A,, S. HARVEY & R. J. FELLOWS. 1990. An evaluation of the environmental fatc and behavior of munitions material (TNT, RDX) in soil and plant systems. Environmental fate and behavior of RDX. Project Order No. 88PP8853, US Army Medical Research and Development Command, Fort Detrick, Frederick, MD

VAN, T. K., W. T. HALLER & G. BOWES. 1976. Comparison of the photosynthetic charac- teristics of three submersed aquatic plants. Plant Physiol. 58: 761-768.

SCHOTT, C. D. & E. G. WORTHLEY. 1974. The toxicity of TNT and related wastes to an aquatic flowering plant: Lcmna perpusilla Torr. Edgewood Arsenal, Technical Report No. EB-TR-74016.

SMOCK, L. A,, D. L. STONEBURNEX& J. R. CLARK. 1976. The toxiceffects of trinitrotoluene (TNT) and its primary degradation products on two species of algae and flathead minnow. Water Res. 10 534-543.

FLETCHER, J. S., P. K. DONNELLY & R. S. HEGDE. 1995. Biostimulation of PCB-degrading bacteria by compounds released Crom plant roots. Proceedings of the Third Interna- tional In situ and On-Site Bioreclamation Symposium.

PALAZZO, A. J. & D. C. LEGETT. 1986. Effect and disposition of TNT in a terrestrial plant. J. Environ. Qual. 15: 49-52.

PENNINGTON, J . 1988. Plant uptake of 2,4,6-trinitrotoluene, 4-amino-2,6-dinitrotoluene, and 2-aniino-4,6-dinitrotoluene using 14C-labeled and unlabeled compounds. Techni- cal Report EL-88-20. Department of thc Army, Waterways Experiment Station, Corps of Engineers, P.O. Box 631, Vicksburg, MS 39181-0631.

CATALDO, D. A., S. HARVEY, R. J. FELLOWS, R. M. BEAN & B. D. MCVEETY. 1989. An evaluation of the environmental fatc and behavior of munitions material (TNT, RDX) in soil and plant systems. Project Order No. 88P8853, US Army Medical Research and Development Command, Fort Detrick, Frederick, MD 21701-5012.

HARVEY, S. D., R. J. FELLOWS, D. A. CAIALDO & R. M. BEAN. 1990. Analysis of 2,4,6- trinitrotoluene and its transformation products in soils and plant tissues by high- performance liquid chromatography. J. Chromatogr. 51& 361-374.

HAKVEY, S. D., R. J. FELLOWS, D. A. CATALDO & R. M. BEAN. 1991. Fate of the explosive hexahydro-1,3,5-trinitro-1,3,.5-triazine (RDX) in soil and bioaccumulation in bush bean hydroponic plants. Environ. Toxicol. Chem. 10: 845-855.

3 0 429-438.

21701 -5010.